R-SPONDIN2+ Mesenchymal Cells Form the Bud Tip Progenitor Niche During Human Lung Development

Renee FC Hein; Joshua H Wu; Emily M Holloway; Tristan Frum; Ansley S Conchola; Yu-Hwai Tsai; Angeline Wu; Alexis S Fine; Alyssa J Miller; Emmanuelle Szenker-Ravi; Kelley S Yan; Calvin J Kuo; Ian Glass; Bruno Reversade; Jason R Spence

doi:10.1016/j.devcel.2022.05.010

. Author manuscript; available in PMC: 2023 Jul 11.

Published in final edited form as: Dev Cell. 2022 Jun 8;57(13):1598–1614.e8. doi: 10.1016/j.devcel.2022.05.010

R-SPONDIN2⁺ Mesenchymal Cells Form the Bud Tip Progenitor Niche During Human Lung Development

Renee FC Hein ¹, Joshua H Wu ², Emily M Holloway ¹, Tristan Frum ², Ansley S Conchola ³, Yu-Hwai Tsai ², Angeline Wu ², Alexis S Fine ², Alyssa J Miller ³, Emmanuelle Szenker-Ravi ⁴, Kelley S Yan ⁵, Calvin J Kuo ⁶, Ian Glass ⁷, Bruno Reversade ⁴, Jason R Spence ^1,^2,^3,^8,^*

PMCID: PMC9283295 NIHMSID: NIHMS1810736 PMID: 35679862

SUMMARY

Human respiratory epithelium is derived from a progenitor cell in the distal buds of the developing lung. These ‘bud tip progenitors’ are regulated by reciprocal signaling with surrounding mesenchyme; however, mesenchymal heterogeneity and function in the developing human lung is poorly understood. We interrogated single-cell RNA sequencing data from multiple human lung specimens and identified a mesenchymal cell population present during development that is highly enriched for expression of the WNT agonist RSPO2, and we found that adjacent bud tip progenitors are enriched for the RSPO2 receptor LGR5. Functional experiments using organoid models, explant cultures, and FACS-isolated RSPO2⁺ mesenchyme show that RSPO2 is a critical niche cue that potentiates WNT signaling in bud tip progenitors to support their maintenance and multipotency.

Graphical Abstract

graphic file with name nihms-1810736-f0001.jpg

eTOC Blurb

Hein et al. present scRNA-seq of the developing human distal lung, showing transcriptionally distinct populations of mesenchyme. Spatial profiling showed that RSPO2⁺ mesenchymal cells are physically adjacent to LGR5⁺ epithelial bud tips. Organoid experiments revealed that RSPO2⁺ cells create a high WNT signaling environment, supportng bud tip identity and multipotency.

INTRODUCTION

Human lung development begins approximately four weeks post conception as the lung buds develop from the ventral anterior foregut endoderm. Soon after, the buds begin branching morphogenesis, leading to a network of epithelial tubes that include the trachea and series of bronchi that become progressively smaller, terminating at the gas-exchanging alveoli (Miller and Spence, 2017; Conway et al., 2020). During branching morphogenesis, the tip of each budding branch possesses a population of transient epithelial progenitor cells called ‘bud tip progenitors.’ In vivo lineage tracing in animal models has shown that bud tip progenitors give rise to all cell types in the lung epithelium, including those that line the airways and alveoli (Rawlins et al., 2009; Yang et al., 2018). More recently, a population of bud tip progenitors in the human lung has been identified, and functional experiments have shown that they have the ability to generate a broad spectrum of lung epithelial cell types (Nikolić et al., 2017; Miller et al., 2018, 2020).

In mice, a specialized niche that supports bud tip progenitors is made up of surrounding mesenchyme that provides physical and biochemical support and determines whether bud tip progenitors will self-renew or differentiate into different epithelial cell types (Shu et al., 2002; Weaver, Batts and Hogan, 2003; Alejandre-Alcázar et al., 2007; Li et al., 2008; Rajagopal et al., 2008; Tsao et al., 2008; Goss et al., 2009; Morrisey and Hogan, 2010; McCulley, Wienhold and Sun, 2015; Zepp and Morrisey, 2019; Riccetti et al., 2020). Genetic gain- and loss-of-function studies have identified many of the signaling pathways important for bud tip progenitor maintenance and for determining cell-fate choices during differentiation (Morrisey and Hogan, 2010). Drawing from these studies, a minimal set of essential signaling cues required to maintain isolated human bud tip progenitor cells in long-term in vitro culture has been described (Nikolić et al., 2017; Miller et al., 2018); however, the specific mesenchymal cells and signaling components that make up the in vivo bud tip progenitor niche are unclear. Moreover, we have only begun to understand similarities and differences between animal models and humans (Danopoulos, Shiosaki and Al Alam, 2019; Conway et al., 2020; Miller et al., 2020).

Here, we investigate mesenchymal cell populations in the developing human distal lung from 8.5 – 19 weeks post-conception, a location where and time when the lung possesses an actively branching bud tip progenitor population. Leveraging single-cell RNA sequencing (scRNA-seq) data and unsupervised clustering analysis, we identified transcriptionally distinct mesenchymal cell populations in the distal lung domain during this time frame, including smooth muscle cells and four, non-smooth muscle mesenchymal cell clusters. Two clusters are highly enriched for expression of the WNT agonist R-SPONDIN 2 (RSPO2). It is notable that Rspo2^-/- mice have mild lung defects compared to lung aplasia that is seen in humans with RSPO2 mutations (Bell et al., 2008; Szenker-Ravi et al., 2018). Indeed, mutations in human RSPO2 are lethal (Szenker-Ravi et al., 2018); however, the specific role of RSPO2 in the developing human lung has not been interrogated.

Based on the severe phenotype linked with RSPO2 mutations in humans as well as the large population of RPSO2⁺ mesenchymal cells identified in scRNA-seq data, we interrogated the spatial localization of RPSO2⁺ cells along with the functional role of RSPO2. By fluorescence in situ hybridization (FISH) and immunofluorescence (IF), we show that RSPO2 is expressed in mesenchymal cells physically located adjacent to bud tip progenitors. On the other hand, SM22⁺ airway smooth muscle cells lack RSPO2 expression and line the newly differentiating proximal (airway) epithelium in domains adjacent to bud tip progenitor cells. In addition, we found that the RSPO2 receptor LGR5, but not other LGR family members, is uniquely expressed in bud tip progenitors, as are other canonical WNT target genes such as AXIN2. Using in vitro lung explant cultures with functional inhibition experiments, we show that blocking endogenous RSPO signaling leads to a loss of the high WNT signaling environment in the bud tip domain and stochastic differentiation of bud tip progenitors into multiple airway, but not alveolar, epithelial cell types. LGR5 knock-down experiments in bud tip progenitor organoids suggest that RSPO2 acts through LGR5 in bud tips to maintain their progenitor fate. Lastly, we identified LIFR as a cell surface marker for RSPO2⁺ cells and show that FACS-isolated LIFR^HI cells support the ability of bud tip organoids to give rise to both airway and alveolar epithelium while LIFR^- cells only support airway differentiation in co-cultures. Collectively, this work identifies an RSPO2-producing niche cell in the human fetal distal lung mesenchyme and reveals a critical RSPO2/LGR5-mediated WNT signaling axis that supports bud tip progenitor cell maintenance and multipotency throughout early lung development.

RESULTS

Single-cell RNA sequencing identifies mesenchymal cell populations in the fetal distal lung

To interrogate the mesenchymal cells present in the developing human lung in the bud tip progenitor domain, we re-analyzed scRNA-seq data from the physically-isolated distal portion of 8.5 – 19 week post-conception lungs (n = 5 lungs) (Figure 1A) (Miller et al., 2020). We used principal component analysis for dimensionality reduction, Louvain clustering, and UMAP for visualization (Wolf, Angerer and Theis, 2018; Becht et al., 2019). Expression analysis of canonical marker genes identified major cell classes within the data, including epithelial, mesenchymal, immune, and endothelial cell types as well as a cluster of proliferating cells (Figures S1A and S1B). To specifically interrogate mesenchymal cell types that may comprise the bud tip progenitor niche, we extracted and re-clustered the non-vascular smooth muscle mesenchymal cell clusters 0 and 1.

Re-clustering identified 5 transcriptionally distinct mesenchymal cell clusters (Figure 1B). Differential expression analysis showed that cluster 4 contains cells expressing canonical markers of smooth muscle cells (e.g., SM22, ACTA2) and is enriched for expression of HHIP, FOXF1, and WNT5A (Figure 1D). Cells in clusters 0 and 1 have the most distinct transcriptome compared to smooth muscle cells and share very similar gene expression profiles, expressing enriched levels of RSPO2, WNT2, CDO1, BMP5, LIFR, and FGFR4, but differ in expression of RPS4Y1, EGR1, and CA3 (Figure 1D). Cluster 2 is uniquely enriched for genes including SOX11, PDGFRα, STC1, and KCKN17 and cluster 3 is uniquely enriched for genes including ELN, MGP, SERPINF1, AGTR2, FGF7, and BMP4.

Of note, transcriptional differences between clusters 0 and 1 may largely be based on gestational age since we observed a nonequivalent distribution over time (Figure 1C). Consistent with this, we found that expression of EGR1, distinguishing cells in clusters 0 and 1, changes over gestational age (Figure S1C). Importantly, clusters 0 and 1 are enriched for the WNT signaling molecules WNT2 and RSPO2 (Figure 1D). Given that the growth of human bud tip progenitor cells in culture requires exogenous WNT stimulation (Nikolić and Rawlins, 2017; Miller et al., 2018), we hypothesized these cells might make up an important part of the bud tip progenitor niche.

RSPO2⁺ mesenchymal cells are localized adjacent to bud tip progenitor cells

To spatially profile RSPO2⁺ mesenchymal cells, we used FISH combined with IF on human fetal lung tissue sections spanning 8.5 – 19 weeks post-conception. RSPO2⁺ mesenchymal cells were co-visualized with airway smooth muscle cells (SM22⁺) because the scRNA-seq data showed that these populations are molecularly distinct. Co-FISH/IF for RSPO2 and SM22 confirmed that these markers are expressed in different mesenchymal cell populations where RSPO2⁺ mesenchymal cells are located physically adjacent to bud tip progenitor cells and SM22⁺ airway smooth muscle cells line the more proximal, bud tip-adjacent epithelium (Figure 1E).

Additional FISH data showed that RSPO2⁺ cells express WNT2 and FGFR4, confirming scRNA-seq data and supporting previous reports revealing expression of WNT2 and FGFR4 in the distal lung near bud tip progenitors (Figure S1C) (Goss et al., 2009; Miller et al., 2012; Danopoulos et al., 2018; Danopoulos, Shiosaki and Al Alam, 2019; Yu et al., 2020). FGFR4 expression was also found in bud tip progenitors (Figure S1C). Also complimenting scRNA-seq data, PDGFRα expression was found in both RSPO2⁺ cells and in SM22⁺ cells and appears particularly enriched in SM22⁺ cells directly adjacent to bud tip regions (Figure S1D, arrowhead). Enriched co-expression of AGTR2 with BMP4 or SERPINF1 (cluster 3 in Figure 1B) is evident in mesenchymal cells adjacent to SM22⁺ smooth muscle cells, especially surrounding larger airways (Figure S1E). Based on scRNA-seq and FISH, the other R-SPONDIN transcripts were detected at much lower levels compared to RSPO2 and not specifically localized near bud tip or differentiating epithelium (Figure S1F and S1J). In contrast to the distal lung domain, RSPO2 expression is nearly absent from lung mesenchyme surrounding the bronchi and trachea and from all epithelia (Figure 1E and S1G). The proximity and specificity of RSPO2⁺ mesenchyme to bud tip progenitors further suggested that RSPO2 comprises an important component of the bud tip progenitor niche.

LGR5 is expressed in bud tip progenitor cells

One mechanism by which R-SPONDIN proteins are known to amplify WNT signaling is by binding to LGR receptors and sequestering ubiquitin ligases RNF43 and ZNRF3, subsequently freeing the WNT receptor Frizzled from protein degradation (de Lau, Snel and Clevers, 2012; Niehrs, 2012; Chen et al., 2013; de Lau et al., 2014; Park et al., 2018; Raslan and Yoon, 2019). To determine which cells RSPO2⁺ mesenchymal cells may signal to, we used FISH to characterize the localization of LGR receptors within the fetal lung. LGR5 (but not LGR4 or LGR6) is highly specific to bud tip progenitor cells in the distal lung (Figure 1F and S1H). LGR5 expression is largely excluded from mesenchymal cells and from differentiated epithelial cell types, except for a subset of basal cells in the more proximal airways (Figure S1H). In contrast, LGR4 is expressed broadly throughout the mesenchyme in the distal and proximal lung and is excluded from the distal epithelium while LGR6 is expressed specifically in airway smooth muscle cells in the distal and proximal lungs (Figure S1H). In addition, the RSPO co-receptors RNF43 and ZNRF3 are bud tip-enriched (Figure S1I). The specific expression of LGR5 in bud tip progenitors suggests that LGR5 may be a cognate receptor for RSPO2 present in the bud tip progenitor niche, and the enriched expression of RNF43 and ZNRF3 suggest these may also play a role in signaling.

WNT target gene expression is enriched in bud tip progenitor cells

Given the expression pattern of RSPO2 and LGR5, we predicted bud tip progenitors would display higher levels of WNT-mediated target gene expression compared to other cell types in the distal lung. Using AXIN2 expression as a read-out for WNT signaling and SOX9 as a marker for bud tips (Rawlins, 2008), quantification of FISH data revealed that AXIN2 is significantly enriched in bud tip progenitor cells compared to other cell types in the distal lung (Figure 1G). Based on our collective data, this further supports our hypothesis that RSPO2 from the mesenchyme may act on bud tip progenitors via LGR5 to support a high WNT signaling domain to maintain bud tip progenitors.

