Differentiation of mouse fetal lung alveolar progenitors in serum-free organotypic cultures

Lung epithelial progenitors differentiate into alveolar type 1 (AT1) and type 2 (AT2) cells. These cells form the air-blood interface and secrete surfactant, respectively, and are essential for lung maturation and function. Current protocols to derive and culture alveolar cells do not faithfully recapitulate the architecture of the distal lung, which influences cell fate patterns in vivo. Here, we report serum-free conditions that allow for growth and differentiation of mouse distal lung epithelial progenitors. We find that Collagen I promotes the differentiation of flattened, polarized AT1 cells. Using these organoids, we performed a chemical screen to investigate WNT signaling in epithelial differentiation. We identify an association between Casein Kinase activity and maintenance of an AT2 expression signature; Casein Kinase inhibition leads to an increase in AT1/progenitor cell ratio. These organoids provide a simplified model of alveolar differentiation and constitute a scalable screening platform to identify and analyze cell differentiation mechanisms.


Introduction
Lung alveolar progenitor cells differentiate at saccular stages (E17 in mouse, 26 weeks in human) into flattened alveolar type 1 (AT1) cells, constituting the surface for gas exchange, and secretory type 2 (AT2) cells, which prevent alveolar collapse by secreting pulmonary surfactant (Morrisey and Hogan, 2010;Chao et al., 2015). Fetal distal lung progenitors are thought to be bipotent and they co-express markers of both AT1 and AT2 cells, including Podoplanin/PDPN and Advanced Glycosylation End product-specific Receptor/RAGE as well as Prosurfactant Protein C/SFTPC, respectively Treutlein et al., 2014). Differentiation of progenitors into AT1 and AT2 cells is regulated by multiple signals including WNT (Wingless and Int-1) and FGF (Fibroblast Growth Factor) from the mesenchyme (Volckaert and De Langhe, 2015;Li et al., 2018), mechanical forces (Li et al., 2018), epigenetic modifications (Wang et al., 2016), and the extracellular matrix (ECM) (Kim et al., 2018).
In particular, WNT signaling was shown to promote the expansion of Axin2+ AT2 cells during alveologenesis and prevent their conversion to the AT1 lineage . Similarly, WNT activity marks lung progenitors in homeostasis and regeneration, and WNT inhibition leads to AT1 differentiation (Nabhan et al., 2018;Zacharias et al., 2018). During mouse fetal development, β-catenin is required for the maintenance of distal lung progenitors (Ostrin et al., 2018); however, a mechanistic understanding of how WNT signaling controls AT1/AT2 cell differentiation or fate selection is currently lacking (Aros et al., 2021). A more detailed picture of how WNT/β-catenin signaling influences early alveolar differentiation could lead to better derivation protocols for alveolar cells and reveal new ways to stimulate differentiation in the fetal lung.
3D cultures of lung-or iPSC-derived alveolar progenitor cells have been used to model respiratory cell fate specification and signaling/cell interactions (reviewed by Gkatzis et al., 2018). However, these models fail to recapitulate the defined epithelial architecture of the mouse lung at canalicular and saccular stages, with HOPX + cells residing in the stalk and SOX9 + /SFTPC + cells in the tip region (Frank et al., 2019). Until recently (Vazquez-Armendariz et al., 2020), most models have displayed a spheroid or unpatterned morphology, making it difficult to assess localized differentiation from progenitors to alveolar cells. In addition, despite the central role of AT1 cells in lung development (Zepp et al., 2021), AT1 cell isolation and generation in vitro have been longstanding obstacles, and the efficiency of AT1 cell generation in vitro, or the properties of the cells obtained in these cultures, have not been analyzed in detail, limiting the relevance of these models for cell differentiation studies. These shortcomings highlight the need for an alveolar organoid model with morphological and cellular features as observed in vivo, which would accelerate the identification of differentiation-promoting factors.
To help address these questions, we developed a serum-free, rapid, and scalable distal lung progenitor organotypic culture system that recapitulates both AT1 cell differentiation and the endogenous tissue architecture of the alveolizing lung, which will facilitate mechanistic investigations of alveolar development.

