Expandable Lung Epithelium Differentiated from Human Embryonic Stem Cells

Abstract

Background

The progenitors to lung airway epithelium that are capable of long-term propagation may represent an attractive source of cells for cell-based therapies, disease modeling, toxicity testing, and others. Principally, there are two main options for obtaining lung epithelial progenitors: (i) direct isolation of endogenous progenitors from human lungs and (ii) in vitro differentiation from some other cell type. The prime candidates for the second approach are pluripotent stem cells, which may provide autologous and/or allogeneic cell resource in clinically relevant quality and quantity.

Methods

By exploiting the differentiation potential of human embryonic stem cells (hESC), here we derived expandable lung epithelium (ELEP) and established culture conditions for their long-term propagation (more than 6 months) in a monolayer culture without a need of 3D culture conditions and/or cell sorting steps, which minimizes potential variability of the outcome.

Results

These hESC-derived ELEP express NK2 Homeobox 1 (NKX2.1), a marker of early lung epithelial lineage, display properties of cells in early stages of surfactant production and are able to differentiate to cells exhibitting molecular and morphological characteristics of both respiratory epithelium of airway and alveolar regions.

Conclusion

Expandable lung epithelium thus offer a stable, convenient, easily scalable and high-yielding cell source for applications in biomedicine.

Supplementary Information

The online version contains supplementary material available at 10.1007/s13770-022-00458-0.

Introduction

The main functions of progenitor cells in the adult airway epithelium are maintenance of homeostasis and reparation of defects in the airway wall. Under normal healthy conditions, the lung epithelium undergoes a constant renewal with complete turnover occurring every 30 to 50 days [1]. However, in many lung diseases such as asthma, chronic obstructive pulmonary disease, obliterative bronchiolitis, and cystic fibrosis, the reparation capacity provided by the endogenous lung epithelial progenitor cells is often insufficient, so that the strategies and tools to tackle this ever growing medical problem are greatly needed [2, 3]. One of the promising cellular instruments can be expandable lung progenitor cells produced in vitro by differentiation of human pluripotent stem cells (hPSC), either embryonic (hESC) or induced (hiPSC), by recapitulating signaling events occurring during normal embryogenesis.

Overall, directed differentiation of hPSC into lung epithelial cell lineage involves specification to required cell fates through precisely dosed and timed activation or inhibition of signaling pathways. The majority of protocols for hESC differentiation into lung epithelium begins with derivation of the definitive endoderm (DE) by Activin A-induced Activin/Nodal signaling [4–9]. For the subsequent anteriorization of definitive endoderm to foregut endoderm (FE), both BMP4 and Activin/Nodal signaling have to be abolished, which is usually carried out by respective inhibitors [5, 10–13].

In order to specify lung lineage from foregut endoderm, several studies have shown an importance of treatment with fibroblast growth factor 7 (FGF7) and FGF10 [5, 14–16]. Others highlight the influence of BMPs normally present within the serum [11, 17–19]. Additionally, previous studies indicate that also canonic activation of WNT signaling has an important role [10, 18, 20, 21]. Others take advantage of FGF2, as a mimic of the developing cardiac mesoderm influence, inducing in the neighboring area the lung epithelial phenotype [19, 22, 23]. In addition, retinoic acid, dexamethasone, and isobutylmethylxanthine have been used to accelerate the maturation of respiratory epithelium progenitors into type II pneumocytes [16, 18, 24].

However, there are also other alternative approaches to derive lung epithelial cells. Protocols generating embryoid bodies (EB) and using commercial culture media like Small Airway Epithelial Cell Growth Medium™ (SAGM™) and Bronchial Epithelial Cell Growth Medium™ (BEGM™), respectively, are being employed [25–28]. In addition, inductive effect of exposure to air for differentiation is used during culture at an air–liquid interface (A-Li) [13, 29, 30].

During the process of human pluripotent stem cell differentiation, human foregut stem cells were derived and those were maintained in vitro while retaining their capacity to differentiate into pancreatic and hepatic cells [31]. Mou and colleagues reported differentiation into NKX2.1 + SOX2 + proximal airway progenitors [11]. From both ESC and iPSC the pneumocytes type II, progenitors of alveolar compartment, have been derived [23, 32]. However, these lung progenitors were used instantly after derivation for their respective applications, without evaluation of their long-term culture potential and maintenance of their properties in vitro. Only few teams so far have been focusing on purifying lung progenitors in various steps of differentiation using fluorescence-activated cell sorting. They found and used surface proteins that correspond to intracellular markers (such as carboxypeptidase M, corresponding to NKX2.1 positivity), or used green fluorescent protein (GFP)-tagged markers [10, 16, 33]. Yet again, in these studies the progenitors were immediately after their isolation applied to subsequent experiments, such as transplantation into injured mice [34, 35].

Yet, methodologies allowing for maintenance of the progenitors in culture for longer periods of time so that they can be used repeatedly without the need to repeat the differentiation protocol are superior for laboratory practice. They would ensure better reproducibility and fast scalability – both much needed for biomedical applications. Recently, several studies reported maintaining lung epithelial cells via formation of organoids—either derived from hESC or harbored from human and/or mouse embryonic lungs [15, 36–41]. But organoid cultures are complex, suffer from higher heterogeneity and lower global standardization of protocols, and thus suboptimal reproducibility [42]. On the other hand, standard monolayer cultures are rather simple, easy to work with and scalable. Yet, to date, there has been no evidence of establishing lung progenitors and maintaining them in standard monolayer culture for a long period of time in a near-homogeneous population.

In this study we derived expandable lung epithelium (ELEP) through differentiation of hESC towards endodermal lineage. Our procedure does not rely on cell sorting, cell enrichment, or generation of 3D organoids. Instead, we generated and stabilized ELEP simply in a monolayer culture. Moreover, our culture protocol enables a long-term cultivation of ELEP for more than 65 passages (corresponding to more than 6 months in culture) while maintaining their capacity to differentiate to mature lung epithelial cells.

