Laminin-511 and -521 Enable Efficient In Vitro Expansion of Human Corneal Endothelial Cells

Abstract

Purpose.: The purpose of this study was to investigate the usefulness of laminin isoforms as substrates for culturing human corneal endothelial cells (HCECs) for clinical application of tissue engineering therapy.

Methods.: Expression of specific laminin chains in human corneal endothelium and Descemet's membrane was analyzed at the mRNA and protein levels. The effect of laminin-511 and -521 on cell adhesion and proliferation was evaluated. Recombinant laminin E8 fragments (E8s), which represent functionally minimal forms of laminins, were also evaluated for their effects on cell density and cellular phenotype. The potential involvement of α3β1 and α6β1 integrins in laminin signal transduction was also investigated using neutralizing antibodies.

Results.: Laminin-511 and -521 were expressed in Descemet's membrane and corneal endothelium. These laminin isoforms significantly enhanced the in vitro adhesion and proliferation, and differentiation of HCECs. A cell density of 2200 to 2400 cells/mm² was achieved when HCECs were cultured on laminin-511 or -521, whereas the density was only 1100 cells/mm² on an uncoated control. E8s also supported HCEC cultivation with a similar efficacy to that obtained with full-length laminin. Functional blocking of α3β1 and α6β1 integrins suppressed the adhesion of HCECs even in the presence of laminin-511.

Conclusions.: Laminin-511 and -521 were the laminin isoforms present in Descemet's membrane, and these laminins modulate the adhesion and proliferation of CECs. Laminin E8s represent an ideal xeno-free defined substrate for the culture of CECs for clinical applications.

The corneal endothelium maintains corneal transparency by pump and barrier functions.¹ However, corneal endothelial cells (CECs) are severely limited in their proliferative ability,² so that any damage to CECs causes a concurrent compensatory migration and enlargement of the remaining cells to cover the damaged area. When the density of CECs decreases due to pathological conditions or aging, the compromised endothelial pump function is compensated by an increase in the metabolic activity via ATP and by an increase in the total number of pump sites.³ However, when the cell density decreases to a critical level, the metabolic pump function and the endothelial barrier can no longer counteract the stromal swelling pressure. Decompensation of the corneal endothelium resulting from various causes (e.g., Fuchs' corneal endothelial dystrophy, surgical trauma, and endotheliitis) ultimately leads to corneal edema, visual disturbance, and blindness. 

Transplantation of donor corneas remains the only therapy for corneal endothelial dysfunction. Endothelial keratoplasty provides a less invasive and more efficient therapy for corneal endothelial disorders, but the worldwide shortage of donor corneas remains a persistent problem.⁴ Several groups, including ours, have endeavored to overcome this shortage of donor tissue by exploring tissue engineering-based approaches using CEC cultivation.^5–13 However, in vitro expansion of CECs poses difficulties due to the limited proliferative ability, undesired transformation to fibroblasts during culture, and senescence of cultured CECs.^14–18 Consequently, a cell culture protocol for clinical application has been investigated in the present study. 

Laminin is a glycoprotein composed of three trimeric (α, β, and γ) chains.^19,20 Fifteen laminin isoforms have been identified and are named according to the specific trimeric combination of their chains (e.g., α1–5, β1–4, and γ1–3).^20,21 Laminins are the major basement membrane components responsible for modulating cellular functions such as migration, survival, proliferation, and differentiation.^19,22 The expression of laminin-411 and -511 in the corneal endothelial basement membrane, the Descemet's membrane, have been reported^23,24; however, the laminin composition and its biological role are not fully understood.²⁵ 

In the present study, we hypothesized that specific laminin isoforms expressed in Descemet's membrane may transduce biological signals to CECs, and that these signals may trigger the in vitro expansion of CECs for clinical use. Therefore, we investigated the expression of specific laminin isoforms in the Descemet's membrane of the human cornea and the potential of these laminins for in vitro expansion of CECs. We also screened the expression of integrins in CECs, as these transmembrane receptors mediate binding and signal transduction between cells and laminins.^19,22,26 In addition, we examined laminin E8 fragments (E8s) as potential culture substrates for clinical applications. E8s are truncated proteins consisting of the C-terminal regions of the α, β, and γ chains and contain the laminin globular domains of the α chain.²⁷ Our findings demonstrated that laminin E8s may represent an important component of a xeno-free defined substrate for human CEC (HCEC) culture and may promote the in vitro expansion of CECs for cell-based therapies. 