Results and discussion
Mouse fetal alveolar epithelial progenitors are maintained in 3D serumfree cultures To generate alveolar organoids by self-organization of endogenous progenitors, we manually dissociated E14.5 mouse lungs and collected distal epithelial fragments. At E14.5, lung pseudoglandular development is almost complete in mouse and NKX2.1 + alveolar progenitors become specified (Frank et al., 2019), while the distal SOX9 + epithelial domain begins to expand compared with the proximal airway (Alanis et al., 2014).
We embedded the epithelial tips into diluted Matrigel/Collagen I domes and established cultures in defined medium supplemented with growth factors ( Figure 1A). Three main soluble growth factors have been implicated in distal lung growth and differentiation: FGF7, FGF10, and BMP4 (Bone Morphogenetic Protein 4) (Bellusci et al., 1997;Weaver et al., 1999;Weaver et al., 2000;Chao et al., 2016;Li et al., 2018;Danopoulos et al., 2019). We tested combinations of these factors and observed that treatment with FGF7 and FGF10 led to epithelial growth and lumenization over three days without inducing hyperproliferation of residual mesenchymal cells (Figure 1-figure supplement 1A, B). We further found that cryopreserved epithelial tips could be expanded in culture and formed organoids in FGF-supplemented medium (52% efficiency compared with 75% using freshly dissociated tissue, Figure 1B), thereby allowing for more experimental flexibility.
To promote progenitor differentiation in longer-term cultures, we developed a six-day culturing protocol with a one-time change of medium on day 3, when the concentration of supplemented growth factors was halved ( Figure 1A). This reduction slowed down organoid expansion and promoted an increase in the number and density of epithelial branches. Morphologically, on day 1, the epithelial tips reorganized to form a closed system (cyst) and began a process of lumen expansion. On day 2, a single large lumen was observed, along with newly formed epithelial branches. By day 6, many epithelial branches had extended into the surrounding matrix ( Figure 1C). Although at this stage mesenchymal cells can be observed, organoid morphogenesis did not appear to depend on their prevalence. A role for interstitial cells in local matrix remodeling and/or epithelial cell proliferation or migration, however, cannot be excluded.
To determine the location and dynamics of progenitor cells in the organoid cultures, we assessed the expression of SOX9, a transcription factor restricted to the distal progenitor population in mouse and human (Rock et al., 2009;Alanis et al., 2014;Nikolić et al., 2017;Frank et al., 2019), and which is required for lung epithelial differentiation (Rockich et al., 2013). SOX9 + cells were prevalent in isolated structures on day 0, and SOX9 expression was maintained throughout organoid lumenization until day 2. During this process, the majority of cells in the organoids were proliferating. By day 6, the tips of newly formed epithelial branches contained fewer SOX9 + cells and also fewer proliferating cells as assessed by KI67 immunostaining ( Figure 1D). These data indicate that the distal progenitor state is initially maintained in the 3D cultures.
Next, we assessed progenitor marker mRNA levels in organoids by comparing multiple culture stages with distal epithelial tips collected at E13.5, one embryonic day prior to sample collection for organoid culture. The distal epithelial markers Id2, Bmp4 and Etv5 displayed higher expression  levels on day 0, possibly reflecting the enrichment in distal characteristics or the endogenous activation of a distal developmental program between E13 and E14 (Frank et al., 2019). A marker of the proximal epithelium, Sox2, showed variable expression on day 0, suggesting that a variable number of proximal epithelial cells was isolated in the different experiments. Sox2 was downregulated by day 2, suggesting that FGF10-supplemented media selectively support distal epithelial progenitors, as described previously (Bellusci et al., 1997;Danopoulos et al., 2019). Although SOX2 expression could be identified in the organoid core throughout the six-day culture period, suggesting that proximal airway cells can be maintained in culture ( , which may explain the progressive reduction in Sox2 mRNA levels. Conversely, other markers of distal lung progenitors, such as Id2, Bmp4, and to a lesser extent Sox9, displayed upregulation by day 2, which may indicate expansion or conversion into distal progenitor cell identity. By day 6, all analyzed distal progenitor markers, as well as Sox2, were downregulated compared with E13.5 epithelial tips, suggesting that the distal progenitor cell identity is progressively lost during culture ( Figure 1D and E). Since on day 6 epithelial branches appear to be patterned in the proximo-distal axis, are morphologically elongated, and have lost most proliferative capacity, similar to distal lung epithelium at late gestation (Frank et al., 2019), we hypothesized that distal epithelial progenitors undergo differentiation in culture.

