Laminin-driven Epac/Rap1 regulation of epithelial barriers on decellularized matrix

Abstract

Decellularized tissues offer a unique tool for developing regenerative biomaterials or in vitro platforms for the study of cell-extracellular matrix (ECM) interactions. One main challenge associated with decellularized lung tissue is that ECM components can be stripped away or altered by the detergents used to remove cellular debris. Without characterizing the composition of lung decellularized ECM (dECM) and the cellular response caused by the altered composition, it is difficult to utilize dECM for regeneration and specifically, engineering the complexities of the alveolar-capillary barrier. This study takes steps towards uncovering if dECM must be enhanced with lost ECM proteins to achieve proper epithelial barrier formation. To achieve this, the epithelial barrier function was assessed on dECM coatings with and without the systematic addition of several key basement membrane proteins. After comparing barrier function on collagen I, fibronectin, laminin, and dECM in varying combinations as an in vitro coating, the alveolar epithelium exhibited superior barrier function when dECM was supplemented with laminin as evidenced by trans-epithelial electrical resistance (TEER) and permeability assays. Increased barrier resistance with laminin addition was associated with upregulation of Claudin-18, E-cadherin, and junction adhesion molecule (JAM)-A, and stabilization of zonula occludens (ZO)-1 at junction complexes. The Epac/Rap1 pathway was observed to play a role in the ECM-mediated barrier function determined by protein expression and Epac inhibition. These findings revealed potential ECM coatings and molecular therapeutic targets for improved regeneration with decellularized scaffolds.

Graphical Abstract

1. INTRODUCTION

Decellularized lung tissue engineering offers a promising regeneration scaffold for patients with a variety of incurable pulmonary diseases. This technology can provide the native structure and biochemical cues of the lung to guide healthy maturation of functional tissue. Decellularized lungs are the gold standard scaffold for whole lung bioengineering; however, current approaches do not produce a long-term, functional lung replacement, indicated by severe edema after implantation. Insufficient recellularization and immature barriers keep this technology from the clinic. Most of the research to date has focused on assessments of stem cell differentiation and attachment [1–4] or mimicking lung development during regeneration [5,6]. More recently, researchers in this field have examined recellularization on a smaller scale by evaluating endothelial barriers [7]; however, there is little research on epithelial barrier formation with recellularized lungs. This research has taken a unique approach to re-epithelialization of whole organ scaffolds by examining specifically alveolar junction formation on a cellular level to then inform future recellularization strategies. Therefore, we have attempted to systematically determine which components are essential to the alveolar barrier function and developed a scaffold coating to replenish these key matrix proteins.

Decellularization techniques utilize several harsh detergents to ensure complete removal of cellular debris, but these detergents can drastically alter the ratio of ECM components that are left behind. Matrix proteins such as collagens, elastin, laminin, fibronectin, and glycosaminoglycans (GAGs) are preserved through the decellularization process, but there are reports of up to a 50% loss of most of these components compared to native ECM [8,9]. Variations in ECM composition and stiffness can have a profound effect on cell phenotype, as well as cell attachment and barrier function [10–12]. The delicate biochemical and physical communication between cells and the ECM environment guides processes from embryogenesis to apoptosis.

The alveolar epithelium that forms gas-exchange barriers resides on a BM, composed of laminin, collagen type IV, nidogen/entactins, fibronectin, and proteoglycans. The composition of the BM is a primary regulator of the epithelial phenotype [13]. High collagen I and fibronectin content within the surrounding matrix promotes cell attachment but at the expense of suboptimal barriers, indicative of epithelial to mesenchymal transition and fibrosis [14–18]. The most abundant components of the BM are laminin and collagen IV that together self-assemble into a dense structure to support the epithelial monolayer [19]. Collagen IV provides the backbone structure of the blood-gas barrier and has been known to guide alveolar development and differentiation [20,21]. Laminin is integral to healthy barrier formation, regulation of ECM production, and epithelial differentiation [22–29]. Other, less abundant ECM proteins such as tenascin-C and fibrillin-2 stimulate epithelial proliferation during alveologenesis [5]. The main focus of this research is to determine the effect of BM composition in the context of post-decellularization on strictly alveolar gas-exchange barrier formation.

