Heterogeneous Mixed-Lineage Differentiation of Mouse Embryonic Stem Cells Induced by Conditioned Media from A549 Cells

Abstract

Conditioned media (CM) of transformed cells, such as the human lung-derived A549 cells, is a useful tool for directing differentiation of embryonic stem cells (ESCs). Previous work indicates that A549-CM induced pulmonary differentiation of mouse ESCs (mESCs). In this study, we compared the effects of A549-CM treatment on the differentiation of mESCs organized in monolayer or embryoid bodies. We analyzed the cultures treated with A549-CM using specific lineage markers by quantitative polymerase chain reaction (qPCR) and lineage-focused PCR arrays and demonstrated heterogeneous CM-induced differentiation. We then constructed bioinformatics-based gene networks to establish correlations between the upregulated lineage-specific genes and proteins in the A549-CM identified by proteomic analysis. Network analysis supported the phenotypic and genotypic heterogeneic differentiation of mESCs into multiple cell lineages via enriched stemness, cardiovascular, neuronal, and lung development gene ontologies (GOs). The significant enrichment toward lung ontologies was specific for treatment with A549-CM, but not CM of liver (HepG2) and pancreas (Capan-1) cells. Based on network analysis, we identified laminin alpha5, prosaposin, lamin A/C, dickkopf homolog 1, clusterin, and calreticulin as the most relevant proteins related to the enrichment of lung GOs. We validated the effects of laminin isoforms on mESC differentiation in vitro and found enriched differential induction of surfactant protein gene expression. Our data suggest that A549-CM can be used for identifying secreted proteins for the heterogeneous mixed-lineage differentiation of mESCs toward a variety of lung-relevant cells. Such a heterogeneous cell population will be required for the in vitro generation of complex lung tissue and mixed cell populations for regenerative pulmonary therapy.

Introduction

Embryonic stem cells (ESCs) may contribute to the repair and regeneration of injured or diseased tissues including lungs [1]. It has been argued that the therapeutic potential of ESCs requires directed organ-specific differentiation into defined homogeneous [2] or heterogeneous populations with distinct cellular phenotypes [3]. Directed homogeneous or heterogeneous ESCs differentiation can be achieved through a variety of published differentiation protocols, including chemically defined cocktails of growth factors and small molecules [4], extracellular matrix components [5], physical cues such as hypoxia [6], matrix stiffness [7], and the use of serum-containing or serum-free conditioned media (CM) [8].

Media conditioned by normal or transformed cells have been used in stem/progenitor cell research as a tool for maintaining stemness [9], but also as a means for inducing differentiation toward specific cell lineages, such as neuronal [10], chondrogenic [11], hematopoietic [12], or distal alveolar epithelial (AE) cells [13]. CM, also referred to as secretome, is a complex collection of secreted proteins, including enzymes, growth factors, cytokines, hormones, and other soluble proteins, which participate in fundamental processes of cell proliferation, differentiation, communication, and migration in an autocrine or paracrine fashion [14]. Detailed proteomic analysis of the secretome of various primary or transformed cell lines is an important tool for the discovery of proteins involved in disease-specific biomarkers or directed stem cell differentiation [15].

CM from A549 cells, an established human lung adenocarcinoma cell line, induces differentiation of mouse and human ESCs (hESCs) toward lung phenotype, as demonstrated in monolayer culture by us [13] and others [16], and by Siti-Ismail et al. in three-dimensional (3D) aggregates [17,18]. The A549-CM promoted differentiation of mouse ESCs (mESCs) toward definitive endoderm (DE) (∼55% of the total cell number) through suppression of visceral endoderm differentiation and yielded a heterogeneic cell population comprising ∼11% and 50% of lung type II AE-like cells, in monolayer [13] and 3D cultures [18], respectively. In both culture conditions the rest of the cell phenotypes remained mainly undetermined.

The focus of this study was to identify differentiation-promoting factors present in A549-CM in order to propose a rationally designed, chemically defined cocktail that can eventually be applied to efficiently derive one or several types of lung-relevant cells from murine and human pluripotent stem cells. We hypothesized that the A549-derived secretome exerts pleiotropic effects in terms of the ability of mESCs to differentiate in both monolayer and embryoid body (EB) cultures. In testing this hypothesis we investigated the effects of A549-CM on the expression of cell-lineage-related gene markers of mESCs cultured as monolayer and EBs. We then used a system biology approach, to describe the complex relationship between select secreted proteins found in A549-CM and a subset of the cell-lineage-related gene markers differentially expressed in mESCs cultured in various spatial organizations (2D vs. 3D) in the presence of A549-CM. Using a network analysis of our genotypic findings, we were able to link some of the proteins in the A549-CM, in particular, an alpha5 chain-containing laminin (LM) isoform, with the differentiation of mESCs into mixed heterogeneic lineages and phenotypes. We then validated in vitro the effects of several LM isoforms on mESC differentiation and found differential induction of surfactant protein gene expression, that is, induction of differentiation of mESCs toward pulmonary cells. This study highlights the complexity of CM-driven mESC differentiation and provides a rational explanation for A549-CM-induced differentiation of mESCs into heterogeneic cell populations with enriched presence of cells capable of differentiating toward pulmonary-related lineages.

Materials and Methods

mESC culture

mESCs, E14-12ΔS (kindly provided by Dr. Paul Gadue, Children's Hospital of Philadelphia, Philadelphia, PA), were cultured as previously described [19]. Briefly, the cells were maintained in a serum-free/feeder-free culture system on 0.1% gelatin-coated (Milipore), tissue-culture-treated plastic. The serum-free media was composed of knockout Dulbecco's modified Eagle's medium (DMEM)/F12 media supplemented with 0.5×N2, 0.5×B27, and 0.05% bovine serum albumin (Invitrogen); 50 IU/mL penicillin and 50 μg/mL streptomycin (Cellgro); 2 mM l-glutamine (Invitrogen); 10 ng/mL human recombinant bone morphogenic protein-4 (BMP-4) (R&D Systems); 1,000 U/mL ESGRO^® mouse leukemia inhibitory factor (LIF) (Millipore); and 0.15 mM 1-thioglycerol (Sigma). The maintenance media was changed daily and cells were split every 2–3 days (upon reaching 80% confluence) at a ratio of 1:6 using TrypLE Express (Invitrogen). The cells were maintained in a humidified incubator at 37°C in 95% air/5% CO₂ atmosphere. The cell culture was evaluated visually and photomicrographs were taken using a Nikon Eclipse TE 2000-U (Nikon) inverted microscope connected to Hitachi KP-D50 digital camera.

Preparation of CM

Transformed cell lines A549 (ATCC CCL-185, human lung adenocarcinoma), HepG2 (ATCC HB-8065, human liver hepatocellular carcinoma), and Capan-1 (ATCC HTB-79, human pancreas adenocarcinoma) were purchased from ATCC and cultured in T-175 tissue culture-treated flasks in a humidified incubator at 37°C in a 5% CO₂ atmosphere. All cells were routinely maintained in DMEM supplemented with 10% fetal bovine serum, 100 IU/mL penicillin and 100 μg/mL streptomycin, and 2 mM l-glutamine. The culture media was exchanged every other day and cultures were passaged upon reaching 80% confluence at a ratio of 1:10 using TrypLE Express (Invitrogen). To collect serum-free CM, 90% confluent flasks were rinsed three times with Hank's balanced salt solution (HBSS) containing calcium and magnesium (Cellgro), followed by three washes for 10 min each with DMEM supplemented with antibiotics only. The efficiency of serum removal was routinely monitored by analyzing CM samples by sodium dodecyl sulfate–polyacrylamide gel electrophoresis. Final bovine serum albumin levels never exceeded 0.10%±0.01% of the maintenance media. To initiate CM collection, 30 mL of serum-free, antibiotics-containing DMEM was added to the washed cell cultures for no longer than 48 h. Thereafter, CM was collected, filtered through sterile low protein-binding 0.2-μm syringe filters (Corning), stored at −20°C, and thawed prior to use. In preliminary experiments we established that the viability of the culture at the end of CM collection process was >90%, that is, comparable to that of the maintenance cultures, as determined using the alamar blue assay (AbD Serotec) [20].