Collagen I promotes AT1 cell differentiation in culture
To assess epithelial differentiation, we performed whole-mount organoid immunostaining for SFTPC and RAGE, marking AT2 and AT1 cells, respectively. Low staining intensity was observed on days 0 and 1, indicating that within the first days of culture, the organoids are mainly composed of undifferentiated progenitors. On day 2, SFTPC/RAGE were co-expressed in a majority of the cells, similar to what is observed with alveolar bipotent progenitors Treutlein et al., 2014). By day 6, two cell populations emerged: SFTPC -/RAGE + cells in the stalk region of the elongated branches, and SFTPC + /RAGEcells in the tip region or at new sites of branching. The tip domain also comprised cells co-expressing SFTPC and RAGE, suggesting the maintenance of a bipotent progenitor (BP) population (Treutlein et al., 2014;Figure 2A). Quantification of the differentiated and progenitor cells on day 6 revealed approximately equal numbers of AT1 and BP cells in epithelial branches while AT2 cells were less prevalent ( Figure 2B). In the absence of Collagen I, the organoids displayed a reduced number of AT1 cells in favor of BP cells, indicating that a collagenous matrix (at 1:1 or 1:3 Matrigel/Collagen I ratio) promotes the differentiation of AT1 cells in vitro ( Figure 2B), consistent with the reported role of the ECM in alveolar epithelial differentiation (Kim et al., 2018). We next asked to what extent the organoids represent mature distal lung tissue, and compared differentiation marker expression with distal lung tissue at E18.5, a time when differentiated alveolar epithelial cells can be unequivocally identified. The expression level of differentiated cell markers, including Sftpc, Ager, Aqp5, and Hopx, progressively increased until day 6, at which point they reached levels comparable with those observed in E18.5 distal lung tissue ( Figure 2C). These data indicate that distal progenitors differentiate within six days of organoid culture, offering a simple model of early alveolar formation.     We further asked whether the AT1 and AT2 cells identified in the organoid cultures displayed morphological hallmarks of a differentiated respiratory epithelium. By immunostaining, we found that SFTPC -/RAGE + cells displayed exclusive basolateral localization of RAGE (Figure 2A and D) and apical localization of Podocalyxin (PODXL) ( Figure 2D), similar to observations in intact lung tissue at E19 (Yang et al., 2016). SFTPC -/RAGE + cells were also marked by uniform plasma membrane expression of Podoplanin/T1alpha (PDPN) ( Figure 2D). Immunostaining for Zonula Occludens-1 (ZO-1), a marker of tight junctions, labeled intercellular adhesions at the luminal/apical surface ( Figure 2D) as previously reported (Yang et al., 2016), suggesting that the epithelial branches form an intact cellular monolayer and display barrier function. In addition, we found that Lysosome-Associated Membrane glycoProtein 3 (LAMP3), a lysosomal marker localized to lamellar bodies in differentiated AT2 cells (Chang et al., 2013;Desai et al., 2014), was localized in apical secretory organelles in SFTPC + cells, indicating that these cells can differentiate into AT2 cells ( Figure 2D). Altogether, these data show that cultured alveolar progenitors differentiate morphologically and molecularly into AT1 and AT2 cells, suggesting that fetal mouse alveolar organoids maintain the distal airway architecture found at late gestation stages (Frank et al., 2019).

