Keratinocyte growth factor enhances barrier function without altering claudin expression in primary alveolar epithelial cells

Abstract

Keratinocyte growth factor (KGF) has efficacy in several experimental models of lung injury; however, the mechanisms underlying KGF's protective effect remain incompletely understood. This study was undertaken to determine whether KGF augments barrier function in primary rat alveolar epithelial cells grown in culture, specifically whether KGF alters tight junction function via claudin expression. KGF significantly increased alveolar epithelial barrier function in culture as assessed by transepithelial electrical resistance (TER) and paracellular permeability. Fluorescence-activated cell sorting of freshly isolated type 1 (AT1) and type 2 (AT2) cells followed by quantitative real-time RT-PCR revealed that more than 97% of claudin mRNA transcripts in these cells were for claudins-3, -4, and -18. Using cultured AT2 cells, we then examined the effect of KGF on the protein levels of the claudins with the highest mRNA levels: -3, -4, -5, -7, -12, -15, and -18. KGF did not alter the levels of any of the claudins tested, nor of zona occludens-1 (ZO-1) or occludin. Moreover, localization of claudins-3, -4, -18, and ZO-1 was unchanged. KGF did induce a marked increase in the apical perijunctional F-actin ring. Actin depolymerization with cytochalasin D blocked the KGF-mediated increase in TER without significantly changing TER in control cells. Together, these data support a novel mechanism by which KGF enhances alveolar barrier function, modulation of the actin cytoskeleton. In addition, these data demonstrate the complete claudin expression profile for AT1 and AT2 cells and indicate that claudins-3, -4, and -18 are the primary claudins expressed in these cell types.

disruption of the alveolar epithelial barrier is a hallmark of acute lung injury (59). Although alveolar barrier function relies on both endothelial and epithelial tight junctions, the epithelial barrier is much less permeable than the endothelial barrier (64). In fact, alveolar epithelial injury is sufficient for pulmonary edema formation, even in the presence of normal endothelial permeability (25). Furthermore, preservation of epithelial barrier function is a common mechanism by which keratinocyte growth factor (KGF) and other clinically relevant biological mediators provide protection in experimental lung injury (15, 18, 30, 38). However, despite the importance of epithelial integrity to outcomes for patients with clinical acute lung injury (34), the mechanisms of alveolar epithelial tight junction regulation remain incompletely understood.

Tight junctions ultimately determine epithelial paracellular permeability (43, 49), but many details about the nature of tight junctions between alveolar epithelial cells in health and disease are uncertain. For example, the specific tight junction proteins expressed by alveolar epithelial type 1 (AT1) and type 2 (AT2) cells have not been entirely elucidated (11, 12, 55), and whether AT1 and AT2 cells express similar junctional proteins has not been previously reported in detail. At baseline, multiple different cell-cell contacts exist in the alveolus, with AT1-AT1 and AT1-AT2 junctions being most abundant. During lung injury, AT2-AT2 junctions likely increase with proliferation of AT2 cells. With respect to paracellular permeability, the claudin family of tetraspan transmembrane proteins is foremost among tight junction constituents (51). Claudins demonstrate tissue-specific expression patterns that account for the different paracellular permeability characteristics of diverse epithelia (3). This is best exemplified in studies of the mouse nephron, where each claudin gene exhibits a unique pattern of expression along the nephron, and the segment-specific expression of different claudins regulates paracellular permeability properties (28). Previous work from our group has shown that a loss of claudins-3 and -4 in the alveolar epithelium impairs alveolar fluid clearance rates in mice and increases susceptibility to lung injury (65). Thus, a more complete understanding of the claudin contribution to paracelullar transport in the lung may have important clinical implications.

Considerable experimental data suggest that KGF may have clinical utility in preventing or ameliorating lung injury (5, 13, 32, 40, 53, 62, 67–69). Notably, recent data show that the protective effect of mesenchymal stem cells in experimental lung injury is attributable to KGF produced by the stem cells (30). KGF has pleiotropic effects in the lung (16, 58), and the mechanisms underlying KGF's protective effect remain incompletely understood. Although KGF is a potent mitogen for AT2 cells in vivo (50), this is unlikely to be the sole mechanism by which KGF attenuates lung injury (63). In fact, multiple non-mitogenic effects for KGF in the lung have been identified, including effects on ion transport (7, 48, 56). The effect of KGF on tight junction function and claudin protein expression has not been previously reported. In addition, KGF has been shown to modulate alveolar epithelial cell phenotype in vitro (8). Specifically, KGF appears to prevent freshly isolated AT2 cells from transdifferentiating toward a more AT1-like phenotype in culture (10, 23).