Mass spectroscopy

A549-CM was analyzed by mass spectroscopy (MS) according to previously established protocols [21] (described in detail in Supplementary Materials and Methods; Supplementary Data are available online at www.liebertpub.com/scd). Briefly, the samples were reduced and digested with sequencing grade-modified trypsin (Promega). The tryptic peptides were desalted, dried, resuspended, and resolved by reverse-phase chromatography on 0.075×200 mm² fused silica capillaries (J&W; Agilent Technologies) packed with Reprosil reversed phase material (Dr. Maisch GmbH). The peptides were eluted with a linear gradient followed by MS on an ion-trap mass spectrometer (Orbitrap; Thermo) in a positive mode repetitively using full MS scans followed by collision-induced dissociation of the seven most dominant ions selected from the first MS scan. The MS data were analyzed using SEQUEST 3.31 software searching against the human section of the UniProt database [22] and filtered according to their Xcore values for any given peptide scan. Proteins were considered positively identified based on at least two peptide hits obtained with a good score, where a good score is 1.2 for single charged, 2.0 for double charged, 2.5 for triple charged, or 3.2 for quadruple charged species. Three separate batches of A549-CM were analyzed by MS, and only proteins that appeared in at least two CM batches were considered for further analysis. To generate a focused secretome, which included only proteins with high confidence of extracellular provenance, the whole secretome was compared for reproducibility with two previously published proteomic analyses of the A549-CM that had been filtered to include extracellular proteins [23,24] and for residing in the extracellular space according to the Ingenuity Pathways Analysis (IPA; Ingenuity^® Systems) and Uniprot databases. To be considered part of the focused secretome, in addition to identification in this study, each protein had to appear in two or more of these sources.

Differentiation experiments

The effects of A549-CM on mESC differentiation were studied in both monolayer and EB cultures. For monolayer cultures, single-cell suspensions of mESCs were plated in maintenance media at a density of 10,000 cells/well in 0.1% gelatin-coated six-well plates (1,000 cells/cm²) and left to adhere overnight. The next day (considered day 0), the maintenance media was replaced with various experimental media and cells were maintained under these conditions for 2 or 10 days, unless mentioned otherwise. The experimental media consisted of either spontaneous (SP) media, which was LIF- and BMP-4-free maintenance media, or 50% of the A549-CM diluted with two-fold-concentrated SP media. LM isoforms 111, 211, 411, 421, 511, and 521 (BioLamina) at a concentration of 500 ng/mL each were prepared in SP media and monolayer cultures were treated for 10 days. EBs were generated in different experimental media using the hanging-drop method, as previously described [25]. Briefly, 1,000 mESCs/33 μL (one drop) were dispensed on the lid of a 10-cm Petri dish, inverted and allowed to form EBs for 48 h in a humidified tissue culture incubator. Upon formation of the EBs (considered as day 2 of differentiation), 40 EBs were frozen at −80°C until processed for RNA isolation and further analysis. In parallel, 20 EBs/well were transferred to 0.1% gelatin-coated six-well plates and cultured for 8 more days, thus being cumulatively exposed to the experimental media for a total of 10 days. Distinct morphological appearances of the cultures either in monolayer or EB configuration under different experimental conditions were visualized by phase-contrast microscopy. Images were digitally acquired from at least 10 different fields in the same well at 100×magnification using a Nikon Eclipse TE 2000-U microscope. Differentiation experiments with mESCs treated with other CMs (HepG2 and Capan-1) were performed in monolayer culture in similar fashion as for A549-CM.

Reverse transcription–quantitative polymerase chain reaction

Total RNA was isolated using RNeasy mini kits (Qiagen) according to the manufacturer's instructions. RNA concentration was determined using a BioPhotometer (Eppendorf). cDNA was reverse transcribed from 1 μg of total RNA using random primers and a high-capacity cDNA reverse transcription kit (Applied Biosystems). Reverse transcription–quantitative polymerase chain reaction (RT-qPCR) was performed in an Eppendorf Mastercycler Ep Realplex II system (Eppendorf) with fast thermal cycling setup using the following Taqman inventoried and validated primers according to manufacturer's instruction (Applied Biosystems): Pou5f1 (Mm03053917), T (brachyury) (Mm01318252), Foxa2 (Mm00839704), Sox17 (Mm00488363), Cxcr4 (Mm01996749), Foxc1 (Mm01962704), Foxg1 (Mm02059886), Sox7 (Mm00776876), Hnf1a (Mm00493434), Pdx1 (Mm00435565), Nkx2.1 (Mm00447558), Sftpa (Mm00499170), Sftpb (Mm00455681), Sftpc (Mm00488144), Aqp5 (Mm00437579), and Scgb1a1 (Mm00442046). For each sample qPCR was performed in triplicate and each experimental condition was independently repeated at least three times. Gene expression levels were normalized to an endogenous reference/house-keeping gene, peptidylprolyl isomerase A (cyclophilin A, Ppia) (Mm02342430_g1). Relative expression levels were calculated as fold of spontaneously differentiating culture (unless otherwise mentioned) using the 2^−ΔΔCT method.

Focused PCR array

The Mouse Cell Lineage Identification PCR Array (PAMM-508A, SABiosciences; see www.sabiosciences.com/rt_pcr_product/HTML/PAMM-508A.html) was used to study the effects of A549-CM on the gene expression profiles of mESCs under various culture conditions. The 96-well PCR array plate contained 84 pathway-focused genes, 5 housekeeping genes, a genomic DNA control, 3 RT controls, and 3 positive PCR controls. Total RNA from EBs and monolayers cultured for altogether 10 days, either in SP media or 50% A549-CM, was isolated in triplicate, using RNeasy mini kits (Qiagen). RT and qPCR gene expression profiling was performed according to the manufacturer's instructions. Briefly, 1 μg of total RNA was reverse transcribed using RT² first-strand kit (SABiosciences). The synthesized cDNA was mixed with RT² SYBR green qPCR master mix (SABiosciences) and a total of 25 μL was loaded into each well of a 96-well PCR array plate. The qPCR was performed in an Eppendorf Mastercycler (Ep Realplex II) according to the manufacturer's instructions with a two-step cycling program starting at 95°C for 10 min for 1 cycle and then 40 cycles of 95°C for 15 s and 60°C for 1 min, followed by generating the dissociation curves to assess the specificity of PCR products.

Array and integrated secretome-transcriptome network analyses

Generation of K-means clustered heat maps

The data from the arrays were initially processed as recommended by the manufacturer. Briefly, the ΔC_t was calculated by subtracting the C_t value of each gene from the mean C_t value of the housekeeping genes on each array, which yielded a coefficient of variance between all analyzed arrays of 2.1%. The expression values of the identified mRNA species were then averaged for biological replicates and used for further analyses. For ease of visualization and to better cluster by expression dynamics, ΔC_t values were standardized to a mean of 0 and a standard deviation of 1 for each gene. A heat map was generated from the standardized expression values and clustered using a K-means clustering algorithm. The K constant was evaluated for the range of 1–15, with the upper limit set to omit clusters composed of a single gene. The optimal K was chosen as the number corresponding to the maximized mean group silhouettes (ie, averaged degree of fitness within each cluster). Genes were color-coded based on standardized expression values on a scale from −1.5 as green to +1.5 as red with further outliers pegged to the maximum or minimum.

Gene expression dynamic profile analysis

The genes grouped into the K-means clusters were further analyzed for effects of SP and CM treatments on gene expression dynamics. Gene expression converging and diverging dynamic profiles were conceptually based on Narayan et al. [26] and further modified and adjusted for our set of PCR array data. To be considered as converging dynamic the following conditions had to be fulfilled: (1) gene expression was upregulated following CM treatment in both EB and monolayer cultures by at least 1.5-fold versus the respective SP culture and (2) gene expression difference between CM-treated groups was at least 50% less than the differences seen under SP conditions. The diverging dynamic was separated according to genes favoring (ie, higher expression values) EB or monolayer culture. Thus, diverging genes favoring EBs were selected according to the following criteria: (1) the difference in folds of expression between EB and monolayer cultures under CM treatment was at least 50% larger than the fold change between the two culture conditions under SP conditions, (2) the expression levels in EBs were at least 50% larger than in monolayer culture under CM treatment, and (3) gene expression was changed following CM treatment in EBs by at least 1.5-fold as fold of SP culture. Divergent genes favoring monolayer culture were selected similarly.

