Regulated gene expression in cultured type II cells of adult human lung

Philip L. Ballard, Jae W. Lee, Xiaohui Fang, Cheryl Chapin, Lennell Allen, Mark R. Segal, Horst Fischer, Beate Illek, Linda W. Gonzales, Venkatadri Kolla, Michael A. Matthay

Abstract

Alveolar type II cells have multiple functions, including surfactant production and fluid clearance, which are critical for lung function. Differentiation of type II cells occurs in cultured fetal lung epithelial cells treated with dexamethasone plus cAMP and isobutylmethylxanthine (DCI) and involves increased expression of 388 genes. In this study, type II cells of human adult lung were isolated at ∼95% purity, and gene expression was determined (Affymetrix) before and after culturing 5 days on collagen-coated dishes with or without DCI for the final 3 days. In freshly isolated cells, highly expressed genes included SFTPA/B/C, SCGB1A, IL8, CXCL2, and SFN in addition to ubiquitously expressed genes. Transcript abundance was correlated between fetal and adult cells (r = 0.88), with a subset of 187 genes primarily related to inflammation and immunity that were expressed >10-fold higher in adult cells. During control culture, expression increased for 8.1% of expressed genes and decreased for ∼4% including 118 immune response and 10 surfactant-related genes. DCI treatment promoted lamellar body production and increased expression of ∼3% of probed genes by ≥1.5-fold; 40% of these were also induced in fetal cells. Highly induced genes (≥10-fold) included PGC, ZBTB16, DUOX1, PLUNC, CIT, and CRTAC1. Twenty-five induced genes, including six genes related to surfactant (SFTPA/B/C, PGC, CEBPD, and ADFP), also had decreased expression during control culture and thus are candidates for hormonal regulation in vivo. Our results further define the adult human type II cell molecular phenotype and demonstrate that a subset of genes remains hormone responsive in cultured adult cells.

  • expression profiling
  • dexamethasone
  • cAMP
  • surfactant
  • pepsinogen C

type ii cells of the alveolar epithelium have a number of known physiological functions. During fetal development they are a progenitor cell for type I cells that constitute the gas exchange surface in alveoli. Following injury to the postnatal lung, adult type II cells proliferate to repopulate the pulmonary epithelium. These cells participate in transepithelial movement of ions and water to clear fetal lung fluid at birth and to maintain normal alveolar fluid homeostasis. Mature type II cells synthesize a variety of immune effector molecules and antioxidant enzymes. A highly specialized function of adult type II cells is the production and secretion of surfactant, the complex mixture of lipid and surfactant proteins that maintains alveolar stability and is required for successful adaptation to extrauterine life (20, 48, 63).

Differentiation of type II cells from undifferentiated precursor epithelial cells has been studied in cultured explants and epithelial cells from midgestation human fetal lung. In these experimental models, cells treated with glucocorticoid plus cAMP/isobutylmethylxanthine (DCI) increase synthesis of phospholipids and surfactant proteins (SP), develop intracellular lamellar bodies, and secrete surface-active surfactant in response to secretagogs; cells cultured without hormones do not undergo differentiation (3032, 49, 51). We previously utilized DNA microarray analysis to profile regulated genes after 3 days of DCI-induced fetal type II cell differentiation in vitro (32, 61). We found that 388 genes, representing ∼7% of expressed genes in these cells, were induced >1.5-fold after hormone treatment. Responsive genes related to surfactant included 13 involved in lipid uptake and phospholipid biosynthesis, SFTPA, SFTPB, SFTPC, PGC (pepsinogen C, which participates in pro-SP-B processing), and genes encoding the lamellar body membrane proteins ABCA3 and DC-LAMP. Induced genes were also overrepresented in the functional categories of ionic channel and cell adhesion.

In the present study, we have expanded characterization of the molecular phenotype to type II cells from adult human lung. Previously, Gonzalez et al. (33) profiled and compared gene expression in type I and type II cells from the adult rat. Although various gene products have been described for adult human type II cells (19, 24, 26, 53, 59, 62), a detailed microarray survey has not been reported. We hypothesized that a subset of DCI-induced genes of human fetal lung epithelial cells would remain responsive to DCI in adult type II cells. This proposal is based on previous observations that content of lamellar bodies and selected induced proteins in fetal cells decreases after removal of hormones (25), that expression of selected genes of cultured adult type II cells is maintained with DCI plus growth factors (62), and that there are known regulatory effects of glucocorticoids in other adult tissues. The goals of this study were to 1) determine whether the gene expression profile in DCI-induced fetal type II cells is similar to that in freshly isolated adult type II cells, 2) identify genes of adult cells that are affected by monolayer culture in the absence of serum and hormones, and 3) ascertain genes of adult cells that are regulated by DCI treatment during culture. These objectives address the fidelity of the in vitro model of hormone-differentiated fetal type II cells, the issue of possible transdifferentiation of cultured adult type II cells into type I cells, and genes of adult human lung that are potentially glucocorticoid- and/or cAMP-regulated in vivo. We report data on the molecular phenotype of freshly isolated adult cells, identify subsets of genes that change expression during cell culture and in response to hormone exposure, and describe 25 genes of adult type II cells that are candidates for hormonal modulation in vivo.

METHODS

Cell isolation.

Type II cells were isolated from donated adult human lungs, which were not used for lung transplantation, from the Northern California Transplant Donor Network. Tissue was obtained from five adult males without lung disease who died after head trauma or cerebrovascular accident; all individuals had a history of smoking. Adult type II cells were isolated by the published method (19). Briefly, a lung lobe was perfused, lavaged, and instilled with elastase (13 U/ml) for 45 min at 37°C. The tissue was minced and filtered, and the cells were partially purified on a discontinuous Percoll density gradient (1.04–1.09 g/ml solution) plus negative selection with CD14-coated magnetic beads and on IgG-coated dishes. The final cell preparation was suspended in DMEM-H21 and F-12 Ham's (1:1) plus 10% FBS, and aliquots were taken for cell counting by hemacytometer and cytocentrifuge preparations to assess purity by immunostaining. Additional aliquots of cells, constituting freshly isolated cells, were centrifuged, and the pellets were frozen for RNA and protein analysis.