Materials and methods

Maintenance of cells

CCTL14 line of hESC [46, XX] was used in this study (for detailed characterization see Human Pluripotent Stem Cell Registry [43]). hESC were routinely maintained in culture with irradiated (50 Gy) mouse embryonic fibroblasts (mEF, derived from 12.5-day-old embryos of CD-1 mice) in culture medium consisting of Dulbecco's modified Eagle medium (DMEM)/F12 supplemented with 15% knockout serum replacement (KSR), 2 mM L-glutamine, 1 × MEM nonessential amino acids, 100 U/ml penicillin, 100 μg/ml streptomycin (all from Invitrogen, Thermo Fisher Scientific), 0.1 mM β-mercaptoethanol (Merck, Prague, Czech Republic) and 10 ng/ml FGF2 (PeproTech, Hamburg, Germany). For feeder-free culture, hESC were cultured on hESC-qualified Matrigel matrix (354,277, Corning, Wiesbaden, Germany) coated dishes in mEF-conditioned medium. For production of mEF-conditioned medium, the KSR-containing medium was collected from the irradiated mEF culture after 24 h, and supplemented with 10 ng/ml FGF2.

H441 human lung papillary adenocarcinoma cells (NCI-H441 ATCC® HTB-174™, American Type Culture Collection, Manassas, VA, USA were cultured in RPMI-1640 medium supplemented with 10% fetal bovine serum (FBS), 2 mM L-glutamine and 100 U/ml penicillin, 100 µg/ml streptomycin (all from Invitrogen, Thermo Fisher Scientific). Cells were passaged with trypsin (Invitrogen, Thermo Fisher Scientific) at 70% confluence.

HepG2 human hepatocellular carcinoma cells (Hep G2 HB-8065™, American Type Culture Collection) were cultured in DMEM supplemented with 10% FBS and 1% penicillin/streptomycin.

Human skin samples were obtained from healthy donors undergoing surgical procedures at the Department of Plastic and Aesthetic Surgery, St. Anne’s University Hospital Brno. All donors were informed of the study goals and procedures and signed informed consent prior to their participation. Collection of skin samples for dermal fibroblast isolation was approved by the Ethics Committee St. Anne’s University Hospital Brno (8 V/2020). Briefly, the skin samples were placed on a 10 cm Petri dish in DMEM/F-12 medium supplemented with 10% fetal bovine serum (FBS), 2 mM L-glutamine and 100 U/ml penicillin, 100 µg/ml streptomycin (all from Invitrogen, Thermo Fisher Scientific) and kept in an incubator at 37 °C for a week. Expanding human dermal fibroblasts (HDF) were then passaged enzymatically using TrypLE™ Express (Gibco, Thermo Fisher Scientific, Prague, Czech Republic) onto a new dish, and the primary culture of HDF was prepared and maintained.

Human lung tissue samples were obtained from therapeutical lung surgery based on the written informed consent by the patient and approval of Ethics Committee of the University Hospital Brno (28–170,621/EK).

Derivation and cultivation of ELEP

hESC grown in a feeder-free culture were dissociated with trypsin into a single-cell suspension and plated on dishes coated with vitronectin (Nucleus Biologics, San Diego, CA, USA) at density of 10⁴ cells/cm². The cells were cultured in DMEM/F12 with 2 mM L-glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin and 10% FBS. After 24 h, the cells were exposed to a new medium containing 50 ng/ml activin A (PeproTech), similarly to Serra et al. [44], for 5 days. On day 5 the cells were passaged with trypsin and seeded at density of 10⁴ cells/cm² on dishes coated with vitronectin. On day 6 the medium was changed to serum-free DMEM/F12 with 2 mM L-glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin, and 1 × insulin/transferrin/selenium (ITS, Gibco, Thermo Fisher Scientific). Differentiating cells were collected on day 8 and seeded on vitronectin-coated dishes at density 5 × 10³ cells/cm² in DMEM/F12 with 2% FBS, 2 mM L-glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin, 1 × ITS, 5 µg/ml heparin (Merck), 10 ng/ml FGF2 and 20 ng/ml epidermal growth factor (EGF) (Peprotech) to differentiate to ELEP. ELEP were then cultured using the above mentioned medium and repeatedly passaged at 70% confluence. Culture medium was changed every 2 days. ELEP were also cryopreserved using the above mentioned medium supplemented with 20% FBS and 10% dimethyl sulfoxide. The passages 0–30 have been designated as low passages (ELEP low) and passages 31–65 as high passages (ELEP high). Doubling time of ELEP was calculated with Roth V. 2006 Doubling Time Computing, available from: http://www.doubling-time.com/compute.php.

Differentiation of ELEP towards terminal phenotypes in monolayer

ELEP were seeded on vitronectin-coated dishes at density of 10⁴ cells/cm² and let grow without passaging in differentiation medium: DMEM/F12 with 10% FBS, 2 mM L-glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin, 10 ng/ml FGF10, 20 ng/ml EGF, and 10 ng/ml FGF7 (all from PeproTech). Culture medium was changed every 2 days.

Differentiation of ELEP towards terminal phenotypes in 3D

The suspension of ELEP in differentiation medium at density 25 × 10⁴ cells/ml was mixed at the ratio 1:1 (50 µl: 50 µl) with Matrigel Matrix Growth Factor Reduced (356,231, Corning) and transferred on a drop of concentrated Matrigel into wells. The gel-embedded cells were overlayered with differentiation media as defined above. Culture medium was changed every 2 days. After 25 days of differentiation, primary 3D organoids were processed for histological analysis. For the secondary organoid formation, the primary 3D organoids were dissociated with Accumax (A7089, Sigma- Aldrich GmbH, Steinheim, Germany) for 5–8 min at 37 °C. The cells were then mixed with Matrigel as above and cultivated in the same culture medium. After 25 days of differentiation, secondary 3D organoids were processed for histological examination.

Transplantation under kidney capsule

10⁵ ELEP were mixed with 5 μl of Matrigel Matrix Growth Factor Reduced and transplanted under the kidney capsule of SCID Hairless Outbred (Crl:SHO-Prkdc^scidHr^hr) mice at 6–12 weeks of age not selected for gender, obtained from the Charles River Laboratories (Erkrath, Germany) and bred in-house. Mice were euthanized with CO₂ 6 weeks after the transplantation and the inoculated and control kidneys were processed for histological analysis. All European Union Animal Welfare lines (EU Directive 2010/63/EU for animal experiments) were respected. Animal experiments were approved by the Academy of Sciences of the Czech Republic (AVCR 13/2015), supervised by the local ethical committee and performed by certified individuals (JR, AR, ZK).

