Generation of lung epithelial-like tissue from human embryonic stem cells
Human embryonic stem cells (hESC) have the capacity to differentiate in vivo and in vitro into cells from all three germ lineages. The aim of the present study was to investigate the effect of specific culture conditions on the differentiation of hESC into lung epithelial cells.
Undifferentiated hESC, grown on a porous membrane in hESC medium for four days, were switched to a differentiation medium for four days; this was followed by culture in air-liquid interface conditions during another 20 days. Expression of several lung markers was measured by immunohistochemistry and by quantitative real-time RT-PCR at four different time points throughout the differentiation and compared to appropriate controls.
Expression of CC16 and NKX2.1 showed a 1,000- and 10,000- fold increase at day 10 of differentiation. Other lung markers such as SP-C and Aquaporin 5 had the highest expression after twenty days of culture, as well as two markers for ciliated cells, FOXJ1 and β-tubulin IV. The results from qRT-PCR were confirmed by immunohistochemistry on paraffin-embedded samples. Antibodies against CC16, SP-A and SP-C were chosen as specific markers for Clara Cells and alveolar type II cells. The functionality was tested by measuring the secretion of CC16 in the medium using an enzyme immunoassay.
These results suggest that by using our novel culture protocol hESC can be differentiated into the major cell types of lung epithelial tissue.
KeywordsHuman Embryonic Stem Cell Preimplantation Genetic Diagnosis Ciliated Cell Alveolar Type Clara Cell
Human embryonic stem cells (hESC)  have the capacity to differentiate in vivo and in vitro into cells from all three germ lineages , which makes them very interesting in developmental biology, regenerative medicine and in vitro pharmacological studies. Furthermore, the derivation of hESC lines carrying a monogenic disease offers the opportunity of creating an in vitro model of the disease. Such lines can be derived from an embryo diagnosed as affected after preimplantation genetic diagnosis, a procedure that allows the detection of a genetic defect at the preimplantation embryo level . The in vitro generation of lung cells and tissues from human embryonic stem cells provides an alternative to lung transplantation in patients with lung injury due to chronic pulmonary disease and inherited genetic diseases such as cystic fibrosis (CF).
The respiratory system originates from the foregut endoderm, differentiating into many kinds of specialized epithelial cells. These include ciliated, secretory, and neuroendocrine cells of the proximal bronchi, and the alveolar cells. The latter can be divided into two types of cells; type I cells, with a highly flattened morphology ideally suited for gas exchange, and the less differentiated, more cuboidal type II cells serving as progenitor cells for type I cells. Furthermore, these alveolar type II cells are important in synthesizing and secreting pulmonary surfactant proteins, a complex mixture of proteins and phospholipids which lower surface tension . One of these proteins, i.e. SP-C, is expressed in the distal epithelium during early lung development, and becomes restricted to alveolar type II cells late in gestation and postnatally.
Nk2 homeobox1 (NKX2.1) (also Thyroid transcription factor 1 (TTF-1)) is the earliest known marker associated with commitment of endodermal cells to pulmonary and thyroid cell lineages; it appears before the definitive lung formation .
During fetal lung development the level of NKX2.1 decreases with advancing age. At first it is expressed in all pulmonary epithelial cells, later it becomes restricted to distal alveolar type II cells and proximal Clara Cells - the nonciliated, nonmucous cells lining the bronchioles of the lung . It is known that these latter cells secrete a variety of proteins, such as the surfactant proteins SP-A, SP-B and SP-D, and most importantly CC16, a Clara cell diffusible 16 kDa protein (previously known as CC10), which appears to be secreted constitutively at a high level throughout life .
Mucociliary clearance is an essential aspect of the innate defence mechanisms of the lung. Mucus secreting cells act in concert with the ciliated cells within the pulmonary epithelium to remove potentially harmful substances from the lung. FOXJ1 is a transcription factor expressed in the ciliated cells of the lung and reproductive tracts. It functions in late-stage ciliogenesis of lung development and is expressed before the appearance of ciliated cells . Many genes expressed in the lung are also expressed in other tissues, such as thymus and pancreas. Lung epithelial-like tissue arising within a complex culture can therefore only be identified by measuring the gene expression of multiple markers.
Few research groups have reported on the successful differentiation towards endodermal lineages such as lung epithelium.