Cell culture.

Freshly isolated cells were plated on collagen-coated dishes as previously described (44) and cultured for 2 days in DMEM-H21 plus F-12 Ham's (1:1) plus 10% FBS. The cells were then cultured for an additional 3 days in serum-free medium (control), or with dexamethasone (10 nM) plus 8-Br-cAMP (0.1 mM) and isobutylmethylxanthine (0.1 mM), a combination that is referred to as DCI. These concentrations maximally induce surfactant components in human lung explant cultures (2).

RNA isolation and DNA microarray analysis.

Total RNA was extracted from control and DCI-treated cells using RNA STAT (Tel-Test, Friendswood, TX) per the manufacturer's instructions. All RNA samples had high integrity and purity as assessed by an Agilent 2100 bioanalyzer (Agilent Technologies, Palo Alto, CA). For microarray analysis, RNA was converted to biotin-labeled cRNA using Affymetrix reagents and protocol (www.affymetrix.com). The cRNA was hybridized to U133A Affymetrix microarray chips that contain 16–20 unique 25-mer oligonucleotide probes for ∼14,500 human genes plus corresponding 12–16 probes with a single nucleotide change (mismatch control). Hybridizations were performed with cRNA prepared from freshly isolated day 5 control and day 5 DCI-treated cells from each of five isolation experiments. A total of 15 chips were analyzed with cells cultured from five individual lungs. Hybridization, washing, staining, and scanning were performed by the Stokes Research Institute Nucleic Acid Core Facility at Children's Hospital of Philadelphia using the procedures described in the Affymetrix GeneChip Expression Analysis technical manual. This facility also performed the microarray analyses on fetal lung cells described in our previous report (61).

Affymetrix Microarray Suite 5.0 was used to quantitate and analyze mRNA content for expressed genes. Default values provided by Affymetrix were applied to all analysis parameters. Probes for 70 control genes on each chip were used to normalize fluorescence intensity between chips, and arrays were scaled to an average intensity of 1,500 fluorescence units and analyzed independently. The Microarray Suite software uses Wilcoxon's signed rank tests to evaluate whether a transcript is detectable on the array (present, marginal, absent) and the probability of a significant change between arrays [i.e., differential expression (DE): DCI-treated vs. control], assigning P values and a change call (increase, decrease, no change) for each probe. When more than one probe set was present for the same gene, data were combined to provide a mean value. Fold-stimulation results are expressed as means ± SD. Data analysis was performed as previously described for our study of gene expression in fetal lung epithelial cells (61) with additional assessment of DE via the limma (55) package, a library of the R Bioconductor suite (27), utilizing false discovery rate-based determinations of numbers of DE genes (58). For gene set enrichment analysis of DE genes, we used the GOStat tool at http://gostat.wehi.edu.au/ (5). Microarray data are available at Gene Expression Omnibus (acc. no. GSE19699) of the National Center of Biotechnology (www.ncbi.nlm.nig.gov/geo/).

Real-time RT-PCR.

cDNA was synthesized from 2 μg RNA samples using the SuperScript First-Strand RT-PCR kit (Life Technologies) according to the manufacturer's instructions. Real-time PCR reactions using a singleplex format were performed using an ABI Prism 7000 (Applied Biosystems, Foster City, CA) in the Real-Time PCR Core Facility of the Children's Hospital of Philadelphia. We used the standard PCR protocol recommended by the manufacturer of Assay-on-Demand kits. The specific primer and probe sequences are shown in Supplemental Table S1 (Supplemental data for this article is available online at the AJP-Lung web site). RNA extracted from human fetal lung epithelial cells maximally induced with DCI was used to prepare the standard curve for each gene.

Antibodies.

Antibodies were obtained from the following sources and used for immunofluorescence staining (1:100) and/or Western analysis (1:1,000–2,000): goat anti-human SP-A (#S8400-01; US Biological, Swampscott, MA); rabbit anti-sheep SP-B (AB3780; Chemicon, Temecula, CA); sheep anti-human pepsinogen II (PGC, ab9013; Abcam, Cambridge, UK); rabbit polyclonal anti-CEBPδ (M-17; Santa Cruz Biotech, Santa Cruz, CA); rabbit polyclonal anti-TTF-1 (H-190; Santa Cruz); rabbit polyclonal anti-Duox1 (gift from F. Miot, Brussels, Belgium); mouse anti-human rDC-LAMP (clone 104.G4; Immunotech, Marseille, France); mouse anti-vimentin (clone V9; Chemicon); rabbit polyclonal anti-fatty acid synthase (gift from S. Smith, Oakland, CA); mouse anti-HTI56 (14) (gift from R. Gonzalez, San Francisco, CA); mouse anti-human CD68 (eBioscience, San Diego, CA); rabbit anti-CC10 (Santa Cruz); and anti-GAPDH (Chemicon).

Western analysis.

We performed immunoblotting using previously described procedures (32) and NuPAGE Bis-Tris gels with MES SDS Running Buffer as per the manufacturer's protocol (Invitrogen, Carlsbad, CA). Proteins were transferred to nitrocellulose (0.45 µm) membrane (BioRad, Hercules, CA) and probed with primary antibodies and appropriate infrared detectable secondary antibodies (Alexa 680-tagged, Molecular Probes, Eugene, OR or IRDye800-tagged, Rockland, Gilbertsville, PA). Signal was detected and quantitated using the Odyssey Imaging System (Licor Biosciences, Lincoln, NE).

Immunofluorescence and microscopy.