Histological analysis

The specimens (3D organoids and kidneys) were fixed with 4% paraformaldehyde, embedded in paraffin, serially sectioned, and stained with hematoxylin–eosin. Bright field images were taken using a Leica microscope (DM5000B, Leica Microsystem GmbH, Wetzlar, Germany) equipped with a Leica camera (DFC480). The histological sections were also used for immunofluorescence and immunohistochemistry analysis.

Immunofluorescence analysis

Cells were fixed with 4% paraformaldehyde for 30 min, permeabilized for 10 min with 0.1% Triton X-100 (Carl Roth GmbH + Co.KG, Karlsruhe, Germany), blocked for 1 h with 1% bovine serum albumin in PBS (pH 7.4), and incubated overnight at 4 °C with primary antibody. The histological sections were first deparaffinized, then permeabilized with Triton X-100 and blocked with 1% BSA and incubated with primary antibody overnight at 4 °C. Next day the cells or sections were washed with PBS (pH 7.4) and incubated for 1 h at room temperature (RT) with appropriate secondary antibody. Cell nuclei were counterstained with DAPI, and samples were mounted in Mowiol® (4–88, Mowiol-containing glycerol-based medium pH 8.5), both from Merck.

Primary antibodies used were: mouse monoclonal antibody against acetylated α tubulin (acTUB, sc-23950, 1:100), goat polyclonal antibody against aquaporin 5 (AQP5, sc-9890, 1:100), both from Santa Cruz Biotechnology (Heidelberg, Germany), rabbit polyclonal antibodies against prosurfactant protein B (pro-SPB, ab15011, 1:500, Abcam, Cambridge, UK).

Immunohistochemistry analysis

The histological sections of the outgrowth on the kidneys of mice were dewaxed in xylene, hydrated through a graded series of alcohols (96, 80, and 70%), and rinsed in deionized water. After antigen retrieval in citrate buffer (pH 6.0 and pH 9.0, respectively) at 98 °C for 30 min, the slides were rinsed in tap and deionized water and washed with 3% H₂O₂ in PBS at RT for 10 min. To block endogenous peroxidase activity, the sections were treated with 10% fetal bovine serum for 30 min. The sections were incubated with the primary antibody against: NKX2.1 (ab76013, 1:200, pH9.0, Abcam), SOX2 (ab97959, 1:2000, pH9.0, Abcam), FOXA2 (sc6554, 1:100, pH 9.0, Santa Cruz Biotechnology), prosurfactant protein C (pro-SPC, AB3786, 1:300, Millipore, Merck), cytokeratin 5 (CK5, 905,504. 1:500, pH 6.0, Biolegend, Amsterdam, The Netherlands), cytokeratin 8 (CK8, 904,806. 1:500, pH 6.0, Biolegend, Amsterdam, The Netherlands) and human nucleoli (ab190710, 1:500, pH6.0, Abcam). The slides were then washed three times in PBS and subsequently incubated with the secondary antibody (En Vision FLEX/HRP) for 20 min. After the last washing step, the slides were incubated in substrate solution (DAB, K3468, Dako, Santa Clara, CA, USA), counterstained in hematoxylin, dehydrated with alcohols and xylene, and mounted in Mowiol®.

Flow cytometry

The cells dissociated with trypsin were fixed in 4% paraformaldehyde for 30 min, permeabilized for 10 min with 0.1% Triton X-100, and incubated with primary anti- NKX2.1 polyclonal antibody (sc-13040, 1:100, Santa Cruz Biotechnology, and ab76013, 1:100, Abcam) overnight at 4 °C. For negative controls the cells were incubated with isotype control antibody (rabbit IgG, sc-2027, 1:200, Santa Cruz Biotechnology). After three washes in PBS, the cells were incubated with secondary anti-rabbit antibody AlexaFluor 488 (A11008, 1:500, Invitrogen, Thermo Fisher Scientific) for 1 h at RT. Stained cells were analyzed on a BD FACSCanto flow cytometer (Becton Dickinson, Franklin Lakes, NJ, USA laser 488). Cell aggregates and debris were excluded from the analysis based on a dual‐parameter dot plot in which the pulse ratio (signal height/y‐axis vs signal area/x‐axis) was displayed. Gates for positivity were set based on isotype controls. Acquired FCS files were exported and analyzed using BD FACSDiva Software v6.1.2.

Western blot analysis

For immunoblot analysis, cell samples were washed twice with PBS (pH 7.4) and lysed in sodium dodecyl sulphate (SDS) lysis buffer (50 mM Tris–HCl, pH 7.5; 1% SDS; 10% glycerol). Protein concentrations were determined using DC Protein assay kit (Bio-Rad, Prague, Czech Republic). Lysates were supplemented with bromphenol blue (0.01%) and β-mercaptoethanol (143 mM), and incubated for 5 min at 95 °C. Equal amounts of total protein (10 µg) were separated with SDS-PAGE and transferred onto polyvinylidene difluoride membrane (Immobilon-P, Merck) using established procedures. The membranes were blocked in TBS (20 mM Tris–HCl pH 7.2, 140 mM NaCl, 0.1% Tween 20) with 5% non-fat milk. Proteins were immunodetected using appropriate primary and secondary antibodies, and visualized by ECL-Plus reagent (Amersham Pharmacia Biotech, Piscataway, NJ, USA) according to manufacturer’s instructions. Primary antibodies were used as follows: rabbit polyclonal antibody against SOX2 (ab97959, 1:1000, Abcam), rabbit monoclonal antibody against NKX2.1 (sc-13040, 1:1000, Santa Cruz Biotechnology, and ab76013, 1:1000, Abcam), and rabbit polyclonal antibody against GAPDH (sc-25778, 1:2000, Santa Cruz Biotechnology), rabbit polyclonal antibody against phospho-Smad1/5/9 (9511, 1:1000, Cell Signaling Technology, Danvers, MA, USA), rabbit polyclonal antibody against phospho-Smad2 (18,338, 1:1000, Cell Signaling Technology), rabbit polyclonal antibody against Smad1 (6944, 1:1000, Cell Signaling Technology) and rabbit polyclonal antibody against Smad2 (5339, 1:1000, Cell Signaling Technology). Secondary antibodies were used anti-Rabbit IgG–Peroxidase antibody produced in goat (A6154, 1:5000, Merck).