We have adapted the approach of Coraux et al. culturing murine ESC in air liquid interface (ALI) conditions, which results in a simple and reliable protocol to differentiate human ESC into a lung epithelial-like tissue. This is to our knowledge the first study to report the differentiation of human ESC into a lung epithelial-like tissue.
Materials and methods
hESC Cultures and differentiation
Maintenance of hESC
hESC were derived, as previously described , from human preimplantation embryos after IVF or preimplantation genetic diagnosis, characterised according to the International Stem Cell Initiative guidelines and karyotyped by array-based comparative genomic hybridization (array-CGH) . The inner cell masses of human blastocysts were obtained by immunosurgery and transferred to 0.1% gelatine-coated culture dishes containing mitomycin C inactivated CF1 mouse embryonic fibroblast feeders (MEF). Upon derivation, cells were routinely cultured on MEF feeders in a humidified atmosphere at 10% CO2 at 37°C in hESC medium that consisted of 80% knockout Dulbecco's modified Eagle's medium (KO-DMEM) supplemented with 20% knockout-serum replacement (KO-SR), 2 mM L-glutamine, 1% nonessential amino acids, 4 ng/ml human recombinant basic fibroblast growth factor (bFGF) and Penicillin/streptomycin (100 U/mL) (all Invitrogen, Carlsbad, CA, USA), and 0.1 mM β-mercaptoethanol (Sigma-Aldrich, St. Louis, MO, USA). Medium was changed daily and passaging was done by mechanical slicing. Five different cell lines were used to check for possible variations in their differentiation potential: VUB03_DM1 carrying a mutation for myotonic dystrophy, VUB04_CF carrying one mutation in the CFTR gene, VUB09_FSHD carrying a mutation for facioscapulohumeral muscular dystrophy, and VUB07 and VUB14, neither carrying any known genetic disease.
The lines were at passages 78-80, 44-47, 73-75, 110-112 and 23-25 respectively, when used for our experiments. The main characteristics of the cell lines, including the chromosomal abnormalities detected by array-CGH, are summarized in additional file 1.
Differentiation of hESC
Collagenase IV dissociated hESC clumps of variable sizes (± 30/insert) were plated on porous membranes of millicell-HA culture inserts (Millipore Corporation, Billerica, MA, USA). These culture inserts were placed in 12-well plates coated with MEF feeders and allowed to expand for four days in hESC medium in liquid-liquid conditions. Differentiation was induced by replacing the medium with differentiation medium (hESC medium without bFGF and β-mercaptoethanol) in liquid-liquid conditions for four days and then switching to air-liquid interface culture for twenty days, in which the medium from the upper compartment was removed (this will be referred to as the "ALI differentiation protocol"). Medium was changed daily. Two types of controls were used in the quantitative real-time RT-PCR experiments: 1) hESC plated on 0.1% gelatine-coated culture dishes on MEF feeders that had differentiated spontaneously in hESC medium (named "hESC spontaneous differentiation control") and 2) hESC plated on porous membranes for four days in hESC medium followed by 24 days in differentiation medium in liquid-liquid conditions (named "NO ALI control"). A third type of control, namely hESC grown on porous membranes in liquid-liquid conditions in hESC medium for two days and referred as "undifferentiated hESC control", was used in the immunostaining experiments.
RNA extraction and quantitative real-time RT-PCR analysis
After removal of the cells from the porous membrane with lysis buffer from the RNeasy kit, RNA was extracted and treated with RNase-free DNase according to the manufacturer's protocol (both Qiagen, Hilden, Germany). Five micrograms of total RNA were reverse transcribed into cDNA using the First Strand cDNA Synthesis Kit (GE Healthcare, Buckinghamshire, UK) with the NotI-d(T)18 primer, according to the manufacturer's instructions.
Real-time PCR primers and probes
Primers (5' to 3')
PCR reactions were performed on an ABI 7500 thermal cycler (Applied Biosystems). Thermal cycling parameters were: 2 min at 50°C, 10 min at 95°C, followed by 40 cycles of 15 s at 95°C and 1 min at 60°C. Measurements were taken at the end of the extension phase at 60°C.
Relative quantification of gene expression between multiple samples was achieved by normalization against two endogenous controls GAPDH and Ubiquitin C (as recommended by Willems et al. ) using the ΔΔCt method of quantification. Relative amount of mRNA was calculated as 2-ΔΔCt and quantitative RT-PCR results were statistically analysed by one-way ANOVA.