Therefore, the primary objective of this study was to further determine the mechanism for the protective effect of KGF on alveolar epithelial barrier function in primary alveolar epithelial cells grown in culture, specifically whether KGF alters tight junction function or claudin expression. Our hypothesis was that KGF treatment would augment barrier function through modulation of claudin expression. As claudins are the critical proteins regulating paracellular transport, a significant focus of this work was to determine which claudins are expressed in freshly isolated, purified populations of primary AT1 and AT2 cells and in cultured AT2 cells that have acquired a more AT1 cell-like phenotype. A second objective was to determine if cultured AT2 cells that adopt an AT1-like phenotype exhibit a different pattern of claudin expression or have different barrier properties from cells with a KGF-induced AT2-like phenotype. We show that KGF augments alveolar epithelial barrier tightness and that this effect is not dependent on cell proliferation. Contrary to our initial hypothesis, the barrier-enhancing effect of KGF is not due to alterations in claudin expression, but due to effects on the actin cytoskeleton. These data further show that AT1 and AT2 cells primarily express three claudins (-3, -4, and -18) and that changes in barrier tightness are not necessarily associated with changes in claudin expression in alveolar epithelial cells.

MATERIALS AND METHODS

Isolation of AT1 and AT2 cells by flow cytometry.

Animal protocols were approved by the local Institutional Animal Care and Use Committee and conform to National Institutes of Health animal use guidelines. Rats (150–200 g) were anesthetized with 4% isoflurane followed by an intraperitoneal injection of ketamine/xylazine (90/10 mg/kg body wt). Lungs were perfused with buffered saline containing heparin, removed en bloc, and lavaged with calcium-magnesium-free PBS containing 1 mM EGTA and 1 mM EDTA (pH 7.4) followed by RPMI containing HEPES. The lungs were instilled with elastase (Roche) for 20 min and minced. Resulting suspensions were filtered through membranes of decreasing porosity to obtain single cell suspensions. After filtration, cells were incubated with 50 μg/ml rat IgG (Sigma) and 50 μg/ml mouse IgG (Sigma) for 10 min at 4°C to block Fc receptor binding sites. Alexa 610 (Invitrogen)-conjugated anti-RT140, an antibody specific to AT1 cells, and anti-RT270, an antibody specific to AT2 cells (both kind gifts of Dr. Leland G. Dobbs, Univ. of California, San Francisco) were used to sort both cell types by flow cytometry as previously described (24). Alexa 488 anti-mouse IgG3 was used to visualize RT270. Differentially labeled rat AT1 and AT2 cells were sorted using FACS Aria (BD Biosciences), and results were analyzed using BD Diva version 6.1.3 (BD Biosciences). AT1 and AT2 cell purities were greater than 99% by immunofluorescence using anti-RT140 and -RT270.

Isolation and primary culture of rat type 2 alveolar epithelial cells.

For cell culture studies, alveolar epithelial type 2 cells were isolated from the lungs of pathogen-free male Sprague-Dawley rats as previously described (14, 22). Lungs were filtered and digested as above. Resulting suspensions were enriched for AT2 cells by differential adherence to IgG-coated plastic dishes. Unattached cells were collected and plated in DME-H21 supplemented with ciprofloxacin and 10% FBS at a density of 1.5 × 10⁶ cells/cm². Cell purity was >90% and cell viability was >95% (trypan blue exclusion). Cells were cultured for 5 days on 0.4-μm polycarbonate membrane Transwell permeable supports (Corning Costar). Cells in the treatment group were given media supplemented with 10 ng/ml KGF (Human Recombinant, Sigma) from the time of initial seeding consistent with previously published work (20). On day 4, serum-free media with or without KGF was added to both the apical and basal compartments.

Barrier function measurements.

Transepithelial electrical resistance (TER) of cells in medium cultured on permeable supports was measured using an ohmmeter (EVOM, World Precision Instruments) as previously described (18) (n = 9 biological replicates). To determine in vitro paracellular permeability to larger solutes, 10 μg/ml of the 0.5 kDa tracer pyranine (Sigma) was added to the apical compartment on day 5, and fluorescence was measured in a sample of medium from the basal compartment at 30 min at excitation and emission wavelengths of 485 and 527 nm, respectively (n = 7 biological replicates). Apparent permeability was calculated (37). Preliminary studies confirmed a linear correlation for pyranine flux between 0 and 2 h in alveolar epithelial cells grown with or without KGF (Supplemental Fig. 1; Supplemental material for this article is available online at the Journal website), consistent with previously published data for similar-sized molecules in primary rat alveolar epithelial cells (37). To investigate the role of the actin cytoskeleton in changes in barrier properties with KGF treatment, TER was measured after 15 min of treating cells with 0.1 μg/ml of the actin depolymerization agent cytochalasin D (Enzo Life Sciences) on day 5 (n = 4 biological replicates) (31, 47). For extracellular calcium depletion, TER was measured 15 min after changing the apical and basolateral chambers to serum-free media containing 1 mM EDTA and EGTA (54).

Determination of cell number.