Materials and Methods

Ethics Statement

The human tissue specimens used in this study were handled in accordance with the tenets set forth in the Declaration of Helsinki. Informed written consent was obtained from the next of kin of all deceased donors with regard to eye donation for research. Human donor corneas were obtained from SightLife (http://www.sightlife.org/, Seattle, WA, USA). All tissue specimens were recovered under the tenets of the Uniform Anatomical Gift Act (UAGA) of the particular state in which the donor consent was obtained and the tissue recovered. 

Cell Culture

All human corneas were stored at 4°C in storage medium (Optisol-GS; Chiron Vision, Irvine, CA, USA) for less than 14 days prior to use. Primary cultures of CECs were established according to published protocols.¹⁶ Briefly, Descemet's membranes containing the CECs were stripped from donor corneas and the membranes were digested with 1 mg/mL collagenase A at 37°C for 12 hours. The resulting CECs were seeded in one well of a 48-well plate. A total of six human donor corneas (from persons >40 years of age) were used for the experiment. Plates had previously been coated with laminin-511, -521, or -211. The culture medium was prepared according to published protocols.¹⁶ First, human bone marrow-derived mesenchymal stem cells (BM-MSCs), provided by JCR Pharmaceuticals Co., Ltd. (Kobe, Japan), were cultured according to previously reported protocols. Briefly, BM-MSCs were plated at a density of 1.3 × 10⁴ cells/cm² and cultured for 24 hours in Dulbecco's Modified Eagle Medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 100 U/mL penicillin, and 100 μg/mL streptomycin. Then, basal medium for human CECs (HCECs) was prepared (OptiMEM-I, 8% FBS, 5 ng/mL epidermal growth factor, 20 μg/mL ascorbic acid, 200 mg/L calcium chloride, 0.08% chondroitin sulfate, 50 μg/mL gentamicin, and 10 μM SB431542) and conditioned by culturing BM-MSCs for 24 hours. Finally, the basal medium conditioned with BM-MSCs was collected for use as the culture medium for HCECs. The HCECs were cultured at 37°C in a humidified atmosphere containing 5% CO₂, and the culture medium was changed every 2 days. Cell density was evaluated by phase contrast images analyzed using the KSS-400EB software (Konan Medical, Inc., Hyogo, Japan). 

Reverse Transcription PCR

Descemet's membranes containing CECs were stripped from three independent human donor corneas. Total RNA was extracted using the RNeasy Mini kit (Qiagen, Hilden, Germany). The quality of the RNA preparations was measured with a NanoDrop (Thermo Fisher Scientific, Inc., Waltham, MA, USA) spectrophotometer. First-strand cDNA was synthesized with 1 μg of total RNA using a ReverTra Ace (Toyobo Corporation, Osaka, Japan) reverse transcriptase kit. The cDNA was subjected to PCR with the specific primers listed in Tables 1 and 2. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used as an internal control for gene analysis. PCR reactions were performed with Extaq DNA polymerase (Takara Bio, Inc., Otsu, Japan) under the following conditions: denaturation at 94°C for 30 seconds, 35 cycles of annealing at 54°C for 30 seconds, and elongation at 72°C for 30 seconds. The PCR products were separated by electrophoresis on 1.5% agarose gels, stained with ethidium bromide, and detected under ultraviolet illumination. 