Integrated secretome-transcriptome network analysis

Data analysis and identification of network formation were initially based on Zoldan et al. [7] and further adapted for our unique approach to integrate proteomic with gene expression data. Upregulated genes of both converging and diverging dynamic profiles were considered for further gene ontology (GO) and integrated secretome-transcriptome analysis. All gene expression processing was performed using MatLab 2012b (MathWorks). To enable an integrated approach, the identified set of upregulated genes in the arrays was connected with proteins identified in A549-CM via protein interaction networks involving no more than two intermediate genes for the whole or the focused secretome, respectively. The integrated network was constructed using the exploratory gene association network (EGAN) software version 1.4 for Mus musculus [27]. To address a potential interspecies mismatch, the human genes were converted into mouse homologs or vice versa using NCBI homologene database and the data were imported according to EGAN default settings using NCBI EntrezGene ID database for upregulated genes and identified proteins. The intermediate nodes on the network were derived from Biogrid, BIND, HPRD, Intact, or MINT database, as included in the EGAN software. The identified protein–gene hybrid networks were overlaid with enriched development-related biological process GOs. To be considered as enriched, at least four members of a particular ontology were present in the network and the ontology was significantly over-represented as determined using Fischer's exact test with P<0.05. Area-proportional Euler diagrams were generated using VennMaster software version 0.38.0 with particle swarm optimizer and 2,000 iterations for each diagram [28]. The raw diagrams were exported to Adobe Photoshop CS3 version 10.0.1 and overlayed using multiply function for final presentation of colors.

Statistical analyses

The data are presented as the mean±SEM, when applicable. Statistically significant differences between experimental groups were determined by analysis of variance (ANOVA) test followed by Bonferroni post-test, unless otherwise mentioned, using InStat statistics program (GraphPad). Significance was considered for P<0.05. Each experiment was independently repeated at least three to five times in triplicates, unless otherwise specified.

Results and Discussion

Effects of A549-CM on mESC lineage marker expression in monolayer and EB cultures

One of the defining features of ESCs is their ability to form 3D EBs, which mimic early stages of embryonic development and lead to a unique gene expression profile that is not recapitulated in 2D monolayer cultures [29]. Therefore, we compared the morphology and gene expression patterns of mESCs under SP conditions or exposed to A549-CM for 10 days either in monolayer or EB culture (Fig. 1). Visual inspection of the two cultures at day 10 under conditions of SP differentiation and exposed to A549-CM indicated the presence of several morphologically distinct phenotypes (Fig. 1A). Under SP conditions, most cells emanating from EBs exhibited elongated, flat morphologies, while a few of the other cells had neurite-like projections. A similar neuronal phenotype was occasionally also seen in SP monolayer cultures (Fig. 1A). In A549-CM-treated EB cultures, most cells exhibited a dense, epithelial-like phenotype, similar to cells in monolayer cultures; however, there were more cells with well-defined neuronal-like projections than in the monolayer (Fig. 1A). The distinct morphologies observed under the various culture conditions (EBs vs. monolayers exposed to either SP or A549-CM treatment) suggest the concomitant presence of heterogeneous cell populations of mixed multiple lineages, which to some extent are unique for each condition. The diverse morphologies seen in the monolayer and EB outgrowth cultures (Fig. 1A) may be explained by differential induction of specific lineages under distinct spatial organization, as previously reported for several hESC lines [30,31].

FIG. 1.

Effect of A549-CM on mESC morphology and differentiation in monolayer and EB cultures. (A) mESCs were cultured in monolayer at low density of 1,000 cells/cm² or 20 EBs/well under LIF- and BMP-4-free (SP) or 50% A549-CM (A549-CM) conditions. Phase-contrast photomicrographs were acquired at day 10 of culture from at least 10 different fields of each well. Shown are representative images depicting the two major distinctive morphologies identified in the same well. The photomicrographs were taken at 100×magnification; scale bar is 100 μm. (B) Quantitative RT-qPCR analysis of Pou5f1, T (brachyury), Foxa2, Sox17, Cxcr4, Foxc1, Foxg1, and Sox7 gene expression at day 10 from SP and 50% A549-CM-treated monolayer (white and light gray bars, respectively) and SP and 50% A549-CM-treated EBs (dark gray and black bars, respectively) cultures. Dashed lines indicate two-fold up- or downregulation. The analysis was performed on at least three different independent experimental samples and was essentially similar. The data are expressed as folds of expression relative to mESC maintenance culture (day 0). Error bars represent one standard deviation from the mean of technical replicates (n=3), *P<0.05 CM versus SP. BMP-4, bone morphogenic protein-4; CM, conditioned media; EBs, embryoid bodies; LIF, leukemia inhibitory factor; mESCs, mouse embryonic stem cells; RT-qPCR, reverse transcription–quantitative polymerase chain reaction; SP, spontaneous.

To study the pleiotropic effects of A549-CM, we evaluated at day 10 of treatment the expression of the following surrogate markers: stemness (Pou5f1) and lineage-specific differentiation such as early mesoendoderm (brachyury T), DE (Foxa2, Sox17, and Cxcr4), mesoderm (Foxc1), ectoderm (Foxg1), and visceral endoderm (Sox 7) [19] (Fig. 1B). At day 10 of A549-CM treatment the levels of Pou5f1 were upregulated as compared with SP cultures by 3.3±0.4- and 18.8±6.7-fold in monolayer and EB cultures, respectively. The persistent presence of a marker for self-renewal in the A549-CM-treated cultures does not negate concomitant progression of a particular differentiation program, as previously reported for media conditioned by HepG2 cells [12,32,33]. Cultures of EBs under A549-CM treatment exhibited a significantly higher expression of ectoderm marker gene (Foxg1) as compared to monolayer, which is in line with the enriched presence of neuronal-like morphologies (Fig. 1A). The marker genes for mesoendoderm (brachyury T) and DE (Foxa2, Sox17, and Cxcr4) were significantly upregulated in A549-CM-treated EBs as compared with EBs under SP conditions and monolayers exposed to A549-CM. However, the expression of DE marker genes in monolayer cultures treated with A549-CM was significantly lower than in monolayers under SP conditions, while the gene marker of mesodermal progenitors (Foxc1) was upregulated by 4.5±0.2-fold in monolayers and not altered in EBs. In parallel, in both monolayer and EB culture conditions, treatment with A549-CM resulted in significant downregulation of the visceral endoderm gene marker (Sox7) by 9.5±1.1- and 6.2±1.9-fold, respectively, as compared with SP cultures (Fig. 1B).

In view of the reported capacity of A549-CM to induce pulmonary differentiation in mESCs [13,18], the downregulation of Sox7 is not unexpected, since the lung contains no cells derived from visceral endoderm [34]. Thus, reduction of visceral endoderm is a desired outcome for generating lung-relevant cell phenotypes [35]. The lower expression of DE markers in monolayer cultures treated for 10 days with A549-CM as compared with SP cultures (Fig. 1B) does not contradict the previously reported A549-CM-driven pulmonary differentiation of mESCs [13,16]. Prior to differentiation toward lung epithelium, the endodermal progenitors undergo anterior–posterior segregation of the primitive foregut and then progress through a primordial progenitor stage [36]. At this stage, the markers of DE are downregulated, while more lung phenotype-specific markers, such as Sftpb, Sftpc, Scgb1a1 (CCSP), Cftr (cystic fibrosis transmembrane conductance regulator), and Foxj1, are upregulated, indicating progression of the cells toward pulmonary epithelium differentiative state [37]. Several recent studies using mESCs indicate that the expression of DE markers prior to pulmonary differentiation is restricted to a narrow “window of opportunity” of 4.5–5.5 days in culture [37,38]. We therefore suggest that the observed reduced expression of DE markers in monolayer cultures treated with A549-CM indicates that in the 2D configuration at day 10 the relevant cell populations may already have differentiated into some of the endoderm-derived pulmonary cell lineages. On the other hand, the fact that our mixed cell populations still contain early progenitors of mesodermal and ectodermal origin is in line with the notion of the slower kinetics of mesoderm and ectoderm differentiation under endoderm-favoring conditions [37].

