Pluripotent stem cell differentiation reveals distinct developmental pathways regulating lung- versus thyroid-lineage specification

Abstract

The in vitro-directed differentiation of pluripotent stem cells (PSCs) through stimulation of developmental signaling pathways can generate mature somatic cell types for basic laboratory studies or regenerative therapies. However, there has been significant uncertainty regarding a method to separately derive lung versus thyroid epithelial lineages, as these two cell types each originate from Nkx2-1⁺ foregut progenitors and the minimal pathways claimed to regulate their distinct lineage specification in vivo or in vitro have varied in previous reports. Here, we employ PSCs to identify the key minimal signaling pathways (Wnt+BMP versus BMP+FGF) that regulate distinct lung- versus thyroid-lineage specification, respectively, from foregut endoderm. In contrast to most previous reports, these minimal pathways appear to be evolutionarily conserved between mice and humans, and FGF signaling, although required for thyroid specification, unexpectedly appears to be dispensable for lung specification. Once specified, distinct Nkx2-1⁺ lung or thyroid progenitor pools can now be independently derived for functional 3D culture maturation, basic developmental studies or future regenerative therapies.

Keywords: Embryo; Endoderm; Lung; Nkx2-1; Pluripotent stem cells; Thyroid.

Fig. 1.

Wnt and BMP promote lung specification of mouse ESC-derived Nkx2-1⁺ endodermal progenitors, whereas FGF and BMP signals promote thyroid specification. (A) Schematic depicting directed differentiation of mESCs into Nkx2-1⁺ endodermal cells, comparing various specification media. (B) Representative sort gates used to purify Nkx2-1^mCherry cells on day 14, showing efficiency of Nkx2-1⁺ reporter induction in each medium. (C) Day 14 Nkx2-1⁺ lung progenitor percentage and yield (per starting day 0 mESC) in each medium. (D) RT-qPCR on day 25 showing fold-change in gene expression over day 0 (2^−ΔΔCt). Data are mean±s.d. The schematic summarizes experimental design. (E) Schematic depicting proposed pathways for generation of thyroid versus lung lineages. See also Fig. S1. *P<0.05, **P<0.005 and *** or ****P<0.001 compared with B+2; one-way ANOVA with Tukey's multiple comparison test. n=3 biological replicates.

Fig. 2.

Nkx2-1⁺ cells specified with BMP4 and Wnt3a can form epithelial spheres in 3D culture that express markers of proximal and distal cells. (A) Fluorescence microscopy of day 30 monolayered sphere derived in 3D Matrigel from Nkx2-1^mCherry+ cells. Quantitation of spheres formed from day 14 Nkx2-1⁺ versus Nkx2-1⁻ sorted populations (right panel). **P<0.05; unpaired t-test. (B) Bright-field photomicrograph of paraffin-embedded section showing an epithelial sphere derived from Nkx2-1⁺ cells stained with Hematoxylin and Eosin; enlarged view shows a multiciliated cell. (C) Immunostaining of day 30 epithelial spheres derived from Nkx2-1⁺ cells specified with BMP4+Wnt3a. Scgb1a1 (club cell marker), tubulin (cilia marker) and DNA staining (DAPI) are shown. Scale bars: 25 µm. (D) Immunostaining of day 30 thyroid follicle organoids derived in 3D Matrigel from Nkx2-1⁺ cells specified with BMP4+FGF2 for Nkx2-1 and thyroglobulin (Tg) with DNA stain (DAPI). Scale bar: 25 µm. (E) RT-qPCR on day 30 showing fold-change in gene expression over day 0 (2^−ΔΔCt), comparing 3D with 2D culture effects on expression of lung epithelial and epithelial mesenchymal transition (EMT) markers. Data are mean±s.d. n=3 biological replicates. *P<0.05, **P<0.01; unpaired t-test. (F) RT-qPCR on day 30 of 2D versus 3D cultures of Nkx2-1⁺ cells specified with Wnt3a+BMP4, showing fold-change in gene expression of thyroid epithelial markers over day 0 (2^−ΔΔCt) with day 30 mESC-derived thyroid outgrowths serving as positive controls. n=3 biological replicates.

Fig. 3.