Cells pelleted on glass slides or cultured on Lab-Tek II four-well collagen I-coated chamber slides (Nunc, Rochester, NY) were fixed with 4% paraformaldehyde and permeabilized with 0.3% Triton X-100 before immunostaining as previously described (32). Secondary antibodies for immunofluorescence detection were tagged with Alexa 488 or Cy3 and used at 1:200 or 1:300. Following immunostaining, some sections were exposed to 4′,6-diamidino-2-phenylindole (DAPI; 0.1 μg/ml, Molecular Probes, Eugene, OR) for 10 min to stain nuclei. Slides were examined for both phase contrast and immunofluorescence with an Olympus 1X70 microscope and Metamorph imaging system (Universal Imaging, West Chester, PA).

For toluidine blue staining and electron microscopy, day 5 cells in monolayer culture on collagen-coated plastic were fixed in situ with 2% glutaraldehyde plus 1% paraformaldehyde in phosphate buffer (pH 7.4), gently scraped from the dish, microfuged into a pellet, and postfixed overnight in 1.5% osmium tetroxide in veronal acetate buffer at 4°C. They were en bloc stained in 1.5% uranyl acetate in maleate buffer, then quickly dehydrated in cold acetone followed by propylene oxide. The tissue was finally infiltrated and embedded in Embed-812 (Electron Microscopy Sciences, Hatfield, PA). Semi-thin (0.5 μm) sections were stained with toluidine blue and examined with a Leitz Orthoplan microscope; ultra-thin sections were stained with 5% uranyl acetate and 0.8% lead citrate and then examined in a Zeiss 10 transmission electron microscope.

RESULTS

Cell isolation and culture.

Postmortem lung tissue from five adult males without lung disease was used for isolation of type II cells by a previously reported procedure (19). Cytospins of freshly isolated cells were immunostained for SP-B and vimentin as markers of type II cells and fibroblasts (Fig. 1) with mean (SE) values of 94 ± 2% and 6 ± 1%, respectively. Additional type II cell type markers were SP-C (95 ± 1%) and DC-LAMP (89 ± 2%); there were no cells positive for CD68 (macrophage marker) and HT156 (type I cell marker); 4.2 ± 0.3% of cells stained positive for CC10 (secretaglobin1A1), a marker for human Clara cells, and also in rodents, bronchoalveolar stem cells (40).

Fig. 1.

Immunostaining of freshly isolated adult type II cell preparation. Most cells are positive for punctate intracellular staining for SP-B (left) and negative for vimentin as a marker for fibroblasts (right). Nuclei are shown by DAPI staining (blue). Images are representative of 5 experiments.

The cells were plated on collagen-coated dishes and cultured for 2 days in medium containing 10% FCS and then for 3 additional days in serum-free medium with or without DCI at concentrations that optimally promote type II cell differentiation in cultures of fetal lung epithelial cells (2). Microarray analyses using Affymetrix U133A chips were performed on freshly isolated cells and cells cultured 5 days with or without hormones. Some results are compared with the expression profile of hormone-induced fetal type II cells of comparable purity, which was determined using the same chips and approach to data analysis as for adult cells (61).

Effect of culture and DCI on cell morphology.

Wang et al. (62) demonstrated that adult human type II cells maintain lamellar bodies after culture on Matrigel in the presence of 1% serum, DCI, and keratinocyte growth factor. We examined lamellar body content in our cultured adult type II cells by imaging and immunostaining (Fig. 2). By immunofluorescence, most DCI-treated cells had punctate staining for SP-B and DC-LAMP, a lamellar body membrane protein, while control cells had less intense staining in fewer cells (Fig. 2, A–D). For comparison, cells were stained for DUOX1, an inducible membrane protein (see Table 5), and demonstrated greater diffuse staining in DCI-treated cells (Fig. 2, E and F). By phase microscopy, treated cells contained numerous refractile inclusions, whereas these were less evident in control cells (Fig. 2, G and H). In a separate experiment, high-power microscopy was performed on toluidine blue-stained cells fixed in situ as a monolayer, and representative images are presented in Fig. 2, I–L. Many cells exposed to DCI contained toluidine blue-positive inclusions consistent with lamellar bodies, and these were much less frequent in control cells. These findings support the conclusion that dexamethasone plus cAMP treatment is sufficient to maintain surfactant production in adult type II cells cultured on collagen.

Fig. 2.

Morphology and staining of induced proteins in day 5 cells. Adult type II cells were cultured for 5 days without (control) or with dexamethasone plus cAMP and isobutylmethylxanthine (DCI) for the last 3 days on collagen-coated chamber slides or culture dishes. Representative areas are shown for punctate immunostaining for SP-B (A, control; B, DCI) and DC-LAMP (C, control; D, DCI) as lamellar body markers, diffuse immunostaining for membrane protein Duox1 (E, control; F, DCI), phase microscopy (G, control; H, DCI), and toluidine blue staining at 2 magnifications (I and K, control; J and L, DCI). Treated cells show more refractile inclusions (G vs. H), prominent lamellar bodies (K vs. L), and increased intensity and proportion of cells staining for SP-B, DC-LAMP, and Duox1 compared with control cells. Counting of cells in the immunostaining experiment (B) indicated that 92% were positive for SP-B.

Electron microscopy was performed on the cells shown in Fig. 3. Most control cells contained membranous vesicles of 1–2 μm diameter containing loosely packed lamellar and/or amorphous material; rarely, lamellar bodies were seen (Fig. 3A). Membranous vesicles were also observed in most DCI-treated cells with an appearance generally similar to that in control cells. In addition, some treated cells contained one or more lamellar bodies with normal ultrastructural appearance (Fig. 3B). The abnormal inclusions in these cells resemble those described by deMello et al. (13) for infants with inherited deficiency of SP-B. We speculate that the ultrastructural findings reflect our culture protocol of 2 days without hormones, which depletes SP-B and alters lamellar body production, followed by only 3 days with hormones, which may be insufficient time to clear abnormal inclusions and to accumulate normal lamellar bodies.