Effects of A549-CM on mESC lineage marker gene expression patterns

To further characterize the effects of the A549-CM on mESC lineage/fate commitment in more detail, the cells were cultured for 10 days, as earlier, either as EBs or in monolayer and exposed to SP or CM. Differential expression of 84 select lineage marker genes was evaluated using a mouse cell lineage identification PCR array (for details see “Materials and Methods” section). We first sought to investigate the trend in global changes in gene expression (ΔCt) via distinguishing between the CM versus SP treatments and the spatial organization, EBs versus monolayer (Fig. 2). Under SP conditions EBs and monolayers displayed diverse patterns of gene expression with poor correlation (R ²=0.5813, Fig. 2A—green symbols), while treatment with A549-CM yielded a higher correlation (R ²=0.8113) and a significant compression (reduction) of the variance (P<0.0001 by F-test, Fig. 2A—blue symbols).

FIG. 2.

Correlations of A549-CM-induced gene expression changes between monolayer and EBs or CM- and SP-treated cultures. (A) Averaged ΔCt values from three independent experiments of cell lineage identification focused PCR array. Green and blue symbols represent SP and A549-CM (CM), respectively. R ²=0.5813 for SP (green line); R ²=0.8113 for CM (blue line; P<0.0001). (B) Averaged −ΔΔCt values from three independent experiments of cell lineage identification focused PCR array comparing A549-CM-treated monolayers and EBs. Linear correlation coefficient R ²=0.0674. For both panels, solid and dashed lines are linear regression and 95% prediction interval, respectively, and the symbols are circles—stemness, triangles—endoderm, squares—mesoderm, diamonds—ectoderm, and stars—neuroectoderm. Color images available online at www.liebertpub.com/scd

Based on these findings we hypothesized that A549-CM effectively suppresses the differences in the expression of the cell lineage markers seen under SP conditions between EB and monolayer cultures and promotes a more uniform signature/trend in expression of the analyzed lineage marker genes. To test this hypothesis, we evaluated the correlation between spatial organization and the effect (ΔΔCt) of the A549-CM on gene expression. The lack of a statistically significant correlation between CM-treated EB and monolayer cultures (R ²=0.0674) in Fig. 2B suggests that the A549-CM-induced differential effects on the expression of lineage marker genes depend on the spatial organization of the mESC cultures. This result is in line with the previously described distinct changes between A549-CM-treated EBs and monolayers, observed by qPCR at the level of a few individual lineage marker genes, such as T, Foxa2, Sox17, Cxcr4, and Foxg1 (Fig. 1B).

To identify genes grouped according to distinct dynamic profiles, the standardized gene expression values from the gene expression arrays were clustered using K-means clustering, a classic algorithm based on Euclidean distance, which is widely used to cluster microarray data [39]. To maximize the clustering efficacy, we measured group silhouettes for K ranging between 1 and 15 (as described in “Materials and Methods” section) and found that the gene expression data were clustered most effectively into seven clusters. The clustered heat map (Fig. 3A) depicts differential gene expression throughout all germ layers between cultures grown under SP and A549-CM treatments in EB or monolayer conditions. The identified seven clusters capture the major patterns in data variability among the samples, with each cluster exhibiting a unique dynamic behavior. These dynamic clusters can be grouped into sets of averaged converging or diverging end-point expression values (Fig. 3B and Supplementary Table S1): (1) convergence to a center point regardless of starting position (clusters 3 and 5), (2) convergence to an upregulated value (clusters 4 and 7), and (3) divergence to different levels of expression (clusters 1, 2, and 6). This analysis substantiates and further extends the notion that A549-CM induced a unique pattern in gene expression, which in turn may explain the lack of a linear correlation between A549-CM-treated EB and monolayer cultures (Fig. 2B).

FIG. 3.

Effect of A549-CM treatment on mESC marker gene expression patterns and dynamic profiles. (A) Heat-map presentation of normalized ΔCt values for EB and monolayer cultures maintained for 10 days under SP conditions or in the presence of 50% A549-CM (CM). The genes are clustered according to K-means clustering algorithm and the clusters are enumerated above the heat map and gene names are listed below the heat map. The data are an average of three independent experiments. (B) Average standardized expression values of genes from cell lineage identification focused PCR arrays organized according to the K-means clusters. Each cluster is subdivided into the SP and A549-CM (CM) treatments. Black and white bars represent monolayer and EB cultures, respectively. The data are mean±standard deviation (n=3) of the standardized gene expression for all genes that are a member of the cluster. Color images available online at www.liebertpub.com/scd

Network analysis-based correlation between A549-CM-derived proteins and enriched developmental GOs

To further characterize the effects of A549-CM on mESC differentiation, we sought to correlate the proteins contained in the A549-CM with the expression of lineage-specific marker genes of mESC differentiation. For this, we first performed a mass spectroscopic analysis of the CM and positively identified 176 proteins in the A549-derived CM (Supplementary Table S2). Second, based on the mESC gene expression results (Fig. 4A), we focused on upregulated converging (Fig. 4D) or diverging genes that were modulated by A549-CM treatment of either EB (Fig. 4B) or monolayer cultures (Fig. 4C) (see Supplementary Table S3). Thereafter, we generated a protein–gene hybrid network connecting all identified proteins with all identified upregulated genes collected from the PCR array, expanded by up to one intermediate (Supplementary Fig. S1). The generated interaction network was further used to detect significantly enriched ontologies (Supplementary Table S4) and identified enrichment in 36 GO terms. However, in-depth analysis of this network indicated that 72% of the identifiable connections were related to intracellular proteins (cytoskeleton components, eg, filamin A; enzymes, eg, alpha-enolase; or nuclear proteins, eg, heterogeneous nuclear ribonucleoprotein L) that are present in the CM, presumably as debris from cell death and inevitably collected during the production of the CM. To exclude interactions, which may be biologically irrelevant in our experimental system, we narrowed the list of identified proteins to 29 entries, which were uniquely classified as “secretome” according to two databases (IPA or Uniprot) or previously published A549 secretome analyses [23,24] (Supplementary Table S2).

FIG. 4.

Convergence and divergence analyses of A549-CM-induced gene expression in mESCs. (A) Correlation plot of averaged values from three independent experiments of cell lineage identification focused PCR arrays. The data, presented as relative expression (Rel. exp.) of cultures under SP and A549-CM (CM) conditions, are expressed as relative expression in EBs as fold of monolayer cultures. Light blue area—diverging genes favoring EBs; light green area—diverging genes favoring monolayer; pink area—converging genes between EBs and monolayer. The solid line indicates 1-fold change (ie, no change) and dashed lines are±1.5-fold change. The symbols are circles—stemness, diamonds—endoderm, squares—mesoderm, upright triangles—ectoderm, and inverted triangles—neuroectoderm. The data from light-color-coded areas in (A) were reanalyzed and presented as relative expression (Rel. exp.) of A549-CM-treated EB and monolayer cultures expressed as fold of respective SP culture. Dark-color-coded areas represent the specifically upregulated subset of genes in CM-treated cultures: (B) diverging genes favoring EBs (dark blue area), (C) diverging genes favoring monolayer (dark green area), and (D) converging genes (dark pink area). Symbols are similar to (A) and lines indicate±1.5-fold change. Color images available online at www.liebertpub.com/scd

This focused secretome was then used to construct more comprehensive A549 secretome–mESC gene expression interaction networks via expanding the narrower network by up to two intermediates for converging genes, diverging genes favoring monolayers, and diverging genes favoring EBs, according to the classification in Fig. 4. Upon overlaying these networks with developmentally related GO terms, we found a total of 60 terms that were significantly enriched (Supplementary Table S5). Comparative analysis of the enriched ontologies in the three networks suggested a high degree of functional similarity; 79% and 70% of the ontologies enriched in the converging network were also enriched in the networks for diverging toward monolayer or EB culture, respectively (Fig. 5A). In addition, the secreted proteins, which exhibited connections with the network of converging genes, had an overlap of 92% and 85% with the proteins exhibiting connections with diverging monolayer or EB network, respectively (Fig. 5A). The converging group consisted of a more diverse set of GO annotations than each diverging group by itself (Supplementary Table S5), providing additional evidence for a specific pattern of gene expression following treatment with A549-CM and compression of variances between the two culture conditions, as discussed earlier. The high percentage of overlapping ontologies assigned for the diverging sets of genes supports the conclusion that A549-CM provides strong and shared differentiation cues to both EB and monolayer cultures.