Casein Kinase modulates epithelial differentiation in lung organoids
We used fetal alveolar organoids to dissect the mechanisms by which WNT signaling controls the differentiation of AT1 and AT2 cells. To this end, we performed a chemical screen of WNT modulators from days 6 to 8 in the absence of supplemented growth factors ( Figure 3A) and determined the relative alveolar cell composition by immunostaining for SFTPC and RAGE ( Figure 3B). In control conditions, the proportion of SFTPC + /RAGE + progenitors decreased from 54% (day 6, Figure 2B) to 40% (day 8, Figure 3C), with a concomitant increase in differentiated cell numbers (more pronounced for AT1 cells), suggesting that progenitors are actively differentiating in culture at the stages chosen for screening.
Unexpectedly, the majority of tested compounds (gray highlight) did not lead to significant changes in the proportion of AT1, AT2, or progenitor cells compared with DMSO controls ( Figure 3C). Belonging to this group are first-in-class WNT pathway inhibitors such as IWR-1, IWP-2, and XAV-939 (Chen et al., 2009;Huang et al., 2009). Active WNT signaling is a feature of alveolar progenitors in homeostasis and regeneration, and blocking the WNT pathway by Ctnnb1 knockout in AT2 cells was shown to induce AT1 differentiation at postnatal stages Nabhan et al., 2018). In addition, an expansion of the AT1 cell population was observed in vitro after treatment of hiPSC-derived AT2 cells with XAV-939 (Kanagaki et al., 2021). Differentiation of cultured adult human alveolar progenitors (TM4SF1 + ) also showed WNT responsiveness, with the AT1 fate induced after XAV-939 treatment. In contrast with these data, the results of our chemical screen suggest that WNT inhibition via Axin stabilization is not sufficient to alter fate selection or cell differentiation of mouse fetal distal lung progenitors in our organoid cultures. This discrepancy with previous findings could be explained in part by differences in the phenotypic outcome of pharmacological versus genetic inhibition of WNT signaling, or by distinct regulation of the WNT pathway between pre-and postnatal stages. Limitations in our screening approach and the cellular composition of the organoid cultures could have influenced the results, as follows. First, a single concentration of 10 µM was used for all tested compounds, possibly leading to false negatives. Second, the heterogeneous cell composition of the organoids, in particular a variable number of mesenchymal cells, may have contributed to some of the variability in the data. Indeed, mesenchymal cells are fundamental players throughout lung development and engage in signaling crosstalk with epithelial cells, notably through FGF and WNT signaling (McCulley et al., 2015;Volckaert and De Langhe, 2015). Because of this variability in the relative cell composition of the organoids, four biological replicates were used for the screening experiments. Six compounds led to an increase in SFTPC + /RAGE + BP cells (yellow highlight), including  several Glycogen Synthase Kinase 3 (GSK3) inhibitors such as CHIR-99021, and a natural flavonoid, Wogonin. In line with published data (Ostrin et al., 2018;de Carvalho et al., 2019), we found that sustained WNT activity was associated with progenitor maintenance. Finally, three compounds induced a greater percentage of AT1 cells (SFTPC -/RAGE + ) at the expense of bipotent progenitors  ( Figure 3C). Mechanistically, these compounds inhibit Casein Kinase (CK) 1 (IC261) and CK2 (CX-4945, DMAT), central regulators of WNT signaling and cell cycle progression (St-Denis and Litchfield, 2009;Cruciat, 2014;Venerando et al., 2014).