SftpcGFP lentiviral reporter identifies alveolar epithelial differentiation of sorted Nkx2-1⁺ cells specified with BMP4+Wnt3a. (A) Schematic of SftpcGFP lentivirus. SPCp, human SPC promoter element; LTR, lentiviral long terminal repeats; RRE, rev responsive element; CPPT, central polypurine tract; WPRE, Woodchuck hepatitis virus post-transcriptional regulatory element; ΔU3, deleted U3 region for in vivo inactivation of the viral LTR promoter; Ψ, Psi lentiviral packaging sequence. (B) Schematic of mESC lung differentiation protocol and timing of SftpcGFP lentiviral infection and 3D Matrigel embedding. (C) Representative micrographs on day 30 showing SftpcGFP expression after 2D or 3D expansion of Nkx2-1⁺ sorted progenitors. Note that in 3D conditions, SftpcGFP⁺ clusters are irregular (arrowheads), whereas larger circular spheres (arrows) do not express the reporter. (D) RT-qPCR showing fold-change in gene expression over day 0 (2^−ΔΔCt) in sorted SftpcGFP⁺ and SftpcGFP⁻ cells. Data are mean±s.d. n=3 biological replicates. *P<0.05, **P<0.01, ****P<0.0001; unpaired t-test. See also Fig. S2. (E) Confocal microscopy of day 30 epithelial spheres derived from Nkx2-1⁺ cells specified with BMP4+Wnt3a for pro-SFTPC immunostaining with DNA stain (Hoechst). Scale bars: 50 µm (upper panel); 25 µm (lower panel).

Fig. 4.

Global transcriptomic profiles of in vitro-derived putative lung and thyroid progenitors. (A) Schematic of experimental design of microarray analysis of global transcriptomic profiles of thyroid versus lung progenitors. (B) Principal component analysis (PCA) across all genes and samples. (C) Table summarizing numbers of differentially expressed genes using a moderated t-test in each indicated comparison (medium effect, mCherry⁺ versus mCherry⁻ status, or interaction effect). (D) Heat map representing unsupervised hierarchical clustering of samples analyzed in the microarray, based on the 1315 differentially expressed transcripts by ‘interaction effect’ of specification medium and Nkx2-1 expression (FDR<0.25 and FC>2), as detailed in the Materials and Methods. Transcripts with similar patterns of gene expression were grouped into nine clusters. (E) Heatmaps of selected transcripts from five of the clusters shown in D. Clusters 2 and 5, differentially expressed lung- or thyroid-specific transcripts; cluster 6, transcripts expressed in both ‘lung’ and ‘thyroid’ conditions; clusters 1 and 3, selected mesenchymal or non-lung, non-thyroid endodermal transcripts differentially expressed in Nkx2-1^mCherry− cells. Table attached to heatmap indicates genes expressed in vivo in the lung or thyroid in human adults or mouse embryos per GTEx Portal (http://www.gtexportal.org/) and GenePaint (www.genepaint.org/).

Fig. 5.

Validation of microarray findings by RT-qPCR of key epithelial, early thyroid- or lung-specific markers. Schematic depicting experimental design. Data are mean±s.d. of fold-change in gene expression over day 0 (2^-ΔΔCt). n=3 biological replicates. *P<0.05, **P<0.01, ****P<0.0001; two-way ANOVA.

Fig. 6.