Homeostatic alveolar epithelial barrier function is maintained by two main structures: tight junctions (TJ) and adherens junctions (AJ). Both junctions provide tension resistance and selective permeability during resting mechanical load. Formation of strong barriers relies on a highly regulated sequential recruitment of several main junction proteins including cadherins, claudins, zonula occludins (ZO), and junctional adhesion molecules (JAM)s [30–32]. Protein kinase A (PKA) and exchange protein directly activated by cAMP (Epac) have both been identified as regulators of barrier function by activation of G-protein-coupled receptor and release of cyclic AMP (cAMP) [33–37]. While PKA has traditionally been studied as the primary mediator of barrier function, inhibition of PKA does not result in substantial barrier disruption [38–40]. This has led to a recent transition into investigating the role of Epac within endothelial [40–45] and less often, epithelial barriers [46–48]. After activation, Epac is rapidly translocated to the plasma membrane from the cytoplasm and nuclear membrane to activate Rap1 to regulate afadin (AF-6) stabilization of AJs, TJs, and integrins [49]. Within epithelial cells, Epac signaling has also been implicated in the inhibition of cell migration and epithelial to mesenchymal transition (EMT), [50] and promoting adhesion to laminin and fibronectin [51–53]. This suggests that epithelial junction mediation may be a function of Epac in response to specific ECM proteins; however, the direct causes and mechanisms of this within the alveolar epithelium are not fully understood.

We hypothesized that the discrepancy between native ECM and decellularized airway surface ECM can alter the epithelial cell-cell junction assembly during recellularization through activation of Epac. Understanding how and which ECM components modulate alveolar barriers, specifically TJs and AJs, is integral to producing a whole lung replacement that is functional on the cellular level. This research was the first step to creating a tailorable ECM environment within decellularized lungs by methodically replenishing basement membrane proteins that promote alveolar barrier formation.

2. MATERIALS AND METHODS

2.1. Decellularization and coating preparation

Male and female porcine lungs were obtained from Smithfield-Farmland slaughterhouse or euthanized research pigs to produce dECM coating solution as previously published by our laboratory [54,55]. Decellularization was achieved with tracheal and vascular detergent perfusion of PBS, triton X-100, sodium deoxycholate, sodium chloride, and DNAse over 3 days. After decellularization, large cartilaginous airways were removed, and the remaining alveolar and small airway structures were lyophilized before milling into a fine powder. Quality control of dECM powders from different donors is performed by picogreen dsDNA quantification to ensure each has below 50 ng of dsDNA per mg of dry tissue. The dECM hydrogel solution is produced by pepsin digestion of 8 mg/mL dECM powder for 4 hrs. The solution was diluted to 0.1 mg/mL for use as a coating. Type I collagen from bovine skin (Sigma) and type IV collagen from Engelbreth-Hom-Swarm lathrytic mouse tumor (Santa Cruz Biotechnology) diluted to 0.1 mg/mL in PBS was used as a control coating solution. Collagen I and dECM solutions were also supplemented with 0.01 mg/mL laminin from Engelbreth-Holm-Swarm murine sarcoma basement membrane (Sigma, L2020) and fibronectin (Sigma, F0895) from human plasma. Collagen I and dECM coating concentrations were determined by literature review [5], and laminin, and fibronectin concentrations were determined by literature review and qPCR dosing experiments (data not shown). TCP was coated overnight at 4°C to avoid gelation, then rinsed 3 times with PBS or saline (for TEER experiments) prior to cell culture.

2.2. Cell culture and reagents

All cells were cultured at 37 °C with 5% CO₂. Mouse alveolar epithelial cell line (MLE12, ATCC) were cultured with HITES medium containing 2% fetal bovine serum according to the manufacturer’s protocols. MLE12s offer rapid doubling times and express lung surfactant proteins B and C. MLE12s were also chosen because primary differentiated human alveolar epithelial cells are not commercially available, and slow proliferation of isolated cells would not be feasible. Primary human airway basal epithelial stem cells (BESCs, identified by integrin α6, KRT5, and KRT14) from healthy donors, in EpiX medium (Propagenix), were also used to demonstrate translatability to recellularization with stem cell populations [56]. Passages between 2 and 6 were used for experimentation. Longer culture periods were used for BESC experimentation to allow for adequate cell spreading. BESCs and MLE12s on hydrogel coatings were treated with either ESI-09 (Sigma), a highly specific Epac inhibitor that does not affect PKA expression, or 8-pCPT-2-O-Me-cAMG-AM (007-AM, TOCRIS), an Epac agonist that does not discriminate between Epac1 or Epac2. 5 nM of each treatment or a DMSO vehicle control was added in media. Concentrations were determined by a dose-response assay (data not shown), testing a range defined by a literature review [37,57] for 8 to 24 h. Antibodies for immunofluorescent staining included, ZO-1 (Invitrogen), DAPI (Invitrogen), Alexa Fluor 488 Phalloidin (Invitrogen), and Alexa Flour 647 secondary antibody (Invitrogen).

2.3. Cell adhesion and spreading assay

Cell Counting Kit-8 (CCK8, Dojindo Molecular Technologies) colorimetric assay was used to determine cell viability and proliferation on each coating. 4 and 24 h after inoculating 3.2 ×10³ MLE12 cells per cm², media and unattached cells were aspirated from each well. The CCK8 and media solution was added to each well according to the manufacturer’s protocol. A microplate reader detected the absorbance of each well. Data are shown as normalized absorbance to collagen I.