To determine the effect of KGF on proliferation, cell number was determined as previously described (19). Isolated primary AT2 cells from rats (n = 4 biological replicates) were plated on 96-well tissue culture-treated plates at a seeding density identical to all Transwell experiments. A standard curve of increasing seeding density was also plated for each isolation. The cells were cultured and treated as for the in vitro permeability studies. At day 5, cells were washed with PBS and fixed in 70% ethanol for 10 min. Cells were then stained with 1% crystal violet (15 min) and thoroughly rinsed with tap water. The stain was extracted from cells by adding 100 μl of cell lysis buffer (1% Triton X-100, 0.1% SDS, 0.5% sodium deoxycholate, and 1% Nonidet P-40). Relative quantification was done by reading absorbance at 570 nm on a plate reader.

RNA isolation and quantitative real-time RT-PCR.

Total RNA was isolated from primary rat AT1 and AT2 cells sorted by flow cytometry (Qiagen) (n = 4 biological replicates). Additional DNase treatment was performed to minimize contamination with genomic DNA (Invitrogen). cDNA was synthesized using random hexamers (SuperScript III, Invitrogen). Quantitative real-time PCR using SYBR green (Applied Biosystems) was performed to amplify claudins-1, -2, -3, -4, -5, -6, -7, -8, -9, -10a, -10b, -11, -12, -14, -15, -16, -17, -18, -19, -20, -22, and -23 using primers as listed in Table 1. Specific products were confirmed by melt curve analysis. DNA standard curves were created for all claudins analyzed, α-tubulin, β-actin, and β2-microglobulin. The gene of interest was amplified and cloned into pCR2.1 (Invitrogen). The mass of one copy for each gene was calculated to estimate mass of plasmid required for specific number of copies in the standard curve. Copy number of the gene of interest was determined by comparison with the external DNA standard curve. Housekeeping genes β-actin, β2-microglobulin, and α-tubulin were amplified as internal controls, and all expression values were also normalized to the expression level of the housekeeping genes. This normalization of gene expression was done using GeNorm (52). For additional validation of the data, probes were used for the quantification of claudins-3, -4, -7, -12, -15, -18, -19, and α-tubulin (Table 1).

Table 1.

Primers and probes used for quantitative real-time PCR in freshly isolated, purified AT1 and AT2 cells

Claudin	Forward Primer	Reverse Primer	Probe
Claudin-1	CTGGGAGGTGCCCTACTTT	CCGCTGTCACACGTAGTCTT
Claudin-2	AACCTGGGCTCCAAAGAAAC	AATGGGTGGCAATTAGAGGA
Claudin-3	AGATGTACGACTCGCTGCTG	CTTGGCCGTCTCATCTTGT	CCTACTGGCAGCCTTCGGGC
Claudin-4	CTCTCGCCTCCACGTTACTC	TCAGTCATCCTCGACACCAA	CGCCCGCGTTTCTGAGCACACCAGG
Claudin-5	CAGGCTCTTGTGAGGACTTG	TGCCCTTTCAGGTTAGCAG
Claudin-6	CTGCAAATCTTGGGGATTGT	AGCAGCGAGTCGTACACCTT
Claudin-7	ATGCTCCTGGATTGGTCATC	CCTGCCCAGCCGATAAAG	AACCCCTTGACGCCCATGAATATT
Claudin-8	CCCCGAGCATATACTCCAAAA	AGACGCTGCGTATTTCCTG
Claudin-9	GGCTGAGAGCTGTGAAGACC	AGCAGCGAGTCGTACACCTT
Claudin-10a	CAGGGTCTGTGGATGAACTG	AGTCCTCTACACGCCTGGAT
Claudin-10b	TCCACACTACCCACCGACTA	CAGGTTGGCAAAATAAGTGG
Claudin-11	AAGGGTTACCAGCGACAATG	AGAGACCCAACCCCTCAACT
Claudin-12	GCGACTCATCACATTCAACA	AGTCACTGCTTCCGTCATACC	CGGGCCTGTGGGTGAAGTGCGCCC
Claudin-14	AAGCTCCCATTTGGTGAATG	AGGAAGCCTAGGAGCTGGAC
Claudin-15	AGCCACCTCTATCCCAATCC	CCAGGGCTGACATGAAGAA	TGGGGCTCAAGCCCCCTTGCGGG
Claudin-16	AGCAAGCACGATGTGTTCAG	AGCTGGAATGCGATTGACTT
Claudin-17	GACCGCCAATATCATCATCC	TGACCTTCCTCTGGCTGTCT
Claudin-18	AAGGGCTCTGGAGGAGTTG	GCCCAGGATGGTGAAGTATG	CGTGCAACAGAGCTCGGGG
Claudin-19	CTGTCAACGCCAGGTATGAA	TGCTGTTGGCTCTCTCAGG	TGGGCTGGCCATCCTGGGCGGT
Claudin-20	TTTCATCCTGGCTGTGTCTG	TGCAGCTAAACATCCCAGTG
Claudin-22	GGAGGATGTGTGCTTCACTG	AGGAGAGCCTACCAAGTCCA
Claudin-23	GTTGCTGCTCAATCTCGTCA	CAGGCTGAGTCCCGAAGTAG
α-Tubulin	GCCCTACAATTCCATCCTCA	CGCTCAATGTCGAGGTTTCT	GCACTCTGATTGTGCCTTCA
β-Actin	CATTGCTGACAGGATGCAGA	CTGGAAGGTGGACAGTGAGG
β2-Microglobulin	TGACCGTGATCTTTCTGGTG	ATCTGAGGTGGGTGGAACTG

All claudin mRNA levels were measured using Sybr Green. Levels of mRNA were also measured for some claudins using fluorescent detection probes.