Conserved pathways induce lung and thyroid cell fate in the developing mouse embryo. (A) Co-staining for Nkx2-1 and Foxa2 in mouse embryos at 6-8 ss and 12 ss. White arrows indicate forebrain (B) and thyroid (T) primordium. (B) Co-staining for Nkx2-1, Sox2 and Cdh1 (Ecad) in mouse foregut explants harvested at 6-10 ss and cultured for 2 days in medium supplemented with vehicle control (DMSO) or BMP signaling inhibitor DMH1. (C) Co-staining for Foxa2 and Nkx2-1 in 6-8 ss mouse foregut explants cultured for 2 days in control medium (DMSO) or media supplemented with FGF inhibitors PD173074, BGJ398 or PD161570. There are three domains of Nkx2-1 staining: brain, thyroid (T) and lung (L). (D) Co-staining for Foxa2 and Nkx2-1 in mouse embryos harvested at the presomitic stage and cultured for 2 days in control medium (DMSO) or medium supplemented with FGF inhibitors PD173074, BGJ398 or PD161570. (E) Mouse embryonic explant culture system where E7.5 anterior endoderm was isolated from a SftpcGFP transgenic mouse embryo and incubated with RA for 24 h, then with the BMP4+Wnt agonist CHIR99021 (CHIR) for 4 days in the presence of FGF chemical inhibitor BGJ398. SftpcGFP reporter expression and branching is induced after recombination with E12 embryonic mouse lung mesenchyme. (F) Mouse embryonic explant culture system in which E7.5 anterior endoderm from a SftpcGFP mouse embryo was incubated with either control medium or medium containing FGF10 and KGF with or without CHIR before recombination with E12 embryonic mouse mesenchyme. Scale bars: 150 µm in A,C,D; 200 µm in B; 500 µm in E,F. Embryos shown in B-F are representative of three repeated independent experiments consisting of a total of 10-12 embryos per condition shown.

Fig. 7.

Conserved pathways induce lung cell fate in human ESC/iPSC-derived endoderm. (A) Schematic of directed differentiation protocol for human ESCs or iPSCs, comparing various specification media (day 6-15). (B) Comparison of lineage specification (NKX2-1+ percentage) induced by the five growth factor standard cocktail (CFKBRA) versus conditions with one factor removed (‘5 factors minus 1’): percentage of NKX2-1⁺ cells induced in each condition is shown as quantified by flow cytometry on day 15 using the C17 iPSC line carrying a GFP reporter targeted to the NKX2-1 locus. Data are mean±s.d. of biological triplicates. (C) Day 15 NKX2-1 induction efficiency in human iPSCs. Representative flow cytometry dot plot showing NKX2-1^GFP reporter expression on day 15, along with average yield±s.d. calculated for the BU3 iPSC line. (D) Box and whiskers plot showing the range, median and quartiles of NKX2-1 induction efficiencies (% of all cells measured by FACS) for each indicated human ESC or iPSC clone after induction with either CFKBRA (blue triangles) or CBRA (black dots). Data represent day 13-16 analyses accumulated over a ∼1 year period of experiments. (E) Violin plots of normalized gene expression measured for each indicated gene by single-cell RNA-Seq of 153 cells on day 15 of differentiation in CFKBRA. See also Fig. S4. (F) Confocal microscopy on day 33 showing NKX2-1 nuclear protein immunostaining of candidate lung progenitors derived with CBRA on days 6-15 followed by 3D culture outgrowth in conditions shown in A, and transmission electron microscopy (TEM) on day 30, indicating lamellar bodies (arrows). Scale bar: 25 µm. (G) Immunofluorescence microscopy of SFTPB protein expression in spherical epithelial cells derived in 3D Matrigel after specification in CBRA followed by differentiation in 3D Matrigel, according to the protocol in A. Scale bars: 25 µm. See also Fig. S3. (H) RT-qPCR showing fold-change in gene expression over day 0 (2^−ΔΔCt) of each indicated lung or thyroid marker gene in progenitors (day 16) or maturing cells (day 36) derived from BU3 iPSCs using ‘lung’ or ‘thyroid’ differentiation media, compared with control human fetal lung epithelium (hLung; 21 weeks gestation) or adult human thyroid tissue (hThyroid). Data are mean±s.d. (I) Experimental design for testing the effect of CHIR with/without chemical inhibition of β-catenin co-activator function (canonical Wnt signaling) on human lung progenitor specification. Inhibitors IQ1 or ICG001 were added to CBRA to block β-catenin interactions with either p300 or CBP transcription factors, respectively. (J) Day 14 NKX2-1 induction efficiency in RUES2 cells exposed to the conditions shown in I from days 6-14. Intracellular staining for NKX2-1 protein was quantified by FACS. Data are mean percentage of NKX2-1⁺ cells±s.d. in biological triplicates. ****P<0.05; one-way ANOVA. (K) Schematic of the mechanism of action of IQ1 and ICG001 chemical inhibitors on canonical Wnt signaling.