Fig. 3.

Transmission electron microscopy of day 5 cells. Adult type II cells were cultured for 5 days without (A, control) or with DCI for the last 3 days (B). Membranous vesicles with loosely packed lamellar or amorphous material were observed in both control and DCI-treated cells, and lamellar bodies (LB) were more common in treated cells. Prominent lysomes (Ly) were observed in some cells.

Expressed genes in freshly isolated adult type II cells.

There was statistically significant expression of 11,176 probes, representing 6,043 annotated unique genes, with a median transcript abundance of 160 fluorimetric units (range 16 to 16,891 f.u.). Overall, approximately one-half of the probes on the chip had a statistically significant signal compared with 35% for fetal lung epithelial cells (61). The 25 most abundantly expressed transcripts, excluding those for 38 ribosomal proteins, are shown in Table 1 along with the Gene Ontology biological process assignment. The list includes SPs A, B, and C, five genes related to protein unfolding and ubiquitination, four genes related to immune response (SCGB1A, CXCL2, B2M, and IL8), two genes involved in protein biosynthesis, both light and heavy ferritin polypeptides, and two genes regulating progression through the cell cycle. Many of these genes were also highly expressed in microarray analysis of adult lung tissue and therefore likely represent ubiquitously expressed genes (34); of note, three genes of Table 1 were expressed at a relatively lower level in lung tissue compared with adult type II cells (IL8, CXCL2, and SFN).

View this table:
Table 1.

Abundantly expressed genes of adult type II cells

Figure 4 compares the abundance signal for 11,176 probes of adult vs. DCI-induced fetal type II cells. There was substantial correlation (r = 0.88), and the slope of the regression line was 0.69, indicating a somewhat overall lower level of gene expression in fetal cells. Expression in adult cells was >10-fold higher (range 10- to 500-fold) than fetal for 187 genes (231 probes), with a fetal signal not detected in 81%, and included five of the highly abundant genes listed in Table 1 (CXCL2, IL8, SCGB1A1, SFN, and SLPI). Of interest, the expression of 118 of the 187 genes (63%) decreased >2-fold during culture of cells under control conditions (as described below). The highest differential expression occurred for secretoglobin 1A1 (uteroglobin, Clara cell protein 10), which was undetectable in fetal cells and had a signal of 10,924 f.u. in adult cells. GO analysis (for biological response, molecular function, and cellular component keywords) was performed for annotated genes: 123 in the 10-fold higher group vs. the 6,043 expressed and annotated genes in adult cells. The differentially expressed genes were overrepresented (P < 0.01) in the categories of immune system response (29.3% vs. 4.7%), chemotaxis (13.6% vs. 0.9%), extracellular space (19.5% vs. 2.2%), chemokine activity (7.3% vs. 0.2%), cytokine activity (11.4% vs. 1.0%), signal transduction (33.3% vs. 18.6%), cell communication (34.1% vs. 20.0%), and serine-type endopeptidase activity (4.9% vs. 0.8%). The highly expressed adult genes were underrepresented in the GO categories of intracellular (46.3% vs. 71.5%) and intracellular organelle (34.1% vs. 57.8%). Thus, genes differentially expressed in adult vs. fetal type II cells largely reflect those with immune-related functions that are not detected in the fetal cells.

Fig. 4.

Transcript abundance in adult compared with fetal type II cells. Mean data from 5 experiments are shown for 11,176 probes that had a significant signal in adult cells. Signals for fetal cells that were called absent were assigned a value of 20 fluorimetric units (f.u.). Colored symbols indicate adult:fetal ratio >10-fold different as indicated. Slope = 0.69 and r = 0.88; the line of identity is shown.

By contrast, only 24 genes (32 probes) were expressed 10-fold higher (range 10- to 81-fold) in fetal cells vs. adult cells; the GO categories of collagen (25% vs. 0.2%) and developmental process (55% vs. 18%) were highly overrepresented in this group, and the collagen genes (COL4A2, COL1A1, COL1A2, AND COL3A1) had the highest levels of differential expression. The full list of differentially expressed genes is available in Tables 2 and 3 of the Supplementary Material.

View this table:
Table 2.

Genes with expression increased >5-fold during cell culture

View this table:
Table 3.

Genes with >10-fold decreased expression during cell culture

Effect of cell culture on gene expression.

Isolated adult type II cells were cultured on collagen-coated dishes with serum present for 2 days (to promote attachment) and then under serum-free conditions for 3 days. Under these control conditions (i.e., no hormones present), expression of 8.1% of the probes increased ≥2-fold (range 2- to 14-fold) and 6.9% decreased ≥2-fold (range 2- to 100-fold). The 30 genes with ≥5-fold increased expression after culture are shown in Table 2. For the entire group of culture-induced genes (total = 580, annotated = 378), significantly (P ≤ 0.01) overrepresented GO categories included cell adhesion (7.9% vs. 3.3%), membrane (43.4% vs. 33.6%), multicellular organismal process (22.0% vs. 14.4%), cytoskeleton (11.1% vs. 6.4%), and collagen (13.2% vs. 0.2%). Underrepresented categories among culture-induced genes included cellular metabolic process (36.5% vs. 52.6%), gene expression (9.0% vs. 23.5%), and membrane-bound organelle (38.1% vs. 52.1%).

The expression level of 499 genes decreased >2-fold during culture; 37 genes with >10-fold decrease are listed in Table 3. The total group of repressed, annotated genes (n = 331) was overrepresented in the following GO categories (P < 0.003): immune response (14.5% vs. 3.4%), cytokine activity (6.6% vs. 1.0%), cell communication (33.8% vs. 20.0%), signal transduction (31.4% vs. 18.6%), chemokine activity (2.7% vs. 0.2%), cell proliferation (11.8% vs. 5.1%), negative regulation of cellular process (16.0% vs. 8.6%), transcription factor activity (10.6% vs. 5.0%), regulation of body fluid level (3.0% vs. 0.7%), and serine type endopeptidase activity (2.4% vs. 0.4%). The gene group was underrepresented in the categories of intracellular location (55.9% vs. 71.5%) and intracellular organelle (45.6% vs. 57.6%).