FIG. 5.

Interaction networks between A549-secretome and enriched mESC GOs. (A) Area-proportional three-component Euler diagrams of statistically significant GOs and proteins derived from converging (red), diverging favoring monolayer (green), and diverging favoring EBs (blue) protein–gene interaction networks. (B) Network analysis of focused secretome and statistically significant enriched major organ development GO terms assigned for significantly affected upregulated converging genes. The color-coded lines indicate the distance between the protein and the GO term: green—immediate member, red—one intermediate, and blue—two intermediates. GO, gene ontology. Color images available online at www.liebertpub.com/scd

Given the concomitant assignment of a large number of identical GO terms and the high degree of overlap of CM-derived proteins, we performed a focused bioinformatics-based analysis of development-related GO terms using the converging gene set, to explore in more detail the connections specific for proteins in the A549-CM and the development of major organs (Fig. 5B). This analysis supports our phenotypic and genotypic findings (Fig. 1) by linking some of the proteins in the A549-CM with the differentiation of mESCs into mixed heterogeneous lineages and phenotypes. In addition, we traced and color-coded the distance between connected proteins and the genes that are members of each GO term (Fig. 5B), where green indicates immediate member, red—one intermediate, and blue—two intermediates. The relative distance indirectly implies the probability for such connection to occur, with green being the highest, red being the middle, and blue being the lowest. Thus, we identified interactions of high probability between the secreted proteins and GO terms relevant for lung, heart, blood vessel, and brain development. Enrichment of brain development is in line with our observations of cells with neurite-like projections under A549-CM treatment (Fig. 1). Of particular interest is the significant enrichment of the lung alveolus development GO term, as it is the most advanced term on the biological-process-related ontology tree, as compared with other enriched annotations. This particular observation suggests a trend toward generating terminally differentiated cells, supporting previous findings that mESCs at day 10 of treatment with A549-CM expressed AE differentiation marker genes [13,18].

In addition to lung development ontologies, the significant enrichment of the blood vessel development GO term is of importance. The presence of vascular progenitors (and vascularization in general) in these mixed cell populations may contribute for lung tissue engineering, since vascularization is critical for generating functional and transplantable bioengineered lung tissue [40]. The enrichment in heart and blood vessel developments may relate to the previously reported proangiogenic capacity of A549-CM [41]. Indeed, preliminary analysis of A549-CM-differentiated mESCs, in monolayer or EB configuration, indicates a significant upregulation of markers of vascular endothelial cells, such as VE-cadherin (Karamil and Lelkes, unpublished data).

Network analysis of the connections between A549-CM-derived proteins and the enrichment of lung development GO

One of our long-term goals is to identify novel factors that may promote lung-specific differentiation. We therefore performed a more detailed network analysis of the lung development GO term. We constructed a protein–gene hybrid network to visualize the distance between potentially involved proteins found in the A549 secretome and mESC genes displaying dynamic profiles of convergence and divergence favoring EB or monolayer genes, which were part of the lung development ontology (Fig. 6A). This analysis led to the identification of high-probability connections for converging and diverging genes favoring EBs or monolayer with six, eight, and five specific proteins in the A549 secretome, respectively (Fig. 6A—green and red lines). The diverging genes favoring monolayer exhibited only high-probability (green and red) connections, due to limited size of this gene set (five genes, Supplementary Table S3), while the diverging genes favoring EBs and the converging gene sets were connected via additional 7 and 11 lower probability connections, respectively (Fig. 6A—blue lines).

FIG. 6.

Comparative analysis between media conditioned by A549, HepG2, and Capan-1 cells. (A) Network analysis of lung development GO enrichment for focused A549-secretome and mESC gene sets of converging, diverging favoring monolayer and EB cultures. The color-coded lines indicate the distance between the protein and the GO term in the connected network: green—immediate member, red—one intermediate, and blue—two intermediates. (B) Comparison of the A549-CM-derived proteins connected to lung development GO term enrichment and proteins identified by proteomic analysis of media conditioned by HepG2 and Capan-1 cells. The presence or absence of a protein in each of the CMs is indicated by orange or light gray colors, respectively. The upper panel represents the distance between the protein and the GO term for converging set of genes as in A. (C) Quantitative RT-PCR analysis of Sftpb, Hnf1a, and Pdx1 gene expression at day 10 from monolayer cultures treated with 50% CM derived from A549 (black bars), HepG2 (dark gray bars), and Capan-1 (light gray bars). Dashed lines indicate twofold up- or downregulation; only values beyond this range were considered as significant compared with SP control cultures. The analysis was performed on at least three different independent experimental samples yielding essentially similar results. The data are expressed as folds of expression relative to SP culture at the same time point. Error bars represent one standard deviation from the mean of technical replicates (n=3), *P<0.05 versus SP. Color images available online at www.liebertpub.com/scd

To explore the potential relevance of our results for hESCs, we performed a hypothetical network analysis connecting the same A549-secreted proteins with lung development ontology, using the same set of upregulated converging homologous human genes. According to this hypothetical analysis, hESCs treated with A549-derived CM exhibit a similar enrichment in lung development ontology. Actually, on average the distance between this ontology and the secreted proteins was shorter for hESCs than for mESCs. Hence, we posit that our results for mESCs can be extrapolated to hESCs.

Of these secreted differentiative proteins, LM containing the alpha5 chain (LAMA5) has specifically been associated with lung development and differentiation (Fig. 6A—green lines). The abnormalities found in LAMA5-knockout mice underscore its essential role in normal early lung development, normal alveolarization, distal epithelial cell differentiation, and late-stage pulmonary maturation [42 –44]. The unique importance of the alpha5-chain containing LM isoform in lung development is further highlighted by the fact that laminin alpha4- or laminin gamma2-null mice develop minimal, if any, obvious lung defects [44].

The remaining secreted proteins are connected to lung development ontology, with high probability, via one intermediate (Fig. 6A—red lines). For example, prosaposin (PSAP) reportedly interacts with Smad2 [45]. Previous studies indicated that PSAP is a potent neurotrophic factor [46], which is in line with the annotation of brain development ontology (Fig. 5B), suggesting that PSAP may contribute to the neuronal-like phenotypes in our cultures (Fig. 1A). On the other hand, PSAP reportedly interacts with Cftr [47] and is identified in the bronchoalveolar lavage fluid [48]. To the best of our knowledge, the role of PSAP in lung physiology or stem cell differentiation remains to be established. Our analysis hints at a possible contribution of PSAP to mESC differentiation toward lung and neuronal phenotypes.

Lamin A/C (LMNA) was assigned to the lung development ontology due to interactions with Cftr [47] and Ctnnb1 [49]. To date, most of the reported activities of LMNA are muscle related and in particular that of the heart [50], suggesting that LMNA may in part contribute to the annotation of the heart development (Fig. 5B—green line) and, to lower degree, also to lung development ontologies (Fig. 6A—red line).

Dickkopf homolog 1 (DKK1), a canonical Wnt antagonist, is connected to the lung development ontology due to interaction with Ctnnb1 [51]. As a Wnt antagonist, DKK1 inhibits lung formation and differentiation [52 –54]. The presence of lung-specific gene markers and ontologies in our experimental setup suggests that the inhibitory effect of DKK1 may be limited by other differentiating cues present in the A549-CM. We surmise that the treatment of mESCs with A549-CM-containing DKK1 may promote other DKK1-specific activities, such as differentiation toward neuroectoderm, which is particularly evident in EB cultures [55 –57] (Fig. 1A), and manifest in the assignment of the brain development ontology (Fig. 5B). In parallel, DKK1 also promotes cardiomyocyte differentiation [58], which supports the assignment of the heart development GO term (Fig. 5B).

Clusterin (CLU) is another secreted protein of the set of five converging genes that has a statistically significant connection with the lung development GO term through a single intermediate, Tgfbr2 [59] (Fig. 6A). In the context of the A549-CM-mediated induction of mixed cell populations, CLU may play also an additional role in lung-specific differentiation as inferred from its effects on branching morphogenesis of terminal lung buds [60].