2.4. Transepithelial electrical resistance (TEER)

96W10idf PET plates (Applied Biophysics) were coated overnight at 4°C with each coating type and rinsed with media. BESCs and MLE12s were seeded at a density of 200,000 cells/cm² to achieve confluence shortly after the start of culture. Cell attachment and cell-cell junction resistances were quantified using electric cell-substrate impedance sensing (ECIS, Applied Biophysics). Wells were coated, and resistances were measured according to manufacturer instructions. Resistance is reported at low frequencies (4000 Hz) and longer time points to highlight cell-cell junction formation and shorter time points to show cell attachment.

2.5. Dextran permeability assay

Alveolar epithelial monolayers formed on 0.4 μm Transwells (Corning) coated with each ECM combination as previously stated prior to seeding of 200,000 cells/cm². After 1 and 3 days, 2 mg/mL of 4 kDa FITC-dextran in PBS (Sigma) was added to the apical chamber of the Transwell for 24 h. 100 uL was taken from the bottom chamber of the well to measure fluorescent intensity using a microplate reader with excitation at 490 nm and emission at 520 nm. The intensity of media without FITC-Dextran was subtracted from each value to determine relative amounts of dextran that has passed through the barrier.

2.6. Gene expression quantification

mRNA was extracted and purified from MLE12s, and BESCs seeded onto each coating for 5 days then treated with a DMSO vehicle control or Epac treatment for 4 h, using a RNeasy Mini kit (Qiagen) and converted to cDNA using iScript cDNA synthesis kit (Biorad). qPCR was performed with primers designed for the desired target genes shown in Table 1.

Table 1.

Primer sequences used in the qPCR analysis.

Target Protein		Mouse Primer Sequence	Human Primer Sequence
E-cadherin	For	AACCCAAGCACGTATCAGGG	--
	Rev		--
		ACTGCTGGTCAGGATCGTTG
Rap1A	For	GTCAACCCAAAATTGGCACCA	ATTGTGTCCCCCACCCTTCA
	Rev		CCAGGGAACTTGTGCAAACC
		GTCAGAGCAGACTTCCCCAC
Afadin (AF-6)	For	CGCCAAGAGACTCGTCTGG	CAGCCAATCCGAACAGACCT
	Rev		CCATGGGAAACACGCAGAAG
		CTGCCCACGCTTCAATAGC
Claudin-18	For	GTGCGATGACGTGATCTG TGA	GGATCATGTCCACCACCACAT
	Rev		TGGTCTGAACAGTCTGCACC
		GTTTATTGGGACTGGGGGCA
F11R/JAMA	For	CTCTGCTCCTCCTGTTCGAC	GACCAAGGAGACACCACCAG
	Rev	ATGCTCCCCTGGAAGACGAT	AAGGTCACC CGTCCTCATA
Epac1	For	GCA ATATTCCCCATGAACG	TCCGGCAAGATGAG AAAGC
	Rev	GGATGGAGCAGACTGCCTTT	TCCAGATCCAAGGGGAGAGG
18s	For	TCCGGCAAGATCGAGAAAGC	GCAATTATTCCCCATGAACG
	Rev	GGATGGAGCAGACTGCCTTT	GGGACTTAATCAACGCAAGC

2.7. Immunofluorescence staining and ZO-1 quantification

MLE12 cells were seeded onto sterilized glass coverslips, or Transwell inserts coated with each of the ECM combinations and were cultured until a confluent monolayer was achieved. Cells were treated with either ESI-09, 007-AM, or the DMSO vehicle control for 4 h. To permeabilize and fix, 0.5% Triton X-100 in 4% paraformaldehyde was applied for 2 min, then replaced with 4% paraformaldehyde for 20 min. After several rinses with PBS, the cells were blocked with 0.1% BSA before labeling with primary antibodies for 30 min at 37°C. 0.1% BSA was applied again before labeling with each with their respective secondary antibodies for 30 min at 37°C. Images will be acquired on a Zeiss LSM 710 confocal microscope using ZEN2011 software.

Quantification of the ZO-1 staining was done using the AngioTool software developed by the NIH Center for Cancer Research. A minimum of 3 images per sample and 3 samples per group were used to create a map of the junction structure and then analyze spatial parameters. The number of cells per image was counted using ImageJ software. Number of junction endpoints or breaks per cell and mean lacunarity were used to characterize junction integrity for each sample. Lacunarity quantifies the degree of spatial order within the junction structure by comparing the variation in both the background and foreground pixel densities [58,59].

2.8. Statistical Analysis

Data are presented as mean +/− standard deviation unless otherwise stated. To determine statistical significance, GraphPad Prism was used to perform a one-way or two-way ANOVA on un-grouped and grouped analyses respectively, followed by a Tukey’s Multiple Comparison Test. When two individual groups were being compared, an unpaired, two-tailed Student’s t-test was used. P < 0.05 was considered statistically significant in sample sizes of three or larger.