RSPO2-mediated WNT signaling in bud tips is required for proximal-distal patterning

To test the necessity of RSPO2-mediated WNT signaling for bud tip progenitor maintenance, we used an adenovirus (ad) expressing the soluble ectodomain of LGR5 (hereafter termed LGR5 ECD ad), which was previously shown to bind and inhibit endogenous RSPO2, leading to reduced levels of WNT signaling in infected tissues (Yan et al., 2017). We infected human fetal lung explants placed in an air-liquid-interface culture system with the LGR5 ECD ad or a control ad encoding murine immunoglobulin IgG2a (Yan et al., 2017) twice over a 4-day culture period. Both the control and the LGR5 ECD ad-infected explants grew over the 4-day culture period, and in both conditions, mesenchymal cells were retained, and the epithelium expanded (Figure 2A). We confirmed successful infection of the virus via antibody staining against murine IgG2a_Fc for the control and against FLAG for the LGR5 ECD ad (Figure S2A). Both viruses appeared to target the mesenchyme more than the epithelium and infect the edge of the explants more strongly than the center (Figure S2A), so we focused our analysis to the periphery of explants when possible.

Figure 2. — (A) Inverted microscope images of 11.5-week human fetal lung explants at the start of air-liquid interface culture (day 0) and after 4 days of culture (day 4) with 2 infections of either a control or LGR5 ECD adenovirus (ad).

(B) Immunofluorescence (IF) for bud tip/distal marker SOX9 and airway/proximal epithelial marker SOX2 on sections from 12-week uncultured human fetal lung tissue, control ad-infected explants, and LGR5 ECD ad-infected explants. Quantification of the SOX9 stain is shown in the bottom right. At the end of the 4-day culture period, SOX9⁺ cells in the control ad-infected explants comprised approximately 36.0% of cells while they only comprised 12.8% in LGR5 ECD ad-infected explants (unpaired Welch’s one-tailed t-test). This quantification was performed in three unique biological specimens with one to three technical replicates and a minimum of three image fields for each specimen.

(C) Fluorescence *in situ* hybridization (FISH) and quantification of *AXIN2* and co-IF for bud tip/distal marker SOX9 on control ad-infected explants and LGR5 ECD ad-infected explants. The average number of *AXIN2* molecules per cell in each image was calculated for epithelial and mesenchymal compartments separately by co-immunofluorescence for the nuclear lung epithelial marker TTF1 and counting *AXIN2* molecules per TTF1⁺ cells (epithelium) and TTF1^- cells (mesenchyme). This quantification was performed in three unique biological specimens for LGR5 ECD ad-infected explants and two unique biological specimens for control ad-infected explants with one to three technical replicates and a minimum of three image fields for each specimen. Statistical tests were carried out using unpaired Welch’s one-tailed t-tests.

(D) FISH of *LGR5* and co-IF for bud tip/distal marker SOX9 on sections from control ad-infected explants and LGR5 ECD ad-infected explants. Boxes on the top mark the inset location shown on the bottom. Insets on the bottom right on the bottom images show a single-channel image for *LGR5*.

(E) FISH and quantification of *RSPO2* and co-IF for smooth muscle marker SM22 on sections from control ad-infected explants and LGR5 ECD ad-infected explants. The level of *RSPO2* expression in LGR5 ECD ad-infected explants is not significantly different than *RSPO2* expression in control ad-infected explants (unpaired Welch’s one-tailed t-test). This quantification was performed in three unique biological specimens with one to three technical replicates and a minimum of three image fields for each specimen.

Following 4 days of culture, the explants infected with the LGR5 ECD ad exhibited reduced staining for the distal bud tip marker SOX9 in the epithelium, with SOX9⁺ cells making up 36.0% of total cells in the control but only 12.8% of cells in the LGR5 ECD ad-infected explants (Figure 2B). In comparison to in vivo, uncultured lung tissue of a similar gestational age and previous reports (Abler et al., 2017; Miller et al., 2018), the control ad-infected explants maintained proper SOX2 and SOX9 epithelial patterning, with SOX9^HI/SOX2^LOW bud tips and SOX9^-/SOX2^HI proximal (airway) epithelium (Figure 2B). Much of the epithelium in the LGR5 ECD ad-infected explants became SOX9^-/SOX2^HI (Figure 2B), indicative of a loss of bud tip identity and differentiation into airway epithelium. WNT target gene expression, measured by AXIN2 FISH, was also reduced in LGR5 ECD ad-infected explants compared to the control (Figure 2C), and non-phosphorylated (active) β-Catenin staining was more robust in control versus LGR5 ECD ad-infected explants (Figure S2B). LGR5, also a WNT target gene, was reduced via FISH (Figure 2D). In control ad-infected explants, LGR5 was expressed and restricted to SOX9⁺ bud tips (Figure 2D). Although LGR5 was still detected in the few remaining bud tip progenitors in the LGR5 ECD ad-infected explants, its endogenous expression appeared reduced relative to bud tip progenitors in control ad-infected explants, consistent with loss of bud tip progenitor identity (Figure 2D). Note that many ectopic LGR5⁺ cells could be detected, revealing abundant expression in strongly-infected cells, which made it difficult to quantify changes in endogenous LGR5 expression (Figure 2D).

RSPO2 expression was not affected by control or LGR5 ECD ad infection as expression was maintained at similar levels in SM22- mesenchyme in both conditions (Figure 2E). There were no significant differences in KI67 expression between the control and LGR5 ECD ad-infected explants (Figure S2C). In some cases, independent of viral infection, explants grew abnormally large, leading to cell death or necrosis towards the center of the explant; however, Cleaved Caspase 3 (CCASP3) staining was low or absent in most explants (Figure S2D).

The bud tip progenitor transcriptional profile is dependent on RSPO2-mediated signaling in bud tips

Human fetal lung explants infected with control or LGR5 ECD ad for 4 days were dissociated, and single cells were sequenced via scRNA-seq. Louvain clustering and UMAP visualization revealed clusters of epithelial, mesenchymal, vascular smooth muscle, endothelial, neuroendocrine, and proliferating cells, which were identified by examining expression of canonical marker genes for these cell types (Figure S3A and S3C). There is also a cell cluster composed of only LGR5 ECD ad-infected cells, which appears to be clustered based on high expression of LGR5 (Figure S3A-C). As a control, we sequenced non-infected explants from the same experiment, which contained the same general cell populations (Figure S3D-F). We noted that most clusters (with the exception of the LGR5⁺ cluster) possessed cells from each sample that were evenly distributed across clusters; however, part of the epithelial cell cluster showed separation between non-infected and control ad-infected cells compared to LGR5 ECD ad-infected cells (Figure S3D-E). This separation was also noticed in the UMAP embedding excluding non-infected cells (Figure S3A-B).

To gain better resolution of changes occurring in the epithelium, the epithelial cell cluster (cluster 1) from the UMAP embedding that includes cells from control and LGR5 ECD ad-infected explants (Figure S3A) was extracted and re-clustered (Figure 3A). We identified cluster 0 as bud tip progenitors based on enriched expression of canonical bud tip progenitor marker genes (Figure 3C). Based on visual inspection of this cluster and expression of bud tip markers, we noted that cells from each sample as well as gene expression were not evenly distributed across the cluster (Figures 3B-C). To gain insights into possible differences between the control and LGR5 ECD ad-infected bud tip cells, cluster 0 was again extracted and re-clustered (Figure 3D). When re-clustered, there were 4 predicted clusters, with sub-cluster 2 primarily consisting of cells from the control ad-infected explants and sub-cluster 0 primarily consisting of cells from the LGR5 ad-infected explants (sub-clusters 1 and 3 contained both conditions) (Figure 3D-E). Bud tip progenitor genes (SOX9, TESC, ETV5, CA2) were more highly expressed in control ad-infected cells (Figure 3F). Individual cells from each sample were evaluated against a panel of the top 22 most differentially expressed genes from published in vivo bud tip progenitors (see Table S1) (Miller et al., 2020), thus assigning a “bud tip progenitor cell score” to every cell (Holloway, Wu, et al., 2020). Consistent with the reduced expression of individual bud tip progenitor genes in LGR5 ECD ad-infected explants, cells from LGR5 ECD ad-infected explants scored lower for having a bud tip progenitor identity (Figure 3G). Together, this data suggests that loss of endogenous RSPO2 activity during human lung development causes a reduction of bud tip progenitor gene expression.

Figure 3. — A) Cluster plot of the epithelial cells (*EPCAM*⁺/*KRT18*⁺/*KRT8*⁺/*CLDN6*⁺) computationally extracted from LGR5 ECD adenovirus (ad)-infected explants and control ad-infected explants sequenced using single-cell RNA sequencing (scRNA-seq) (re-cluster of cluster 1 from Figure S3A). Each dot represents a single cell and cells were computationally clustered based on transcriptional similarities. The plot is colored and numbered by cluster.

B) UMAP plot corresponding to Figure 3A. Each dot represents a single cell and dots/cells are colored by the sample from which they came from.

C) UMAP feature plots corresponding to the cluster plot in Figure 3A and displaying expression levels of the known bud tip progenitor markers *SOX9*, *TESC*, and *ETV5*. The color of each dot indicates log-normalized and z-transformed expression level of the given gene in the represented cell.

D) Cluster plot of bud tip-like cells (re-cluster of cluster 0 from Figure 3A). Each dot represents a single cell and cells were computationally clustered based on transcriptional similarities. The plot is colored and numbered by cluster.

E) UMAP plot corresponding to Figure 3D. Each dot represents a single cell and dots/cells are colored by the sample from which they came from.

F) Violin plots corresponding to the cluster plot in Figure 3D and displaying expression known of bud tip progenitor markers *SOX9*, *TESC*, *ETV5*, and *CA2* in control ad- and LGR5 ECD ad-infected cells.

G) Violin plot and UMAP feature plot corresponding to the cluster plot in Figure 3D and displaying bud tip progenitor cell score, calculated as the average expression of the top 22 enriched genes in *in vivo* bud tip progenitor cells (see methods), for cells in LGR5 ECD ad-infected explants and control ad-infected explants. The color of each dot in the feature plot indicates log-normalized and z-transformed expression level of the set of bud tip genes in the represented cell.

RSPO2-potentiated signaling in bud tips prevents differentiation into airway cell types

Based on the reduced bud tip progenitor transcriptional profile in LGR5 ECD ad-infected explants compared to control ad-infected explants (see Figure 3), and because epithelial SOX2 expression was higher in the LGR5 ECD ad-infected explants compared to the control via IF (Figure 2), we hypothesized that the bud tips from these explants might be differentiating into airway lung cell types. Using scRNA-seq data from control and LGR5 ECD ad-infected explants and complementary IF and FISH stains on tissue sections from these explants, we evaluated expression of proximal (airway) and distal (alveolar) differentiated cell type markers. When possible, we calculated cell type scores (Holloway, Wu, et al., 2020) for specific cell types by evaluating the average expression of the published 50 most differentially expressed genes from in vivo cells of the listed cell type (Table S1) (Miller et al., 2020). We limited the scRNA-seq analysis to cells within the bud tip cluster (Figure 3) to specifically determine how the bud tips in the LGR5 ECD ad-infected cells are changing relative to controls.

The biggest difference in cell type marker expression from the control and LGR5 ECD ad-infected explants was expression of secretory cell genes. By FISH, the secretory cell marker SCGB3A2 was properly localized to airway structures and differentiating, bud tip-adjacent cells in control ad-infected explants, as seen in in vivo tissue (Miller et al., 2020) (Figure 4Ai.). In contrast, remaining SOX9⁺ bud tip regions in LGR5 ECD ad-infected explants expressed SCGB3A2 mRNA, suggesting airway differentiation (Figure 4Ai.). In agreement with this, scRNA-seq data revealed that LGR5 ECD ad-infected cells from the bud tip cluster had higher cell type scores for secretory progenitor and club cells, and to a lower extent, for goblet cells compared to the control (Figure 4Aii. – v.). Additionally, there was an increase in TP63⁺ cell numbers in the LGR5 ECD ad-infected explants compared to the controls (Figure 4Bi. – ii.), although basal cell scores for LGR5 ECD and control ad-infected cells were similar (Figure 4Biii.). Although the cell type scores for neuroendocrine and multiciliated cells were moderately increased in cells from LGR5 ECD ad-infected explants, CHGA⁺ cells and FOXJ1⁺ cells were only sparsely detected by IF in both conditions (Figure S4A). Overall, LGR5 ECD ad-infected explants had higher expression of airway lineage markers compared to control ad-infected explants, particularly with respect to secretory cell lineages.

Figure 4. — A) Expression of airway secretory cell type markers in LGR5 ECD adenovirus (ad)-infected explants and control ad-infected explants. **(i.)** Fluorescent *in situ* hybridization staining of the secretory cell marker *SCGB3A2* and co-immunofluorescence (IF) for the bud tip marker SOX9 and airway marker SOX2 on sections from control ad-infected explants and LGR5 ECD ad-infected explants. DAPI is shown in gray. Boxes on the top mark the inset location shown on the bottom. **(ii. - iv.)** Violin plots displaying the cell score for the listed airway secretory cell type, calculated from single-cell RNA sequencing (scRNA-seq) data as the average expression of the top 50 enriched genes in *in vivo* fetal secretory cell types (see methods), for cells in LGR5 ECD ad-infected explants and control ad-infected explants. **(v.)** Log-normalized and z-transformed expression level of the airway marker *SOX2* in scRNA-seq data from LGR5 ECD ad-infected explants and control ad-infected explants.

B) Expression of airway basal cell type markers in LGR5 ECD ad-infected explants and control ad-infected explants. **(i.)** IF staining for the airway progenitor/basal cell marker TP63 on sections from control ad-infected explants and LGR5 ECD ad-infected explants. **(ii.)** Quantification of basal cell marker TP63 on sections from control ad-infected explants and LGR5 ECD ad-infected explants. The number of TP63⁺ cells in LGR5 ECD ad-infected explants is increased but not significantly different than the number of TP63⁺ cells in control ad-infected explants (unpaired Welch’s one-tailed t-test). This quantification was performed in three unique biological samples with one to three technical replicates and a minimum of three image fields for each sample. **(iii.)** Violin plot displaying the basal cell score for cells in LGR5 ECD and control ad-infected explants, calculated from scRNA-seq data as the average expression of the top 50 enriched genes in *in vivo* fetal basal cells (see methods), for cells in LGR5 ECD ad-infected explants and control ad-infected explants.