The presence of calreticulin (CALR) in the A549-CM is of particular interest since this protein is found in high levels in newborn lungs and decreases with age [61]. In addition, analysis of CALR-deficient mESCs indicated inter alia a significant effect on the developmental network of the respiratory system [62]. In our experimental system CALR was uniquely assigned to the enrichment of the lung alveolus development GO term (Fig. 5B), which is a subgroup of the lung development ontology, via Nkx2.1 [63], a marker of early lung differentiation [64]. Moreover, CALR exhibited a significant connection with all three sets of genes, leading to the enrichment in the lung development GO term (Fig. 6A).

To validate the in silico findings in vitro, we focused on LM 511, which is present in copious amount in the A549-CM [65]. We treated the mESC monolayers for 10 days with SP media supplemented with 500 ng/mL of LM 111, 211, 411, 421, or 511 isoform. Differentiation was assessed by measuring the expression levels of several of the lung-specific differentiation markers for early lung progenitors (Nkx2.1), AE cell type II [surfactant proteins A (Sftpa), B (Sftpb), and C (Sftpc)], alveolar type I cells [aquaporin type 5 (Aqp5)], club cells (club cell secreted protein, Scgb1a1), and other endoderm-derived organs, such as liver (Hnf1a) and pancreas (Pdx1) [13]. All LM isoforms significantly increased the expression of AE type II marker Sftpa, while minimally affecting the expression of Sftpb, Sftpc, or Aqp5, marker genes for AE type II or I cells, respectively (Supplementary Fig. S2). By contrast, Scgb1a1, marker gene for club cells, was strongly downregulated by all LM isoforms. None of the LM isoforms affected the expression of marker genes for liver and pancreas (Supplementary Fig. S2). The fact that all LM isoforms, with the exception of LM 421, reduced the expression of an early lung progenitor marker gene Nkx2.1 and enhanced the gene expression of surfactant protein A (Supplementary Fig. S2) suggests that LMs are involved in mESC differentiation toward distal lung phenotype, as predicted from our in silico analysis.

Finally, we sought to identify whether the A549-CM-driven pulmonary differentiation of mESCs was unique, or whether this trait was also shared by media conditioned by other cells derived from different organs of endodermal origin, such as liver and pancreases. For this, we first compared the list of proteins in the A549 secretome that are involved in lung development (Fig. 6A) with two recently published high-fidelity analyses of proteins in the A549-CM [23,24] and with proteins found in media conditioned by HepG2 [66,67] and Capan-1 [68] cells, which are derived from human liver hepatocellular carcinoma and human pancreas adenocarcinoma, respectively (Fig. 6B). Overall, this comparative analysis revealed a high degree of similarity in terms of protein contents between the different CMs. Our proteomic analysis of A549-CM was in line with the reports by others, with the exception of CALR, which we identified here for the first time in A549-CM (Fig. 6B).

To compare the differentiative effects of A549-derived CM to media conditioned by other (endoderm-derived) cell lines, HepG2 and Capan-1, of liver and pancreas origins, respectively, we treated monolayer cultures of mESCs for 10 days with different CMs and measured the expression levels of some of the organ-specific differentiation markers of endoderm derivatives, such as lung, liver, and pancreas. As seen in Fig. 6C, only A549-CM induced a significant and selective increase in lung-specific marker gene expression, Sftpb, by 7.4±0.3-fold as compared with SP cultures, while expression of Nkx2.1, Sftpa, Sftpc, Aqp5, and Scgb1a1 was similar to SP cultures. The apparent discrepancy between LM-induced upregulation of Sftpa and A549-CM-induced upregulation of Sftpb is most probably due to presence of additional effector proteins in the A549-CM, as mentioned earlier. Based on this finding we speculate that differentiation protocols based on defined LM isoforms will result in AE type II cells predominantly expressing Sftpa, while cultures treated with A549-CM will lead to enriched expression of Sftpb by the AE type II cells.

Treatment of mESCs with the two other CMs, derived from liver and pancreas cell lines, did not change Sftpb expression by comparison to SP cultures. The expression of the hepatic lineage marker gene Hnf1a in mESCs treated with media conditioned by A549 and HepG2 cells was reduced by 14.3±2.1- and 4.5±0.4-fold, respectively (Fig. 6C). The latter observation is in line with prior reports that exposure of mESCs to HepG2-CM leads to mesodermal differentiation [11,12]. Interestingly, none of the CMs induced pancreatic differentiation. To the best of our knowledge, this study is the first attempt to utilize Capan-1-CM for modulating mESC differentiation. In our hands this CM had no effects on differentiation toward lung, liver, or pancreas. Taken together, our data indicate that among the three studied CMs, only the media conditioned by lung-derived A549 cells is capable to induce mESC differentiation toward the organ of origin.

Based on the similarities in composition of CMs from several endodermal-organ-derived cell lines (Fig. 6B), we speculate that one possible explanation for the observed differential induction of a pulmonary phenotype (Fig. 6C) may be due to dissimilar levels of the proteins identified to-date. On the other hand, we cannot exclude the presence of other, as yet, unidentified minor component(s) in the A549-CM. Identification of such minor components will have to await further advancements in mass spectroscopic technologies. For example, vascular endothelial growth factor [69], fibroblast growth factor-2 [69], and nerve growth factor [70] were all detected by ELISA in A549-CM at concentrations of ∼2, 0.2, and 0.1 ng/mL, respectively. However, none of these growth factors was positively identified by proteomics analyses in the A549 secretome, neither by us nor by others [23,24]. Previous studies had shown that growth factors in the single-digit ng/mL range can elicit specific cellular responses only in cells overexpressing specific cognate receptors, beyond physiological levels [71]. Therefore, we tend to conclude that though low concentration of additional proteins may be present in the A549-CM, their biological relevance maybe limited.

Conclusions

In this study, using proteomic and genomic integrated approach, we identified six proteins of the A549 secretome—LAMA5, PSAP, LMNA, DKK1, CLU, and CALR—as candidate proteins that may lead to enrichment in lung, neuronal, and blood vessel development-relevant GO terms. Based on the concepts laid out in this study, the ability of the A549-derived CM to induce commitment/differentiation of mESCs toward all three lineages in general and lung-relevant pulmonary progenitor cells in particular provides a useful experimental tool to generate defined heterogeneous cell populations of mixed lineage for the proper functioning of engineered whole lungs. The availability of defined heterogeneous lung-specific cell populations may be also preferable over single-cell-type-based treatments in pulmonary regenerative applications.

Footnotes

Acknowledgments

The authors are grateful to Dr. P. Gadue, Children's Hospital of Philadelphia, Philadelphia, PA, for providing the mESCs used in this study. P.I.L. is the Laura H. Carnell Professor for Bioengineering; C.T.S. is a NASA graduate student research program (GSRP) fellow. The authors acknowledge the Smoler Proteomics Center (Faculty of Biology, Technion, Israel) for technical assistance with the proteomic experiments and Mrs. Zehava Cohen for graphics design.

Author Disclosure Statement

No competing financial interests exist.

References

Weiss

, Bertoncello

, Borok

, Kim

, Panoskaltsis-Mortari

, Reynolds

, Rojas

, Stripp

, Warburton

and Prockop

. (2011). Stem cells and cell therapies in lung biology and lung diseases. Proc Am Thorac Soc, 8:223–272.

Varanou

, Page

and Minger

. (2008). Human embryonic stem cells and lung regeneration. Br J Pharmacol, 155:316–325.

Miller

, Knapp

and Eaves

. (2013). Heterogeneity in hematopoietic stem cell populations: implications for transplantation. Curr Opin Hematol, 20:257–264.

Chetty

, Pagliuca

, Honore

, Kweudjeu

, Rezania

and Melton

. (2013). A simple tool to improve pluripotent stem cell differentiation. Nat Methods, 10:553–556.

Watt

and Huck

. (2013). Role of the extracellular matrix in regulating stem cell fate. Nat Rev Mol Cell Biol, 14:467–473.

Hawkins

, Sharp

and McKay

. (2013). The role of hypoxia in stem cell potency and differentiation. Regen Med, 8:771–782.

Zoldan

, Karagiannis

, Lee

, Anderson

, Langer

and Levenberg

. (2011). The influence of scaffold elasticity on germ layer specification of human embryonic stem cells. Biomaterials, 32:9612–9621.