3. RESULTS

3.1. Decellularized lung ECM requires supplementation for proper epithelial adhesion and barrier formation

Alveolar epithelial cell barrier formation on each dECM combination compared to dECM alone were examined by TEER and Transwell permeability of MLE12 cells cultured onto matrix-coated plates (Fig. 1A–B). The first 12 h indicated resistance predominately caused by attachment and spreading of the cells on each matrix and showed initially increased resistance by cells cultured on laminin alone, dECM with laminin, and dECM with fibronectin coatings. Cell viability at 4 and 24 h using CCK8 also showed a significant increase of MLE12 attachment to fibronectin-enhanced matrices compared to all other substrates (Supplemental Fig.1). This trend persisted through 24 h when the highest resistance monolayer were produced by cells cultured on the combination of dECM and laminin. dECM with laminin had a significantly higher barrier function compared to collagen I and fibronectin for the entire 3-day culture. Collagen IV exhibited lower resistance compared to laminin, but similar strength compared to the combination of dECM and laminin (Supplemental Fig. 2). Collagen I was chosen as the control for all further experimentation because of the relevance to decellularized lung scaffolds that predominantly contain collagen I [60] and its use as a common cell culture substrate.

Fig. 1.

The effect of dECM composition on alveolar epithelial barrier formation. TEER of MLE12 cells on electrode-embedded 96W10 idfPET plates coated with collagen I (C), dECM (D), laminin (L), fibronectin (F) or a combination of two for 72 h is shown by (A) a representative real-time graph of the average of 3 experimental replicates and (B) binned at 12, 24, 48 and 72 h to show statistical differences. Data are normalized to the collagen I control and presented as mean +/− standard deviation. N ≥ 3 for each coating condition. All groups are significantly different from each other with p < 0.05 unless denoted with n.s. MLE12 monolayer permeability to small molecules with respect to ECM coating after (C) 2 and (D) 4 days of culture was determined by a 4 kDa FITC-Dextran Transwell permeability assay. Normalized dextran flux is shown as absorbance values from the bottom chamber of the Transwell, correlating to dextran transported across the membrane. Data are normalized to the collagen I control and presented as mean +/− standard deviation. N ≥ 3 for each coating condition. * and ** indicates p < 0.05 and p < 0.001 respectively between the groups specified by brackets.

Small molecule permeability with 4 kDa dextran by MLE12 monolayers formed onto matrix-coated Transwell supports was also examined. At 2 days of culture, both dECM + laminin and laminin alone have less permeability compared to collagen I, but not significantly different from dECM (Fig. 1C). It is only after four days that cells cultured on laminin alone were less permeable compared to cells cultured on both collagen I and dECM (Fig. 1D).

3.2. Analysis of ECM-mediated AJ and TJ protein recruitment within the alveolar epithelium

Increases in barrier function with laminin coatings led us to investigate the underlying mechanisms further. MLE12s were seeded on ECM coatings for 5 days to evaluate mRNA expression with qPCR (Fig. 2A–F). A laminin concentration of 0.01 mg/mL was chosen based on decreased junction gene expression with higher coating concentrations (data not shown). We concluded that 0.01 mg/mL laminin is the most effective concentration for further experimentation. Gene expression of Claudin 18, Epac, and AF-6 all increased with cells cultured onto laminin-supplemented matrices. E-cadherin and JAM-A gene expression were highest with the combination of laminin and dECM. Rap1 expression only had significant differences between cells cultured on dECM with laminin and dECM with fibronectin. Expression of all genes was either decreased or not significantly different by cells cultured on coatings containing laminin compared to all other coating groups. Additionally, there was an increasing trend in expression of AF-6, JAM-A, and Rap1 with cells cultured on dECM compared to on collagen I alone, suggesting that collagen I alone was not enough to stimulate this pathway and the remnants of laminin or other proteins trigger Epac/Rap1 activation.

Fig 2.

dECM coatings dictated alveolar epithelial junction regulation. Alveolar epithelial cells were cultured on TCP coated with collagen I (C), dECM (D), laminin (L), fibronectin (F) or a combination of two for 5 days before collecting mRNA. Gene expression of the (A-C) junction proteins, claudin-18, E-cadherin, and JAM-A, and (D-F) junction regulators, Epac, AF-6, and Rap1 were quantified using qPCR and normalized to collagen I coating controls. Immunofluorescent staining of ZO-1 (white) and DAPI (blue) showed stabilization of the junctions depending on coating type by representative confocal images (G, scale bar = 20 μm). Arrows indicate breaks within the cell junctions. Quantification of confocal images from 3 experimental replicates using AngioTool determined the (H) lacunarity index and (I) the number of junction interruptions per cell. Data are presented as mean +/− standard deviation. N ≥ 3 experiments with 2–3 technical replicates for each coating condition. n.s, *, and ** indicate p > 0.05, p < 0.05, and p < 0.001, respectively. Significance is with respect to all other groups, unless brackets are shown to denote significance between 2 specific groups.