Experiments were performed at least three times, and at least two different samples were taken for each measurement to provide biological replicates. Experiments were repeated on four different cell lines (VUB03_DM1, VUB04_CF, VUB07 and VUB14). Each PCR reaction was performed in duplicate as a technical replicate. No-template negative controls were included. Samples were compared to hESC spontaneous differentiation control samples and/or NO ALI control samples. Primers and probes were designed to avoid murine mRNA amplification. Nevertheless, to exclude possible contamination with murine mRNA from MEF, samples with cDNA derived from mouse fibroblasts were included in the different assays. No signal could be measured by quantitative real-time RT-PCR, except for Vimentin.
hESC grown on porous membranes following the ALI differentiation protocol were fixed in 3.7% formaldehyde (Merck, Darmstadt, Germany) when still on the membrane, dehydrated and embedded as such in paraffin. For SP-A staining, tissue was embedded in a perpendicular direction as well. Undifferentiated hESC control samples were treated in the same way as a control. Sections (4 μm thick) were rehydrated in gradual stages. For SSEA4, TRA1-60 and TRA 1-81, an additional step with antigen retrieval was performed by heating at 95° for 75 min in freshly made 0.01 M citrate buffer, pH 6.0. Non-specific binding was blocked for 2 h with PBS containing 0.25% Triton-X (Sigma Aldrich) and 2% of the appropriate normal serum (Sigma Aldrich).
Sections were incubated with goat polyclonal antibody to CC16 (Santa Cruz Biotechnology, Santa Cruz, CA; dilution 1/100), goat anti-human polyclonal antibody to SP-C (Santa Cruz: dilution 1/100), SP-A (Chemicon, Temecula, CA; dilution 1/500), SSEA4 (Millipore Corporation, Billerica, MA, USA), dilution 1/100), TRA1-60 (Millipore; dilution 1/100) or TRA1-81 (Millipore; dilution 1/100) overnight at 4°C. After washing for 30 min in PBS, sections were incubated with the corresponding Alexa 488/633 conjugated secondary antibodies for 2 h. After washing in PBS, sections were mounted with Prolong® Gold Antifade reagent with DAPI (all Invitrogen)
Double stainings were performed by sequentially adding the two sets of primary and secondary antibodies.
Immunostaining for light microscopy
Immunostaining for Vimentin was performed on both sections of fetal lung tissue (positive control) and hESC grown on porous membranes following the ALI differentiation protocol. Hydrogen peroxide (0.3%) was used to block endogenous peroxidase (15 min). The staining required antigen retrieval (95°C for 75 min in freshly prepared 0.01 M citrate buffer, pH 6.0). Vimentin monoclonal mAb (Dako, Denmark) was used (dilution 1/200) and incubated overnight at 4°. The day after the HRP-labelled secondary antibody was applied for 1 h and staining was finalised with a diaminobenzidine solution (Dako REAL Envision Detection Systems Peroxidase/DAB Rabbit/Mouse, Dako, Denmark) so as to detect the sought after antigen. Slides were extensively washed with PBS between the different stages and counterstained with hematoxylin before mounting.
Both for fluorescence and light microscopy immunostainings, fetal lung and kidney tissue (from a 22- week old fetus) were included as a positive and negative control respectively. Other negative controls were performed by omitting the primary antibodies or replacing them with antibody isotype controls from the appropriate host species at the same concentration.
Analysis of CC16 protein secretions in the medium was performed using a CC16 Elisa Kit (DiaMed Eurogen, Turnhout, Belgium) according to the manufacturer's instructions and the absorbancy values at 450 nm were determined. The minimum detectable concentration was <50 pg/mL (as defined by the manufacturer). Medium was collected after 24 h and proteins were concentrated by centrifugation for 45 min on 4000 rpm in a Amplicon Ultra-15 Centrifugal filter unit with Ultracel-3 membranes (Millipore). Experiments were performed on three different cell lines (VUB04_CF, VUB07 and VUB09_FSHD).
Standard sera samples were used to generate a standard curve, from which the concentration of CC16 in the samples was interpolated.
Differentiation medium that had not been in contact with any cells was used as a negative control.
Characterising the differentiated cells by quantitative real-time RT-PCR
The control samples containing only MEF were found to be positive for Vimentin as well. Potential contamination with murine mRNA in samples would be similar and would not lead to an underestimation or an overestimation of the Vimentin mRNA level in ALI conditions versus NO ALI control samples.