To determine if bud tip progenitors in LGR5 ECD ad-infected explants were undergoing general differentiation or differentiating specifically to airway cell types, we examined LGR5 ECD ad-infected bud tips for the presence of differentiated alveolar cell types. By scRNA-seq, SFTPB was increased in LGR5 ECD ad-infected explants (Figure S4Bi.). It has recently been shown that SFTPB is expressed in some airway secretory cell types in addition to alveolar cells (Miller et al. 2020), which could account for higher expression in the LGR5 ECD ad-infected explants. Remaining alveolar cell type markers were similarly expressed between control and LGR5 ECD ad-infected explants (Figure S4Bii. – vii.). Overall, there does not appear to be a differentiation bias of bud tips into alveolar cell types between the two conditions, rather, we see an increase in differentiation towards airway cell types upon LGR5 ECD ad-infection.

LGR5 can respond to RSPO2 to maintain bud tip cell fate

Based on the specificity of LGR5 expression in bud tips, we hypothesized that exogenous RSPO2 would be able to maintain the bud tip cell fate in human fetal-derived bud tip progenitor organoids (BTOs) (Miller et al., 2018). Therefore, we cultured BTOs in the presence of recombinant human RSPO2, WNT3a-afamin conditioned media (Mihara et al., 2016; Nanki et al., 2018), or both, and included a positive control (+CHIR99021, shown to maintain BTOs) (Miller et al., 2018) and negative control (-CHIR99021; CHIR). After 4 days of culture, both conditions containing RSPO2 maintained both cystic structures (similar to the positive control) and denser structures (similar to the negative control) while bud tips cultured in WNT3a alone contained mostly dense structures (Figure 5A). RT-qPCR analysis confirmed that the addition of RSPO2 maintained expression of bud tip genes and AXIN2 at levels close to the positive control and at higher levels than the negative control or WNT3a-only condition; however, airway genes SCGB3A2 and TP63 were increased in all conditions compared to the positive control (Figure 5B). This suggests RSPO2 can act to maintain bud tip progenitors in BTOs but is less robust at doing so than the positive control containing CHIR.

Next, we confirmed that LGR5 is expressed in BTOs when cultured in published media containing FGF7, ATRA, and CHIR (Miller et al., 2018) (Figure 5C). To more directly test the requirement of LGR5 for maintaining a high WNT environment in bud tip cells, we used a published shRNA against LGR5 (Jha et al., 2017) to knock-down LGR5 in BTOs (Figure 5D). Cohorts of scrambled controls and LGR5 shRNA BTOs were switched into media containing FGF7 and ATRA either with or without CHIR (positive/negative control) or with RSPO2 in place of CHIR. After 4 days of culture, RT-qPCR analysis of bud tip markers and airway markers revealed that removing CHIR in scrambled controls led to a loss of bud tip markers (NPC2, TESC, SFTPC) that could be rescued by the addition of exogenous RSPO2. In contrast, RSPO2 was unable to rescue loss of bud tip markers in LGR5 shRNA BTOs. No differences were observed for expression of airway markers (SCGB3A2, TP63) (Figure 5D). This data suggests that bud tip progenitors with normal levels of LGR5 can respond to RSPO2 to maintain the bud tip cell fate, whereas LGR5 knockdown renders bud tips unable to respond to exogenous RSPO2.

Isolated RSPO2⁺ mesenchymal cells support bud tip multipotency in organoid co-cultures

To determine how the RSPO2 and SM22 mesenchymal cell populations differentially regulate bud tip progenitor cell behavior in culture, we performed 3D co-cultures with BTOs (Miller et al., 2018) and isolated RSPO2⁺ or SM22⁺ mesenchyme in Matrigel. In order to use fluorescence activated cell sorting (FACS) to isolate these populations, we used published approaches to find putative cell surface markers that are co-expressed in RSPO2⁺ cells [SurfaceGenie – (Waas et al., 2020)], which identified LIFR. We confirmed that LIFR is strongly co-expressed with RSPO2 in mesenchymal cells by scRNA-seq (Figure S5A). To specifically enrich for LIFR^HI (RSPO2⁺) and LIFR^- (SM22⁺) mesenchyme, we used a combinatorial staining approach that allowed for the isolation of non-epithelial (EPCAM^-), non-endothelial (CD31^-), LIFR^HI or LIFR^- cells using FACS (Figure S5B). RT-qPCR analysis on the isolated populations confirmed that genes enriched in RSPO2⁺ mesenchyme (RSPO2, FGFR4) were enriched in the isolated LIFR^HI population and genes enriched in airway smooth muscle cells (SM22, FOXF1) were enriched in the isolated LIFR^- population (Figure S5C). We also confirmed that epithelial cells and endothelial cells were successfully depleted from the LIFR^HI and LIFR^- populations (Figure S5C).

BTOs were cultured with LIFR^HI or LIFR^- mesenchymal cells directly after FACS-isolation in a media including FGF7 and ATRA (Miller et al., 2018), but excluding any WNT ligands or small molecule activators. A positive control without mesenchyme but with CHIR as well as a negative control without mesenchyme or CHIR were included in each experiment. After 10 – 11 days of bud tip/mesenchyme co-culture, the mesenchyme caused condensation of the Matrigel and bud tips into a more tightly compacted structure compared to bud tips cultured in Matrigel without mesenchyme; however, this was much more pronounced in the LIFR^- co-cultures (Figure 6A). Positive control (+CHIR) bud tips retained a normal, cystic phenotype while negative control (-CHIR) bud tips became dense, as expected (Figure 6A) (Miller et al., 2018).

By FISH, RSPO2⁺ cells were found in the LIFR^HI co-culture after 11 days of culture (Figure 6Bi. top panel). Surprisingly, SM22⁺ cells were also detected in the LIFR^HI co-culture at this time point, but at lower numbers and expression levels compared to the LIFR^- co-culture (Figure 6Bi. top panel). Within LIFR^HI co-cultures, RSPO2 expression was largely absent from SM22⁺ cells while RSPO2 expression was almost entirely absent from the LIFR^- co-culture, as expected (Figure 6Bi. top panel). PDGFRα expression was found throughout the mesenchyme in both co-cultures, similar to the in vivo lung (Figure 6Bi. and S1D). Interestingly, when isolated LIFR^HI cells are cultured in 2D, they largely maintain their expression profile in short-term cultures (6 hours) but gain expression of airway smooth muscle cell markers by 7 days in culture, whereas isolated LIFR^- cells maintain smooth muscle cell markers in short-term and long-term cultures (Figure S5D-E). This suggests that LIFR^HI/RSPO2⁺ cells can give rise to LIFR^-/SM22⁺ cells in 2D culture conditions. To corroborate this, we also performed pseudotime analysis on scRNA-seq data of 8.5 – 19-week human fetal distal lung mesenchyme and found predicted lineage trajectories from RSPO2⁺ cells to smooth muscle cells (Figure S5F) (Haghverdi et al., 2016).

By IF, we observed that aSMA⁺ cells (a marker for smooth muscle) in the LIFR^HI co-culture surrounded epithelium that had low levels of SOX9 and had high expression of the airway marker SOX2, while epithelial cells surrounded by RSPO2⁺ cells retained SOX9 expression (Figure 6Bi. bottom panel and 6Bii. – iii.). Epithelium in the LIFR^- co-cultures was SOX2^HI and SOX9^{- or LOW} (Figure 6B), indicative of bud tips undergoing airway differentiation. Bud tip genes SOX9, ETV5, NPC2, and LGR5 were increased in the LIFR^HI co-cultures compared to the LIFR^- co-cultures by RT-qPCR (Figure S6Bi.).

Next, we wanted to investigate if differentiated proximal (airway) or distal (alveolar) cell types were present in each co-culture. Airway cell types, including secretory, basal, multiciliated, and goblet cell markers were detected by IF and RT-qPCR at levels above the positive control BTOs in both co-culture conditions but were more significantly upregulated in LIFR^- co-cultures (Figure S6A and S6Bii.). Interestingly, the alveolar cell type markers SFTPC, ABCA3, and RAGE were upregulated in the LIFR^HI co-culture compared to all other conditions by RT-qPCR (Figure S6Biii.). At the protein level, SFTPC was detected at high levels in the LIFR^HI co-culture, and some SFTPC⁺ cells co-expressed other alveolar type II proteins, including SFTPB and HTII-280 (Figure S6Ai. – ii. and vi.). The alveolar type I marker RAGE was also sporadically detected in the LIFR^HI co-culture (Figure S6Av.). This data suggests some SOX9⁺ bud tips progenitors in the LIFR^HI co-culture have begun alveolar differentiation. Abundant SFTPB⁺ cells were detected in the LIFR^- co-culture and in the negative control, but they co-expressed SCGB3A2 (Figure S6Aii.), indicative of an airway secretory cell (Miller et al., 2020). Together, this data shows that smooth muscle cells (LIFR^-/SM22⁺) provide cues for differentiation into airway cell types while LIFR^HI/RSPO2⁺ mesenchymal cells support bud tip differentiation into alveolar cell types and airway cell types, potentially through RSPO2⁺ differentiation into smooth muscle cells for the latter.

In addition to differentiation, we wanted to determine if there were changes in cell proliferation between the two co-culture conditions. We found that KI67 expression was significantly upregulated in the LIFR^HI co-culture (Figure 6Ci. and 6Ciii.). This correlated with higher AXIN2 expression in the LIFR^HI co-culture compared to the LIFR^- co-culture and negative control (Figure 6Cii. and 6Civ.). This data suggests that RSPO2⁺ mesenchymal cells support a high WNT signaling niche conducive for self-renewal (proliferation) and differentiation of bud tip progenitor cells into both airway and alveolar epithelium.

DISCUSSION

The lung development field has well-established literature interrogating the diversity of cell types and functions in the epithelium; however, less is known with respect to the developing lung mesenchyme, and particularly in the context of the human lung. Most of what is known about the role of the lung mesenchyme during development has come from animal models and has primarily focused on understanding airway smooth muscle cells or mesenchymal heterogeneity and function during alveolar and later stages of development (Torday, Torres and Rehan, 2003; Chen et al., 2012; McQualter et al., 2013; El Agha and Bellusci, 2014; Li et al., 2015, 2018, 2020; Green et al., 2016; Zepp et al., 2017; Endale et al., 2017; Kishimoto et al., 2018; Wu et al., 2018; Noe et al., 2019; Goodwin et al., 2019, 2020; Guo et al., 2019; Han et al., 2019; Bridges et al., 2020; Riccetti et al., 2020; Yin and Ornitz, 2020; Gouveia et al., 2020; Negretti et al., 2021). There are also far fewer similar studies in humans (Rehan et al., 2006; Danopoulos, Shiosaki and Al Alam, 2019; Du et al., 2019; Goodwin et al., 2019; Leeman et al., 2019; Shiraishi, Nakajima, et al., 2019; Shiraishi, Shichino, et al., 2019; Danopoulos et al., 2020; Guney et al., 2020). Nevertheless, recent advances in single-cell analytical tools and in vitro human-specific model systems have allowed us to begin addressing unknowns in human lung development (Treutlein et al., 2014; Brazovskaja, Treutlein and Camp, 2019; Du et al., 2019; Kishimoto et al., 2019; Travaglini et al., 2019; Danopoulos et al., 2020; Miller et al., 2020; Yu et al., 2020). Here, we aimed to understand how mesenchymal cells are involved in creating an epithelial bud tip progenitor cell niche.

Single-cell RNA sequencing analysis predicted airway smooth muscle cells and four additional non-smooth muscle populations that have similar but not identical gene expression profiles in the developing human lung. We observed that two mesenchymal cell clusters identified by scRNA-seq are enriched for expression of the WNT-agonist RSPO2 throughout the developmental time frame analyzed, and spatial localization showed that RSPO2 is expressed adjacent to the bud tip domain. It is known that high WNT signaling conditions are necessary for the maintenance of the bud tip progenitor cell state in vitro (Nikolić et al., 2017; Miller et al., 2018; Rabata et al., 2020), which made this cell population a strong bud tip-associated mesenchymal cell candidate.

By using human tissue specimens and human in vitro model systems to explore this cell population further, we provide evidence that RSPO2⁺ mesenchymal cells signal to bud tip progenitors through LGR5 to maintain a high WNT signaling zone in the bud tips. We show that this RSPO2/LGR5-mediated WNT signaling niche provides support for the bud tips to maintain their progenitor state and give rise to both alveolar and airway cell types. The process of RSPO ligands signaling through LGR receptors to maintain high WNT signaling in progenitor and stem cells has been described in other organs and tissues, such as in the intestinal crypt, skin, hair follicle, and mammary tissue (Barker et al., 2007; Jaks et al., 2008; Trejo et al., 2017; Yan et al., 2017; Dame et al., 2018; Baulies, Angelis and Li, 2020; Holloway, Czerwinski, et al., 2020). However, an alternative mechanism where RSPO2 acts independently of LGRs, and instead through RNF43/ZNRF3, has been described in the context of limb development (Szenker-Ravi et al., 2018). It is possible that other receptors may act independent of or synergistically with LGR5 to promote WNT signaling in the bud tips, and further genetic gain- and loss-of-function experiments would be required to test this.

Our data suggests a potential differentiation bias towards airway secretory cells when the RSPO2/LGR5-mediated WNT signaling axis is disrupted in bud tips. However, previous data using cultured bud tip progenitors shows that removal of the WNT component of the media causes an upregulation of secretory cell markers, suggesting secretory cells may just be a default differentiation state (Miller et al., 2018). Moreover, in mice, multiciliated, goblet, and neuroendocrine cells appear later in development compared to secretory and basal cell types (Rock et al., 2009; Treutlein et al., 2014; Ardini-Poleske et al., 2017; Montoro et al., 2018; Miller et al., 2020). Our short-term (4-day) cultures may not have provided enough time for non-secretory cell types to emerge.