qRT-PCR analysis

Total RNA was extracted using RNeasy Mini Kit (QIAGEN, Hilden, Germany). Human brain RNA sample was obtained commercially (636,530, Takara, Japan). Complementary DNA was synthesized according to the manufacturer's instructions using High-Capacity RNA-to-cDNA™ Kit (Applied Biosystems™, Thermo Fisher Scientific, Prague, Czech Republic). qRT-PCR was analyzed with LightCycler® 480 (Roche, Prague, Czech Republic) using the following program: initial activation step at 95 °C for 5 min, followed by 40 cycles at 95 °C for 10 s, annealing temperature for 10 s, and 72 °C for 10 s. Gene expression values for each sample were expressed in terms of the threshold cycle normalized to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) expression (ΔΔC_t method). Sequences of primers and annealing temperatures are in Table 1. For analysis of FOXG1 expression QuantiTect Primer Assays (QIAGEN) was used.

Table 1.

Sequences of primers used in quantitative RT-PCR

Gene	Oligonucleotide	Ta
AQP5	GCCACCTTGTCGGAATCTACT	61.5 °C
	GGCTCATACGTGCCTTTGATG
CDX2	GTCCCTCGGCAGCCAAGTGAAAAC	60.3 °C
	TCCGGATGGTGATGTAGCGACTGTAG
FOXA2	GCCCGAGGGCTACTCCTCC	62.1 °C
	GCCCACGTACGACGACAT
FOXJ1	CACCTGGCAGAATTCAATCCG	60.9 °C
	CTTGAAAGCGCCGCTCAGT
PAX8	CCGAGGTGTCCAGTTCTAGC	62.6 °C
	GAGTATTCACTTCCTGCCACCAT
PDX1	AGTGGGCAGGCGGCGCCTAC	64.8 °C
	GGCGCGGCCGTGAGATGTAC
SFTPB	CCCCATTCCTCTCCCCTAT	58.3 °C
	GCGCACCCTTGGGAATCA
NKX2.1	GCTTCCCCGCCATCTCCC	65.0 °C
	GCCATGTTCTTGCTCACGTC
GAPDH	TCTGCTCCTCCTGTTCGACA	59.9 °C
	CCCAATACGACCAAATCCGT

Transmission electron microscopy (TEM)

Cells were harvested and fixed with 3% glutaraldehyde in 0.1 M cacodylate buffer for 2 h. After three washing steps with cacodylate buffer, samples were post-fixed with 1% OsO₄ solution in the same buffer for 1 h. After post-fixation, cells were washed three times and embedded into small blocks of 1% agar. Agar blocks were dehydrated in increasing concentrations of ethanol (50, 70, 96, and 100%), treated with 100% acetone, and embedded in epoxy resin Durcupan. All reagents were from Merck Ultrathin sections were prepared on an ultramicrotome (Leica EM UC6, LKB 8802A, Leica Microsystems), contrasted with uranyl acetate and lead citrate solutions, and examined with Morgagni 268 TEM (FEI, Hillsboro, OR, USA) and images were taken with a Veleta CCD Camera (Olympus, Prague, Czech Republic).

Telomerase activity assay

Cells were collected at ~ 80% confluence and extracted with CHAPS lysis buffer from the TRAPeze®XL Telomerase Detection Kit (Merck) using the manufacturer’s recommendations. Telomerase activity was determined by real-time RT-PCR. Relative TRAP quantification was performed using internal control templates. The internal control template TSK2 was amplified by the K2 and TS primers from the kit to generate a 56-bp product. Some of the telomerase activity assays were carried out using primers TS and NT to amplify TSNT oligonucleotide as an internal control template [45]. The products were then resolved on 10% polyacrylamide gels, followed by SYBR Green staining (Roche) (as described in [45]). Protein concentrations of cell lysates were determined using Bio-Rad protein micro assay with a 96-wells microplate reader (Sunrise™, Tecan, Männedorf, Switzerland) and BSA as a standard.

Intact cell mass spectrometry (MS) and analysis of mass spectra

Cells were enzymatically detached from culture dish and disaggregated using TryPLE Express (1x) for 2 min at 37 °C, washed three times in 1 × PBS and counted. To eliminate traces of PBS, the cells were centrifuged (200 g, 2 min) and washed in 1 ml of 150 mM ammonium bicarbonate buffer. Finally, the cells were resuspended in 150 mM ammonium bicarbonate buffer. The cell suspension was then mixed with an acidified matrix containing sinapinic acid (30 mg/ml) and 2,2,2-trifluoroacetic acid (7.5% vol/vol) in a 2:1 sample to matrix ratio. 25 × 10³ cells were applied in technical pentaplicates onto the target plate, which was purified in an ultrasonic bath before spotting the samples.

The mass spectra were recorded in linear positive ion mode over the 2,000–20,000 m/z range using 7090 mass spectrometer from Shimadzu Biotech (Kratos, UK) equipped with a solid-state UV laser (355 nm). The laser frequency was 200 Hz, 50 shots were recorded for a single spectral profile. In total, 100 spectral profiles were collected from one sample. Pulse extraction was set to 12 500 Da and blanking mass to 2000 Da. The irradiated spot size was approximately 100 µm in diameter. An external calibration was performed using standard mixtures of peptides and proteins (TofMix and PepMix from LaserBio, Valbonne, France) and standard bacterial extracts (BTS standard) from Bruker BioSpin AG (Fällanden, Switzerland). The spectral data were exported as ASCII files and further processed in R software environment. The raw mass spectra were aligned using the peak alignment by the Loess regression model and reduced to a uniform distribution of the 100 data points over the 2000–20,000 m/z range. The baseline was subtracted, negative values were set to zero, and an average spectrum was calculated. Afterward, the normalization to the area under the curve or to the total ion count was performed. Pearson’s correlation and principal component analysis (PCA) were performed in STATISTICA 12 (StatSoft Inc., Tulsa, OK, USA) and R software.