It has been previously reported that adult type II cells from rat (79, 12, 15, 17, 50) and human (62, 64) undergo dedifferentiation and/or transdifferentiation to type I cells during monolayer culture under control conditions. Related to this topic, we examined expression levels of 21 genes involved in surfactant production. Expression decreased significantly (2- to 25-fold) during control culture for genes encoding SP-A/B/C, glycerol kinase α, sterol-CoA desaturase, phosphatidic acid phosphatase 2B, fatty acid binding proteins ⅘, adipose differentiation-related protein, and pepsinogen C. There was no significant change in levels of mRNAs for the lipogenic enzymes fatty acid synthetase, fatty acid desaturase 1, low density lipoprotein receptor, choline phosphotransferase 1, and lipoprotein lipase, or for the lamellar body membrane proteins ABCA3 and DC-LAMP. mRNA expression increased for fatty acid desaturase 3 (2.1-fold), phosphatidic acid phosphatase 2A and 2C (1.7- and 2.9-fold), and TTF-1 (Nkx2.1, 3.9-fold). Thus, these findings identify candidate genes responsible for decreased surfactant production during monolayer culture of adult type II cells.

Effect of DCI on gene expression.

Glucocorticoid plus cAMP treatment of cultured epithelial cells from fetal lung induces differentiation into type II cells, and this process is associated with altered expression of a subset of genes (32, 61). To test hormonal responsiveness in adult type II cells, we examined gene expression of cells cultured 3 days with or without DCI exposure. One-hundred forty-eight mRNAs were decreased >2-fold (range 2- to 5-fold) by hormone exposure. The median abundance for repressed genes in treated cells was 94 f.u. (range 14–1,324). Genes decreased >3-fold are listed in Table 4. Compared with all expressed annotated genes in adult type II cells (n = 6,053), repressed genes (n = 111) were overrepresented (all P < 0.005) in the GO categories of developmental process (38.7% vs. 18.0%), cell adhesion (14.4% vs. 3.3%), response to external stimulus (14.4% vs. 3.7%), coagulation (4.5% vs. 0.5%), cell surface receptor-linked signal transduction (15.3% vs. 6.5%), cytoskeleton (18.9% vs. 6.4%), proteinaceous extracellular matrix (9.0% vs. 1.1%), and collagen (4.5% vs. 0.2%). DCI-repressed genes were underrepresented in the categories of metabolic process (33.3% vs. 57.1%), gene transcription (7.2% vs. 23.5%), and intracellular membrane-bound organelle (21.6% vs. 52.1%).

View this table:
Table 4.

Genes repressed >3-fold by dexamethasone plus cAMP

Dexamethasone plus cAMP treatment increased 316 transcripts ≥1.5-fold (range 1.6- to 59.1-fold) with abundance values of 40–6,765 f.u. (median 375 f.u.), which was significantly higher than for repressed genes. We chose 1.5-fold as the cut-off for this analysis because this value was used in the earlier microarray study of fetal cells (61). The 316 induced genes represent ∼4% of genes expressed in adult type II cells as assessed by the Affymetrix U133A chip. The 35 identified genes induced >4-fold are listed in Table 5, with the full list shown in Supplemental Table 4. Highest induction (>10-fold) occurred for PGC, ZBTB16, DUOX1, PLUNC, and CIT. On analysis of GO categories, the set of hormone responsive genes was overrepresented in 11 major categories (Table 6). In addition to containing all three genes assigned to liquid surface tension (SPs), induced genes were preferentially related to lipid synthesis, metal ion (in particular calcium and cadmium) and cofactor binding, organ and cell development, and metabolism involving oxidoreductase activity and monocarboxylic acid. There were 21 induced genes that encode for SPs or proteins related to lipogenesis. Figure 5 compares transcript level for these 21 genes in DCI-treated cells with that in freshly isolated cells; 19 of the 21 surfactant-related genes had similar or greater expression levels after culture with hormones, which is consistent with continuing surfactant production in treated cells.

View this table:
Table 5.

Genes induced >4-fold by dexamethasone plus cAMP

View this table:
Table 6.

GO categories overrepresented in hormone-induced genes of adult type II cells

Fig. 5.

Surfactant-related genes induced by DCI: transcript abundance on day 5 compared with freshly isolated day 0 cells. The gene list includes 4 surfactant proteins, PCG (pro-SP-B processing), 9 genes involved in de novo lipogenesis, and 7 genes related to lipid substrate or uptake. For the 21 genes, the induced transcript level on day 5 is greater than for day 0 for 7 genes, similar for 12 genes and less than day 0 for 2. Data are means + SE (n = 5); *P ≤ 0.01.

Of the 316 induced genes, 127 (40%) were also hormonally induced in fetal lung epithelial cells (61), indicating that hormonal regulation during cell culture was maintained postnatally for a subset of genes. We were particularly interested in identifying genes that potentially would be responsive to hormonal modulation in vivo, consistent with a role in mediating homeostatic responses. We reasoned that such genes would have decreased expression during culture under control, serum-free conditions as a result of loss of circulating hormones (e.g., cortisol, epinephrine). Of the 316 induced genes, expression of 64 (20%) decreased by at least 50% during control culture: 19 of these were hormonally induced to the day 0 level or greater levels, 28 reached levels between 50 and 90% of day 0, and 17 were ≤50% of day 0 after exposure to DCI. Thus, for the majority of induced genes, transcript abundance equaled or exceeded the level in freshly isolated cells.