Co-culture experiments with bud tip organoids and isolated RSPO2⁺ mesenchymal cells or SM22⁺ smooth muscle cells support the idea that RSPO2⁺ mesenchymal cells provide cues for bud tip maintenance and differentiation towards alveolar cells while smooth muscle cells provide cues for bud tip differentiation into airway cells. Additional experiments culturing these mesenchymal populations separately in 2D short and long-term suggest that RSPO2⁺ mesenchymal cells differentiate to SM22⁺ airway smooth muscle cells, and pseudotime analysis on scRNA-seq data propose this could also be true in vivo. Together, this suggests a potential model by which RSPO2⁺ cells support distal (bud tip and alveolar) cell fates and differentiate to smooth muscle to promote airway cell fates. Nevertheless, our experiments do not exclude the possibility that SM22⁺ mesenchymal cells captured in the LIFR^HI FACS-isolated population due to sorting error were able to expand over the culture period. We believe this is unlikely because RSPO2⁺ mesenchymal cells are more proliferative and were not detected in the FACS-isolated LIFR^- population.

Of particular interest for the current study, RSPO2 mutations in humans are lethal at birth, causing nearly complete lung aplasia with lung development ceasing just after the primary lung buds emerge from the trachea (Szenker-Ravi et al., 2018), supporting a critical role for RSPO2⁺ mesenchymal cells during human lung development. Although non-functional Rspo2 in mice is also lethal at birth, with mutant lungs exhibiting reduced branching, laryngeal-tracheal defects, and reduced Wnt signaling in the bud tips, mutant lungs undergo some branching and do not show nearly as severe lung aplasia seen in humans (Bell et al., 2008). There are multiple possible explanations for why humans and mice have different severities of developmental defects caused by RSPO2 mutations. First, it is possible that Rspo1, Rspo3, or Rspo4 can compensate for the loss of Rspo2 in mice, which may not exist in humans. Our scRNA-seq and FISH data for RSPO1, RSPO3, and RSPO4 in the human distal lung indicate that they are expressed in the same broad cell population as RSPO2, but at much lower levels. It would be valuable to determine the expression patterns of the other Rspo transcripts and proteins in the murine lung and determine if other RSPOs can replace the role of RSPO2 in the murine and human lungs. The advent of in vitro model systems of the developing human lung provides an excellent opportunity to explore these questions further (Conway et al., 2020).

Another possibility for the mouse/human phenotype difference is if RSPO2 is necessary for initiating branching morphogenesis during the earliest stages of human, but not mouse, lung development. In mice, deletion of Wnt2 and Wnt2b together inhibit lung progenitors from ever being specified, and deletion of Wnt2 and Wnt7b together result in defective branching and non-localized SOX9 expression (Goss et al., 2009; Miller et al., 2012). The possible necessity of RSPO2 to promote WNT signaling that may be necessary for maintaining lung progenitors, and preventing precocious airway differentiation, in the primary lung buds and/or to initiate branching morphogenesis in humans could explain why lungs in humans lacking functional RSPO2 fail to develop past the primary lung bud stage. Through infection of human fetal lung explants with an LGR5 ECD adenovirus that reduces RSPO2-mediated WNT signaling, we show that premature differentiation of bud tip progenitors into airway lineages occurs. Although we show this phenomenon after branching morphogenesis has already begun, because of RSPO2’s persistence throughout all the time points sequenced in this study, this could be the case beginning at the primary lung buds. Therefore, if premature differentiation of lung progenitors occurs at the primary lung bud stage, the lung could fail to develop further.

Limitations of the current study:

The findings in this study have also prompted additional questions. It is known that WNT signaling is an important regulator of human bud tip progenitor maintenance (Miller et al., 2018); however, for the first time, we can appreciate the much larger signaling network involved in maintaining WNT signaling in the bud tip niche. We interrogated scRNA-seq data for many WNT signaling components and found an enrichment of many components throughout multiple cell types, suggesting a complex WNT signaling environment in the developing human lung. How other cell types and signaling pathways may be integrated to control bud tip progenitor behavior is a fascinating avenue of future exploration. For example, BMP4 is known to be critical for the bud tip niche and branching morphogenesis (Morrisey and Hogan, 2010; Hines and Sun, 2014; Miller et al., 2018), but it was recently shown that RSPOs can also antagonize BMP signaling (Lee et al., 2020). How BMP is mechanistically involved in bud tip progenitor behavior will be important to investigate. Additionally, although RSPO2⁺ mesenchymal cells are localized adjacent to bud tip progenitor cells, expression of RSPO2 extends far beyond the cells that sit near bud tip progenitors. Combined with the unique expression pattern of LGR4 throughout the mesenchyme and LGR6 in airway smooth muscle cells, it would be interesting to understand the role that RSPO2⁺ cells have a role in regulating the behavior of other mesenchymal cells. The role of RSPOs may also extend beyond the distal lung and into the proximal lung.

STAR METHODS

Resource Availability

Lead Contact

Please contact Jason R. Spence at spencejr@umich.edu if you would like to request materials used in this study.

Materials Availability

This study did not generate any new reagents or tools.

Data and Code Availability

Sequencing data used in this study is deposited at EMBL-EBI ArrayExpress. Single-cell RNA sequencing of human fetal lung and human fetal lung explants: human fetal lung (ArrayExpress: E-MTAB-8221) (Miller et al., 2020), human fetal lung explants (ArrayExpress: E-MTAB-10662) (this study). Accession numbers for deposited data are also provided in the Key Resources Table. Code used to process data can be found at: https://github.com/jason-spence-lab/Hein_2021. CellProfiler pipelines used in this study will be made freely available to readers upon request. Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

KEY RESOURCES TABLE

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Rabbit anti-TAGLN (SM22)	Abcam	Cat#ab14106 AB_443021
Goat anti-SOX9	R&D Systems	Cat#AF3075 AB_2194160
Rabbit anti-SOX9	Millipore	Cat#AB5535 AB_2239761
Rabbit anti-SOX9	Santa Cruz	Cat#sc-20095 AB_661282
Mouse anti-ECAD	BD Biosciences	Cat#610181 AB_397580
Goat anti-SOX2	R&D Systems	Cat#AF2018 AB_355110
Rabbit anti-SOX2	Seven Hills Bioreagents	Cat#WRAB-1236 AB_2715498
Rabbit anti-Non-phospho (Active) β-Catenin	Cell Signaling Technology	Cat#8814 AB_11127203
Rabbit anti-TTF1	Abcam	Cat#ab76013 AB_1310784
Goat anti-TP63	R&D Systems	Cat#BAF1916 AB_2207173
Rabbit anti-CHGA	Abcam	Cat#ab15160 AB_301704
Mouse anti-RAGE	Abcam	Cat#ab54741z AB_2242462
Mouse anti-ABCA3	Seven Hills Bioreagents	Cat#WMAB-17G524 AB_577285
Mouse anti-FOXJ1	Seven Hills Bioreagents	Cat#WMAB-319 AB_451719
Rabbit anti-FLAG	Sigma	Cat#F2555 AB_796202
Rabbit anti-KI67	Thermo Fisher	Cat#RM-9106-S1 AB_149792
Rabbit anti-CCAS3	Cell Signaling	Cat#9664 AB_2070042
Rabbit anti-pro-SP-C	Seven Hills Bioreagents	Cat#WRAB-9337 AB_2335890
Mouse anti-SP-B	Seven Hills Bioreagents	Cat#WMAB-1B9 AB_451723
Rabbit anti-SCGB3A2	Abcam	Cat# ab181853
Mouse anti-HTII-280	Terrace Biotech	Cat#TB-27AHT2-280 AB_2832931
Mouse anti-alpha smooth muscle actin	Sigma	Cat#C6198 AB_476856
Mouse anti-MUC5AC	Abcam	Cat#ab79082 AB_1603327
LIFR alpha PE	R&D Systems	Cat#FAB249P AB_2136109
IgG1 PE isotype control	R&D Systems	Cat#IC002P AB_357242
CD326 (EPCAM) FITC	Miltenyi	Cat#130-113-825 AB_2726341
REA control IgG1 FITC	Miltenyi	Cat#130-113-437 AB_2733689
CD31 APC	Miltenyi	Cat#130-117-314 AB_2727917
REA control IgG1 APC	Miltenyi	Cat#130-113-438 AB_2733893
Mouse anti-CD31 647	BD Biosciences	Cat#558094 AB_397020
Goat anti-mouse IgG2a FC APC	Thermo Fisher	Cat#31983
Bacterial and virus strains
LGR5 ECD adenovirus	(Yan et al., 2017)	N/A
Control adenovirus	(Yan et al., 2017)	N/A
pLKO.1 all-species scrambled non-target shRNA lentivirus	University of Michigan Vector Core; shRNA originally from Sigma Aldrich	Cat#SHC016
pLKO.1 human LGR5 shRNA lentivirus	University of Michigan Vector Core; shRNA originally from Sigma Aldrich	Clone: TRCN0000011586

Biological samples
Human fetal lung (8.5 – 19 weeks post-conception)	University of Washington Laboratory of Developmental Biology	N/A




Chemicals, peptides, and recombinant proteins
Collagen Type I	Thermo Fisher	Cat#A10483-01
Human fibronectin	Corning	CaT#356008
B-27 Supplement	Thermo Fisher	Cat#17504044
N-2 Supplement	Thermo Fisher	Cat#17502048
1-Thioglycerol	Sigma	Cat#M1753
Red Blood Cell Lysis Buffer	Sigma	Cat#11814389001
Normal donkey serum	Sigma	Cat#D9663
Tween 20	Thermo Fisher	Cat#BP337
Trisodium citrate	Sigma	Cat#S1804
CHIR99021	APExBIO	Cat#A3011
Recombinant human FGF7	R&D Systems	Cat#251-KG
All trans retinoic acid	Sigma	Cat#R2625
L-Ascorbic acid	Sigma	Cat#A4544
Y-27632	APExBIO	Cat#A3008
Recombinant human RSPO2	R&D Systems	Cat#3266-RS
Afamin/WNT3a conditioned media	MBL International Corporation	Cat#J2-001
Dispase (for bud tip organoid establishment)	Corning	Cat#354235
Dispase (for FACS)	Thermo Fisher	Cat#17105041
Collagenase Type II	Thermo Fisher	Cat#17101015
TrypLE Express Enzyme	Thermo Fisher	Cat#12605010
Polybrene Transfection Reagent	Sigma	Cat#TR-1003-G
Puromycin dihydrochloride	Sigma	Cat#P9620
Critical commercial assays
Neural Tissue Dissociation Kit	Miltenyi	Cat#130-092-628
RNAscope Multiplex Fluorescent Reagent Kit v2	ACD Bio	Cat#323100
MagMAX-96 Total RNA Isolation Kit	Thermo Fisher	Cat#AM1830
PicoPure RNA Isolation Kit	Thermo Fisher	Cat#AM1830
SuperScript VILO cDNA Kit	Thermo Fisher	Cat#11754250
QuantiTect SYBR Green PCR Kit	Qiagen	Cat# 204145
Deposited data
Raw scRNA-seq data from in vivo human fetal lung	(Miller et al., 2020)	ArrayExpress: E-MTAB-8221
Raw scRNA-seq data from lung explants	This manuscript	ArrayExpress: E-MTAB-10662



Experimental models: cell lines
Human fetal lung explants originated from fresh tissue sourced from the University of Washington Laboratory for Developmental Biology	N/A	N/A
Bud tip organoids were derived from fresh tissue sourced from the University of Washington Laboratory for Developmental Biology	N/A	N/A
Human fetal lung mesenchymal cells were isolated from fresh tissue sourced from the University of Washington Laboratory for Developmental Biology	N/A	N/A

Experimental models: organisms/strains






Oligonucleotides
For primer sequences, see Table S2	IDT	N/A




Recombinant DNA





Software and algorithms
CellProfiler	(Lamprecht, Sabatini and Carpenter, 2007)(Jones et al., 2008; Erben et al., 2018)	https://cellprofiler.org
Prism	GraphPad	https://www.graphpad.com/scientific-software/prism/
Sony Sorter MA900 associated software	Cytosens	https://www.cytosens.com/2020/03/25/how-you-can-easily-sort-fluorescent-protein-expressing-cells/
Scanpy	(Wolf, Angerer and Theis, 2018)	https://github.com/theislab/scanpy
CellRanger	10x Genomics	https://support.10xgenomics.com/single-cell-gene-expression/software/pipelines/latest/what-is-cell-ranger
Python	Python	Python.org
Methods and code for scRNA-seq analysis	GitHub	https://github.com/jason-spence-lab/Hein_2021
Other
Nucleopore Track-Etched Membranes	Sigma	Cat#WHA110414
Matrigel	Corning	Cat#354234
RNAse Away	Thermo Fisher	Cat#700511
Histo-Clear II	National Diagnostics	Cat#HS-202
ProLong Gold	Thermo Fisher	Hs-LGR6
TSA Plus Cyanine 5	Akoya Biosciences	Cat#NEL705001KT
TSA Plus Cyanine 3	Akoya Biosciences	Cat#NEL744001KT
Hs-RSPO2-02	ACD Bio	Cat#423991
Hs-LGR5	ACD Bio	Cat#311021
Hs-AXIN2	ACD Bio	Cat#400241
Hs-PDGFRA-C2	ACD Bio	Cat#604481-C2
Hs-FGFR4-no-XMm-C2	ACD Bio	Cat#443431-C2
Hs-WNT2	ACD Bio	Cat#584071
Hs-EGR1-C2	ACD Bio	Cat#457671-C2
Hs-RSPO1	ACD Bio	Cat#446921
Hs-RSPO3-01	ACD Bio	Cat#429851
Hs-RSPO4	ACD Bio	Cat#510201
Hs-LGR4	ACD Bio	Cat#460551
Hs-LGR6	ACD Bio	Cat#410461
Hs-SCGB3A2	ACD Bio	Cat#549951
Hs-SFTPB	ACD Bio	Cat#544251
Hs-AGTR2	ACD Bio	Cat#459141
Hs-BMP4-C2	ACD Bio	Cat#45301-C2
Hs-SERPINF1-C3	ACD Bio	Cat#564391-C3

Open in a new tab

Experimental Models and Subject Details

Human Lung Tissue

Research involving human lung tissue (8.5 – 19 weeks post conception) was approved by the University of Michigan Institutional Review Board. All human lung tissue used in these experiments was normal, de-identified tissue obtained from the University of Washington Laboratory of Developmental Biology. Specimens from both male and female sexes were used. The tissue was shipped overnight in Belzer-UW Cold Storage Solution (Thermo Fisher, Cat#NC0952695) on ice, and all experiments were performed within 24 hours in Belzer solution.