Immunoblot.

After 5 days in culture, cells were harvested and lysed in complete cell lysis buffer (10 mM Tris·HCl, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% Triton X-100, 0.1% SDS, 0.5% sodium deoxycholate, 1% Nonidet P-40, protease and phosphatase inhibitor cocktails, Sigma). Equal amounts of protein were resolved by SDS-PAGE under reducing conditions and blotted onto PVDF or nitrocellulose membranes. Membranes were blocked for 1 h at room temperature with 5% nonfat milk and probed for claudin-3, -4, -5, -7, -12, -15, -18, occludin, and ZO-1 (all from Invitrogen) overnight at 4°C. HRP-conjugated anti-mouse and anti-rabbit were used as secondary antibodies (Cell Signaling). β-Tubulin (Santa Cruz Biotechnology) was used as the loading control for densitometry analysis (ImageJ) in all conditions except for immunoblots of freshly isolated cells. In this instance, β-actin (Cell Signaling) was used as the loading control, and cells were harvested immediately after FACS analysis and sorting as described above.

Immunofluorescence staining.

Freshly isolated, unpurified cell suspensions were deposited onto glass slides using a Shandon Cytospin (Thermo Fisher Scientific). Samples were fixed in 4% paraformaldehyde, and antigen retrieval was done using citrate buffer at pH 6.0 (Dako Cytomation). Slides were incubated with 2% goat serum in all reactions to block nonspecific binding of antibodies. Cells were stained with antibodies against RT140, RT270, claudin-3, claudin-4, claudin-15, and claudin-18. Samples were incubated with primary antibody followed by the appropriate goat anti-mouse or goat anti-rabbit IgG Alexa 488-, Alexa 647-, or Alexa 594-secondary antibodies at 1:200 (Invitrogen). For cultured, primary rat AT2 cells, Transwells were fixed in ice-cold methanol:acetone (1:1) or room temperature 4% paraformaldehyde on day 5. Samples were permeabilized by exposure to 0.5% Triton X-100 in 1% BSA (EMD Chemicals) for 5 min and then blocked in 10% goat serum for 1 h. Transwells were incubated for 1 h at room temperature in primary antibodies against RT140, RT270, claudin-3, claudin-4, claudin-18, or ZO-1 followed by the appropriate goat anti-mouse or goat anti-rabbit IgG Alexa 488- or Alexa 647-secondary antibodies at 1:200. For filamentous actin detection, Transwells were stained with 1:40 Alexa 568 phalloidin (Invitrogen) for 1 h. All samples were mounted using Vectashield mounting media for fluorescence (Vector Labs). Confocal images were taken using LSM 510 Meta (Carl Zeiss) with EC Plan-Neofluar ×40 or Plan Apochromat ×63 oil lenses at a resolution of 1,024 × 1,024. Images were acquired using HAL-100 and LSM 510 Meta software. Epifluorescence images were obtained using Nikon Eclipse E600 (Nikon) with either a plan ×40 or ×20 objective. Images were acquired using Spot (RT Color Diagnostics Instruments) and Spot Advanced software. All images were processed using ImageJ or Photoshop (Adobe) and Illustrator (Adobe).

Statistics.

Differences between groups were compared using unpaired t-tests or ANOVA with post hoc Bonferroni correction for multiple comparisons. Each n represents cell isolation from a single rat. A P value ≤0.05 was considered significant. Data are reported as means ± SD or means ± SE where indicated.

RESULTS

Effect of KGF on barrier properties, cell phenotype, and proliferation.

KGF (10 ng/ml) significantly increased TER in alveolar epithelial cell monolayers consistent with enhanced barrier function (710 ± 51 Ω·cm² vs. 1,310 ± 130 Ω·cm², n = 9 biological replicates, P < 0.001; Fig. 1A). In addition, KGF treatment produced a small but significant decrease in paracellular permeability to the 0.5 kDa tracer pyranine (apparent permeability, P_APP, 1.53 ± 0.13 vs. 1.11 ± 0.13 × 10⁻⁷ cm/s, n = 7 biological replicates, P < 0.05; Fig. 1B). These values were less than 1% of P_APP across cell-free Transwells. Consistent with previous reports (8), KGF-treated alveolar epithelial cells were more AT2-like as indicated by lack of RT140 staining after 5 days in culture (Fig. 2A). There was no difference in cell number with KGF treatment (1.67 ± 0.27 vs. 1.76 ± 0.26 × 10⁵, n = 4 biological replicates, P = NS; Fig. 2B).

Fig. 1.