Hannoun

, Fletcher

, Greenhough

, Medine

, Samuel

, Sharma

, Pryde

, Black

, Ross

, et al. (2010). The comparison between conditioned media and serum-free media in human embryonic stem cell culture and differentiation. Cell Reprogram, 12:133–140.

Mannello

and Tonti

. (2007). Concise review: no breakthroughs for human mesenchymal and embryonic stem cell culture: conditioned medium, feeder layer, or feeder-free; medium with fetal calf serum, human serum, or enriched plasma; serum-free, serum replacement nonconditioned medium, or ad hoc formula? All that glitters is not gold!. Stem Cells, 25:1603–1609.

10.

Arien-Zakay

, Lecht

, Bercu

, Amariglio

, Rechavi

, Galski

, Lazarovici

and Nagler

. (2009). Interferon-gamma-induced neuronal differentiation of human umbilical cord blood-derived progenitors. Leukemia, 23:1790–1800.

11.

Kang

, Nagy

, Polak

and Mantalaris

. (2009). Proteomic characterization of the conditioned media produced by the visceral endoderm-like cell lines HepG2 and END2: toward a defined medium for the osteogenic/chondrogenic differentiation of embryonic stem cells. Stem Cells Dev, 18:77–91.

12.

, Kardel

, O'Connor

and Eaves

. (2009). Enhanced generation of hematopoietic cells from human hepatocarcinoma cell-stimulated human embryonic and induced pluripotent stem cells. Exp Hematol, 37:924–936.

13.

Roszell

, Mondrinos

, Seaton

, Simons

, Koutzaki

, Fong

, Lelkes

and Finck

. (2009). Efficient derivation of alveolar type II cells from embryonic stem cells for in vivo application. Tissue Eng Part A, 15:3351–3365.

14.

Mustafa

, Hoheisel

and Alhamdani

. (2011). Secretome profiling with antibody microarrays. Mol Biosyst, 7:1795–1801.

15.

Dowling

and Clynes

. (2011). Conditioned media from cell lines: a complementary model to clinical specimens for the discovery of disease-specific biomarkers. Proteomics, 11:794–804.

16.

Rippon

, Lane

, Qin

, Ismail

, Wilson

, Takata

and Bishop

. (2008). Embryonic stem cells as a source of pulmonary epithelium in vitro and in vivo. Proc Am Thorac Soc, 5:717–722.

17.

Siti-Ismail

, Bishop

, Polak

and Mantalaris

. (2008). The benefit of human embryonic stem cell encapsulation for prolonged feeder-free maintenance. Biomaterials, 29:3946–3952.

18.

Siti-Ismail

, Samadikuchaksaraei

, Bishop

, Polak

and Mantalaris

. (2012). Development of a novel three-dimensional, automatable and integrated bioprocess for the differentiation of embryonic stem cells into pulmonary alveolar cells in a rotating vessel bioreactor system. Tissue Eng Part C Methods, 18:263–272.

19.

Gadue

, Gouon-Evans

, Cheng

, Wandzioch

, Zaret

, Grompe

, Streeter

and Keller

. (2009). Generation of monoclonal antibodies specific for cell surface molecules expressed on early mouse endoderm. Stem Cells, 27:2103–2113.

20.

Lecht

, Arien-Zakay

, Kohan

, Lelkes

and Lazarovici

. (2010). Angiostatic effects of K252a, a Trk inhibitor, in murine brain capillary endothelial cells. Mol Cell Biochem, 339:201–213.

21.

Barnea

, Sorkin

, Ziv

, Beer

and Admon

. (2005). Evaluation of prefractionation methods as a preparatory step for multidimensional based chromatography of serum proteins. Proteomics, 5:3367–3375.

22.

UniProt Consortium. (2012). Reorganizing the protein space at the Universal Protein Resource (UniProt). Nucleic Acids Res, 40:D71–D75.

23.

Luo

, Liu

, Wang

, Hu

, Zeng

and Chen

. (2011). A high-quality secretome of A549 cells aided the discovery of C4b-binding protein as a novel serum biomarker for non-small cell lung cancer. J Proteomics, 74:528–538.

24.

Caccia

, Zanetti Domingues

, Micciche

, De Bortoli

, Carniti

, Mondellini

and Bongarzone

. (2011). Secretome compartment is a valuable source of biomarkers for cancer-relevant pathways. J Proteome Res, 10:4196–4207.

25.

Bidez

, Kresh

, Wel

and Lelkes.

(2011). Enhanced cardiac differentiation of mouse embryonic stem cells by electrical stimulation. In: Stem Cell Engineering. Artmann

, Minger

, and Hescheler

, eds. Springer, Berlin, pp 119–141.

26.

Narayan

, Sylvia

, Malhotra

, Yin

, Martens

, Vallerskog

, Kornfeld

, Xiong

, Cohen

, et al. (2012). Intrathymic programming of effector fates in three molecularly distinct gammadelta T cell subtypes. Nat Immunol, 13:511–518.

27.

Paquette

and Tokuyasu

. (2010). EGAN: exploratory gene association networks. Bioinformatics, 26:285–286.

28.

Kestler

, Muller

, Kraus

, Buchholz

, Gress

, Liu

, Kane

, Zeeberg

and Weinstein

. (2008). VennMaster: area-proportional Euler diagrams for functional GO analysis of microarrays. BMC Bioinformatics, 9:67

29.

Jackson

, Taylor

, Jones

and Forrester

. (2010). The culture of mouse embryonic stem cells and formation of embryoid bodies. Methods Mol Biol, 633:1–18.

30.

Pal

, Totey

, Mamidi

and Bhat

. (2009). Propensity of human embryonic stem cell lines during early stage of lineage specification controls their terminal differentiation into mature cell types. Exp Biol Med (Maywood), 234:1230–1243.

31.

Osafune

, Caron

, Borowiak

, Martinez

, Fitz-Gerald

, Sato

, Cowan

, Chien

and Melton

. (2008). Marked differences in differentiation propensity among human embryonic stem cell lines. Nat Biotechnol, 26:313–315.

32.

Pal

, Mamidi

, Das

, Gupta

and Bhonde

. (2012). A simple and economical route to generate functional hepatocyte-like cells from hESCs and their application in evaluating alcohol induced liver damage. J Cell Biochem, 113:19–30.

33.

Hwang

, Randle

, Bielby

, Polak

and Mantalaris

. (2006). Enhanced derivation of osteogenic cells from murine embryonic stem cells after treatment with HepG2-conditioned medium and modulation of the embryoid body formation period: application to skeletal tissue engineering. Tissue Eng, 12:1381–1392.

34.

Kadzik

and Morrisey

. (2012). Directing lung endoderm differentiation in pluripotent stem cells. Cell Stem Cell, 10:355–361.

35.

Ali

, Edgar

, Samadikuchaksaraei

, Timson

, Romanska

, Polak

and Bishop

. (2002). Derivation of type II alveolar epithelial cells from murine embryonic stem cells. Tissue Eng, 8:541–550.

36.

Cardoso

. (2004). Lung morphogenesis, role of growth factors and transcription factors. In: The Lung: Development, Aging and the Environment. Harding

, Pinkerton

, Plopper

, eds. Elsevier Ltd., Oxford, UK, pp 3–12.

37.

Longmire

, Ikonomou

, Hawkins

, Christodoulou

, Cao

, Jean

, Kwok

, Mou

, Rajagopal

, et al. (2012). Efficient derivation of purified lung and thyroid progenitors from embryonic stem cells. Cell Stem Cell, 10:398–411.

38.

Huang

, Islam

, O'Neill

, Hu

, Yang

, Chen

, Mumau

, Green

, Vunjak-Novakovic

, Bhattacharya

and Snoeck

. (2014). Efficient generation of lung and airway epithelial cells from human pluripotent stem cells. Nat Biotechnol, 32:84–91.

39.

Ulloa-Montoya

, Kidder

, Pauwelyn

, Chase

, Luttun

, Crabbe

, Geraerts

, Sharov

, Piao

, et al. (2007). Comparative transcriptome analysis of embryonic and adult stem cells with extended and limited differentiation capacity. Genome Biol, 8:R163

40.

Prakash

and Stenmark

. (2012). Bioengineering the lung: molecules, materials, matrix, morphology, and mechanics. Am J Physiol Lung Cell Mol Physiol, 302:L361–L362.