Lung Explants

Three unique human tissue samples spanning 11 – 13 weeks post-conception were used. For each unique tissue, one to three explants were included for each type of analysis.

Culture Establishment

For air-liquid-interface culture, Nucleopore Track-Etched Membranes (13mm, 8µm pore, polycarbonate) (Sigma, Cat#WHA110414) placed in 24-well tissue culture plates (Thermo Fisher, Cat#12565163) were pre-coated with 20µg/cm² Collagen Type I (Thermo Fisher, Cat#A10483-01) diluted in 0.01N ice-cold acetic acid for 30 minutes on ice followed by 2 hours at 37°C. The membranes were then washed with 1X PBS directly before use. To prepare the explants, the lung was placed in a petri dish in ice-cold 1X PBS and approximately 1mm² pieces of tissue were cut from the most distal edge of the lung under a stereomicroscope using forceps and a scalpel. 500µL culture media containing Advanced DMEM/F-12 (Thermo Fisher, Cat#12634010), 100U/mL penicillin-streptomycin (Thermo Fisher, Cat#15140122), 2mM L-Glutamine (Thermo Fisher, Cat#25030081), 10mM HEPES (Corning, Cat#25060CI), 1 bottle B-27 Supplement (Thermo Fisher, Cat#17504044), 1 bottle N-2 Supplement (Thermo Fisher, Cat#17502048), and 0.4µM 1-Thioglycerol (Sigma, Cat#M1753) was added to each well in the plate underneath the membrane. One explant per membrane was placed directly on the center of the membrane. Media was changed every 2 days.

Infection with adenovirus

Immediately following placement of the explants on the Nucleopore Track-Etched Membranes, 10¹⁰ pfu of control or LGR5 ECD adenovirus previously described in Yan et al. (Yan et al., 2017) was pipetted directly on top of each explant under a stereo microscope using a p10 pipette. A maximum of 2µL adenovirus was added to each explant at a time to prevent the adenovirus from running off the explant. The explants were re-infected every 2 – 3 days. Infection of the tissue was confirmed by immunofluorescence for FLAG and the murine IgG2a Fc fragment for the LGR5 ECD adenovirus and control adenovirus respectively.

Bud Tip Organoid and Co-Cultures

Establishment of Bud Tip Organoid Lines

Human fetal lung bud tip organoids were derived as previously reported (Miller et al., 2018). In short, the lung was placed in a petri dish in ice-cold 1X PBS, and approximately 1cm² pieces of tissue were cut from the most distal edge of the lung under a stereomicroscope using forceps and a scalpel. The tissue was enzymatically digested using dispase (Corning, Cat#354235) for 30 minutes, then was placed in 100% FBS (Sigma, Cat#12103C) for 15 minutes. In DMEM/F-12 (Corning, Cat#10-092-CV) supplemented with 10% FBS and 100U/mL penicillin-streptomycin (Thermo Fisher, Cat#15140122), the tissue was vigorously pipetted with a P1000 and subsequently a P200 to dissociate the epithelium from the mesenchyme, then was washed multiple times in the media described to obtain as pure a population of epithelial bud tips as possible. Bud tips were plated in ~20µL 8mg/mL Matrigel (Corning, Cat#354234) droplets in 24-well tissue culture plates (Thermo Fisher, Cat#12565163) and fed every 3 – 4 days in previously published bud tip media and passaged by shearing through a 27-gauge needle as previously described (Miller et al. 2018). All experiments involving bud tip cultures were derived from lungs 8 – 15 weeks post-conception for experiments involving culture with RSPO and WNT ligands or 16.5 – 18 weeks post-conception for co-cultures.

Bud Tip Organoid Growth Factor Experiments

Bud tips were established for a minimum of 1 passage to ensure epithelial-only cultures before experiments began. Bud tip organoids were passaged 3 days prior to the start of the experiment. On the day of the experiment, media conditions were changed from the original, published bud tip media (Miller et al. 2018) consisting of DMEM/F-12 (Corning, Cat#10-092-CV), 100U/mL penicillin-streptomycin (Thermo Fisher, Cat#15140122), 1 bottle B-27 supplement (Thermo Fisher, Cat#17504044), 1 bottle N-2 supplement (Thermo Fisher, Cat#17502048), 0.05% BSA (Sigma, Cat#A9647) and final concentrations of 50µg/mL L-ascorbic acid (Sigma, Cat#A4544), 0.4µM 1-Thioglycerol (Sigma, Cat#M1753), 50nM all trans retinoic acid (Sigma, Cat#R2625), 10ng/mL recombinant human FGF7 (R&D Systems, Cat#251-KG), and 3µM CHIR99021 (APExBIO, Cat#A3011). Positive control bud tips were given the described media and negative control bud tips had CHIR99021 removed. Experimental conditions were given either 500ng/mL recombinant human R-Spondin 2 Protein (R&D Systems, Cat#3266-RS) and/or 1X Afamin/WNT3a conditioned media (MBL International Corporation, Cat#J2-001) in place of CHIR99021. Cultures were fed again on day 3 and collected on day 4 for RT-qPCR analysis. Three unique biological specimens with 1 technical replicate per condition each were used for RT-qPCR analysis.

LGR5 Knock-down in Bud Tip Organoids

3 wells of 20µL-size Matrigel drops of bud tip organoids (approximately 200,000 cells/well), 5 – 10 days post-harvest, were placed into a microcentrifuge tube and removed from Matrigel by pipetting with a P1000. Bud tip organoids were digested to small fragments by incubation with TrypLE (Thermo Fisher, Cat#12605010) at 37°C for 7–10 minutes, with mechanical digestion with a P1000 pipette at the end. Trypsinization was quenched with DMEM/F-12 (Corning, Cat#10-092-CV). Cells were centrifuged at 300g for 5 minutes and resuspended in approximately 10⁶ TU/mL lentivirus carrying shRNA pLKO.1 plasmid vectors expressing either a non-target scrambled sequence (Sigma, Cat#SHC016) or a sequence against human LGR5 (Sigma, #11586) with 6µM polybrene (Sigma, Cat#TR-1003-G) and 10µM Y-27632 (APExBIO, Cat#A3008). Cells were placed at 37°C for 6 hours with gentle agitation every 30 – 60 minutes. Cells were then centrifuged at 300g for 5 minutes, resuspended in 60µL Matrigel (Corning, Cat#354234), and plated in 20µL droplets. Cells were grown in bud tip organoid maintenance media previously described (Miller et al. 2018) plus 10µM Y-27632 for the first 24-hours. Cells were given 72 hours to reform organoids prior to antibiotic selection using 1µM puromycin (Sigma, Cat#P9620) for 3 days. After 3 days, bud tip organoids were switched to media containing 0.25µM puromycin for the remainder of the experiment. Bud tip organoids were passaged as described above 2 – 4 days post-antibiotic selection. 3 – 4 days after passaging, experimental media conditions were added. Conditions are as described above (positive control: 3µM CHIR, negative control: no CHIR, RSPO2 only: 500ng/mL recombinant human RSPO2 in place of CHIR). Cultures were fed again on day 3 and collected on day 4 for RT-qPCR analysis. Two unique biological specimens with 1 – 3 technical replicates per condition each were used for RT-qPCR analysis.

FACS of Mesenchymal Cells

The lung (10 – 11.5 weeks post-conception) was placed in a petri dish and approximately 1 gram was cut from the most distal edge using forceps and a scalpel. The tissue was minced as much as possible using dissecting scissors, then was placed into a 15mL conical tube containing 9mL 0.1% (w/v) filter-sterilized Collagenase Type II (Thermo Fisher, Cat#17101015) in 1X PBS and 1mL filter-sterilized 2.5 units/mL dispase (Thermo Fisher, Cat#17105041) in 1X PBS. The tube was placed at 37°C for 60 minutes with mechanical dissociation using a serological pipette every 10 minutes. After 30 minutes, 75µL DNase I was added to the tube. 5mL isolation media containing 78% RPMI 1640 (Thermo Fisher, Cat#11875093), 20% FBS (Sigma, Cat#12103C), and 100U/mL penicillin-streptomycin (Thermo Fisher, Cat#15140122) was added. Cells were passed through 100µm and 70µm filters, pre-coated with isolation media, and centrifuged at 400g for 5 minutes at 4°C. 1 – 2mL Red Blood Cell Lysis Buffer (Sigma, Cat#11814389001) and 0.5 – 1mL FACS buffer (2% BSA, 10µM Y-27632 (APExBIO, Cat#A3008), 100U/mL penicillin-streptomycin) was added to the tube, and the tube was rocked for 15 minutes at 4°C. The cells were centrifuged at 500g for 5 minutes at 4°C, washed twice in 2mL FACS buffer, re-suspended in FACS buffer, and counted. 10⁶ cells were placed into FACS tubes (Corning, Cat#352063) for all control tubes (no antibody, DAPI only, isotype controls, individual antibodies/fluorophores) and 8 × 10⁶ cells were placed into a FACS tube for cell sorting. Primary antibodies were added at room temperature (30 minutes for LIFR and corresponding isotype, 10 minutes for CD324 and CD31 and corresponding isotypes) (see Table S2 for antibody dilutions). 3mL FACS buffer was added to each tube, then tubes were centrifuged at 300g for 5 minutes at 4°C. Cells were washed twice with 3mL FACS buffer, centrifuging at 300g for 5 minutes at 4°C between washes. Cells were resuspended in FACS buffer and 0.2µg/mL DAPI was added to appropriate tubes. FACS was performed using a Sony MA900 cell sorter and accompanying software. LIFR^HI/CD324^-/CD31^- cells and LIFR^-/CD324^-/CD31^-cells were collected in 1mL isolation media. LIFR^HI cells were gated highest 30% of LIFR expression.

Bud Tip Organoid and Mesenchyme Co-cultures

Established bud tip organoids (at least 1 passage post-derivation) were placed into a microcentrifuge tube and removed from Matrigel by pipetting with a P1000. Bud tips that had not been passaged within 10 days were also passed 1x through a 27-guage needle. The bud tips were then centrifuged for ~10 seconds in a microcentrifuge and the media and Matrigel was removed under a stereomicroscope. Freshly FACS-isolated mesenchymal cells were immediately counted using a hemocytometer and enough cells were pelleted to reach approximately 150,000 mesenchymal cells per well. On ice, atrigel (Corning, Cat#354234) was added to the tubes containing bud tip organoids. For co-cultures, the bud tips in Matrigel were transferred to the tubes containing the mesenchymal cell pellets. The bud tip organoids and mesenchyme were thoroughly mixed in the Matrigel by pipetting and swirling a P200 with the tip cut off. ~20µL droplets of Matrigel with bud tip organoids +/- mesenchyme were placed into the center of wells of a 24-well tissue culture plate (Thermo Fisher, Cat#12565163). The plate was inverted and placed in an incubator at 37°C for 20 minutes. For co-culture and negative control wells, 0.5mL media consisting of DMEM/F-12 (Corning, Cat#10-092-CV), 100U/mL penicillin-streptomycin (Thermo Fisher, Cat#15140122), 1 bottle B-27 supplement (Thermo Fisher, Cat#17504044), 1 bottle N-2 supplement (Thermo Fisher, Cat#17502048), 0.05% BSA (Sigma, Cat#A9647) and final concentrations of 50µg/mL L-ascorbic acid (Sigma, Cat#A4544), 0.4µM 1-Thioglycerol (Sigma, Cat#M1753), 50nM all trans retinoic acid (Sigma, Cat#R2625), and 10ng/mL recombinant human FGF7 (R&D Systems, Cat#251-KG) were added to each well. For positive control wells, 3µM CHIR99021 (APExBIO, Cat#A3011) was added to the above media. The cultures were fed every 3 – 4 days and were cultured for a total of 10 – 11 days. Each biological replicate shown is from a unique bud tip line co-cultured with mesenchymal cells from a unique human tissue specimen, and all three experiments were performed with four technical replicates.

Mesenchyme 2D Cultures

After FACS of LIFR^HI and LIFR^- mesenchymal cells as described above, 100,000 cells/well (for 6-hour timepoint) or 10,000 cells/well (for 7-day timepoint) were added to 24-well tissue culture plates (Thermo Fisher, Cat#12565163) pre-coated with 5µg/cm² human fibronectin (Corning, Cat#356008). Cells were cultured in serum-free media consisting of 50% IMDM (Thermo Fisher Cat#12440-053), 50% F12 (Thermo Fisher, Cat#11765-054), 1% lipid mixture (Sigma, Cat#L0288), 1X insulin-tranfserrin-selenium-ethanolamine (Thermo Fisher, Cat#51500-056), and 3% BSA (Sigma, Cat#A9647). Culture media was changed every 3 days.