Effect of keratinocyte growth factor (KGF) on barrier function of alveolar epithelial cells. Freshly isolated alveolar epithelial type 2 (AT2) cells were cultured for 5 days in control media or media supplemented with 10 ng/ml KGF. A: transepithelial electrical resistance (TER) was significantly greater in KGF-treated cells. *P < 0.001 compared with control, n = 9 biological replicates. B: apparent permeability (P_APP) to the 0.5 kDa tracer molecule pyranine was significantly decreased in KGF-treated cells. *P < 0.05 compared with control, n = 7 biological replicates. These values were less than 1% of P_APP across cell-free Transwells. Data are expressed as means ± SE.

Fig. 2.

Effect of KGF administration on alveolar epithelial cell phenotype and proliferation. Freshly isolated AT2 cells were cultured for 5 days in control media or media supplemented with KGF. A: at day 5, cells were fixed and immunolabeled for the AT2 marker RT270 (green) and the AT1 marker RT140 (red). Although cells expressed RT270 under both conditions, KGF-treated cells were more AT2-like given the absence of RT140 staining. Scale bars, 50 μm. B: cell number was measured in confluent monolayers of control and KGF-treated cells at day 5 in culture. There was no difference in cell number between the 2 groups as measured by quantification of crystal violet staining (n = 4 biological replicates). Data are expressed as means ± SE.

Claudin expression profile in type 1 and type 2 alveolar epithelial cells.

To determine the claudin expression profile in alveolar epithelial cells, we used fluorescence-activated cell sorting to isolate AT1 and AT2 cells with >99% purity (24) (Fig. 3) (n = 4 biological replicates). For all known claudins, mRNA expression was compared by quantitative real-time PCR using external standard curves such that mRNA copy numbers could be compared among genes. The mRNA levels for claudins-1, -2, -6, -8, -10a, -14, -16, and -17 were below the limits of detection in both cell types. AT1 cells expressed primarily claudins-18, -3, and -4 (P < 0.05 compared with all other claudins expressed in AT1 cells) with detectable levels of claudins-15, -12, -7, -5, -19, -22, -11, -10b, -9, -20, and -23 in decreasing order (Fig. 4A). In contrast, freshly isolated AT2 cells (Fig. 4A) expressed primarily claudins-3, -4, and -18, with detectable levels of claudins-5, -7, -10b, -12, -19, -9, -15, -22, -20, and -23 in decreasing order (P < 0.05 for claudins-3 and -4 compared with all other claudins expressed in AT2 cells). Claudin-11 was below the limit of detection in AT2 cells. The majority (97%) of claudin mRNA transcripts in both cells were for claudins-3, -4, and -18, although the relative amount of each of these was not the same (Fig. 4B). Comparison of claudin expression in both cell types together by analysis of variance followed by Bonferroni correction for multiple comparisons demonstrated that claudin-3 was the only transcript expressed at a statistically significantly different level between AT1 and AT2 cells with 10-fold more claudin-3 mRNA transcript in AT2 cells (P < 0.05). Probe-based and Sybr Green-based assays of mRNA expression yielded similar results.

Fig. 3.

Isolation of AT1 and AT2 cells by fluorescence-activated cell sorting. Lung cells liberated by digestion with elastase were stained with anti-RT140 (AT1 cell marker, red) and anti-RT270 (AT2 cell marker, green) as described in text. A: cytocentrifuged preparation of mixed cell population before FACS analysis and sorting showing AT1 (red) and AT2 (green) cells together. B–E: cytocentrifuged preparations of essentially pure (>99%) AT1 (B and C, red) and AT2 (D and E) cells after FACS analysis and sorting. Magnification, ×10 (B and D) and ×40 (A, C, and E).

Fig. 4.

Claudin mRNA expression levels in freshly isolated, purified AT1 and AT2 cells. To determine which claudins are expressed in these cell types, mRNA expression was compared in AT1 and AT2 cells for all known claudins using quantitative real-time PCR as described in text. A: AT1 cells (left) expressed primarily claudins-18, -3, and -4 (*P < 0.05 compared with all other claudins expressed in AT1 cells) with detectable levels of claudins-15, -12, -7, -5, -19, -22, -11, -10b, -9, -20, and -23 in decreasing order. AT2 cells (right) expressed primarily claudins-3, -4, and -18, with detectable levels of -5, -7, -10b, -12, -19, -9, -15, -22, -20, and -23 in decreasing order (*P < 0.05 for claudins-3 and -4 compared with all other claudins expressed in AT2 cells). B: >97% of claudin mRNA transcripts in AT1 (left) and AT2 (right) cells were for claudins-3, -4, and -18, although the relative amounts of each of these varied with claudin-18 being most highly expressed in AT1 cells and claudin-3 being the dominant transcript in AT2 cells. After analysis of variance followed by Bonferroni correction for multiple comparisons, claudin-3 was the only transcript expressed at a statistically significantly different level between AT1 and AT2 cells with 10-fold less claudin-3 in AT1 cells (P < 0.05). Data are expressed as means ± SE.

To assess claudin protein expression levels, immunofluorescence staining of freshly isolated, unpurified mixed whole lung cell preparations before flow cytometry was done (Fig. 5A). Consistent with the mRNA expression data, claudin-3 appeared more abundant in AT2 cells than AT1 cells. Claudins-4, -15, and -18 were detected in both AT1 and AT2 cells. As this immunofluorescence staining was done without fluorescence-activated cell sorting, cells negative for both AT1 and AT2 cell markers were present. Comprehensive immunoblot analysis was technically limited due to the quantity of material obtained from AT1 and AT2 cells after fluorescence-activated cell sorting and the limited availability of suitable antibodies. Thus, quantitative analysis of protein expression was not done on the freshly isolated cells. However, claudin-3, the only claudin with a statistically significantly different mRNA transcript level between AT1 and AT2 cells, was expressed at a 17-fold higher level in AT2 cells (P < 0.001, Fig. 5B) consistent with the mRNA expression data.

Fig. 5.

Claudin protein expression in AT1 and AT2 cells. A: cytocentrifuged preparations of lung cells before FACS analysis and sorting were stained for RT140 (red, lower left), RT270 (green, top left), claudin-3, -4, -15, or -18 (all blue, top right). Coexpression of specific claudins with RT140 and RT270 was also assessed (bottom right). Claudin-3 was primarily expressed in AT2 cells, although staining in AT1 cells was also detected. Claudins-4, -15, and -18 were expressed in AT1 and AT2 cells. Of note, cells negative for both AT1 and AT2 markers in the unpurified cell preparations stained positive for claudins-15, -4, and -3. Magnification, ×40 for all images. B: immunoblot analysis of claudin-3 in isolated, purified AT1 and AT2 cells after flow cytometry. β-actin was also labeled to normalize for total sample protein content. Claudin-3 protein levels were 17-fold greater in AT2 cells (*P < 0.001, n = 3 biological replicates). Data are expressed as means ± SE.

Effect of KGF on claudin expression and localization in cultured AT2 cells.

KGF did not significantly change the protein level of any tested tight junction proteins in cultured alveolar epithelial cells: claudins-3, -4, -5, -7, -12, -15, -18, occludin, and ZO-1 (n = 3 biological replicates, densitometry P = not significant for all groups; Fig. 6). These claudins were selected based on the results of the claudin expression profile determined in freshly isolated primary AT1 and AT2 cells (Fig. 4). There were also no differences in the localization of ZO-1 and claudins-3, -4, and -18 between control and KGF-treated cells (Fig. 7). All of these tight junction proteins were predominantly localized to the cell membrane at cell-cell contacts.

Fig. 6.

Effect of KGF administration on the expression of tight junction proteins in alveolar epithelial cells. Freshly isolated AT2 cells were cultured on permeable supports for 5 days in control media or media supplemented with KGF and then harvested and resolved by SDS-PAGE. Levels of claudin-3, -4, -5, -7, -12, -15, -18, zonula occludens (ZO)-1, and occludin were determined by immunoblot. Immunoblots were also labeled for β-tubulin to normalize for total sample protein content. Densitometry revealed no significant differences in expression levels of these tight junction proteins in KGF-treated cells compared with control (n = 3 biological replicates). Data are expressed as means ± SE.

Fig. 7.

Localization of tight junction proteins cultured for 5 days in the absence or presence of KGF. Freshly isolated AT2 cells were plated on permeable supports in control media or media supplemented with KGF, and then cultured for 5 days, fixed, and immunostained for claudin-3, claudin-4, claudin-18, and ZO-1. There were no differences in localization noted for these tight junction proteins in KGF-treated cells compared with control. All proteins were predominantly localized to the membrane at cell-cell contacts. Scale bars, 20 μm.

Mechanism of KGF effect on alveolar epithelial barrier properties.

As changes in actin organization have been associated with changes in barrier permeability in epithelia (36, 47), the effect of KGF on the actin cytoskeleton was investigated. Cells cultured in the presence of KGF showed a marked increase in the apical perijunctional F-actin ring compared with controls (Fig. 8, A and B, and Supplemental Fig. 2). To determine if this change in cytoskeletal structure was associated with the KGF effect on TER, we treated the alveolar epithelial cell monolayers with the actin depolymerization agent cytochalasin D. Cytochalasin D completely abolished the barrier-enhancing effect of KGF on cultured AT2 cells without significantly decreasing TER in control cells (Fig. 8C). This decrease in TER in KGF-treated alveolar epithelial cells after cytochalasin D treatment was associated with disruption of the perijunctional F-actin ring (Fig. 8D). Moreover, tight junctions were still maintained after cytochalasin D treatment given the relatively preserved TER in control and KGF-treated alveolar epithelial cells compared with cells treated with calcium-free media, an intervention that results in the dissolution of tight junctions (Fig. 8C).

Fig. 8.