Casein Kinase inhibition leads to downregulation of AT2 marker gene expression and WNT-dependent transcription
Casein Kinase (CK) acts in stem cell maintenance by phosphorylating transcriptional regulators, WNT signaling components, and cell cycle factors (St-Denis and Litchfield, 2009;Cruciat, 2014). We found that CK inhibition by IC261, CX-4945, and DMAT led to the downregulation of markers of the AT2 fate, including Sftpc and Abca3, compared with control organoids ( Figure 4A), suggesting that upon CK inhibition, progenitor cells lose AT2 fate potential and convert to AT1 cells. Similarly, WNT signaling activation was reduced in organoids treated with the same CK inhibitors for 6 h as assessed by measuring mRNA levels of Axin2 ( Figure 4B), a β-catenin-dependent WNT target gene. Pharmacological CK inhibition could also significantly reduce the transcriptional response to WNT/β-catenin signaling in HEK293T cells, as shown by SuperTOPFlash luciferase reporter assays ( Figure 4C). Collectively, these data indicate that progenitor cells convert to AT1 cells by combined loss of β-catenindependent WNT activity and AT2 marker expression, and suggest that Casein Kinases antagonize cell differentiation by modulating WNT signaling transduction and transcriptional regulation. However, since Casein Kinases mediate a plethora of housekeeping cellular functions, determining which specific CK targets are involved in alveolar differentiation requires further investigation. Recently, CK2 has emerged as a prominent target of SARS-CoV-2 infection in the lung, and its overactivation was associated with increased viral transmission by cytoskeletal remodeling and formation of cell protrusions mediating virus budding (Bouhaddou et al., 2020). Combined with our data, these recent findings suggest that SARS-CoV-2 infection could affect lung cell fate establishment and regeneration of the alveolar compartment directly by targeting CK2 activity. In summary, our fetal mouse lung alveolar organoids constitute an ideal system to assess cell differentiation at the single-cell level and will allow the investigation of this process by live imaging and chemical screening. This new model should also facilitate the identification of modulators of alveolar cell differentiation, and help discover treatments to promote lung development and maturation in premature newborns.
Media were prepared fresh just before tissue isolation.

Lung epithelial tip isolation
E14.5 C57BL/6 J mouse embryos (staged according to Theiler, 1989) were harvested and washed in ice-cold PBS. Embryos were quickly decapitated and lungs extracted and washed in fresh ice-cold PBS. Lobes were isolated by cutting them away from the proximal bronchi using microscissors (Fine Science Tools 15003-08). Using a P1000 pipette and tips coated with fresh 2% BSA (Sigma A1595), lobes from three to five lungs were transferred to freshly prepared ice-cold DM in a 60 mm petri dish. Lobes were pipetted up and down using a coated-tip P1000 pipette until fragmented. Next, if necessary, a coated-tip P200 pipette was used to triturate the tissue more finely. The degree of dissociation was monitored using a Zeiss Stemi 305 stereomicroscope. The petri dish was placed over ice-cold black metal blocks, allowing for simultaneous temperature control and contrast for visual inspection.
Fragments including one to threeepithelial tips were collected using a coated-tip P20 pipette and transferred in fresh DM on ice, to dilute out dissociated mesenchymal cells. Epithelial tips were re-collected, counted and transferred into a 1.5 ml tube on ice (about 100-120 per isolation). Tips were allowed to sink to the bottom of the tube on ice for 1-2 min and then washed twice in 100 µl DM and twice in 100 µl CM.

Organoid embedding
During epithelial tips washes, thawed Collagen I (500 µl) was neutralized by adding 57 µl 10 x DMEM (Sigma D2429) and approximately 16-17 µl 1 M NaOH (Sigma S2770). Once mixed to homogeneity, neutralized collagen was allowed to polymerize on ice for 10 min. Next, neutralized collagen was mixed to an isovolume of GFR Matrigel and kept on ice.
Washed epithelial tips were resuspended in 500 µl ice-cold CM, to which 500 µl of Matrigel/ Collagen solution was added. Epithelial tips were carefully and thoroughly resuspended on ice and plated using cooled P20 tips onto a 48-well tissue culture plate (20 µl/well) over a 39 °C heat block. Domes were allowed to solidify over the heat block for 30 min and plates were moved to a standard tissue culture incubator at 37 °C for 45 min-1 h before CM with growth factors (CM+) was added as follows.

Organoid culture
A total of 50 ml CM was warmed up to 37 °C and human recombinant FGF proteins were added (50 ng/ml FGF10,. 300 µl CM with growth factors (CM+) was added to each culture well.
At day 6, organoid medium was changed to CM without supplemented growth factors. For long-term maintenance of organoid cultures, organoid medium was changed every two days starting at day 6. Organoids could be maintained in media devoid of growth factors for up to 30 days.
Organoids with overgrown mesenchymal cells (attached to the plastic) were excluded from further analysis.