Method Details

scRNA-seq Tissue Processing

All tubes and pipette tips were pre-washed with 1% BSA in 1X HBSS (all HBSS in protocol is with Mg²⁺ and Ca²⁺) to prevent cell adhesion to the plastic. The tissue was placed in a petri dish in ice-cold 1X HBSS, and the tissue was minced under a stereomicroscope using scissors. For uncultured lung tissue, roughly 1cm² of the most distal portion of the lung was isolated, and for lung explants, 4 explants were collected per condition. The whole explants were used for sequencing. The minced tissue was transferred to a 15mL conical tube with the HBSS, centrifuged at 500g for 5 minutes at 10°C, and the HBSS was removed. Mix 1 from the Neural Tissue Dissociation Kit (Miltenyi, Cat#130-092-628) was added to each tube, and the tube was placed at 37°C for 15 minutes, then Mix 2 was added and the cells were incubated for 10 minutes at 37°C. The cells were agitated by harshly pipetting with a P1000. The incubation/agitation step was repeated every 10 minutes until the cells looked to be a single-cell suspension, approximately 30 minutes. The cells were then filtered through a 70µm filter, pre-coated with 1% BSA in 1X HBSS, into a 15mL conical tube. The filter was rinsed 3x with 1mL 1% BSA in 1X HBSS. The cells were centrifuged at 500g for 5 minutes at 10°C, and the supernatant was removed. 1mL Red Blood Cell Lysis Buffer (Sigma, Cat#11814389001) and 0.5mL 1% BSA in 1X HBSS was added to the tube, and the tube was rocked for 15 minutes at 4°C. The cells were centrifuged at 500g for 5 minutes at 10°C, washed twice in 2mL 1% BSA in 1X HBSS and centrifuged again. The cells were resuspended in 200µL 1% BSA in 1X HBSS, counted using a hemocytometer, centrifuged at 500g for 5 minutes at 10°C, and resuspended to reach a concentration of 1,000 cells/µL. Approximately 100,000 cells were put on ice and single-cell libraries were immediately prepared on the 10x Chromium at the University of Michigan Sequencing Core with a target of 10,000 cells.

scRNA-seq Quantification and Statistical Analysis

Overview

To visualize distinct cell populations within the single cell RNA sequencing dataset, we employed the general workflow outlined by the Scanpy Python package (Wolf, Angerer and Theis, 2018). This pipeline includes the following steps: filtering cells for quality control, log normalization of counts per cell, extraction of highly variable genes, regressing out specified variables, scaling, reducing dimensionality with principal component analysis (PCA) and uniform manifold approximation and projection (UMAP) (McInnes, Healy and Melville, 2018), and clustering by the Louvain algorithm (Blondel et al., 2008).

Sequencing data and processing FASTQ reads into gene expression matrices

All single-cell RNA sequencing was performed at the University of Michigan Advanced Genomics Core with an Illumina Novaseq 6000. The 10x Genomics Cell Ranger v3 pipeline was used to process raw Illumina base calls (BCLs) into gene expression matrices. BCL files were demultiplexed to trim adaptor sequences and unique molecular identifiers (UMIs) from reads. Each sample was then aligned to the human reference genome (hg19) to create a filtered feature bar code matrix that contains only the detectable genes for each sample.

Quality Control

To ensure quality of the data, all samples were filtered to remove cells expressing too few or too many genes (Figure 1, S1, S5 - <750, >3000; Figure 3, 4, S3, S4 - <750, >10000) with high UMI counts (Figure 1, S1, S5 - >15000; Figure 3, 4, S3, S4 - >50000), or a fraction of mitochondrial genes greater than (Figure 1, S1, S5 - >0.05, Figure 3, 4, S3, S4 - >0.1)

Normalization and Scaling

Data matrix read counts per cell were log-normalized, and highly variable genes were extracted. Using Scanpy’s simple linear regression functionality, the effects of total reads per cell and mitochondrial transcript fraction were removed. The output was then scaled by a z-transformation.

Variable Gene Selection

Highly variable genes were selected by splitting genes into 20 equal-width bins based on log normalized mean expression. Normalized variance-to-mean dispersion values were calculated for each bin. Genes with log normalized mean expression levels between 0.125 and 3 and normalized dispersion values above 0.5 were considered highly variable and extracted for downstream analysis.

Batch Correction

We have noticed batch effects when clustering data due to technical artifacts such as timing of data acquisition or differences in dissociation protocol. To mitigate these effects, we used the Python package BBKNN (batch balanced k nearest neighbors) (Polański et al., 2019). BBKNN was selected over other batch correction algorithms due to its compatibility with Scanpy and optimal scaling with large datasets. This tool was used in place of Scanpy’s nearest neighbor embedding functionality. BBKNN uses a modified procedure to the k-nearest neighbors’ algorithm by first splitting the dataset into batches defined by technical artifacts. For each cell, the nearest neighbors are then computed independently per batch rather than finding the nearest neighbors for each cell in the entire dataset. This helps to form connections between similar cells in different batches without altering the PCA space. After completion of batch correction, cell clustering should no longer be driven by technical artifacts.

Dimension Reduction and Clustering

Principal component analysis (PCA) was conducted on the filtered expression matrix followed. Using the top principal components (Figure 1, S3, S5 – 20; Figure 3, 4, S4 – 15; Figure S1 – 30), a neighborhood graph was calculated for the nearest neighbors (Figure 1, S3, S5 – 30; Figure 3, 4, S4 – 20; Figure S1 – 50). BBKNN was implemented when necessary and calculated using the top 50 principal components with 3 neighbors per batch. The UMAP algorithm was then applied for visualization on 2 dimensions. Using the Louvain algorithm, clusters were identified at set resolutions (Figure 1, S5 – 0.55; Figure 3, 4, S3, S4 – 0.4, Figure S1 – 0.225). 43,079 cells were included in Figure S1, 10,751 cells in Figure S3A – C, and 14,900 cells in Figure S3D – F.

Sub-clustering

After annotating clusters within the UMAP embedding, specific clusters of interest were identified for further sub-clustering and analysis. The corresponding cells were extracted from the original filtered but unnormalized data matrix to include 30,841 cells in Figure 1 and S5, 2,279 cells for Figure 3A – C, and 859 cells for Figure 3D – G, 4, and S4. The extracted cell matrix then underwent log-normalization, variable gene extraction, linear regression, z-transformation, and dimension reduction to obtain a 2-dimensional UMAP embedding for visualization.

Cluster Annotation

Using canonically expressed gene markers, each cluster’s general cell identity was manually annotated. The list of genes can be found in Figure S1B.

Diffusion Pseudotime

Diffusion pseudotime analysis was used to predict temporal order of cells (Figure S5F) (Haghverdi et al., 2016). Using the Scanpy implementation, a diffusion map with 20 components was calculated using log-normalized data. Pseudotime analysis was then performed with cluster 0 set as the root cell type.

Cell Scoring

Cells were scored based on expression of a set of 22 – 50 marker genes per cell type. Gene lists were compiled based on the previously-published top 22 – 50 most differentially expressed genes from in vivo cells of the cell type of interest (Miller et al., 2020). See Table S1 for gene lists. After obtaining the log-normalized and scaled expression values for the data set, scores for each cell were calculated as the average z-score within each set of selected genes.

Tissue Processing, Staining, Quantification

All fluorescent images were taken using a NIKON A1 confocal microscope, an Olympus IX83 fluorescence microscope, or an Olympus IX71 fluorescence inverted microscope and were assembled using Photoshop CC 2021. Imaging parameters were kept consistent for images in the same experiment and post-image processing was performed equally on all images in the same experiment.

Tissue Processing

Whole tissue and organoids (removed from Matrigel by gentle pipetting) were immediately fixed in 10% Neutral Buffered Formalin for 24 hours at room temperature on a rocker, washed 3x, for 15 minutes each, with UltraPure DNase/RNase-Free Distilled Water (Thermo Fisher, Cat#10977015), and dehydrated for 1 hour in each of the following alcohol series diluted in UltraPure DNase/RNase-Free Distilled Water: 25% MeOH, 50% MeOH, 75% MeOH, 100% MeOH, 100% EtOH, 70% EtOH. Tissue was processed into paraffin blocks in an automated tissue processor with 1-hour solution changes in the following series: 70% EtOH, 80% EtOH, 95% EtOH x2, 100% EtOH x3, Xylene x3, Paraffin x3. Tissue was then embedded into paraffin wax blocks. For FISH, all equipment was sprayed with RNase AWAY (Thermo Fisher, Cat#700511) prior to sectioning. Paraffin blocks were sectioned into 4µm-thick sections for FISH (no longer than one week prior to performing FISH) or 4 – 7µm-thick sections for IF onto charged glass slides. Slides were baked for 1 hour in a 60°C dry oven (within 24 hours of performing FISH). Slides were stored at room temperature in a slide box containing a silicone desiccator packet and with the seams sealed with parafilm.

IF Protein Staining

Tissue slides were rehydrated in Histo-Clear II (National Diagnostics, Cat#HS-202) 2x for 5 minutes each, then put through the following solutions for 2x for 2 minutes each: 100% EtOH, 95% EtOH, 70% EtOH, 30% EtOH. Then, slides were put in double-distilled water (ddH20) 2x for 5 minutes each. Antigen retrieval was performed by steaming slides in 1X Sodium Citrate Buffer (100mM trisodium citrate (Sigma, Cat#S1804), 0.5% Tween 20 (Thermo Fisher, Cat#BP337), pH 6.0) for 20 minutes and subsequently cooling and washing quickly (moving slides up and down 5x) 2x in ddH20 and 2x in 1X PBS. Slides were incubated in a humidified chamber at room temperature for 1 hour with blocking solution (5% normal donkey serum (Sigma, Cat#D9663) in PBS with 0.1% Tween 20). Slides were then incubated in primary antibodies diluted in blocking solution in a humidified chamber at 4°C overnight. Slides were washed 3x in 1X PBS for 10 minutes each. Slides were incubated with secondary antibodies and DAPI (1µg/mL) diluted in blocking solution and placed in a humidified chamber at room temperature for 1 hour, then were washed 3x in 1X PBS for 10 minutes each. Slides were mounted in ProLong Gold (Thermo Fisher, Cat#P369300) and imaged within 2 weeks. Stained slides were stored in the dark a 4°C. All primary antibody concentrations are listed in Table S2. Secondary antibodies were raised in donkey, purchased from Jackson Immuno, and were used at a dilution of 1:500.

Mesenchymal cells in 2D tissue culture plates were immediately fixed in 4% paraformaldehyde for 10 minutes at room temperature, then washed 3x with 1X PBS for 10 minutes prior to blocking, primary antibody incubations, and secondary antibody incubations as described above. 1X PBS was added to the plates for imaging, and the plates were sealed with paraffin.

FISH

The FISH protocol was performed according to the manufacturer’s instructions (ACD Bio; RNAscope multiplex fluorescent manual protocol) with a 6-minute protease treatment and 15-minute antigen retrieval in a steamer. For IF protein co-stains, the last step of the FISH protocol (DAPI) was skipped. Instead, the slides were washed 1x in PBS followed by the IF protocol above, beginning with the blocking step.

Quantification of IF and FISH images

All FISH images for quantification were taken at 40x magnification. For IF, 20x or 40x magnification was used. Nuclear stains and punctate FISH stains were analyzed using unbiased automated signal detection and quantification using CellProfiler (Lamprecht, Sabatini and Carpenter, 2007; Jones et al., 2008; Erben et al., 2018). Punctate per image, number of cells per image, punctate associated with specific nuclear stains, and numbers of specific positive nuclear stains were quantified using CellProfiler. Statistical analysis was performed using ordinary one-way ANOVA or unpaired Welch’s one-tailed t-test using the GraphPad Prism software.

RNA Extraction and qRT-PCR

Three biological replicates as well three technical replicates from the same biological specimen were included in each analysis. mRNA was isolated using the MagMAX-96 Total RNA Isolation Kit (Thermo Fisher, Cat#AM1830) or the PicoPure RNA Isolation Kit (Thermo Fisher, Cat#KIT0204) for FACS-sorted cells and LGR5 knock-down bud tip organoids, and RNA quality and yield was measured on a Nanodrop 2000 spectrophometer just prior to cDNA synthesis. cDNA synthesis was performed using 100ng RNA from each sample and using the SuperScript VILO cDNA Kit (Thermo Fisher, Cat#11754250). qRT-PCR was performed on a Step One Plus Real-Time PCR System (Thermo Fisher, Cat#43765592R) using QuantiTect SYBR Green PCR Kit (Qiagen, Cat# 204145). Primer sequences can be found in Table S2. Expression of genes in the measurement of arbitrary units was calculated relative to GAPDH or ACTB (for bud tip organoid experiments involving exogenous RSPO2) using the following equation:

2^{(H o u s e k e e p i n g_{C t} - G e n e_{C t})} \times 10, 000

Quantification and Statistical Analysis

Graphs and statistical analysis for RT-qPCR and FISH/IF quantification were performed in GraphPad Prism software. Quantification of FISH and IF were done using CellProfiler software. See figure legends for the number of replicates used, the statistical test performed, and the p-values used to determine significance (if p-values are not reported in the figure) for each analysis.

Supplementary Material

Supplemental Figures

NIHMS1810736-supplement-Supplemental_Figures.pdf^{(47.3MB, pdf)}

Table S1

Supplementary Table 1. Gene lists for cell scoring analysis, Related to Figures 3, 4, and S4

Genes used to calculate cell type scores in Figures 3, 4, and S4. The listed genes are the previously published top 22 or 50 most differentially expressed genes in in vivo cells from the human fetal lung of the listed cell type on each tab (Miller et al., 2020).

NIHMS1810736-supplement-Table_S1.xlsx^{(22.4KB, xlsx)}

Highlights.

scRNA-seq of developing human distal lung mesenchyme identified cellular heterogeneity
RSPO2⁺ mesenchymal cells lie adjacent to LGR5⁺ epithelial bud tip progenitors
Blocking RSPO2/LGR5 in vitro reduced WNT signaling and lead to airway differentiation
RSPO2⁺ mesenchyme provides a niche for bud tips in co-cultures

ACKNOLEDGMENTS

Financial support: This project has been made possible in part by grant number CZF2019-002440 from the Chan Zuckerberg Initiative DAF, an advised fund of Silicon Valley Community Foundation, and in part by the NIH-NHLBI (R01HL119215) funding to J.R.S. R.F.C.H. was supported by a NIH Tissue Engineering and Regenerative Medicine Training Grant (NIH-NIDCR T32DE007057) and by a Ruth L. Kirschstein Predoctoral Individual National Research Service Award (NIH-NHLBI F31HL152531). A.J.M. was supported by a Ruth L. Kirschstein Predoctoral Individual National Research Service Award (NIH-NHLBI F31HL142197). E.M.H. was supported by a Ruth L. Kirschstein Predoctoral Individual National Research Service Award (NIH-NHBLI F31HL146162). T.F. was supported by a NIH Tissue Engineering and Regenerative Medicine Training Grant (NIH-NIDCR T32DE007057). A.S.C. was supported by the T32 Michigan Medical Scientist Training Program (5T32GM007863-40). I.G. and the University of Washington Laboratory of Developmental Biology was supported by NIH award number 5R24HD000836 from the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD).