Of the 127 genes induced by dexamethasone plus cAMP in both fetal epithelial and adult type II cells, 25 also had decreased expression during control culture of adult cells. Figure 6 illustrates the interrelationships between adult and fetal inducible genes, and Table 7 lists the 25 genes of interest. In analysis of GO categories, only surfactant proteins (respiratory gaseous exchange) were overrepresented compared with all expressed genes (P < 0.01). However, the 25 genes contain a total of six genes that have known relevance to surfactant: SFTPA, SFTPB, SFTPC, PGC, CEBPD (SP synthesis), and ADFP (lipid synthesis). The greater fold induction of SP genes in fetal vs. adult cells (Table 7) reflects the very low expression of these genes in undifferentiated fetal epithelial cells. Most other genes in this set had similar levels of induction in fetal compared with adult cells with the exceptions of Duox1 (6-fold greater in adult) and AGR2 (6-fold greater in fetal).

Fig. 6.

Venn diagram of regulated genes in fetal and adult type II cells. Approximately 300 genes were induced by hormones in both fetal and adult cells, with 127 genes in common; 25 of these genes were also downregulated during control culture of adult cells and represent candidate genes for hormonal regulation in vivo.

View this table:
Table 7.

Twenty-five genes of adult cells with decreased expression in control culture and hormone induction in both fetal and adult cells: candidate genes for hormonal responsiveness in vivo

Validation of selected regulated genes.

In separate experiments, with additional preparations of adult type II cells, we examined nine genes to confirm induction by qPCR (Fig. 7A); data for parallel experiments with fetal cells are shown for comparison. Induction was demonstrated for eight genes of adult cells, in agreement with the microarray data. TTF-1 mRNA, which is induced in fetal cells (Fig. 7B and Ref. 61), was not significantly increased by DCI in adult cells, by either microarray or qPCR analysis. Fold induction of the other genes by qPCR and microarray was generally similar except for SP-A (greater by qPCR) and Duox1 (higher by microarray). The relative content of the mRNAs (normalized to 18S rRNA) was similar in DCI-treated adult and fetal cells except for CEBPD (greater in fetal) and DUOX1 (greater in adult).

Fig. 7.

Validation of selected induced genes. Adult type II cells and fetal lung epithelial cells were cultured with or without DCI for qPCR (A) and Western (B and C) analysis. A: mRNA content for control and DCI-treated cells normalized to 18S rRNA. All transcripts were significantly induced by DCI except for TTF-1 in adult cells. mRNA content of adult DCI-treated cells was similar to that in fetal type II cells except for CEPBD (decreased) and Duox1 (increased). Data are means + SE, n = 3; *P < 0.05 adult DCI vs. fetal DCI. B: immunoreactive protein is shown for a representative Western of adult (left) and fetal (right) cells with GAPDH used as loading control. A stronger signal is observed for DCI-treated cells vs. control (C) for each protein; day 0 = day 0 cells. Nonspecific bands are seen above PGC and below Duox1. C: fold induction by densitometry in 3 experiments for proteins of panel B. Data are means + SE; *P < 0.05 for DCI-treated vs. control.

Western analysis was performed for proteins with specific antibodies available. Representative results for six proteins are shown in Fig. 7B, comparing adult vs. fetal cells, and data for three experiments are presented in Fig. 7C. DCI induction was demonstrated for each protein. Of note, content of TTF-1 was significantly induced in adult cells (3-fold), whereas mRNA content was not altered by DCI, suggesting a developmental change in mode of regulation. TTF-1 was not detected by Western blot in day 0 adult cells, which may represent lability of the protein under the stress of tissue anoxia and/or cell isolation. We also found FAS to be induced (2.9 ± 1.4-fold, n = 3) similar to the 4-fold increase in fetal cells on the same blot (data not shown). These findings regarding the general reliability of microarray data are in agreement with our previous data validating induction of specific genes in fetal type II cells (61).

DISCUSSION

In this study we profiled gene expression in type II cells of adult human lung. To our knowledge, this is the first description of the phenotype of these cells by DNA microarray in the human. The cells expressed approximately one-half of the genes probed by the chip and had high levels of expression for surfactant proteins as expected. Specific subsets of genes had altered expression during cell culture and in response to DCI treatment. A small group of genes was identified as hormone responsive in both fetal and adult cells with decreased expression in adult cells during control culture. We propose that these genes of adult type II cells are candidates for hormonal modulation in vivo in response to stress, injury, or disease.

Do human adult type II cells transdifferentiate into type I cells during culture?

It is recognized that type II cells of adult rat change phenotype during monolayer cell culture. In particular, expression of surfactant-related genes decreases and lamellar bodies disappear (7, 16, 45, 62). Recently, this process has also been observed in human fetal type II cells after removal of DCI (25). This process may represent transdifferentiation of adult type II cell into a more type I cell-like phenotype similar to the proposed precursor role of type II cells in vivo during development (1) and after lung injury (18). This topic has been addressed recently in microarray experiments that compared gene expression in freshly isolated, purified adult rat type I and type II cells (10, 33). Between the two studies, a total of 30 differentially expressed (≥5-fold) genes, representing potential type I cell marker genes, were identified. In our study with human cells, mRNA abundance for 7 of these genes (CAV1, CAV2, VEGF, LGALS1, LAMA3, AGRN, AQP1) increased 2- to 3-fold in control cultures compared with day 0 expression. In addition, expression of LOX was not detected in day 0 cells but was present on day 5. Only two of the eight genes (LGALS1 and LOX) were modestly suppressed by DCI treatment during culture (∼2-fold). These findings are consistent with some increased expression of candidate type I cell genes during culture but do not allow definitive conclusions regarding the extent of transdifferentiation in vitro. By contrast, the pattern of suppressed genes during culture indicates dedifferentiation of cultured adult type II cells in the functional categories of surfactant production (10 genes), immune response, cytokine or chemokine production (118 genes), and fluid regulation (15 genes).