RT-qPCR
Total RNA was isolated from pooled organoids (9-15 domes) or tissue using the NucleoSpin RNA kit (Macherey-Nagel 740955.50). cDNA was synthesized from total RNA using the Superscript III Reverse Transcriptase system (Thermo Scientific 18080093). qPCR was performed using the DyNAmo Color-Flash SYBR green qPCR kit (Thermo Scientific F416XL) on a CFX Connect Real-Time System (Bio-Rad). qPCR reactions were performed in technical duplicates; data from at least three biological replicates were collected. Gene expression values were normalized to the mouse Actb gene. Heatmaps were generated using R/Bioconductor and ggplot2 packages. Mean values among technical replicates were color-coded and each biological replicate was represented.

Organoid whole-mount immunostaining
Matrix domes were quickly washed in PBS and fixed in 4% PFA for 15 min at room temperature (RT). Domes were detached by the well bottom with a flat microspatula and transferred to tubes or staining baskets. After several PBS washes, domes were permeabilized for 1 h at RT in sterile-filtered permeabilization solution: 0.5% Triton-X100, 0.5% Tween20, 3% Donkey Serum (Jackson ImmunoResearch 017-000-121), 1% BSA. Samples were incubated with primary antibodies in permeabilization solution diluted 1:1 in PBS for two days at 4 °C, with mild agitation. Domes were then washed in PBST (PBS + 0.05% Tween20) for 6 h at RT or overnight at 4 °C with mild agitation. Next, domes were incubated with secondary antibodies, Alexa488-Phalloidin (Invitrogen A12379, both at 1:500) and DAPI (1 µg/ml) in diluted permeabilization solution (1:1 in PBS) for 3 h at RT, followed by 3-4 h wash in PBST. Domes were mounted in mounting medium (Dako S3023) on glass slides using a thin vacuum grease ring as coverslip spacer. After overnight incubation at 4 °C, imaging was performed on a Leica SP8 or Zeiss LSM 800 Observer confocal microscope, ×25 and ×40 magnification.

Cryostorage
Freshly isolated E14.5 epithelial tips were mixed with 500 µl of CryoStor CS10 (Stem Cell Technologies 07959) and stored at -80 °C for up to 6 months.

Chemical screen
Organoids were plated into 48-well plates and incubated from day 6 for 48 h in the presence of compounds. Chemicals were dissolved in CM devoid of additional growth factors at a final concentration of 10 µM (from a DMSO stock). Screened compounds belong to the Stem Cell Signaling Compound Library (MCE HY-L017). Treatment with 0.1% DMSO (Sigma D2650) was used as internal control in each organoid plate. On day 8, organoids were washed and fixed in 4% PFA and three or four technical replicates were pooled for immunostaining. Confocal scans were followed by cell counting as detailed below. Four biological replicates (12 tip cultures in total) were analyzed per compound.
We thank Emma Rawlins, Saverio Bellusci and Chi-Chung Wu for discussions, and Saverio Bellusci, Chi-Chung Wu, Felix Gunawan and Simon Perathoner for suggestions and comments on the manuscript. We thank Hyun-Taek Kim, Alessandra Gentile and Till Lautenschläger for technical suggestions and help. Paolo Panza and Didier Stainier are recipients of a CPI flexible outbreak project grant. Research in the Stainier lab is supported in part by the Max Planck Society, the DFG (Sonderforschungsbereich) (SFB 834), the Leducq Foundation, and the European Research Council (AdG project: ZMOD 694455).

Additional information
Competing interests Didier YR Stainier: Senior editor, eLife. The other authors declare that no competing interests exist.

Funder
Grant reference number Author • Transparent reporting form

Data availability
All data generated or analyzed during this study are included in the manuscript and supporting files. Source data files have been provided for Figures 1, 2, 3, 4 and Figure 1-figure supplement 1.