We would like to thank Judy Opp and the University of Michigan Advanced Genomics core for the operation of the 10X Chromium single cell capture platform, the University of Michigan Microscopy core for providing access to confocal microscopes, the Flow Cytometry core for providing access to flow cytometers, and the Vector core for providing lentivirus production and adenovirus purification and expansion services. We would also like to thank the University of Washington Laboratory of Developmental Biology. Lastly, we would like to thank Michael Dame and the Translational Tissue Modeling Laboratory (TTML) for providing helpful suggestions regarding the use of WNT3a Afamin conditioned media as well as Lindy K. Brastrom for her careful technical editing of the manuscript.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

DECLARATION OF INTERESTS

The authors have no competing interests.

REFERENCES

Abler LL et al. (2017) ‘Human embryonic lung epithelial tips are multipotent progenitors that can be expanded in vitro as long-term self-renewing organoids’, Development American Physiological Society, 4(1), p. e26575. doi: 10.7554/eLife.26575. [DOI] [PMC free article] [PubMed] [Google Scholar]
El Agha E and Bellusci S (2014) ‘Walking along the Fibroblast Growth Factor 10 Route: A Key Pathway to Understand the Control and Regulation of Epithelial and Mesenchymal Cell-Lineage Formation during Lung Development and Repair after Injury.’, Scientifica Hindawi, 2014, p. 538379. doi: 10.1155/2014/538379. [DOI] [PMC free article] [PubMed] [Google Scholar]
Alejandre-Alcázar MA et al. (2007) ‘Temporal and spatial regulation of bone morphogenetic protein signaling in late lung development’, Developmental Dynamics John Wiley & Sons, Ltd, 236(10), pp. 2825–2835. doi: 10.1002/dvdy.21293. [DOI] [PubMed] [Google Scholar]
Ardini-Poleske ME et al. (2017) ‘LungMAP: The Molecular Atlas of Lung Development Program’, American Journal of Physiology-Lung Cellular and Molecular Physiology American Physiological Society Bethesda, MD, 313(5), pp. L733–L740. doi: 10.1152/ajplung.00139.2017. [DOI] [PMC free article] [PubMed] [Google Scholar]
Barker N et al. (2007) ‘Identification of stem cells in small intestine and colon by marker gene Lgr5’, Nature Nature Publishing Group, 449(7165), pp. 1003–1007. doi: 10.1038/nature06196. [DOI] [PubMed] [Google Scholar]
Baulies A, Angelis N and Li VSW (2020) ‘Hallmarks of intestinal stem cells’, Development (Cambridge, England) NLM (Medline). doi: 10.1242/dev.182675. [DOI] [PubMed] [Google Scholar]
Becht E et al. (2019) ‘Dimensionality reduction for visualizing single-cell data using UMAP’, Nature Biotechnology Nature Publishing Group, 37(1), pp. 38–44. doi: 10.1038/nbt.4314. [DOI] [PubMed] [Google Scholar]
Bell SM et al. (2008) ‘R-spondin 2 is required for normal laryngeal-tracheal, lung and limb morphogenesis.’, Development (Cambridge, England) The Company of Biologists Ltd, 135(6), pp. 1049–58. doi: 10.1242/dev.013359. [DOI] [PubMed] [Google Scholar]
Blondel VD et al. (2008) ‘Fast unfolding of communities in large networks’, Journal of Statistical Mechanics: Theory and Experiment, 2008(10). doi: 10.1088/1742-5468/2008/10/P10008. [DOI] [Google Scholar]
Brazovskaja A, Treutlein B and Camp JG (2019) ‘High-throughput single-cell transcriptomics on organoids’, Current Opinion in Biotechnology Elsevier Current Trends, 55, pp. 167–171. doi: 10.1016/J.COPBIO.2018.11.002. [DOI] [PubMed] [Google Scholar]
Bridges JP Lung et al. (2020) ‘Glucocorticoid Regulates Mesenchymal Cell Differentiation Required for Perinatal Lung Morphogenesis and Function’, American Journal of Physiology-Cellular and Molecular Physiology American Physiological Society; Rockville, MD: , p. ajplung.00459.2019. doi: 10.1152/ajplung.00459.2019. [DOI] [PMC free article] [PubMed] [Google Scholar]
Chen L et al. (2012) ‘Dynamic regulation of platelet-derived growth factor receptor α expression in alveolar fibroblasts during realveolarization.’, American journal of respiratory cell and molecular biology American Thoracic Society, 47(4), pp. 517–27. doi: 10.1165/rcmb.2012-0030OC. [DOI] [PMC free article] [PubMed] [Google Scholar]
Chen PH et al. (2013) ‘The structural basis of R-spondin recognition by LGR5 and RNF43’, Genes and Development Genes Dev, 27(12), pp. 1345–1350. doi: 10.1101/gad.219915.113. [DOI] [PMC free article] [PubMed] [Google Scholar]
Conway RF et al. (2020) ‘Understanding Human Lung Development through In Vitro Model Systems’, BioEssays John Wiley and Sons Inc., 42(6). doi: 10.1002/bies.202000006. [DOI] [PMC free article] [PubMed] [Google Scholar]
Dame MK et al. (2018) ‘Identification, isolation and characterization of human LGR5-positive colon adenoma cells’, Development (Cambridge) Company of Biologists Ltd, 145(6). doi: 10.1242/dev.153049. [DOI] [PMC free article] [PubMed] [Google Scholar]
Danopoulos S et al. (2018) ‘Human lung branching morphogenesis is orchestrated by the spatiotemporal distribution of ACTA2, SOX2, and SOX9’, American Journal of Physiology-Lung Cellular and Molecular Physiology American Physiological Society; Bethesda, MD: , 314(1), pp. L144–L149. doi: 10.1152/ajplung.00379.2017. [DOI] [PMC free article] [PubMed] [Google Scholar]
Danopoulos S et al. (2020) ‘Transcriptional characterisation of human lung cells identifies novel mesenchymal lineage markers’, European Respiratory Society, 55(1), p. 1900746. doi: 10.1183/13993003.00746-2019. [DOI] [PMC free article] [PubMed] [Google Scholar]
Danopoulos S, Shiosaki J and Al Alam D (2019) ‘FGF Signaling in Lung Development and Disease: Human Versus Mouse’, Frontiers in Genetics, 10, p. 170. doi: 10.3389/fgene.2019.00170. [DOI] [PMC free article] [PubMed] [Google Scholar]
Du Y et al. (2019) ‘Integration of transcriptomic and proteomic data identifies biological functions in cell populations from human infant lung’, American journal of physiology. Lung cellular and molecular physiology NLM (Medline), 317(3), pp. L347–L360. doi: 10.1152/ajplung.00475.2018. [DOI] [PMC free article] [PubMed] [Google Scholar]
Endale M et al. (2017) ‘Temporal, spatial, and phenotypical changes of PDGFRα expressing fibroblasts during late lung development’, Developmental Biology Academic Press, 425(2), pp. 161–175. doi: 10.1016/J.YDBIO.2017.03.020. [DOI] [PMC free article] [PubMed] [Google Scholar]
Erben L et al. (2018) ‘A Novel Ultrasensitive In Situ Hybridization Approach to Detect Short Sequences and Splice Variants with Cellular Resolution’, Molecular Neurobiology Humana Press Inc., 55(7), pp. 6169–6181. doi: 10.1007/s12035-017-0834-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
Goodwin K et al. (2019) ‘Smooth muscle differentiation shapes domain branches during mouse lung development.’, Development (Cambridge, England) Oxford University Press for The Company of Biologists Limited, p. dev.181172. doi: 10.1242/dev.181172. [DOI] [PMC free article] [PubMed] [Google Scholar]
Goodwin K et al. (2020) ‘Patterning the embryonic pulmonary mesenchyme’, bioRxiv Cold Spring Harbor Laboratory, p. 2020.08.20.259101. doi: 10.1101/2020.08.20.259101. [DOI] [Google Scholar]
Goss AM et al. (2009) ‘Wnt2/2b and β-catenin signaling are necessary and sufficient to specify lung progenitors in the foregut’, Developmental cell NIH Public Access, 17(2), p. 290. doi: 10.1016/J.DEVCEL.2009.06.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
Gouveia L et al. (2020) ‘Lung developmental arrest caused by PDGF-A deletion: consequences for the adult mouse lung’, American Journal of Physiology-Lung Cellular and Molecular Physiology, p. ajplung.00295.2019. doi: 10.1152/ajplung.00295.2019. [DOI] [PMC free article] [PubMed] [Google Scholar]
Green J et al. (2016) ‘Diversity of Interstitial Lung Fibroblasts Is Regulated by Platelet-Derived Growth Factor Receptor α Kinase Activity.’, American journal of respiratory cell and molecular biology American Thoracic Society, 54(4), pp. 532–45. doi: 10.1165/rcmb.2015-0095OC. [DOI] [PMC free article] [PubMed] [Google Scholar]
Guney TG et al. (2020) ‘Epithelial-stromal cell interactions and ECM mechanics drive the formation of airway-mimetic tubular morphology in lung organoids’, bioRxiv doi: 10.1101/2020.12.03.408815. [DOI] [PMC free article] [PubMed] [Google Scholar]
Guo M et al. (2019) ‘Single cell RNA analysis identifies cellular heterogeneity and adaptive responses of the lung at birth’, Nature Communications Nature Publishing Group, 10(1), p. 37. doi: 10.1038/s41467-018-07770-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
Haghverdi L et al. (2016) ‘Diffusion pseudotime robustly reconstructs lineage branching’, Nature Methods Nature Publishing Group, 13(10), pp. 845–848. doi: 10.1038/nmeth.3971. [DOI] [PubMed] [Google Scholar]
Han L et al. (2019) ‘Single cell transcriptomics reveals a signaling roadmap coordinating endoderm and mesoderm lineage diversification during foregut organogenesis’, bioRxiv Cold Spring Harbor Laboratory, p. 756825. doi: 10.1101/756825. [DOI] [Google Scholar]
Hines EA and Sun X (2014) ‘Tissue Crosstalk in Lung Development’, Journal of Cellular Biochemistry John Wiley & Sons, Ltd, 115(9), pp. 1469–1477. doi: 10.1002/jcb.24811. [DOI] [PMC free article] [PubMed] [Google Scholar]
Holloway EM, Wu JH, et al. (2020) ‘Differentiation of Human Intestinal Organoids with Endogenous Vascular Endothelial Cells’, Developmental Cell Cell Press, 54(4), pp. 516–528.e7. doi: 10.1016/j.devcel.2020.07.023. [DOI] [PMC free article] [PubMed] [Google Scholar]
Holloway EM, Czerwinski M, et al. (2020) ‘Mapping Development of the Human Intestinal Niche at Single-Cell Resolution’, Cell Stem Cell Cell Press. doi: 10.1016/j.stem.2020.11.008. [DOI] [PMC free article] [PubMed] [Google Scholar]
Jaks V et al. (2008) ‘Lgr5 marks cycling, yet long-lived, hair follicle stem cells’, Nature Genetics Nat Genet, 40(11), pp. 1291–1299. doi: 10.1038/ng.239. [DOI] [PubMed] [Google Scholar]
Jha R et al. (2017) ‘Downregulation of LGR5 Expression Inhibits Cardiomyocyte Differentiation and Potentiates Endothelial Differentiation from Human Pluripotent Stem Cells’, Stem cell reports Stem Cell Reports, 9(2), pp. 513–527. doi: 10.1016/J.STEMCR.2017.07.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
Jones TR et al. (2008) ‘CellProfiler Analyst: Data exploration and analysis software for complex image-based screens’, BMC Bioinformatics BioMed Central, 9(1), p. 482. doi: 10.1186/1471-2105-9-482. [DOI] [PMC free article] [PubMed] [Google Scholar]
Kishimoto K et al. (2018) ‘Synchronized mesenchymal cell polarization and differentiation shape the formation of the murine trachea and esophagus’, Nature Communications Nature Publishing Group, 9(1), pp. 1–13. doi: 10.1038/s41467-018-05189-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
Kishimoto K et al. (2019) ‘Induction of tracheal mesoderm and chondrocyte from pluripotent stem cells in mouse and human’, bioRxiv Cold Spring Harbor Laboratory, p. 758235. doi: 10.1101/758235. [DOI] [Google Scholar]
Lamprecht MR, Sabatini DM and Carpenter AE (2007) ‘CellProfiler™: Free, versatile software for automated biological image analysis’, BioTechniques Biotechniques, 42(1), pp. 71–75. doi: 10.2144/000112257. [DOI] [PubMed] [Google Scholar]
de Lau WBM, Snel B and Clevers HC (2012) ‘The R-spondin protein family.’, Genome biology BioMed Central, 13(3), p. 242. doi: 10.1186/gb-2012-13-3-242. [DOI] [PMC free article] [PubMed] [Google Scholar]
de Lau W et al. (2014) ‘The R-spondin/Lgr5/Rnf43 module: Regulator of Wnt signal strength’, Genes and Development Genes Dev, pp. 305–316. doi: 10.1101/gad.235473.113. [DOI] [PMC free article] [PubMed] [Google Scholar]
Lee H et al. (2020) ‘R-spondins are BMP receptor antagonists in Xenopus early embryonic development’, Nature Communications Nature Research, 11(1). doi: 10.1038/s41467-020-19373-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
Leeman KT et al. (2019) ‘Mesenchymal Stem Cells Increase Alveolar Differentiation in Lung Progenitor Organoid Cultures’, Scientific Reports Nature Publishing Group, 9(1), p. 6479. doi: 10.1038/s41598-019-42819-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
Li C et al. (2015) ‘Progenitors of Secondary Crest Myofibroblasts Are Developmentally Committed in Early Lung Mesoderm’, STEM CELLS, 33(3), pp. 999–1012. doi: 10.1002/stem.1911. [DOI] [PMC free article] [PubMed] [Google Scholar]