Does the pattern of gene expression in DCI-treated fetal epithelial cells resemble that in adult type II cells?

With regard to surfactant, treated fetal cells have upregulated expression of surfactant proteins and selected lipogenic enzymes, increased rate of phosphatidylcholine synthesis, altered phospholipid composition, appearance of lamellar bodies, and secretion of surface active surfactant in response to secretagogs (3032, 49, 51). By microarray analysis, ∼7% of expressed genes were hormone-induced (61). Evaluating all expressed genes, there was a strong correlation in expression levels between fetal and adult type II cells with somewhat lower expression in fetal cells. Of particular note, mRNA abundance for SP-A/B/C and 13 proteins involved in lipid synthesis was similar for fetal and adult type II cells. These findings are in agreement with previous conclusions that DCI treatment of undifferentiated fetal lung epithelial cells causes a robust induction of surfactant production, and they support the utility of this fetal model for studies of surfactant.

Adult type II cells differed from the fetal cells with markedly higher expression of 187 genes, many of which were involved in immune and inflammatory responses. Most of these genes also had decreased expression during control culture of adult cells. We propose three possibilities for this difference in phenotype between fetal and adult type II cells. First, some genes of this set may be under developmental regulation that is separate from the influence of DCI, perhaps reflecting evolutionary segregation of a category of genes that are not immediately required for survival on air breathing (in contrast to, for example, surfactant components and proteins related to fluid homeostasis). The decreased expression during control culture may reflect a requirement for circulating mediators that promote expression; however, we cannot rule out an artifactual effect of monolayer culture per se. A second possibility is that some genes were induced by chronic exposure to tobacco smoke. All of the tissue specimens were obtained from donor individuals who smoked. Previous studies have reported altered gene expression in lung tissue, bronchial epithelium, and macrophages of smokers (37, 56, 57, 69), including gene sets related to immune defense; of interest, both PLUNC and SCGB1A1, which were expressed at high levels in our freshly isolated cells, were detected at lower levels in bronchial epithelial cells of smokers compared with nonsmokers (56). We favor a third possibility that many of these genes represent acute phase reactants that are upregulated secondary to hypoxia associated with brain injury (23) and/or death, postmortem storage of tissue and/or the cell isolation procedure. Of note, however, fetal lung tissue is exposed to similar insults during cell isolation, and these genes are not expressed, implying a developmental change in responsiveness to these stresses. We conclude that the phenotype of hormonally differentiated fetal type II cells is closely, but not completely, similar to adult cells as isolated in our procedure.

The finding of high mRNA content for CC10 (secretoglobin 1A1) in adult type II cells was unexpected. This protein is restricted to Clara cells in mouse lung, present in Clara cells and some alveolar cells of rat, and observed in both cell types of rabbit (47). In one previous report for human lung, CC10 was rarely observed in normal alveoli by RNA-RNA in situ hybridization, a finding independent of smoking history (38). We observed immunostaining for CC10 in a small percentage of freshly isolated cells, which is inconsistent with the high level of transcript. Of note, abundance of CC10 mRNA decreased many-fold during culture, and the protein was detected at a low level by Western blot on day 5 (data not shown). We propose the following scenario to explain these observations: under normal conditions, the CC10 gene is either not expressed in human type II cells or expression is restricted to a small subset of alveolar cells, perhaps representing bronchoalveolar stem cells (40). However, CC10 mRNA is highly induced in freshly isolated type II cells in response to the asphyxia associated with death and/or the stress of cell preparation. The same conditions provoke an endoplasmic reticulum unfolded protein response that transiently blocks translation of stress-induced transcripts, including CC10, and depletes content of other proteins with short half-lives (such as TTF-1) (68). Transcript levels of CC10 (and other acute phase reactants) decreases as cells adjust to culture conditions. It is possible that expression of CC10 also occurs in type II cells in vivo under conditions of stress or infection. Although these findings and interpretation require future studies for confirmation, they suggest caution in equating transcript and protein levels for specific genes in type II cells isolated from postmortem tissue. Nevertheless, our observations indicate that CC10 transcript can be expressed in adult type II cells under some conditions.

Hormonally regulated genes.

Three-hundred sixteen genes were induced >1.5-fold by DCI treatment. SPs A/B/C along with 18 genes related to lipid synthesis were induced, suggesting coordinate regulation of key genes for surfactant production. Moreover, hormone exposure maintained (or increased) the day 0 level of expression for most of these genes. It should be noted, however, that expression levels reported for day 0 cells could have been influenced by the cell isolation procedure. In support of the microarray findings, microscopy of day 5 cells showed increased number and size of lamellar bodies and greater SP-B immunostaining in DCI-treated compared with control cells.

PGC was the most highly induced gene (∼60-fold by both microarray and qPCR) compared with day 5 control cells in this subset. This gene encodes a type II cell-specific aspartic protease involved in proteolytic processing of pro-SP-B to the mature form (28). Because mature SP-B is required for normal lamellar body genesis, surfactant function, and processing of pro-SP-C, we propose that decreased PGC expression in control cultured cells is a major reason for loss of surfactant production, and conversely, that induction of PGC is an important contributor to maintenance of lamellar body production with DCI treatment. Moreover, decreased pepsinogen C levels in vivo in response to lung injury or disease could produce SP-B deficiency and dysfunctional surfactant.

Many of the genes induced by DCI in fetal lung epithelial cells are dependent on induction of TTF-1 (42). This transcription factor binds and activates the SP-B promoter (6) and is also required for early lung morphogenesis in mice (41). Induction of this gene in human fetal cells involves increased mRNA content and rate of gene transcription. Thus, we were surprised to find that DCI treatment of adult cells did not significantly increase TTF-1 mRNA content under conditions that induced known target genes such as LAMP3 and SP-A/B/C. However, we did find increased TTF-1 by Western analysis, consistent with a role for TTF-1 in some of the hormonal effects observed in adult cells. Further studies are required to explore regulatory mechanisms and the role of TTF-1 in adult type II cells.