ZO-1 localization by immunofluorescence staining (Fig. 2G) showed qualitative increases in the amount of ZO-1 stabilizing the junctions of MLE12s cultured on dECM with laminin and laminin alone compared to collagen I. Compared to cells cultured on collagen I, there was also more uniformity in cell and junctional morphology of MLE12s on both dECM and dECM with laminin coatings. Junctions formed on dECM or collagen I coatings both exhibited gaps within the ZO-1 lining the junction, (shown by arrows, Fig. 2G), and a zipper arrangement of the ZO-1 at the junction. Alternatively, laminin coatings appeared to induce a more continuous, linear ZO-1 formation at the junction. Quantitative analysis by Angiotool shows no significant difference in the lacunarity index (Fig. 2H), corresponding to cell junction shape uniformity, and a decrease in the number of junctions per cell with laminin added to the coating before culture (Fig. 2I).

3.3. Epithelial progenitor cell barrier formation

To determine if similar findings would translate to a progenitor cell population that has been previously used for lung tissue engineering, basal epithelial stem cell (BESC) junction formation through the Epac/Rap1/AF-6 pathway was probed. BESCs were cultured onto collagen I, dECM, and dECM with laminin-coated TCP for a week prior to examination with TEER (Fig. 3A–B). Initially, there were few differences in resistance means between BESCs cultured on dECM and dECM with laminin coatings. After 9 days of culture, the barrier function of cells cultured on dECM with Laminin began to reach higher resistances overall. Barriers formed on collagen I matrices showed very low resistance throughout the entire culture.

Fig 3.

BESC barrier function and Epac/Rap1 pathway gene expression are enhanced with dECM + laminin. TEER of BESCs cultured on TCP coated with collagen I (C), dECM (D), laminin (L) or a combination of two was examined over 9 days after being cultured on collagen I, dECM or dECM + Laminin oatings for 1 week. (A) A representative graph with adjusted means and (B) binned into 24-hour periods with raw resistances is shown. (C) Gene expression of AF-6, Rap1, JAM-A, Epac, and Claudin-18 after 5 days of culture on each respective coating quantified with qPCR. Unless otherwise stated, all data are presented as mean +/− standard deviation. N ≥ 3 for each coating condition. * and ** indicates p < 0.05, and p < 0.001, respectively.

Similar to MLE12s cultured on laminin coatings, dECM caused significant upregulation of JAM-A and Claudin-18, but not Epac. Contrasting MLE12 gene expression, BESCs did not show significant increases in AF-6 expression when cells are cultured on dECM and dECM with laminin compared to collagen I (Fig. 3C).

3.4. Role of Epac in laminin-mediated barrier formation

To fully understand the effects of the Epac/Rap1 upregulation on laminin-rich matrices, Epac inhibition with ESI-09 or Epac activation with 007-AM was used to evaluate laminin-mediated barrier reinforcement. TEER analysis of MLE12 cells treated with Epac inhibitor or agonist for 24 h at 48 h of culture, showed significantly decreased barrier resistance with Epac inhibition after 24 h compared to addition of a vehicle control. Alternatively, significant barrier strengthening was seen by cells cultured on dECM with laminin coatings with Epac activation over 24 h compared to cells treatment with the vehicle control (DMSO). (Fig. 4A). The effects of the Epac treatments began to diminish upon media change at 72 hours.

Fig. 4.

Epac drives laminin-medicated barrier function. (A) MLE12 alveolar epithelial cells were cultured for 48 h on TCP coated with collagen I (C), dECM (D), laminin (L), or a combination of two before treatment with either an Epac inhibitor (ESI-09) or Epac agonist (007-AM) for 24 hours to examine TEER. Data are presented as mean +/− standard deviation. N ≥ 3 for each coating condition. n.s., *, **, and **** indicates significance of p> 0.05, p < 0.05, p < 0.01, and p < 0.0001, respectively, between all other groups within each binned time point. (B) Cells cultured on dECM with laminin and laminin only coatings were treated with Epac inhibitor to determine if laminin-medicated barrier selectivity to 4 kDa molecules could be abolished. (C) 4 kDa FITC-dextran permeability of cells cultured on collagen I and dECM with Epac agonist treatment was evaluated to determine whether barrier permeability of 4 kDa molecules could be further reduced. All data are presented as mean +/− standard deviation. N ≥ 3 for each coating condition. * and ** indicates significance of p < 0.05 and p < 0.001, respectively, between all other groups within each coating type.