Characterising the differentiated cells by immunofluorescence staining
During differentiation, cells became covered with a tough mucus-like layer, which dissolved during fixation. On sections, bronchiolar-like structures were seen within the cell clumps.
Expression of Vimentin by immunostaining for light microscopy
Quantitative determination of secreted Clara Cell 16 kD Protein by ELISA
The "NO ALI control" only gave a detectable secretion level after 20 days of ALI culture conditions, while no secretion could be measured at all in the "spontaneous differentiation control" samples. (Figure 11B)
This study describes a protocol to obtain lung epithelial-like tissue from human embryonic stem cells (hESC). The protocol relies on an air-liquid interface (ALI) system to mimic the conditions of an adult trachea and has been adapted from a protocol on murine ESC . The specificity of the differentiation into lung epithelial cells was supported by expression data from quantitative real-time RT-PCR, immunohistochemistry and ELISA. Although most of the markers are not exclusively expressed in the lung, their cellular co-expression can be considered as sufficient to evaluate the generation of lung epithelial-like tissue.
Our protocol uses porous membranes to create an ALI to mimic the conditions of an adult trachea. It was previously shown on murine ESC  that these conditions support the differentiation into airway epithelial tissue. However, transfer of this protocol to hESC did not work. The protocol we use is a one step protocol (hESC plated directly on the porous membranes) instead of a two step protocol on murine ESC (plated on collagen before dissociation and transfer to the membrane). Furthermore the medium used in the two studies is different. Nevertheless, it should be mentioned that although our approach on human ESC is finally quite different from the protocol used on murine ESC, the results are similar. In both instances, the presence of CC16 mRNA early in the differentiation process (day 8 and day 10) was shown and there was evidence of differentiation into the main types of lung cells.
Although variation in mRNA expression level could be detected between the differentiating hESC lines, the expression pattern remained similar for all lines tested. The expression of CC16, a marker for Clara Cells, the nonciliated nonmucous cells lining the bronchioles of the lung, peaked significantly after ten days of differentiation with the ALI protocol. The same expression peak at day 10 was seen for NKX2.1, an early distal marker, while SP-C expression started to increase around day 20. This pattern resembles the lung development in vivo, where the level of NKX2.1 decreases with advancing age and the expression of the target genes such as SP-A and SP-C increases in association with differentiation of alveolar type II cells during the saccular-alveolar stage of lung development . This was also confirmed by immunodetection of cytoplasmic SP-C and localization of SP-A at one side of the cells, corresponding with alveolar type II cells.
A significant increase in β-tubulin IV and FOXJ1 mRNA expression could be detected after twenty days of culture following the ALI differentiation protocol. In vivo, the FOXJ1 transcription factor functions in late-stage ciliogenesis of lung development and is expressed before the formation of ciliary 9+2 axonemes . It therefore marks a lineage of cells distinct from those dependent upon NKX2.1. Both markers thus clearly show evidence for the differentiation of hESC into ciliated cells.
In addition, Aquaporin 5, a gene coding for a major water channel and in lung only expressed in alveolar type I cells, is also significantly increased after 20 days of culture, providing proof of cell maturation. The immunohistochemistry staining pattern for SP-A resembled the pattern seen in fetal lung tissue. This gene is expressed in lung alveolar type II cells and altered SP-A levels in Clara cells seem to play a role in diseases such as cystic fibrosis and respiratory distress syndrome [18, 19, 20, 21]. In our study, we not only checked mRNA expression and protein localisation of several lung-specific genes, we also showed that one of these, CC16, is secreted by the cells and could be measured in the medium in our ALI culture conditions, but not in NO ALI control samples nor in spontaneous differentiated cell controls.
Vimentin is one of the mammalian intermediate filament proteins. It is expressed in cells of mesenchymal origin and is a feature of proliferating cells at the fetal stage. In our experiments Vimentin mRNA expression was measured both in our ALI cultures and in NO ALI control samples and both were compared to spontaneous differentiated control samples. No significant difference could be detected. Using immunostaining we showed the presence of Vimentin proteins in the ALI cultured tissue, ressembling Vimentin expression in fetal lung tissue.
The mRNA levels of the pluripotent stem cell markers POU5F1 and NANOG. showed a strong decrease, but were still detectable after twenty days of ALI culture. A decrease of these hESC-specific markers is generally observed upon differentiation. The RT-PCR analyses were confirmed by immunostaining for other stemness markers SSEA-4, TRA1-60 and TRA1-81; small groups of undifferentiated cells could be detected scattered amongst the tissue, indicating that only a small fraction of the cells remained undifferentiated after twenty days of ALI culture.
Other studies use small airway growth medium (SAGM), a commercial lung specific medium containing factors known to promote pneumocyte differentiation such as retinoic acid. One study on murine ESC describes a three-step protocol using Activin A in combination with low serum concentration to enhance the specification of the endodermal germ layer in embryoid bodies (EBs) and the final application of SAGM on plated EBs . Results showed at least one cluster of SP-C expressing cells in the majority of the adhered embryoid bodies. Furthermore, the authors were able to measure the expression of several lung markers. Results of the same group  on human ESC were positive, but were only focused on alveolar type II cells. Here, hESC were allowed to differentiate spontaneously as EBs before plating, followed by changing the medium to SAGM ten days after plating. They have shown that hESC have the capacity to differentiate into lung-specific cells, expressing SP-C in vitro. Other research groups developed a transfection and culture procedure, which facilitates, via genetic selection the differentiation of hES cells into a relatively pure population (>99%) of alveolar type II cells .
A recent study on human amniotic fluid cells also shows their potential to differentiate into epithelial lung lineages . For a review on these studies, we refer to a recent review by Rippon et al. .
In this study we demonstrate that hESC can differentiate into lung epithelial-like tissue without specific growth factors or embryoid body formation. The air-liquid interface on a porous membrane in combination with low serum is sufficient to prime the cells to form an airway epithelial-like tissue. It has been proven before that physical effects can direct stem cells .
Nevertheless, efforts will be made to further improve this novel culture protocol, trying to increase the number of differentiated cells or to guide the differentiation into particular cell types by adding certain growth factors to this system, such as fibroblast growth factors (FGF). These are important in the developing lung, and their addition to the culture medium may have an influence on the differentiation pattern .
The differentiation protocol was also tested on VUB04_CF, carrying one mutation in the CFTR gene. At first sight, no significant difference was noted between this line and other hESC lines. In the future, VUB04_CF and the newly derived VUB22_CF carrying the p.F508del mutation in both alleles can be used as model system to study differentiation effects on mRNA and protein levels.
It is remarkable that, although some chromosomal abnormalities could be found by array-based comparative genomic hybridization in the stem cell lines used in this study , results were comparable for all lines. These abnormalities seem to have no or only minor influence on the differentiation of hESC into lung epithelial-like tissue. It should be noted that our lines were checked by array-CGH, while most groups use G-banding obtained by Giemsa staining as a regular monitoring method, a technique that cannot detect small chromosomal aberrations.
To summarize, we developed a convenient protocol using ALI culture conditions that stimulates differentiation of human ESC into lung epithelial-like tissue. The combination of the mRNA expression results as well as results of protein expression and protein localisation showed the presence of the major cell types of lung epithelial tissue.
The authors wish to thank the Fund Alphonse & Jean Forton and the Association Française contre les Myopathies (AFM) for financial support. We gratefully acknowledge the assistance of Ileana Mateizel, Mieke Geens, Dr. Bing Chen, Dr. Claudia Spits and Anna Seriola Petit for maintaining the hESC lines in culture and for their critical comments. We also wish to extend our thanks to M. Whitburn of the VUB language centre for proofreading our text.
- 3.Sermon K, Van Steirteghem A, Liebaers I: Preimplantation genetic diagnosis. Lancet 2004, (363):1633–1641.Google Scholar
- 13.Mateizel I, De Temmerman N, Ullmann U, Cauffman G, Sermon K, Ven de Velde H, De Rycke M, Degreef E, Devroey P, Lieabaers I, Van Steirteghem A: Derivation of human embryonic stem cell lines from embryos obtained after IVF and after PGD for monogenic disorders. Hum Reprod 2006,21(2):503–11.CrossRefPubMedGoogle Scholar
- 22.Carraro G, Perin L, Sedrakyan S, Giuliani S, Tiozzo C, Lee J, Turcatel G, De Langhe S, Driscoll B, Bellusci S, Minoo P, Atala A, De Filippo RE, Warburton D: Human amniotic fluid stem cells can integrate and differentiate inot epithelial lung lineages. Stem Cells 2008,26(11):2902–11.CrossRefPubMedPubMedCentralGoogle Scholar
- 24.Stolberg S, McCloskey K: Can shear stress direct stem cell fate? Biotechnol Prog 2009.Google Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.