Effect of KGF administration on the actin cytoskeleton. Freshly isolated AT2 cells were plated on permeable supports in control media or media supplemented with KGF, and then cultured for 5 days, fixed, stained for F-actin with Alexa 568 phalloidin, and imaged under epifluorescence (A and B) or confocal microscopy (D). A and B: KGF-treated cells (B) demonstrated significant reorganization of the actin cytoskeleton with a marked increase in the apical perijunctional F-actin ring compared with control cells (A). Inset is higher magnification image (n = at least 4 biological replicates). Scale bars, 50 μm on ×20 image and 20 μm on inset ×40 image. C: actin depolymerization with cytochalasin D treatment decreased TER to control levels (P < 0.05) without a significant decrease in control TER (n = 4 biological replicates consisting of 4–6 technical replicates). Treatment of control and KGF-treated AT2 cells with calcium-free media resulted in near complete loss of TER. Data are expressed as means ± SE. D: the perijunctional F-actin ring in KGF-treated alveolar epithelial cells (top) was disrupted by treatment with cytochalasin D (bottom) (n = at least 4 biological replicates). Scale bars, 10 μm.

DISCUSSION

Contrary to our initial hypothesis, the major finding of this study is that KGF enhances alveolar epithelial tightness by altering the actin cytoskeleton rather than by changing claudin expression or trafficking. A second major finding is that AT1 and AT2 cells primarily express claudins-3, -4, and -18 at the mRNA level: these claudins account for 97% of transcripts in each cell type. However, the two cell types express these claudins in differing proportions. AT2 cells express 10-fold more claudin-3 mRNA and 17-fold more claudin-3 protein than AT1 cells. To our knowledge, this is the first report of the complete claudin transcript profile of primary AT1 and AT2 cells. A third finding is that although KGF contributes to the preservation of certain features of the AT2 phenotype in cultured AT2 cells, which normally adopt a more AT1-like phenotype, KGF does not alter claudin expression.

The preservation of intact alveolar epithelial barrier function is associated with improved clinical outcomes in patients with acute lung injury (34, 57). Successful recovery from lung injury requires the proliferation and differentiation of AT2 cells and formation of tight junctions (2, 33). KGF is a potent epithelial mitogen in vivo (50) that has been shown to ameliorate lung injury in multiple experimental models: bleomycin (13, 69), acid installation (67), hyperoxia (40), hydrogen peroxide (66), pseudomonas infection (53), radiation (69), and ventilator-induced lung injury (62). KGF also appears to be critical to the therapeutic capacity of human mesenchymal stem cells to restore alveolar epithelial fluid transport in acute lung injury (30). Although KGF induces AT2 cell proliferation in vivo, this effect has been less consistent in in vitro studies (4, 17, 20, 41). Alveolar epithelial type 2 cells have a low proliferative index when cultured in vitro (4, 21, 24), and our results are consistent with others that have found no increase in cell number following KGF treatment in vitro (20). KGF also has multiple non-mitogenic effects that are of potential importance (17). For instance, KGF upregulates active ion transport in alveolar epithelial cells by increasing Na⁺-K⁺-ATPase expression (7), which explains in part the ability of KGF to increase alveolar fluid clearance through non-mitogenic mechanisms (56). KGF also augments pulmonary surfactant production (48) and enhances alveolar epithelial repair (4). In this study, we demonstrate an additional mechanism by which KGF may preserve alveolar epithelial barrier function: modulation of tight junction function via changes in the actin cytoskeleton.

KGF promoted alveolar barrier tightness as measured by electrical resistance and permeability to a 0.5 kDa tracer. The increase in TER with KGF treatment in this study is particularly notable given that this growth factor is known to increase transcellular active ion transport in alveolar epithelial cells in culture (7) and in vivo (26). Although KGF is known to influence the actin cytoskeleton in different model systems (6, 9, 20, 35, 39, 42, 46, 60), previous studies have not reported that KGF increases TER in cultured AT2 cell via modulation of the actin cytoskeleton. However, Waters and colleagues (60) demonstrated that KGF pretreatment prevented hydrogen peroxide-induced disruption of the perijunctional F-actin ring in airway epithelia, thus preventing an increase in paracellular protein permeability. Savla and Waters (42) similarly demonstrated stabilization of the F-actin cytoskeleton with KGF treatment and resultant protection of airway epithelial barrier function after radiation exposure. Data from the present study extend these prior studies by showing that KGF treatment results in the formation of a more pronounced perijunctional actin ring and decreased permeability in cultured AT2 cells.

The importance of the F-actin cytoskeleton in influencing alveolar epithelial barrier function has been previously described but is not well characterized. For instance, although thrombin induces actin stress fiber formation and loss of barrier integrity in lung endothelial cells, it has the opposite effect in alveolar epithelial monolayers: increased perijunctional F-actin formation and an increase in TER (27). Importantly, soluble factors secreted by human mesenchymal stem cells, namely KGF, angiopoietin-1 (Ang1), and IL-1 receptor antagonist have recently been demonstrated to augment alveolar epithelial barrier function in models of acute lung injury and inflammation (15, 30, 38). Whereas the effect of KGF on paracellular permeability in these models is currently not well understood (30), Ang1 and human mesenchymal stem cells decreased epithelial permeability following cytomix-induced injury of cultured AT2 cells in part through cytoskeletal reorganization (15). Specifically, Ang1 promoted apical peripheral F-actin formation similar to the effect of KGF in our study. Therefore, modulation of the actin cytoskeleton may represent a common mechanism by which soluble factors secreted by human mesenchymal stem cells augment alveolar epithelial barrier function. As these cells are being studied for potential therapy in acute lung injury (61), the mechanism of cytoskeletal reorganization should be investigated further.