Transepithelial electrical resistance (TEER)

For A-Li culture, ELEP, ELEP differentiating for 16 days, and H441 cells, respectively, were seeded onto vitronectin-coated Falcon™ Cell Culture Inserts (0.33 cm² polyester, 1.0-μm pore size; 353,104, Falcon, Göteborg, Sweden) at a density of 25 × 10⁵ cells/cm² and cultured until respective measurement time points. Subsequently, the integrity of cell layer was evaluated by TEER measurement using a Millicell-ERS 2 V-Ohmmeter (Millipore Co., Bedford, MA, USA) at specific time points (1, 5, 10 days) according to manufacturer´s instructions. The electrode was soaked in 70% ethanol and rinsed with culture medium prior to use. TEER was determined as resistance (Ωcm²) = (R_sample–R_blank) related to effective membrane area (cm²).

Transepithelial dextran permeability assay

ELEP, ELEP differentiating for 16 days, and H441 cells, respectively, were seeded at a density of 25 × 10⁵ cells/cm² on inserts and provided with medium lacking phenol red. At specific time points (6 and 10 days), the medium in the apical chamber was replaced with the same medium containing 1 mg/ml fluorescein isothiocyanate (FITC)-conjugated dextran (Sigma-Aldrich GmbH) and incubated for 20 min at RT. Subsequently, 100 μL of media from each well from the basolateral chamber were transferred to 96-well plates. Fluorescence intensity was measured on the Fluostar Reader (BMG Labtech Ortenberg, Germany) using 485/520 nm excitation/emission maxima.

Statistical analysis

Statistical analysis was performed using GraphPad Prism 8 software (San Diego, CA, USA). Outliers were identified with GraphPad Outlier Calculator. The results are presented as the mean + standard error of the mean (SEM) of at least three independent replicates. Gene expression data were log transformed for statistical analysis, as per the general recommendations [46], graphs of gene expression show non-transformed data. Statistically significant differences among the groups were determined by one-way ANOVA followed by Tukey´s multiple comparisons test. Single, double and triple asterisks indicate statistical significance of p < 0.05, p < 0.01 and p < 0.001, respectively.

Results

Derivation of expandable lung epithelium from human embryonic stem cells

To derive expandable lung epithelium (ELEP) hESC were first differentiated towards foregut endoderm (FE). In order to do so, we adopted recent protocols involving treatment with Activin A to differentiate hESC first into definitive endoderm (DE) cells with high efficiency (Fig. 1) [5, 10, 11]. Next, to generate FE from DE, Activin A was withdrawn. The inactive TGFβ signaling, is crucial for endoderm anteriorization [47]. Our protocol does not involve any inhibitors of BMP4 and Activin/Nodal signaling. Instead, it uses simple exclusion of the respective inductive growth factors, mediated by the use of serum-free medium. Abolishment of both BMP4 and Activin A/Nodal signaling was confirmed by western blot (Supplementary Fig. S1). In particular, decreased levels of phosphorylated SMAD 1/5/9 and 2 were found in DE cultivated in serum free media as opposed to high phosphorylated SMAD levels in DE cultivated in media containing serum and/or BMP4 and activin A, respectively. Of note is that serum concentration as low as 2% is already able to activate SMAD phosphorylation. qRT-PCR and Western blot analysis further confirmed that the use of serum-free medium is sufficient to generate SOX2⁺ FOXA2⁺ FE cells (Figs. 1B, 2).

Fig. 1.

Generation of ELEP from hESC. A Schematic representation of directed differentiation of hECS to ELEP. The plot indicates timing schedule of hESC treatment with inducing factors: fetal bovine serum (FBS), Activin A, insulin/transferrin/selenium (ITS), FGF2, EGF. During these steps the cells differentiated into definitive endoderm (DE) cells by day 6, and subsequently into foregut endoderm (FE) cells by day 8. The FE cells were then treated with growth factors and repeatedly passaged to generate ELEP. B Relative expression of FOXA2 during differentiation of hESC into FE. Log transformed data were used for statistical analysis, graphs show non-transformed data. C Flow cytometry analysis of the expression of NKX2.1 in nondifferentiated hESC, ELEP in low (1–30) and high (31–65) passages, and H441 (positive control). Gates for positivity were set based on isotype controls. ΔMFI represents the median fluorescence intensity (MFI) of isotype control subtracted from the sample MFI. D Abundance of NKX2.1 positive cells in population of hESC, ELEP low and high, and H441. The plot is graphical representation of the results from flow cytometry analyses in (C). E Telomerase activity of ELEP low and high, and primary human fibroblasts, relative to nondifferentiated hESC. Data are presented as the mean + SEM of at least three independent experiments. *p < 0.05, **p < 0.01 and ***p < 0.001. F Intact cell mass spectrometry and analysis of mass spectra. Principal component analysis (PCA) of the dataset containing mass spectra recorded from hESCs, immature differentiation stages harvested at day 1, day 5 and day 7, and from ELEP. The cells were harvested, washed and analyzed by mass spectrometry without previous extraction or fractionation. Mass spectra were recorded in pentaplicates, processed as described previously [49, 50], and used as inputs for the PCA. Each point in the PCA plot represents a unique biological sample

Fig. 2.

Analysis of lineage marker expression during differentiation of hESC into ELEP. A-D Relative gene expression of NKX2.1 (lung and thyroid marker, A), CDX2 (liver and intestinal marker, B), PDX1 (pancreatic marker, C), and PAX8 (thyroid marker, D) in nondifferentiated hESC, definitive endoderm cells (DE), foregut endoderm cells (FE), ELEP low and ELEP high, normalized to GAPDH and related to hESC. E Western blot analysis of NKX2.1 and SOX2 expression in hESC, DE, FE and ELEP (low and high). F Quantification of NKX2.1 expression at protein level as determined by optical density of western blot, normalized to GAPDH. G Quantification of SOX2 expression at protein level as determined by optical density of western blot, normalized to GAPDH. Data are presented as the mean + SEM of at least three independent experiments. Log transformed data were used for statistical analysis of gene expression, graphs show non-transformed data. *p < 0.05, **p < 0.01, ***p < 0.001

FE cells were then further cultured in medium containing FGF2 and EGF, giving rise, by repeated passaging, to population of cells with high proliferative potential. We determined that cells established by such manipulation were able to double their average population in 26 h and 27 min. Because the respiratory cell lineages of foregut are firstly recognized by the expression of NKX2.1 [48], we performed flow cytometric analysis to determine the percentage of NKX2.1-positive cells in this cell population. We detected 75.41 ± 6.29% of NKX2.1-positive cells in low passage cells (passage number 3–30) and 75.72 ± 12.91% of NKX2.1-positive cells in high passage cells (passage number 31 to 65) (Fig. 1C, D). These data suggest that the derived progenitors belong to lung epithelium. In addition, NKX2.1 positivity in population of ELEP is stable even during long-term cultivation (Fig. 2A, E, F).