The precocious maturation of the fetal surfactant system with hormone treatment of animals appears to mimic the effect of endogenous glucocorticoid/cAMP in modulating the developmental pattern in vivo (3). However, there are limited previous data for hormonal responsiveness of type II cells of adult lung. In a study with type II cells isolated from suckling rats, glucocorticoid plus cAMP treatment maintained expression of SPs and enhanced stimulated secretion of phosphatidylcholine (4). In vivo treatment of adult rats with dexamethasone is reported to increase content of vitamin K carboxylase activity (29), P-selectin (52), cJUN (43), antioxidant enzymes (39), SP-A/B/C/D (54, 66), and total phospholipid plus disaturated phosphatidylcholine (67), and to decrease endothelial ICAM1 (11), IL-17 (65), and retinoic acid receptors (35). None of these proteins other than the SPs were induced in our study, which may reflect localization to cell types other than type II cells, species differences and/or different in vitro vs. in vivo responsiveness.

With type II cells isolated from adult human lung, Wang et al. (62) found that cells cultured on matrigel/collagen in the presence of keratinocyte growth factor plus DCI contained lamellar bodies and had an elevated rate of acetate incorporation into phosphatidylcholine compared with untreated cells cultured in serum-containing medium alone. They also found that levels of SPs and fatty acid synthase decreased during culture and were maintained at least in part by the combination of agents; our data are in agreement with these changes. Hormonal responsiveness for cells cultured on collagen alone, as used in our study, was not reported. Our findings indicate that DCI in the absence of keratinocyte growth factor and matrigel is sufficient to promote lamellar body production and maintain or enhance expression of a subset of genes related to surfactant. These results indicate the potential utility of adult human type II cells cultured in the presence of DCI as a model system for studies of surfactant and other biochemical functions of these cells. To our knowledge, there have been no human studies of in vivo glucocorticoid effects on protein or mRNA profile of lavage fluid, bronchial cells, or lung tissue for comparison with our results.

The 25 genes that were induced in both fetal and adult cells, and also had decreasing mRNA content during control culture of adult cells, are of particular interest as candidates for hormonal modulation of expression in vivo. We studied a subset of these genes in more detail and confirmed hormonal induction of mRNA and/or protein. This type of validation was also performed previously for selected genes of fetal type II cells (61). Approximately 25% of these genes relate to surfactant or lipid synthesis, and these may represent critical genes for enhancing surfactant content and function in vivo. In addition, induction of WIF may be related to surfactant based on its interaction with the TGF-β/β-catenin pathway and the inhibitory effect of TGF-β on surfactant production (36, 46). The remaining genes reflect a variety of functional roles (transcription, signaling, transport, adhesion, metabolism, and inflammatory) and are not overrepresented in any specific functional category. The gene DUOX1 is highly induced in adult cells (22-fold) compared with fetal cells (3-fold), and the mRNA is more abundant (959 f.u. vs. 88 f.u.); induction was confirmed by Western analysis and immunostaining. This membrane protein is a member of the NADH oxidase family and catalyzes formation of intracellular protons and extracellular hydrogen peroxide with a proposed role in bacteriostasis and signaling (21). In fetal lung, Duox1 is type II cell specific, developmentally regulated, and synergistically induced by glucocorticoid and cAMP (22). The current studies establish regulated expression in type II cells of adult lung consistent with a role for Duox1 in response to alveolar insults.

Limitations.

The precision of our expression profiling is influenced by biological variability between individual donors, including smoking history, and is likely further affected by variables such as extent of hypoxemia prior to death, use of mechanical ventilation, time from death to cell isolation, duration of the isolation procedure, and cell purity. The variability in lung gene expression between individuals is recognized and was similar for our experiments (median coefficient of variation of 32%) and those with postmortem human lung tissue (34). All donor lungs in our study were from males, which may have a minor effect on our results; in a gender-specific microarray study of lung tissue, 11% of X-linked genes were expressed at >1.5-fold different level between genders (60). The presence of some fibroblasts (∼5%) is a limitation in interpretation of results. For example, we found that ADFP is present and hormone responsive in both epithelial and fibroblast cells of fetal lung (data not shown), and the increased expression of some collagens in fetal vs. adult cells and after control culture of adult cells may reflect production primarily by fibroblasts. Cellular localization will need to be determined for specific genes of future interest, particularly for those of relatively low transcript abundance. The Affymetrix U133A chip was chosen to allow direct comparison between adult cells and previous results with fetal cells. However, this chip does not probe the entire human genome, and there are likely other type II cell genes of interest that are not included in this analysis. It should also be noted that our results for adult human type II cells may not necessarily reflect the expression level of specific genes in type II cells of other species. In a preliminary comparison of our human and previous rat type II cell (33) profile data, for example, a number of genes probed in both data sets were expressed at significantly different levels.

GRANTS

This research was supported by National Heart, Lung, and Blood Institute Grants HL-086323 (H. Fischer), HL-088193 (P. L. Ballard), HL-51856 and HL-51854 (M. A. Matthay), and HL-093026 (J. W. Lee) and the Foundation of Anesthesia Education and Research (J. W. Lee).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

ACKNOWLEDGMENTS

We thank E. Rapport for performing microarray analyses and P. Wang for technical assistance. We appreciate discussions and advice given by S. Guttentag, L. Dobbs, and R. Gonzalez.

REFERENCES

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
  51. 51.
  52. 52.
  53. 53.
  54. 54.
  55. 55.
  56. 56.
  57. 57.
  58. 58.
  59. 59.
  60. 60.
  61. 61.
  62. 62.
  63. 63.
  64. 64.
  65. 65.
  66. 66.
  67. 67.
  68. 68.
  69. 69.
View Abstract