In many cell types, cytoskeletal disassembly disrupts barrier function, an effect mediated in part by the interaction of the apical junction complex and actin cytoskeleton (36, 44, 49). In this study, we used cytochalasin D, an actin depolymerization agent that leads to shortening and severing of existing microfilaments (45), to assess the contribution of increased perijunctional F-actin to KGF-induced increases in TER (47). Cytochalasin D treatment caused the TER of KGF-treated cells to return to baseline but did not significantly change TER in control cells, which had a much smaller perijunctional actin ring. This decrease in TER in KGF-treated alveolar epithelial cells was associated with disruption of the perijunctional F-actin ring as demonstrated by confocal microscopy. Furthermore, dissolution of existing tight junctions via extracellular calcium depletion resulted in near complete loss of TER in both groups. These data support the conclusion that the KGF-induced augmentation of barrier function is at least in part dependent on the changes in perijunctional actin.

Claudin expression in the alveolar epithelium can be regulated by multiple factors in the cell microenvironment, including the extracellular matrix (29) and proinflammatory cytokines (15). Given that KGF has been previously demonstrated to alter connexin expression and gap junctional communication in cultured alveolar epithelial cells (1), we hypothesized that KGF would also regulate claudin expression. However, KGF administration in vitro was not associated with changes in claudin expression in this study. An additional reason we elected to study the effects of KGF on tight junction structure and function was its ability to promote an AT2-cell like phenotype in culture. We hypothesized that maintaining a more AT2 cell-like phenotype in culture with KGF would allow modeling of the AT2 cell-cell contacts that potentially occur during acute lung injury. KGF did promote a more AT2-like phenotype as measured by absence of RT140 expression, but claudin expression was not affected. This is consistent with the inability of KGF to maintain all components of the AT2 phenotype (48). Cultured type 2 cells acquire many characteristics of type 1 cells, but are distinct from freshly isolated type 1 cells based on mRNA expression data (23).

Although the present studies identify an additional mechanism by which KGF may exert its beneficial effects in acute lung injury, another important aspect of this work is the complete description of the claudin mRNA expression profile in AT1 and AT2 cells. Claudins are essential in regulating paracellular permeability (51), and delineation of the claudin profile of the lung is necessary for the study of alveolar epithelial barrier function in acute lung injury. Previous studies are consistent with the findings of this study, but prior studies have not reported the complete claudin mRNA expression profile of primary type 1 and type 2 cells (12, 55). However, a limitation of this study is that protein levels for most claudins in freshly isolated AT1 and AT2 cells can only be inferred from mRNA levels. An experimental approach focused on detecting mRNA levels was selected given it provided the best ability to quantitatively compare expression for all claudins in the two cell types, even those for which suitable antibodies are not readily available. Comprehensive immunoblot analysis was also technically limited due to the quantity of material obtained from AT1 and AT2 cells after fluorescence-activated cell sorting. It is likely that there are some differences between claudin mRNA expression levels and actual protein levels in these cell types. However, for claudin-3, the only claudin with a statistically significantly different transcript level between AT1 and AT2 cells, the mRNA and protein data were consistent. In addition, immunostaining of mixed populations of AT1 and AT2 cells prior to FACS analysis and sorting supported the mRNA data on claudin expression levels. Moreover, that claudins-3, -4, and -18 make up more than 97% of the claudin transcripts in both AT1 and AT2 cells could potentially guide future experiments on paracellular transport in the lung. However, claudins expressed at low levels may have important effects on barrier properties that would not be predicted by their levels alone. The physiological consequences of higher claudin-3 levels in AT2 cells are uncertain; however, prior studies have shown that claudin-3 can bind with different claudins on opposing cell membranes, a quality not shared with all claudin family members (56).

In conclusion, we report an additional non-mitogenic mechanism by which KGF exerts a beneficial effect in acute lung injury, specifically enhanced alveolar barrier function through reorganization of the F-actin cytoskeleton that does not require a change in claudin expression. In addition, we describe the complete claudin mRNA expression profile in purified, freshly isolated rat AT1 and AT2 cells. These data will be important in future studies of the regulation of paracellular permeability in the alveolar epithelium during acute lung injury. There are important clinical implications to these findings. Because KGF shows promise as a potential therapeutic agent in acute lung injury, these data elaborate on the barrier-specific effects of this pleiotropic agent and identify what may be a common mechanism by which several barrier-enhancing mediators exert their effects. Further characterization of the barrier-enhancing properties of KGF independent from its other effects could lead to a targeted treatment of acute lung injury.

GRANTS

This work was supported by National Heart, Lung, and Blood Institute Grant HL-088440.

DISCLOSURES

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