While TEER of cells cultured on dECM with Laminin was reduced with ESI-09 Epac inhibition, 4 kDa dextran permeability did not change significantly compared to the DMSO vehicle control (Fig. 4B). However, barrier disruption was observed when the Epac inhibitor was added to MLE12 culture on laminin alone after 4 days, shown by an increase in dextran flux (Fig. 4B). Increased barrier resistance was also observed when the Epac inhibitor was added to cells cultured on collagen I. The reason for this is unknown, but we speculate that it could be caused by a feedback loop when laminin is not present. To determine if laminin-mediated barrier function could be induced with Epac activation in cells cultured without laminin-supplemented coatings, similar dextran permeability assays were also conducted on collagen I and dECM with the Epac agonist, 007-AM (Fig. 4C). Treatment of cells on collagen I and laminin coatings with the Epac agonist after 2 days of culture were the only groups to show significant decreases in junction permeability.

Delving further into the mechanisms of Epac, the junctional gene and protein expression with Epac inhibition was also examined. All junctional gene expression decreased significantly compared to the vehicle control, DMSO, except for Rap1, which showed a decreasing trend when treated with the Epac inhibitor, ESI-09 (Fig. 5A). ZO-1 localization to the junction was also disrupted with Epac inhibition (Fig. 5B). Additionally, F-actin arrangement and cell morphology were drastically altered with increased stress fiber formation and junctional degradation. Quantification by image analysis (Fig. 5C–D), showed no significant change in mean lacunarity or endpoints per cell. As seen in the images of the mask created by the ZO-1 analysis (Fig. 5B), there were so few junctions formed and subsequently masked by the program with Epac inhibition leading to overall similarities in lacunarity within the collagen I coating group and an overall decrease in junction end points that distort the quantification within the collagen I group.

Fig. 5.

Inhibiting Epac barrier strengthening by cells on dECM with laminin matrices. (A) Gene expression of claudin-18, E-cadherin, JAM-A, Epac, AF-6, and Rap1 by MLE12 alveolar epithelial cells cultured onto TCP coated with dECM with laminin for 5 days before treatment with the Epac inhibitor (ESI-09) was quantified and normalized to the vehicle control (DMSO). (B) Upon confluency, MLE12s were treated for 4 h with the Epac inhibitor and stained for ZO-1 (white) and F-actin (green, scale bar = 20 μm). Representative confocal images show disorganized junctions with Epac inhibition. Quantification of confocal images from 3 experimental replicates using AngioTool determined (C) the lacunarity index and (D) the number of junction interruptions per cell with ESI-09 treatment. Data is presented as mean +/− standard deviation from 3 experiments and *, **, **** indicates p < 0.05, p< 0.01, and p < 0.0001, respectively, compared to the DMSO counterpart within the same coating group.

Conversely, Epac activation with 007-AM increases gene expression of both Rap1 and JAM-A (Fig. 6A). Immunofluorescence staining cells cultured on collagen I matrices with and without Epac activation shows ZO-1 localization to the junctions and f-actin cortical organization (Fig. 6B), more resembling junctions formed on laminin-rich matrices. Image quantification characterizes these changes by finding a increase in cell shape uniformity (lacunarity, Fig. 6C) and a decrease in the number of junction-breaks per cell (Fig. 6D). These data show that JAM-A increases expression while other AJ and TJ maturation may be caused by stabilization of ZO-1 or translocation to the junction.

Fig. 6.

Epac agonist rescue of barriers formed on collagen matrices. (A) Gene expression of claudin-18, E-cadherin, JAM-A, Epac, AF-6, and Rap1 by MLE12 alveolar epithelial cells cultured onto TCP coated with collagen I for 5 days before treatment with 007-AM was quantified and normalized to the vehicle control. (B, scale bar = 20 μm) Upon confluency, MLE12s were treated for 4 h with 007-AM and stained for ZO-1 (white) and F-actin (green). Representative confocal images show more organized junctions with 007-AM treatment. Quantification of confocal images from 3 experimental replicates using AngioTool determined (C) the lacunarity index or the uniformity of junction formation and (D) the number of junction interruptions per cell with 007-AM treatment. Data is presented as mean +/− standard deviation from 3 experiments with a minimum of 2 technical replicates each and *, **, *** indicates p < 0.05, p< 0.01, and p < 0.001, respectively, compared to the DMSO counterpart within the same coating group.

4. DISCUSSION

This research aims to expand the basic knowledge of epithelial barrier physiology and to determine how decellularization and further ECM processing can alter engineered alveolar barriers. Using a common decellularization protocol adapted from Andrew et al. [54,61] including a 3 day perfusion regimen of Triton X-100, sodium deoxycholate (SDC), and DNAse followed by protease digestion, we have attempted to improve alveolar epithelial barriers formed on dECM by systematically replenishing laminin or fibronectin. Laminin and fibronectin were chosen for their abundance within epithelial basement membranes and for the drastic effects both can have on epithelial differentiation and propagation of disease pathologies if concentrations are dysregulated [24,62–66]. The decellularization detergent combination was previously developed to preserve proteoglycans, fibronectin, and laminins of the BM, compared to previous protocols containing harsher detergents such as 3-cholamidopropyl dimethylammonio 1-propanesulfonate (CHAPS); however all components are diminished with either protocol [60,67,68]. These findings may be altered with different detergents and should be reinvestigated to tailor ECM pretreatment strategies to the decellularization protocol. These findings have identified a pathway by which cells sense ECM composition and direct barrier formation through the activation of Epac. The connection between integrin ECM binding and Epac activation by GPCRs suggests crosstalk between the two, which has also been seen in several other systems [69,70].

Through TEER and dextran permeability studies, we have identified laminin as a key matrix protein for alveolar junction formation over collagen I, collagen IV, and fibronectin. Fibronectin caused a prolonged decrease in barrier function but an increase in initial attachment rate, seen both in TEER before 12 h and CCK8 viability assays at 4 and 24 h. These results are consistent with previous studies showing increased cell attachment onto fibronectin and laminin-rich matrices, but only laminin enhances barrier maturation within decellularized lungs and in vitro [10,71]. Analysis of epithelial barrier formation on dECM shows some similarities in the characteristics of barriers formed on both fibronectin and laminin, yet at a fraction of the resistance. At earlier time points, dECM had a similar attachment rate to fibronectin, but a long-term slope resembled cells on laminin coatings. This confirms that previously identified concentrations of both laminin and fibronectin within dECM [72] are promoting initial attachment and steady increases in barrier resistance by alveolar epithelial cells. Nevertheless, to achieve prolonged barrier maturation, laminin would need to be added. Collagen IV was also compared because of its abundance within the BM [73]. Barriers formed on collagen IV matrices were similar to dECM with laminin suggesting collagen IV may also be contributing to barrier strength; however, laminin had significantly higher resistance and was the main focus of the present study.

Similarly, small molecule permeability with laminin coatings also decreased; however, dECM and dECM with laminin did not show significant changes from that of collagen I, mainly after 4 days of culture. Dextran permeability showed unexpected differences from the TEER results, but we postulate that after 4 days, collagen I and dECM may have formed tight enough barriers to regulate the movement of larger molecules, although the changes in TEER may give more insight into prolonged ion movement associated with edema [74]. Further study of edema would need to be performed in co-culture study of alveolar epithelium along with endothelium in order to fully form the air liquid interface.

To further understand how laminin was increasing the barrier function of the alveolar epithelium, we sought to identify which junction structures are being regulated. Increases in JAM-A and Rap1 gene expression when dECM was present and not laminin would suggest that the expression of JAM-A may be regulated by Rap1 but not solely by laminin. Other components, such as other collagens or vitronectin, within dECM may be causing this upregulation because JAM-A has been previously linked to integrin β1 activity [75] and integrin αVβ5 [76] previously. Conversely, Claudin-18, Epac, E-cadherin, and AF-6 increased expression in response to laminin matrices and not dECM alone or fibronectin. Upregulation of Epac and AF-6 with laminin coatings suggests activation of the Epac/AF-6 pathway that led to increases in E-cadherin and Claudin-18 expression. This was expected from previous literature linking Epac activation to TJ and AJ expression and stabilization [75,77–80].

Upregulation of both Claudin 18 and E-cadherin with laminin or dECM with laminin matrices, suggests both TJ, AJ, and focal adhesion complexes are potentially altered to achieve the overall barrier strengthening with laminin-coated substrate. We believe this was mostly due to the stabilization of multiple junction structures by the scaffolding protein ZO-1, seen by translocation of ZO-1 to the junction in a continuous band with laminin enriched coatings. Analysis of ZO-1 immunofluorescent staining at the cell junction showed that there was a significant decrease in the number of breaks in the ZO-1 junction staining per cell. There was, however, not a significant decrease in the lacunarity between cells cultured on any of the matrices, however, the junction appears to have a more linear junction morphology when cells were cultured on laminin. These results would suggest that significant disruptions to the ZO-1 continuity around the junction and not junction shape uniformity may be the cause of decreases in barrier strength when laminin is not present. ZO-1 recruitment to the cell junction in an organized morphology is a strong correlation to alveolar permeability because it is required for assembly of AJs, TJs, and the actin cytoskeleton [7,28,81–83].