Li M et al. (2008) ‘Mesodermal Deletion of Transforming Growth Factor-β Receptor II Disrupts Lung Epithelial Morphogenesis: CROSS-TALK BETWEEN TGF-β AND SONIC HEDGEHOG PATHWAYS*’, The Journal of Biological Chemistry American Society for Biochemistry and Molecular Biology, 283(52), p. 36257. doi: 10.1074/JBC.M806786200. [DOI] [PMC free article] [PubMed] [Google Scholar]
Li R et al. (2018) ‘Pdgfra marks a cellular lineage with distinct contributions to myofibroblasts in lung maturation and injury response.’, eLife Elife, 7. doi: 10.7554/eLife.36865. [DOI] [PMC free article] [PubMed] [Google Scholar]
Li R et al. (2020) ‘Myofibroblast contraction is essential for generating and regenerating the gas-exchange surface’, Journal of Clinical Investigation doi: 10.1172/JCI132189. [DOI] [PMC free article] [PubMed] [Google Scholar]
McCulley D, Wienhold M and Sun X (2015) ‘The Pulmonary Mesenchyme Directs Lung Development’, Current opinion in genetics & development NIH Public Access, 32, p. 98. doi: 10.1016/J.GDE.2015.01.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
McInnes L, Healy J and Melville J (2018) ‘UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction’, arXiv. arXiv Available at: http://arxiv.org/abs/1802.03426 (Accessed: 13 January 2021). [Google Scholar]
McQualter JL et al. (2013) ‘TGF-β signaling in stromal cells acts upstream of FGF-10 to regulate epithelial stem cell growth in the adult lung’, Stem Cell Research Elsevier, 11(3), pp. 1222–1233. doi: 10.1016/J.SCR.2013.08.007. [DOI] [PubMed] [Google Scholar]
Mihara E et al. (2016) ‘Active and water-soluble form of lipidated wnt protein is maintained by a serum glycoprotein afamin/α-albumin’, eLife eLife Sciences Publications Ltd, 5(FEBRUARY2016). doi: 10.7554/ELIFE.11621. [DOI] [PMC free article] [PubMed] [Google Scholar]
Miller AJ et al. (2018) ‘In Vitro Induction and In Vivo Engraftment of Lung Bud Tip Progenitor Cells Derived from Human Pluripotent Stem Cells.’, Stem cell reports Elsevier, 10(1), pp. 101–119. doi: 10.1016/j.stemcr.2017.11.012. [DOI] [PMC free article] [PubMed] [Google Scholar]
Miller AJ et al. (2020) ‘In Vitro and In Vivo Development of the Human Airway at Single-Cell Resolution’, Developmental Cell Elsevier, 0(0). doi: 10.1016/j.devcel.2020.01.033. [DOI] [PMC free article] [PubMed] [Google Scholar]
Miller AJ and Spence JR (2017) ‘In vitro models to study human lung development, disease and homeostasis’, Physiology American Physiological Society, pp. 246–260. doi: 10.1152/physiol.00041.2016. [DOI] [PMC free article] [PubMed] [Google Scholar]
Miller MF et al. (2012) ‘Wnt ligands signal in a cooperative manner to promote foregut organogenesis’, Proceedings of the National Academy of Sciences of the United States of America, 109(38), pp. 15348–15353. doi: 10.1073/pnas.1201583109. [DOI] [PMC free article] [PubMed] [Google Scholar]
Montoro DT et al. (2018) ‘A revised airway epithelial hierarchy includes CFTR-expressing ionocytes’, Nature Nature Publishing Group, 560(7718), pp. 319–324. doi: 10.1038/s41586-018-0393-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
Morrisey EE and Hogan BLM (2010) ‘Preparing for the first breath: genetic and cellular mechanisms in lung development.’, Developmental cell NIH Public Access, 18(1), pp. 8–23. doi: 10.1016/j.devcel.2009.12.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
Nanki K et al. (2018) ‘Divergent Routes toward Wnt and R-spondin Niche Independency during Human Gastric Carcinogenesis’, Cell Cell Press, 174(4), pp. 856–869.e17. doi: 10.1016/J.CELL.2018.07.027. [DOI] [PubMed] [Google Scholar]
Negretti NM et al. (2021) ‘A Single Cell Atlas of Lung Development’, bioRxiv Cold Spring Harbor Laboratory, p. 2021.01.21.427641. doi: 10.1101/2021.01.21.427641. [DOI] [Google Scholar]
Niehrs C (2012) ‘The complex world of WNT receptor signalling’, Nature Reviews Molecular Cell Biology Nature Publishing Group, 13(12), pp. 767–779. doi: 10.1038/nrm3470. [DOI] [PubMed] [Google Scholar]
Nikolić MZ et al. (2017) ‘Human embryonic lung epithelial tips are multipotent progenitors that can be expanded in vitro as long-term self-renewing organoids.’, eLife eLife Sciences Publications, Ltd, 6. doi: 10.7554/eLife.26575. [DOI] [PMC free article] [PubMed] [Google Scholar]
Nikolić MZ and Rawlins EL (2017) ‘Lung Organoids and Their Use To Study Cell-Cell Interaction’, Current Pathobiology Reports Springer US, 5(2), pp. 223–231. doi: 10.1007/s40139-017-0137-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
Noe N et al. (2019) ‘Mesenchyme-specific deletion of Tgf-β1 in the embryonic lung disrupts branching morphogenesis and induces lung hypoplasia’, Laboratory Investigation doi: 10.1038/s41374-019-0256-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
Park S et al. (2018) ‘Differential activities and mechanisms of the four R-spondins in potentiating Wnt/β-catenin signaling.’, The Journal of biological chemistry American Society for Biochemistry and Molecular Biology, 293(25), pp. 9759–9769. doi: 10.1074/jbc.RA118.002743. [DOI] [PMC free article] [PubMed] [Google Scholar]
Polański K et al. (2019) ‘BBKNN: fast batch alignment of single cell transcriptomes’, Bioinformatics Edited by Berger B. Oxford University Press, 36(3), pp. 964–965. doi: 10.1093/bioinformatics/btz625. [DOI] [PMC free article] [PubMed] [Google Scholar]
Rabata A et al. (2020) ‘3D cell culture models demonstrate a role for FGF and WNT signaling in regulation of lung epithelial cell fate and morphogenesis’, bioRxiv Cold Spring Harbor Laboratory, p. 2020.04.04.024943. doi: 10.1101/2020.04.04.024943. [DOI] [PMC free article] [PubMed] [Google Scholar]
Rajagopal J et al. (2008) ‘Wnt7b stimulates embryonic lung growth by coordinately increasing the replication of epithelium and mesenchyme’, Development, 135(9), pp. 1625–1634. doi: 10.1242/dev.015495. [DOI] [PMC free article] [PubMed] [Google Scholar]
Raslan AA and Yoon JK (2019) ‘R-spondins: Multi-mode WNT signaling regulators in adult stem cells’, International Journal of Biochemistry and Cell Biology Elsevier Ltd, pp. 26–34. doi: 10.1016/j.biocel.2018.11.005. [DOI] [PubMed] [Google Scholar]
Rawlins EL (2008) ‘Lung Epithelial Progenitor Cells: Lessons from Development’, Proceedings of the American Thoracic Society, 5(6), pp. 675–681. doi: 10.1513/pats.200801-006AW. [DOI] [PMC free article] [PubMed] [Google Scholar]
Rawlins EL et al. (2009) ‘The Id2+ distal tip lung epithelium contains individual multipotent embryonic progenitor cells’, Development, 136(22), pp. 3741–3745. doi: 10.1242/dev.037317. [DOI] [PMC free article] [PubMed] [Google Scholar]
Rehan VK et al. (2006) ‘EVIDENCE FOR THE PRESENCE OF LIPOFIBROBLASTS IN HUMAN LUNG’, Experimental Lung Research Taylor & Francis, 32(8), pp. 379–393. doi: 10.1080/01902140600880257. [DOI] [PubMed] [Google Scholar]
Riccetti M et al. (2020) ‘The elephant in the lung: Integrating lineage-tracing, molecular markers, and single cell sequencing data to identify distinct fibroblast populations during lung development and regeneration’, Matrix Biology Elsevier B.V., 91–92, pp. 51–74. doi: 10.1016/j.matbio.2020.05.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
Rock JR et al. (2009) ‘Basal cells as stem cells of the mouse trachea and human airway epithelium.’, Proceedings of the National Academy of Sciences of the United States of America National Academy of Sciences, 106(31), pp. 12771–5. doi: 10.1073/pnas.0906850106. [DOI] [PMC free article] [PubMed] [Google Scholar]
Shiraishi K, Nakajima T, et al. (2019) ‘In vitro expansion of endogenous human alveolar epithelial type II cells in fibroblast-free spheroid culture’, Biochemical and Biophysical Research Communications Academic Press, 515(4), pp. 579–585. doi: 10.1016/J.BBRC.2019.05.187. [DOI] [PubMed] [Google Scholar]
Shiraishi K, Shichino S, et al. (2019) ‘Mesenchymal-Epithelial Interactome Analysis Reveals Essential Factors Required for Fibroblast-Free Alveolosphere Formation’, iScience Elsevier, 11, pp. 318–333. doi: 10.1016/J.ISCI.2018.12.022. [DOI] [PMC free article] [PubMed] [Google Scholar]
Shu W et al. (2002) ‘Wnt7b regulates mesenchymal proliferation and vascular development in the lung.’, Development (Cambridge, England), 129(20), pp. 4831–42. Available at: http://www.ncbi.nlm.nih.gov/pubmed/12361974 (Accessed: 21 July 2019). [DOI] [PubMed] [Google Scholar]
Szenker-Ravi E et al. (2018) ‘RSPO2 inhibition of RNF43 and ZNRF3 governs limb development independently of LGR4/5/6’, Nature Nature Publishing Group, p. 1. doi: 10.1038/s41586-018-0118-y. [DOI] [PubMed] [Google Scholar]
Torday JS, Torres E and Rehan VK (2003) ‘The role of fibroblast transdifferentiation in lung epithelial cell proliferation, differentiation, and repair in vitro.’, Pediatric pathology & molecular medicine, 22(3), pp. 189–207. Available at: http://www.ncbi.nlm.nih.gov/pubmed/12746170 (Accessed: 2 October 2019). [DOI] [PubMed] [Google Scholar]
Travaglini KJ et al. (2019) ‘A molecular cell atlas of the human lung from single cell RNA sequencing’, bioRxiv Cold Spring Harbor Laboratory, p. 742320. doi: 10.1101/742320. [DOI] [Google Scholar]
Trejo CL et al. (2017) ‘Lgr5 is a marker for fetal mammary stem cells, but is not essential for stem cell activity or tumorigenesis’, npj Breast Cancer Nature Publishing Group, 3(1), p. 16. doi: 10.1038/s41523-017-0018-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
Treutlein B et al. (2014) ‘Reconstructing lineage hierarchies of the distal lung epithelium using single-cell RNA-seq’, Nature Nature Publishing Group, 509(7500), pp. 371–375. doi: 10.1038/nature13173. [DOI] [PMC free article] [PubMed] [Google Scholar]
Tsao P-N et al. (2008) ‘Gamma-secretase activation of notch signaling regulates the balance of proximal and distal fates in progenitor cells of the developing lung.’, The Journal of biological chemistry American Society for Biochemistry and Molecular Biology, 283(43), pp. 29532–44. doi: 10.1074/jbc.M801565200. [DOI] [PMC free article] [PubMed] [Google Scholar]
Waas M et al. (2020) ‘SurfaceGenie: A web-based application for prioritizing cell-type-specific marker candidates’, Bioinformatics Oxford University Press, 36(11), pp. 3447–3456. doi: 10.1093/bioinformatics/btaa092. [DOI] [PMC free article] [PubMed] [Google Scholar]
Weaver M, Batts L and Hogan BLM (2003) ‘Tissue interactions pattern the mesenchyme of the embryonic mouse lung.’, Developmental biology, 258(1), pp. 169–84. Available at: http://www.ncbi.nlm.nih.gov/pubmed/12781691 (Accessed: 21 July 2019). [DOI] [PubMed] [Google Scholar]
Wolf FA, Angerer P and Theis FJ (2018) ‘SCANPY: Large-scale single-cell gene expression data analysis’, Genome Biology BioMed Central Ltd., 19(1), p. 15. doi: 10.1186/s13059-017-1382-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
Wu J et al. (2018) ‘Role of Fibroblast Growth Factor 10 in Mesenchymal Cell Differentiation During Lung Development and Disease’, Frontiers in Genetics Frontiers, 9, p. 545. doi: 10.3389/fgene.2018.00545. [DOI] [PMC free article] [PubMed] [Google Scholar]
Yan KS et al. (2017) ‘Non-equivalence of Wnt and R-spondin ligands during Lgr5+ intestinal stem-cell self-renewal.’, Nature NIH Public Access, 545(7653), pp. 238–242. doi: 10.1038/nature22313. [DOI] [PMC free article] [PubMed] [Google Scholar]
Yang Y et al. (2018) ‘Spatial-Temporal Lineage Restrictions of Embryonic p63+ Progenitors Establish Distinct Stem Cell Pools in Adult Airways’, Developmental Cell Cell Press, 44(6), pp. 752–761.e4. doi: 10.1016/J.DEVCEL.2018.03.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
Yin Y and Ornitz DM (2020) ‘FGF9 and FGF10 activate distinct signaling pathways to direct lung epithelial specification and branching.’, Science signaling American Association for the Advancement of Science, 13(621). doi: 10.1126/scisignal.aay4353. [DOI] [PMC free article] [PubMed] [Google Scholar]
Yu Q et al. (2020) ‘An organoid and multi-organ developmental cell atlas reveals multilineage fate specification in the human intestine’, bioRxiv Cold Spring Harbor Laboratory, p. 2020.07.24.219147. doi: 10.1101/2020.07.24.219147. [DOI] [Google Scholar]
Zepp JA et al. (2017) ‘Distinct Mesenchymal Lineages and Niches Promote Epithelial Self-Renewal and Myofibrogenesis in the Lung’, Cell Cell Press, 170(6), pp. 1134–1148.e10. doi: 10.1016/J.CELL.2017.07.034. [DOI] [PMC free article] [PubMed] [Google Scholar]
Zepp JA and Morrisey EE (2019) ‘Cellular crosstalk in the development and regeneration of the respiratory system’, Nature Reviews Molecular Cell Biology Nature Publishing Group, 20(9), pp. 551–566. doi: 10.1038/s41580-019-0141-3. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials