Tissue Engineering & Regenerative Medicine Centre, Chelsea & Westminster Campus, Imperial College, London, United Kingdom
Tissue Engineering & Regenerative Medicine Centre, Imperial College Faculty of Medicine, Chelsea & Westminster Campus, Fulham Road, London SW10 9NH, U.K. Telephone: 44-208-237-2670; Fax: 44-208-746-5619
Various means have been used to encourage the differentiation of embryonic stem cells (ESCs) toward specific lineages, including growth factor administration, genetic modification, and coculture with relevant cells/tissues. Cell extract–based reprogramming has recently been used to derive mature cells from nonrelated phenotypes. In this communication, we tested whether this in vitro reprogramming approach can be used to direct ESC differentiation. Permeabilized murine ESCs exposed to extracts of murine type II pneumocytes showed increased expression of surfactant protein C and its corresponding mRNA, reflecting enhanced differentiation of pneumocytes. Subsequent differentiation to a type I phenotype was demonstrated by expression of aquaporin 5. Pneumocyte formation occurred quicker than with growth factor–induced differentiation. Our findings establish that ESCs can be differentiated in vitro using cellular extracts. This model provides a tool for analysis of the key factors involved in the differentiation of ESCs to type II pneumocytes.
Embryonic stem cells (ESCs) have been shown to differentiate in vitro to cell phenotypes arising from each of the three germ layers [1–4]. A variety of means has been used to try to direct the differentiation of these cells toward specific cell lineages, in particular for the purposes of tissue engineering [1–4]. In the distal lung, type II pneumocytes undergo proliferation and differentiation to the type I phenotype following injury and are crucial to the natural regenerative process of the alveoli [5, 6]. Thus, as part of our strategy to produce pulmonary cell types for tissue engineering purposes, we have derived type II pneumocytes from ESCs using soluble factors [7, 8]. However, as this process is lengthy and does not provide high yields of the target phenotype, we have sought a more efficient method.
A method of cell reprogramming, based on the exposure of permeabilized cells to a nuclear and cytoplasmic extract derived from a “target” cell, has been used successfully to derive mature cells from nonrelated phenotypes, including T cells  and pancreatic β (insulin-producing) cells  from 293T cells and fibroblasts, and cardiomyocytes from adipose stem cells . We hypothesized, therefore, that this in vitro reprogramming technology could be used to differentiate ESCs to type II pneumocytes. In this study, a murine ESC line (E14tg2α) was stably transfected with a construct containing an enhanced green fluorescent protein (eGFP) reporter gene under the control of the surfactant protein C (SPC) promoter, a specific marker of type II pneumocytes. Reversibly permeabilized ESCs were reprogrammed with extracts of “target cells” (transformed murine pneumocytes). The reprogrammed ESCs became GFP-positive cells and coexpressed SPC, specific for type II pneumocytes, and thyroid transcription factor-1 (TTF-1) that is found in immature pulmonary epithelium. In accordance with the well-known, in vitro differentiation of type II pneumocytes to the type I phenotype [12, 13], we were able to show that SPC gene expression declines over time and is accompanied by an increase in aquaporin 5 mRNA and protein, a marker for type I pneumocytes . These findings demonstrate that differentiation of murine ESCs can be driven by using an extract-based, in vitro reprogramming approach.
Materials and Methods
Murine E14Tg2α ESCs were grown in an undifferentiated state, without feeder layers, on tissue culture plates with 1000 U/ml of leukemia inhibitory factor (LIF; Chemicon, Temecula, CA, http://www.chemicon.com). Cells were maintained in ESC medium, comprising Dulbecco's modified Eagle medium (DMEM), supplemented with 10% fetal bovine serum, 2 mM L-glutamine, 100 U/μg penicillin/streptomycin, and 0.1 mM 2-mercaptoethanol. After removal of LIF to allow differentiation of the ESCs, cells were maintained in the same ESC culture medium. MLE-12 cells  (American Type Culture Collection, Manassas, VA, http://www.atcc.org/Home.cfm) were grown on plates in HITES medium, comprising Ham's F12K medium 50% and DMEM 50%, supplemented with 2% fetal bovine serum, 10 mM HEPES, 2 mM L-glutamine, 100 U/μg penicillin/streptomycin, 0.005 mg/ml insulin, 0.01 mg/ml apo-transferrin, 30 nM sodium selenite, 10 nM hydrocortisone, and 10 nM β-estradiol. Murine embryonic fibroblasts (MEFs), isolated from 11-week BALB/C embryos, were cultured in ESC medium without LIF. All cultures were maintained in an incubator at 37°C in a humidified atmosphere of 5% CO2. For the undifferentiated ESC cultures, medium was changed every day and, for other cells, on alternate days.
A 4.8-kb murine SPC promoter/GFP construct was transfected into undifferentiated E14tg2α, MLE-12 (positive control) and MEF (negative control) cells using LipofectamineTM 2000 (Invitrogen, Paisley, U.K., http://www.invitrogen.com/), according to the manufacturer's instructions. The transfected ESCs were selected in ES (embryonic stem) culture medium containing 300 ng/ml geneticin (Invitrogen) for 2 weeks. To test transfection efficiency, a GFP construct (pTracer-GFP, Invitrogen) was also transfected into all three cell types using same protocol.
MLE-12 Cell Extract Preparation
MLE-12 extract was prepared from 80% confluent MLE-12 cells. MLE-12 cells were trypsin-digested, washed in ice-cold phosphate-buffered saline (PBS), and resuspended in one volume of lysis buffer (10 mM HEPES, pH 8.2; 50 mM NaCl; 5 mM MgCl2; and 1 mM dithioreitol, a cocktail of protease inhibitors and 0.1 mM phenylmethylsulfonyl fluoride; Sigma, Poole, U.K., http://www.sigmaaldrich.com). The cells were snap-frozen in liquid nitrogen for 2 minutes, thawed, and disrupted by vortexing. After centrifugation (15,000g for 15 minutes at 4°C), the supernatant (extract) was used fresh or frozen in liquid nitrogen and stored at −80°C . Protein concentration of the extract was 30 mg/ml (Bradford). A similar extract was prepared from MEFs for use as a negative control.
Reversible Permeabilization of EB Cells
To determine the optimal degree of permeabilization, embryoid body (EB) cells were exposed to increasing concentrations of streptolysin O (SLO) (0–1000 mg/ml) and permeabilization was assessed by uptake of a 70,000-Mr Texas Red–conjugated dextran (Molecular Probes Inc., Eugene, OR, http://probes.invitrogen.com), as described previously .
To initiate the differentiation, transfected ESCs were cultured in suspension for 10 days to allow EB formation and the resulting 10-day EBs were cultured in tissue culture T75 flasks for 3 days. These cells were then permeabilized at 5 × 105 cells per 500 μl with 800 ng/ml SLO in Ca2+-Mg2+-free Hanks' balanced salt solution (Gibco-BRL, Invitrogen) for 50 minutes at 37°C. SLO was replaced with 100 μl of MLE extract containing an ATP-generating system (1 mM ATP, 1 mM GTP, 1 mM NTP, 10 mM phosphocreatine, and 25 μg/ml creatine kinase, Sigma) and incubated for 60 minutes at 37°C. Controls comprised the MEF extract and the ATP-generating system alone. To reseal plasma membranes, 2 mM CaCl2 was added to ES culture medium, and cells were cultured overnight at 37°C . Extract-treated cells were cultured in ES medium for up to 14 days. When indicated, the extent of permeabilization was assessed by uptake of 50 μg/ml Texas Red–conjugated 70,000-Mr dextran after overnight recovery of the cells.
Total RNA was extracted from the cultured cells using RNeasy Mini kit (Qiagen, Crawley, West Sussex, U.K., http://www.qiagen.com), according to the manufacturer's instructions. For reverse-transcription–polymerase chain reaction (RT-PCR), ThermoScriptTM RT-PCR (Invitrogen) was used to synthesize cDNA from 1 μg total RNA. Semi-quantitative real-time PCR was performed on the Applied Biosystems GeneAmp 5700 using the SYBR Green PCR Master Mix (Applied BioSystems, Foster City, CA, http://www.appliedbiosystems.com). Real-time PCR primer sequences were as follows: SPC forward 5′-CTCCACTGGCATCGTTGTGT-3′, SPC reverse 5′-GGTTCCTGGAGCTGGCTTATAG-3′; aquaporin 5 forward 5′-CCACAGCTCAGACCTCAGAGATT-3′ reverse CCTTTCTGGGATGGTCCTGAA; eGFP forward 5′-CTGCTGCCCGACAACCA-3′, eGFP reverse 5′-ACCCAGTCCGCCCTGAGCAAAGAC-3′ and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) forward 5′-CAGAAGGAGATTACTGCTCT-3′, GAPDH reverse: 5′- GGAGCCACCACACA-3′. The lengths of amplicons designed for real-time PCR were short: SPC, 67 bp; aquaporin 5, 50 bp; eGFP, 73 bp, and GAPDH, 78 bp.
For real-time PCR analysis, the expression levels of each gene were normalized to the GAPDH gene and presented as the mean fold difference relative to control values. Data were analyzed in triplicate from three separate experiments.
Cells grown in multi-chamber slides were washed twice in PBS and fixed in 4% paraformaldehyde for 30 minutes. Immunofluorescence staining of cultured cells was carried out using polyclonal goat antibodies to pro-SPC (1:100; Santa Cruz Biotechnology Inc., Santa Cruz, CA, http://www.scbt.com), polyclonal rabbit antibodies to aquaporin 5 (1:200; Chemicon Europe Ltd., Chandlers Ford, Hampshire, U.K.), and murine monoclonal antibody to TTF-1 (1:100; Novocastra Ltd., Newcastle upon Tyne, U.K., http://www.novocastra.co.uk). Cells were incubated overnight at 4°C. The second layers comprised Texas Red–conjugated rabbit anti-goat IgG, rhodamine-conjugated donkey anti-rabbit IgG, and goat anti-mouse IgG (all 1:100; Santa Cruz Biotechnology Inc.), as appropriate. Nuclei were counterstained with 4,6-diamindino-2-phenylindole-2- (DAPI) (Vectashield; Vector Laboratories, Burlingame, CA, http://www.vectorlabs.com).
Preparations were viewed under a BX-60 epifluorescence microscope (Olympus, Southall, Middlesex, U.K., http://www.olympus.co.uk), and images were captured with an Axiocam digital camera (Carl Zeiss, Jena, Germany, http://www.zeiss.com) and analyzed using KS-300 software (Imaging Associates, Thame, U.K., http://www.imas.co.uk). GFP was viewed at an excitation wavelength of 490 nm, Texas Red at 570 nm, and nuclear counterstaining with DAPI at 360 nm. SPC/GFP-positive reprogrammed cells were counted at a magnification of ×100. For each of the six wells in plates from five independent experiments, 10 fields were randomly selected and a total of 2,000 cells was counted.
Reprogrammed cells at day 7 of culture were rinsed in 0.1 M phosphate buffered 0.15 M saline and fixed in 2.5% (v/v) glutaraldehyde in 0.1 M phosphate buffer for 25 minutes. Cells were scraped from culture wells and then post-fixed in 1% (v/v) osmium tetroxide in 0.1 M cacodylate buffer. After two rinses in washing buffer, the cells were scraped from wells, transferred to microfuge tubes, and centrifuged. Pellets were dehydrated through graded alcohols, followed by propylene oxide, and embedded in Araldite resin. Ultra-thin sections of silver-gold interference color were stained in saturated uranyl acetate in 50% (v/v) ethanol and Reynolds lead citrate and examined in a Philips CM10 transmission electron microscope (FEI Company, Eindhoven, Netherlands, http://www.feicompany.com).
Differences between groups were determined using a Student's t-test. A value of p < .05 was taken as significant.
Reversible Permeabilization of EB Cells
In the absence of SLO, few EB cells were labeled with Texas Red (Fig. 1A). With 200–700 ng/ml SLO, the cells were not sufficiently permeabilized, and only approximately 10% of cells were found to be labeled with Texas Red (Fig. 1B). At 800–900 ng/ml SLO, 80% of the cells were labeled with Texas Red (Figs. 1C, 1E, 1F), indicating that a high proportion of EB cells could take up the dextran and be successfully replated. Increasing the SLO concentration to 1000 ng/ml appeared to damage the cells, and fewer were labeled (Fig. 1D).
An eGFP reporter gene driven by the pneumocyte-specific murine SPC promoter was transfected into undifferentiated ES, MEF, and MLE-12 cells. As expected, no SPC/GFP expression was detected in transfected ESCs (Fig. 2A) or MEFs (Fig. 2B). However, SPC/GFP was expressed in transfected MLE-12 cells (Fig. 2C). All cell types expressed eGFP transfected alone, indicating that transfection efficiency was satisfactory (Figs. 2G–I). These results demonstrate that the SPC/GFP construct is specifically expressed in type II pneumocytes.
Induction of SPC/GFP Expression in Differentiated Cells
EB cells reversibly permeabilized with 800 ng/ml SLO were incubated for 1 hour in a nuclear and cytoplasmic extract of MLE-12 pneumocytes. Reprogramming was assessed by the induction of SPC promoter-driven eGFP expression in the EB cells. In cells exposed to MLE-12 extract, SPC/GFP expression was detected as early as day 3 of exposure to the extract, and the proportion of GFP-positive cells increased up to day 7 to 7.53 ± 1.14% (mean + SE) (Fig. 3Ac). No expression of SPC/GFP could be detected in cells reprogrammed with ATP-regenerating buffer (Fig. 3Aa) or MEF extract (Fig. 3Ab). Induction of SPC gene expression was also analyzed by real-time PCR. At day 7 of exposure to MLE-12 extract, SPC mRNA expression was markedly higher than that of control cells incubated in buffer only or MEF extract (Fig. 3B). SPC gene expression in cells exposed to buffer only or MEF extract was also found to be higher (3.28 ± 0.74 fold) than that of undifferentiated ESCs, as some of these cells had probably undergone spontaneous differentiation to pneumocytes. Furthermore, SPC mRNA expression of ESCs treated with 800 ng/ml SLO was higher (p < .005) than that of cells treated with 200 or 1000 ng/ml (Fig. 3C), corroborating our Texas Red–dextran uptake observations. Collectively, these results indicate that the MLE-12 extract is capable of inducing expression of a pneumocyte-specific promoter in ESCs.
The phenotype of SPC/GFP-transfected reprogrammed cells was examined by immunocytochemistry and electron microscopy. Pro-SPC and TTF-1, two markers for type-II pneumocytes, were detected in SPC/GFP-expressing reprogrammed cells (Figs. 4A–C, 4G–I). Both markers were also detected in control MLE-12 cells (tMLE-12) transfected with SPC/GFP, (Figs. 4D–F, 4J–L). In all cells, immunoreactivity of pro-SPC was found in the cytoplasm and that of TTF-1, a transcription factor, was found only in the nucleus. The phenotype of the pneumocytes was confirmed by electron microscopy that revealed the presence of organelles with the appearance of lamellar bodies in the cytoplasm of reprogrammed cells harvested after 7 days of culture (Figs. 4M, 4N).
At day 7, pro-SPC–expressing reprogrammed cells, assessed by GFP fluorescence and immunostaining of pro-SPC, reached a maximal proportion (5%–10%), with some of the cells showing very strong SPC/GFP expression. During the second week of culture, the proportion of SPC/GFP-expressing cells started to decrease, although some cells were still positive at day 14. Real-time PCR analysis showed a concomitant gradual decrease in SPC mRNA expression (Fig. 5B) with a more rapid decrease of GFP expression after day 7 (Fig. 5A). Our contention that this decline in SPC over 14 days of culture was due to differentiation of type II pneumocytes to the type I phenotype, was supported by the finding of an increase in aquaporin 5 mRNA, a marker for type I pneumocytes, during that period (Fig. 5C). The cells also changed their morphology during this time, becoming larger and thinner, an appearance consistent with that of type I pneumocytes, and were found to express aquaporin 5 protein (Figs. 5D and 5E).