Recently we have introduced intact cell mass spectrometry (MS) as a sensitive quality control tool for identifying minute changes to cell phenotypes [49, 50]. Here, we applied this methodology to delineate the differentiation trajectory leading from undifferentiated hESC to ELEP. To do so, we subjected undifferentiated hESC, cells at days 1, 5, and 7 of their differentiation, and also the fully stabilized ELEP to intact cell MS and processed the final normalized spectral datasets as described above in the M&M section. Principal component analysis (PCA) clearly distinguished cluster of undifferentiated hESC from the committed precursors at days 1, 5 and 7, and also from the ELEP (Fig. 1F). From these multivariate data we may elute that differentiation of hESC towards ELEP is a stepwise synchronized process that produces, at defined stages of its progression, well distinguished cell populations, with phenotypically stable ELEP at the end.

Expandable lung epithelium possesses high proliferative potential in vitro

The infinite life span of tumor cells and of stem cells in turning over tissues has been attributed to highly active telomerase [51]. Cell differentiation and propagation in cell culture lead to a decrease of telomerase activity and to replicative senescence after a predictable number of cell divisions [52].

Therefore, to test the capability of hESC-derived ELEP to continuously proliferate we analyzed their telomerase activity. We found that telomerase activity in ELEP is lowered for only about one third compared to undifferentiated hESC and is several-fold higher than in primary human fibroblasts, specifically to 71.15% (± 5.78) in low passage ELEP and to 68.24% (± 5.86) in high passage ELEP (Fig. 1E). These numbers document high activity of telomerase in ELEP and are consonant with their capability to avoid Hayflick limit, similar to various tissue stem cells and progenitors [53, 54].

Since cells have to be stored or transported in frozen state, ELEP were also cryopreserved in order to investigate the effect of these conditions. As such, ELEP maintained their proliferative and differentiation capacity (as discussed below) even after being subjected to repeated freezing and thawing cycles.

The gene expression profile of expandable lung epithelium shows their lung and stem/progenitor identity

To verify the relation of ELEP to lung lineage we first evaluated their separation from more posterior derivatives of FE. We examined the expression of CDX2 and PDX1 corresponding to liver, intestinal and pancreatic lineages, respectively. We found that mRNA levels of CDX2 and PDX1 did not change in ELEP relative to those in hESC (Fig. 2B, C), whereas the level of lung-specific NKX2.1 mRNA significantly increased (Fig. 2A). To exclude the possibility that we derived NKX2.1-positive cells belonging to thyroid lineage we analyzed PAX8 expression. Quantitative RT-PCR data did not show any significant increase in PAX8 mRNA level in ELEP compared to hESC (Fig. 2D). Moreover, we confirmed that NKX2.1 positive ELEP concurrently do not belong to ectodermal forebrain neuronal lineage by determining the FOXG1 mRNA levels (Supplementary Fig. S2). In addition, the Western blot analysis has proven significantly increased expression of NKX2.1 and SOX2 in ELEP also at protein level (Fig. 2E–G). Importantly, NKX2.1 and SOX2 proteins co-expression in ELEP further suggests that they belong to early-stage lung lineage.

The subcellular composition of ELEP document their lung epithelial progenitor phenotype

Detailed morphology of ELEP was analyzed by transmission electron microscopy. Electron microphotographs showed ovoid-shaped medium-sized cells with nuclei containing only euchromatin and prominent nucleoli of reticular type, which indicate high metabolic activity and progenitor attribute of the cells (Fig. 3). The FE cells contained only a few mitochondria and other cell organelles were rare, with only some cisternae of rough endoplasmic reticulum, a small Golgi-complex, and few little multivesicular bodies and lamellar bodies (surfactant storage organelles) formed close to Golgi apparatus (Fig. 3A–D). In contrast to FE cells, lamellar bodies in ELEP were present in higher proportion of cells, were larger and also more abundant per cell. The presence of multivesicular and lamellar bodies and large glycogen deposits suggests an early (initial) phase of surfactant production (Fig. 3E–T). In conclusion, ELEP exhibit more advanced differentiation compared to FE cells in terms of number and morphologies of organelles and inclusions, while they still maintain ultrastructural signature typical for progenitor cells.

Fig. 3.

Subcellular analysis of cells by transmission electron microscopy. A–D Representative images of foregut endoderm (FE) cells. FE cells exhibit high nucleocytoplasmic ratio, pale euchromatin-rich nuclei (nu) containing large reticular nucleoli (re) with fibrillary centers (fc, arrowheads). In cell cytoplasm, the mitochondria (mi) are mainly scattered and lamellar bodies (arrow) are rare. E–T Representative images of expandable lung epithelium (ELEP) cells in different passages (5–51). ELEP show high nucleocytoplasmic ratio, an increased amount of cell organelles, such as mitochondria (mi) and Golgi apparatus (ga) with forming vesicles (ve), multivesicular bodies (mvb) or typical lamellar bodies (arrows). There are short cisternae of rough endoplasmic reticulum (rer), free ribosomes, centrioles (ce) and cell inclusions, such as lipid droplets (li) and glycogen, in cell cytoplasm