MLE12 cells offer transformed surfactant expressing alveolar epithelial cells and high expansion rates for investigating these fundamental questions concerning alveolar epithelial junctions; however, lung bioengineering requires progenitor cell populations that can repopulate with more than 40 types of cells within the lungs [84,85]. Several groups have identified basal epithelial stem cells (BESCs) as a promising multipotent stem cell candidate for lung regeneration [4,86] that would potentially differentiate into ciliated, club and pneumocyte populations under the right culture conditions [4,86–88]. To demonstrate the potential translation of laminin treatments into stem cell repopulation of bioengineered lungs, we conducted similar coating experiments with BESCs with TEER and qPCR analysis. Like the MLE12 cells, BESCs formed higher resistance barriers on dECM and dECM with laminin compared to a complete lack of barrier formation on collagen I-coated plates. Gene expression of JAM-A, EPAC, and Claudin-18 were increased with both dECM and laminin compared to collagen I alone. These data suggest that laminin is a stimulator of Epac and junction formation within both stem cell and fully differentiated lung epithelial cell populations. The effect of replenishing these components on the lineage differentiation of these cells has not been determined by this study. An extensive understanding of the differentiation of BESCs by laminin and other BM components is a complex endeavor and warrant further investigation with longer culture periods, air-liquid interface culture, growth factor supplementation, or co-culture with endothelium.

Increases in MLE12 Epac expression and barrier function in response to laminin matrices led us to investigate further Epac’s role in laminin-mediated barrier function by interfering with the interaction of Epac with downstream effectors. Epac inhibition abolished laminin-mediated barrier resistance, as observed by TEER and ZO-1 immunofluorescent staining. Quantification of ZO-1 localization found that Epac inhibition did not significantly change ZO-1 recruitment, however there is an apparent difference in junction morphology between dECM with laminin treated with Epac inhibitor or the vehicle control. The quantified lacunarity identifies the inhomogeneity throughout the image [89], but we have concluded that since ZO-1 organization with Epac inhibition caused ZO-1 to arrange randomly in short lines along the junction, the software is finding uniformity in the line arrangement, when they are essentially not comparable to the vehicle control ZO-1 arrangement. Advancements in the image analysis tools would help to accurately quantify this phenomenon.

Also of translational interest is how we can apply Epac activation to enhance barrier formation without the addition of laminin. An Epac agonist, 007-AM, was applied to collagen I coatings, significantly increasing JAM-A and Rap1 gene expression, rearranging ZO-1 and the actin cytoskeleton cortically, and decreasing barrier permeability. In addition, through immunofluorescent imaging, we saw cells were more densely packed and smaller on collagen and dECM with laminin treated with the Epac inhibitor, ESI-09, compared to the size of cells cultured on other substrates. This may be due to a lack of cell spreading that has been found to also be driven by laminin through the Epac/Rap1 pathway [51]. The Epac agonist, 007-AM, has recently been established as a long-lasting treatment for endothelial barriers within bioengineered lungs [7] and treatment of vascular permeability within acute lung injury [37,90]. These data now confirm that Epac activation and laminin-enriched ECM coatings can both benefit the epithelial barriers to produce functional lung biomaterials and potentially serve as therapies for edema-related diseases.

This research provided some understanding to high-level tissue-engineering challenges concerning alterations to ECM composition with processing on alveolar epithelial junction complexes. These overarching findings also unearth detailed questions about specifics of epithelial AJ and TJ physiology and lung recellularization that require further complex, in depth investigations to fully understand. Future studies to confirm that these findings can be translated to ex vivo lung tissue engineering and in vivo treatment of diseases must also consider multicellular cross-talk and stem cell differentiation. This would include investigations into co-culture of organotypic mixtures of cells and effects on endogenous laminin expression [91] and influences on Epac signaling. This high throughput and definitive in vitro approach has allowed us to answer basic questions about junction physiology that would not be possible within a whole lung model. Also, while these experiments have definitively established the role of both laminin and fibronectin in alveolar epithelial barrier formation, the role of other critical epithelial differentiators such as collagen IV, nidogens, GAGs, other glycoproteins will need to be investigated.

Statement of Significance.

Efforts to produce a transplantable organ-scale biomaterial for lung regeneration has not been entirely successful to date, due to incomplete cell-cell junction formation, ultimately leading to severe edema in vivo. To fully understand the process of alveolar junction formation on ECM-derived biomaterials, this research has characterized and tailored decellularized ECM (dECM) to mitigate reductions in barrier strength or cell attachment caused by abnormal ECM compositions or detergent damage to dECM.

These results indicate that laminin-driven Epac signaling plays a vital role in the stabilization of the alveolar barrier. Addition of laminin or Epac agonists during alveolar regeneration can reduce epithelial permeability within bioengineered lungs.

Abbreviations:

ECM

extracellular matrix

dECM

decellularized extracellular matrix

TEER

trans-epithelial electrical resistance

JAM

junction adhesion molecule

ZO

zonula occludens

TJ

tight junction

AJ

adherens junction

PKA

Protein kinase A

Epac

exchange protein directly activated by cAMP

cAMP

cyclic AMP

AF-6

Afadin

EMT

Epithelial to mesenchymal transition

GAGs

glycosaminoglycans

Rap

repressor activator protein

SDC

Sodium Deoxycholate

BESC

basal epithelial stem cell

MLE12s

mouse alveolar epithelial cells