This study demonstrates that murine ESCs can be encouraged to differentiate to lung-specific epithelium using an adaptation of a published method for in vitro cell reprogramming [9–11] incorporating exposure of the ESCs to an extract of alveolar epithelial cells. These findings present a potentially new approach to the control of ESC differentiation.
The specificity of the differentiation of ESCs into pneumocytes was supported by several sets of data. There was a significant increase in SPC mRNA expression, a specific marker for type II pneumocytes, in ESCs exposed to MLE-12 extract but not in ESCs exposed to buffer or fibroblast (MEF) extract. Variations in the extent of ESC plasma membrane permeabilization, and hence exposure to MLE extract, were paralleled by changes in the level of SPC mRNA expression. Acquisition of a pneumocyte phenotype was also confirmed by immunodetection of cytoplasmic SPC and nuclear TTF-1 and the presence in the cytoplasm of lamellar bodies, organelles specific to type II pneumocytes. This reflected genuine differentiation, and not direct uptake from the extract, because the high percentage of permeabilized cells (>80%) contrasted with the lower proportion of SPC/GFP-positive cells (5%–10%). Also, if uptake of SPC or another protein had occurred, SPC expression should have been detected immediately after exposure to the extract, as happened with the uptake of Texas Red-dextran, and without concomitant expression of GFP. Moreover, SPC/GFP expression cannot be accounted for by a putative uptake of SPC from the extract, but rather, occurs from a transfected ESC genome. Consistent with our results, recent results from Håkelien et al.  suggest that direct uptake of a phenotype-specific protein from the extract is unlikely to occur.
Continued culture of the reprogrammed cells up to 14 days resulted in a decrease in the expression of SPC RNA. Our contention that this was due to the well-known phenomenon of type II pneumocyte differentiation to the type I phenotype [12, 13] was confirmed by observations that the levels of aquaporin 5 RNA, a marker for type I cells , rose over the same period of time and thin, membranous cells showing immunoreactivity for the corresponding protein appeared in the cultures. As well as accounting for the decrease in SPC RNA expression, these findings reinforce the phenotyping of the type II pneumocytes by showing that they are able to differentiate to type I cells.
The exact mechanism by which the cell extract–mediated reprogramming or differentiation occurs is not yet understood. In our experiments, the pattern of SPC expression in ESCs exposed to MLE-12 extract suggests a contribution of transcription regulators from the extract. However, our results suggest that refining the approach, by analyzing the composition of MLE extract and modifying factors, such as extract volume and time of exposure, will form the basis of a relatively efficient and reproducible means to derive pneumocytes. Indeed, compared with our previous methods for deriving type II pneumocytes from murine ESCs [7, 8], in vitro reprogramming provides a major improvement in terms of both speed and pneumocyte yield.
In conclusion, our results show that cell extract–based, in vitro reprogramming provides a new, relatively efficient means to promote the differentiation of ESCs to a specific phenotype. With the aim of deriving alveolar epithelium, the extract may constitute a tool to enhance our understanding of the key factors and mechanisms involved in the differentiation of ESCs to type II pneumocytes.
This work was supported by the Medical Research Council (Cooperative Group G9900355 and Component Grant G0300106). The authors thank Prof. Austin G. Smith, Institute of Stem Cell Research, University of Edinburgh, Scotland, U.K., for the E14Tg2α ES cells; Prof. Jeffrey A. Whitsett, The Children's Hospital, Cincinnati, for the 4.8-kbmurine SPC promoter/GFP construct; Mrs. Margaret Mobberley, Histopathology Department, Charing Cross Campus, Imperial College, for expert electron microscopy, and Prof. Richard L. Lubman, Keck School of Medicine of the University of Southern California, for his advice on the manuscript. Authors Qin and Tai contributed equally to this article.