41.

, Wang

, Chen

, Lu

, Jiang

, Liang

, Xu da

and Li

. (2009). Involvement of Akt1/protein kinase Balpha in tumor conditioned medium-induced endothelial cell migration and survival in vitro. J Cancer Res Clin Oncol, 135:1543–1550.

42.

Nguyen

, Miner

, Pierce

and Senior

. (2002). Laminin alpha 5 is required for lobar septation and visceral pleural basement membrane formation in the developing mouse lung. Dev Biol, 246:231–244.

43.

Kadoya

, Mochizuki

, Nomizu

, Sorokin

and Yamashina

. (2003). Role for laminin-alpha5 chain LG4 module in epithelial branching morphogenesis. Dev Biol, 263:153–164.

44.

Nguyen

and Senior

. (2006). Laminin isoforms and lung development: all isoforms are not equal. Dev Biol, 294:271–279.

45.

Colland

, Jacq

, Trouplin

, Mougin

, Groizeleau

, Hamburger

, Meil

, Wojcik

, Legrain

and Gauthier

. (2004). Functional proteomics mapping of a human signaling pathway. Genome Res, 14:1324–1332.

46.

Leonova

, Qi

, Bencosme

, Ponce

, Sun

and Grabowski

. (1996). Proteolytic processing patterns of prosaposin in insect and mammalian cells. J Biol Chem, 271:17312–17320.

47.

Wang

, Venable

, LaPointe

, Hutt

, Koulov

, Coppinger

, Gurkan

, Kellner

, Matteson

, et al. (2006). Hsp90 cochaperone Aha1 downregulation rescues misfolding of CFTR in cystic fibrosis. Cell, 127:803–815.

48.

Lee

, Chen

, Tsai

, Su

and Liao

. (2006). Proteomics analysis revealed changes in rat bronchoalveolar lavage fluid proteins associated with oil mist exposure. Proteomics, 6:2236–2250.

49.

Choi

, Park

, Costantini

, Jho

and Joo

. (2004). Adenomatous polyposis coli is down-regulated by the ubiquitin-proteasome pathway in a process facilitated by Axin. J Biol Chem, 279:49188–49198.

50.

, Jaalouk

, Vartiainen

and Lammerding

. (2013). Lamin A/C and emerin regulate MKL1-SRF activity by modulating actin dynamics. Nature, 497:507–511.

51.

Chamorro

, Schwartz

, Vonica

, Brivanlou

, Cho

and Varmus

. (2005). FGF-20 and DKK1 are transcriptional targets of beta-catenin and FGF-20 is implicated in cancer and development. EMBO J, 24:73–84.

52.

Shu

, Guttentag

, Wang

, Andl

, Ballard

, Lu

, Piccolo

, Birchmeier

, Whitsett

, Millar

and Morrisey

. (2005). Wnt/beta-catenin signaling acts upstream of N-myc, BMP4, and FGF signaling to regulate proximal-distal patterning in the lung. Dev Biol, 283:226–239.

53.

De Langhe

, Sala

, Del Moral

, Fairbanks

, Yamada

, Warburton

, Burns

and Bellusci

. (2005). Dickkopf-1 (DKK1) reveals that fibronectin is a major target of Wnt signaling in branching morphogenesis of the mouse embryonic lung. Dev Biol, 277:316–331.

54.

Chen

, Cao

, Qian

, Shao

, Niederreither

and Cardoso

. (2010). A retinoic acid-dependent network in the foregut controls formation of the mouse lung primordium. J Clin Invest, 120:2040–2048.

55.

Watanabe

, Kamiya

, Nishiyama

, Katayama

, Nozaki

, Kawasaki

, Watanabe

, Mizuseki

and Sasai

. (2005). Directed differentiation of telencephalic precursors from embryonic stem cells. Nat Neurosci, 8:288–296.

56.

Kong

and Zhang

. (2009). Dickkopf (Dkk) 1 promotes the differentiation of mouse embryonic stem cells toward neuroectoderm. In Vitro Cell Dev Biol Anim, 45:185–193.

57.

Verani

, Cappuccio

, Spinsanti

, Gradini

, Caruso

, Magnotti

, Motolese

, Nicoletti

and Melchiorri

. (2007). Expression of the Wnt inhibitor Dickkopf-1 is required for the induction of neural markers in mouse embryonic stem cells differentiating in response to retinoic acid. J Neurochem, 100:242–250.

58.

Rai

, Walthall

, Hu

and Hatzopoulos

. (2012). Continuous antagonism by Dkk1 counter activates canonical Wnt signaling and promotes cardiomyocyte differentiation of embryonic stem cells. Stem Cells Dev, 21:54–66.

59.

Reddy

, Karode

, Harmony

and Howe

. (1996). Interaction of transforming growth factor beta receptors with apolipoprotein J/clusterin. Biochemistry, 35:309–314.

60.

French

, Chonn

, Ducrest

, Baumann

, Belin

, Wohlwend

, Kiss

, Sappino

, Tschopp

and Schifferli

. (1993). Murine clusterin: molecular cloning and mRNA localization of a gene associated with epithelial differentiation processes during embryogenesis. J Cell Biol, 122:1119–1130.

61.

Higashino

and Kageyama

. (2008). Development-dependent expression of calreticulin in the brain and other tissues of the Japanese monkey, Macaca fuscata. J Med Primatol, 37:303–310.

62.

Faustino

, Chiriac

, Niederlander

, Nelson

, Behfar

, Mishra

, Macura

, Michalak

, Terzic

and Perez-Terzic

. (2010). Decoded calreticulin-deficient embryonic stem cell transcriptome resolves latent cardiophenotype. Stem Cells, 28:1281–1291.

63.

Perrone

, Tell

and Di Lauro

. (1999). Calreticulin enhances the transcriptional activity of thyroid transcription factor-1 by binding to its homeodomain. J Biol Chem, 274:4640–4645.

64.

Moodley

, Thompson

and Warburton

. (2013). Stem cells: a recapitulation of development. Respirology, 18:1167–1176.

65.

Lecht

, Stabler

, Rylander

, Chiaverelli

, Schulman

, Marcinkiewicz

and Lelkes

. (2014). Enhanced reseeding of decellularized rodent lungs with mouse embryonic stem cells. Biomaterials, 35:3252–3262.

66.

Chen

, Chi

, Tsai

, Tseng

, Lin

, Chen

, Huang

and Lin

. (2012). Stable isotope labeling with amino acids in cell culture (SILAC)-based quantitative proteomics study of a thyroid hormone-regulated secretome in human hepatoma cells. Mol Cell Proteomics, 11:M111 011270.

67.

Slany

, Haudek

, Zwickl

, Gundacker

, Grusch

, Weiss

, Seir

, Rodgarkia-Dara

, Hellerbrand

and Gerner

. (2010). Cell characterization by proteome profiling applied to primary hepatocytes and hepatocyte cell lines Hep-G2 and Hep-3B. J Proteome Res, 9:6–21.

68.

Makawita

, Smith

, Batruch

, Zheng

, Ruckert

, Grutzmann

, Pilarsky

, Gallinger

and Diamandis

. (2011). Integrated proteomic profiling of cell line conditioned media and pancreatic juice for the identification of pancreatic cancer biomarkers. Mol Cell Proteomics, 10:M111 008599.

69.

Ogawa

, Takayama

, Takakura

, Kitano

and Ueno

. (2002). Anti-tumor angiogenesis therapy using soluble receptors: enhanced inhibition of tumor growth when soluble fibroblast growth factor receptor-1 is used with soluble vascular endothelial growth factor receptor. Cancer Gene Ther, 9:633–640.

70.

Fox

, Patel

, Barnes

and Belvisi

. (2001). Release of nerve growth factor by human pulmonary epithelial cells: role in airway inflammatory diseases. Eur J Pharmacol, 424:159–162.

71.

Katzir

, Shani

, Regev

, Shabashov

and Lazarovici

. (2002). A quantitative bioassay for nerve growth factor, using PC12 clones expressing different levels of trkA receptors. J Mol Neurosci, 18:251–264.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.08 MB

0.00 MB

0.03 MB

0.61 MB

0.22 MB

0.02 MB

0.03 MB