The ELEP are capable of further differentiation towards lung epithelial cells

The airway epithelium consists of specialized cell types, including basal cells, goblet/secretory cells, ciliated columnar cells and Club cells in the conducting part, and pneumocytes I and II in the alveolar part. To confirm the ability of ELEP to differentiate into various mature lung epithelial cells we seeded ELEP on vitronectin-coated dishes and cultured them in lung differentiation promoting medium containing FGF7 and FGF10. As shown in Fig. 4A–C, further differentiation of ELEP was accompanied by time-dependent increase in mRNA for forkhead box J1 (FOXJ1; marker of ciliated cells), aquaporin 5 (AQP5; marker of pneumocytes I), and surfactant specific protein B (SFTPB; marker of pneumocytes II), collectively representing cells of both lung anatomical regions. The epithelial nature of differentiating ELEP was further evaluated by determining epithelial barrier integrity and paracellular permeability, using transepithelial electrical resistance (TEER) and flux of dextran (Supplementary Fig. S3) as the measures. The increase in TEER was accompanied by a reduction in paracellular permeability of FITC-dextran in the differentiating ELEP as compared to undifferentiated ELEP. In differentiating H441 cells the TEER increased until day 5 with concurrent low dextran flux on day 6, yet it was replaced by a subsequent decline in resistance and increase in dextran permeability as seen on day 10.

Fig. 4.

ELEP differentiation in 2D and 3D. A–C Real-time qPCR analysis of gene expression of FOXJ1 (ciliated cell marker, A), AQP5 (pneumocyte I marker, B) and SFTPB (pneumocyte II marker, C) in ELEP differentiating on the dish in 2D for 7, 18, or 28 days, normalized to GAPDH and related to hESC. Three independent samples of both ELEP low and high were used for the differentiation and the data pooled to collectively represent ELEP. Data are presented as the mean + SEM of at least three independent experiments. Log transformed data were used for statistical analysis of gene expression, graphs show non-transformed data. *p < 0.05, **p < 0.01 and ***p < 0.001. D Hematoxylin–eosin staining highlights a formation of airway-like tubular structures and alveolar-like simple epithelium (arrow). E, F Immunofluorescence analysis of ciliated cell marker acetylated tubulin (acTUB), pneumocyte I marker AQP5, and pneumocyte II marker pro-SPB in 25 days differentiated organoids. Nuclear DNA was stained with DAPI in primary (E) and secondary (F) organoids

Next, we differentiated ELEP in 3D culture to also investigate self-organizing capacity of differentiating cells. Under 3D conditions, differentiating ELEP formed primary organoids containing alveolar-like structures with squamous epithelium, and tubular structures with airway-like epithelium (Fig. 4D). To test the expandability of primary organoids, secondary organoids were also prepared by dissociation and further culture of the primary ones. As determined by indirect immunofluorescence, epithelial lining inside these organoids included cells producing acetylated tubulin (acTUB) as well as cells producing aquaporin 5 (AQP5) and pro-surfactant B (pro-SPB), speaking for the presence of both ciliated cells and type I/type II pneumocytes in both primary and secondary organoids (Fig. 4E, F).

The ELEP produce lung-like structures upon their transplantation under kidney capsule

We have finally placed ELEP under kidney capsule of immunodeficient mice to test their differentiation and morphogenetic capabilities under the most physiological conditions. We performed three independent transplantations into a total of 19 mice. Six weeks post transplantation we observed outgrowth on the kidney (Fig. 5A) in 15 mice, 4 mice died prior to the harvesting timepoint. Histological analysis of the outgrowths showed branched tubular structures with decreasing height of epithelium, and more differentiated regions with low cuboidal to squamous epithelium similar to alveolar structures (Fig. 5B, Supplementary Fig. S4A). There were also regions with nondifferentiated cells in the outgrowths. Very importantly, these histologically visualized morphologies, were perfectly consonant with the expression of cell type-specific molecules in both airway-like and alveoli-like structures. Specifically, as shown in Fig. 5C, D, larger tubules were lined by cells expressing acTUB that is typical for ciliated cells of airway epithelium, while smaller tubules and alveoli-like structures with thin epithelium contained cells expressing pro-SPB and pro-SPC that is typical for surfactant-producing type II pneumocytes (Fig. 5E, F, Supplementary Fig. S4, respectively), as well as cells expressing AQP5 that is typical for type I pneumocytes (Fig. 5G, H). Immunohistochemistry analysis showed concurrent FOXA2, SOX2 and NKX2.1 positivity which changed depending on the region (Fig. 5I–L) and corresponds to distinct developmental events and timing [38, 48, 55, 56]). In particular, FOXA2 was expressed in almost all epithelial cells. SOX2 and NKX2.1 positivity was observed mainly in the epithelium of airway-like structures. Epithelial nature of ELEP was confirmed by the expression of cytokeratin 5 and 8 (Supplementary Fig. S4B, C). Also, notably, outgrowths formed from ELEP did not show any signature of cell types not belonging to lung tissues. Importantly, all the outgrowth-forming structures are of human origin, as determined by staining for human nucleolar antigen (HuNuc) (Supplementary Fig. S4E).

Fig. 5.

Growth and differentiation of ELEP in vivo. ELEP were transplanted under mouse kidney capsule and the outgrowths were analyzed 6 weeks post transplantation. A Macroscopic view of outgrowth formed from ELEP. B Histology analysis revealed that ELEP outgrowths contain both undifferentiated tissue and differentiated tissue with visible alveolar- and airway-like structures, including squamous epithelium and branched tubular structures with decreasing height of epithelium from pseudostratified to columnar and cuboidal, respectively. C–H Immunofluorescence analysis of lung lineage markers acetylated tubulin (acTUB; cilliated cells, C, D), pro-SPB (pneumocytes II, E, F), and AQP5 (pneumocytes I, G, H, in airway-like and alveoli-like regions of outgrows. Nuclear DNA was stained with DAPI. I–L Immunohistochemistry analysis of FOXA2, NKX2.1 and SOX2 in various airway-like regions of kidney outgrows. FOXA2 was expressed in almost all epithelial cells. SOX2 and NKX2.1 positivity was observed mainly in the epithelium of airway-like structures

Discussion

Here, we established a method, which utilizes only a standard monolayer culture combined with exposure to or withdrawal of, respectively, a set of growth factors at specific timepoints, for efficient generation of cells with phenotype of expandable lung epithelium from hESC. Such progenitors can grow in culture for long periods of time, minimally for 65 passages that equals to more than 6 months, without losing their key properties. In particular, we followed a widely accepted protocol and firstly derived definitive endoderm-like cell population by treatment of hESC with Activin A [5, 8, 10]. For further specification of endodermal cells towards foregut phenotype, dual inhibition of TGF-β and BMP signaling is widely used [5, 18, 57]. As an alternative to the use of inhibitors, which are quite costly, we simply used serum-free medium to exclude these factors. We confirmed that the serum-free conditions themselves were able to induce anteriorization of endodermal cells and thus generate SOX2- and FOXA2-positive cells.

Combined treatment with FGF2 and EGF was previously used for isolation and maintenance of stem/progenitor cells of numerous tissues, such as neural stem cells, epidermal progenitors, muscle progenitors, and also for isolation of stem/progenitor cells from adult human lung tissue [58–61]. It has also been shown by Mou and colleagues that FGF2 signaling is required for NKX2.1 induction [11]. Here, we built on these grounds and applied FGF2 and EGF to foregut endodermal cells to lock these cells in proliferative state while still promoting further anterior specification elicited by the low serum conditions, as we have shown that even the 2% FBS activated the BMP4 signaling (Fig. S1). We found that these simple conditions were able to produce lung phenotype as demonstrated by NKX2.1 positivity and by repression of the liver, intestinal and pancreatic lineage markers, CDX2 and PDX1. Given that thyroid gland as an FE structure also expresses NKX2.1, we needed to exclude the possibility that ELEP may also take the thyroidal state. Indeed, this was not the case since ELEP completely lacked the expression of thyroid-specific PAX8. Additionally, ELEP also do not belong to ectodermal forebrain neuronal lineage as determined by the FOXG1 levels.

We consider ELEP as in vitro equivalents of transit-amplifying progenitors of lung epithelial lineage. Progenitors belonging to various cell lineages, including progenitors to the airway epithelia of the future trachea, bronchi, and bronchioles during the process of branching morphogenesis express high levels of SOX2 [11, 62]. The same high level of SOX2 is typical also for undifferentiated hESC [63–65], and SOX2 generally is considered as molecule involved in driving self-renewal of cells. The level of SOX2 observed here in ELEP is even higher than that in hESC, thus further implying progenitor nature of this in vitro stably growing cell population. Progenitor nature of ELEP is also evidenced by their high telomerase activity (about two thirds of that measured in hESC), building on the well proven fact that high and low telomerase activity distinguishes immortal and terminally differentiated cells, respectively, between each other [62, 66].

Molecular signatures described above collectively speak for developmental restriction of self-renewing ELEP to cells of lung epithelium. We have extensively studied the detailed morphology and differentiation properties of ELEP to further document their belonging solely to lung epithelial lineage. The features of ELEP that we have confirmed by multiple experiments at various passage numbers after their establishment are as follows: (i) ELEP contain lamellar bodies and deposits of glycogen in their cytoplasm, as signs of their preparedness for surfactant production [67], (ii) ELEP display epithelial nature and readily differentiate into cells expressing markers of ciliated cells as well as type I and type II pneumocytes, both in 2D and 3D, with appropriate morphologies, (iii) ELEP exposed to differentiation media and cultured at an air–liquid interface increase steadily their transepithelial resistance and concurrently decrease permeability for dextran, both speaking of establishment of epithelial barrier functions, and (iv) ELEP placed in vivo respond to live tissue-born signals by forming only organized tubular and vesicular structures lined by near-to-genuine airway and alveolar epithelium cells, as determined by immunocytochemistry.

It is of note that the findings discussed above were confirmed by several independent repeats, both technical and biological, so that we are confident that the strategy to derive and maintain ELEP is very efficient and robust.

Here, we devised a protocol for derivation, from hESC, and for long-term propagation of cells, which possess multiple features of progenitors to lung epithelial lineage. This protocol is simple, relies only on monolayer culture without a need of 3D culture conditions and/or cell sorting steps, which minimizes a potential variability of its outcome. The resulting cells, expandable lung epithelium (ELEP), can be propagated in culture for long periods of time and can be subjected to cryopreservation without a loss of their potential to differentiate to cells exhibiting molecular and morphological characteristics of respiratory epithelium of airway and alveolar regions. With these properties, the ELEP offer an easily accessible and high-yielding source of cells for various biomedical applications.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

HK: Conception and design, collection and/or assembly of data, data analysis and interpretation, manuscript writing, final approval of manuscript. MC: Collection and/or assembly of data, data analysis and interpretation, manuscript writing. VP: Collection and/or assembly of data, data analysis and interpretation, manuscript writing. JD: Collection and/or assembly of data, data analysis and interpretation, manuscript writing. ZK: Collection and/or assembly of data, data analysis and interpretation, manuscript writing, other (animal work). JR: Collection and/or assembly of data, other (animal work). KS: Conception and design, manuscript writing. ZG: Collection and/or assembly of data. VS: Collection and/or assembly of data, data analysis and interpretation, manuscript writing. AR: Collection and/or assembly of data other (animal work). PV: Collection and/or assembly of data, data analysis and interpretation, manuscript writing. LM: Collection and/or assembly of data, data analysis and interpretation. LP: Collection and/or assembly of data, data analysis and interpretation. VP: Collection and/or assembly of data. MK: Telomerase assay, collection and/or assembly of data. LS: Surgical samples of human skin, collection and/or assembly of data. JH: Data analysis and interpretation, manuscript writing. AH: Conception and design, financial support, administrative support, manuscript writing, final approval of manuscript.

Declarations

Conflict of interest

The authors have no financial conflicts of interest.

Ethical statement

Human lung tissue samples were obtained from therapeutical lung surgery based on the written informed consent by the patient and approval of Ethics Committee of the University Hospital Brno (28–170621/EK). Human skin samples were obtained from healthy donors undergoing surgical procedures based on the written informed consent by the patient and approval of Ethics Committee St. Anne’s University Hospital Brno (8 V/2020). For animal experiments, all European Union Animal Welfare lines (EU Directive 2010/63/EU for animal experiments) were respected. Animal experiments were approved by the Academy of Sciences of the Czech Republic (AVCR 13/2015), supervised by the local ethical committee and performed by certified individuals (JR, AR, ZK).