Table 1

 

Oligonucleotide Sequences of Laminins for RT-PCR

 

Table 2

 

Oligonucleotide Sequences of Integrins for RT-PCR

 

Immunofluorescence

Corneas obtained from three independent donors were embedded in OCT compound, sectioned at a thickness of 6 μm, and fixed in 4% paraformaldehyde for 5 minutes at room temperature. HCECs cultured in a 48-well cell culture plate were fixed in 4% paraformaldehyde for 10 minutes at room temperature. Nonspecific binding was blocked using 1% bovine serum albumin (BSA) for 30 minutes at room temperature. Primary antibodies against laminin α2 (1:200), laminin α5 (1:100), laminin β1 (1:200), laminin β2 (1:200), laminin β3 (1:500, Thermo Fisher Scientific, Inc.), laminin γ1 (1:200), Ki67 (1:400, Sigma-Aldrich Corp., St. Louis, MO, USA), Na⁺/K⁺-ATPase (1:200, Upstate Biotechnology, Lake Placid, NY, USA), and ZO-1 (1:200, Zymed Laboratories, South San Francisco, CA, USA) were used. Antibodies against laminin α2, α5, β1, β2, and γ1 were a gift from Lydia Sorokin.²⁸ Either Alexa Fluor 488-conjugated goat anti-mouse (Life Technologies Corp., Carlsbad, CA, USA) or Alexa Fluor 594-conjugated goat anti-rabbit IgG (Life Technologies) were used as secondary antibodies at a 1:1000 dilution. Nuclei were stained with DAPI (Vector Laboratories, Burlingame, CA, USA). The slides were examined with a fluorescence microscope (TCS SP2 AOBS; Leica Microsystems, Wetzlar, Germany, or BZ-9000; Keyence, Osaka, Japan). 

Cell Adhesion Assay

HCECs were seeded in 96-well plates and the numbers of adhered cells were measured 24 hours after seeding, using a CellTiter-Glo luminescent cell viability assay (Promega Corporation, Madison, WI, USA) according to the manufacturer's instructions. The number of adhered cells was determined using a Veritas microplate luminometer (Promega Corporation). Culture plates were coated with laminin-511, -521, and -211 (BioLamina, Sundbyberg, Sweden), type I collagen (Nitta Galatin, Inc., Osaka, Japan), and fibronectin (Wako Pure Chemical Industries, Ltd., Osaka, Japan; 20 μg/μL). The culture plate was also coated with FNC Coating mix (Athena Environmental Sciences, Inc., Baltimore, MD, USA), a commonly used coating reagent for cultivation of HCECs. The effect of laminin fragments on cell adhesion was examined using plates coated with laminin-521 or laminin E8 fragments (iMatrix-511; Nippi, Inc., Tokyo, Japan; 1.0, 2.0, 4.0 μg/cm²). The effect of the interaction between cellular integrins and laminins on cell adhesion was evaluated by seeding HCECs (5 × 10³ cells/well) in 96-well plates coated with laminin-511 or 211 in the presence or absence of integrin-neutralizing antibodies (2 μg/mL): anti-α₃ integrin (Merck Millipore, Billerica, MA, USA), anti-α₆ integrin (Merck Millipore), and anti-β₁ integrin (R&D Systems, Inc., Minneapolis, MN, USA). Three hours after seeding, the numbers of adherent cells were determined with the CellTiter-Glo luminescent cell viability assay, as described above. 

Cell Proliferation Assay

Cells were seeded at a density of 5000 cells/well in a 96-well plate and cultured for 24 hours, followed by incubation in serum-free media for an additional 24 hours (n = 6). DNA synthesis was detected as incorporation of 5-bromo-2′-deoxyuridine (BrdU) into the Cell Proliferation Biotrak ELISA system, version 2, according to the manufacturer's instructions (Sigma-Aldrich Corp.). Briefly, cells were incubated with 10 mol/L BrdU for 24 hours at 37°C in a humidified atmosphere containing 5% CO₂. The cultured cells were incubated with 10 μM BrdU labeling solution for 2 hours, and then incubated with 100 μL of monoclonal antibody against BrdU for 30 minutes. The BrdU absorbance was measured directly using a spectrophotometric microplate reader at a test wavelength of 450 nm. 

Flow Cytometry

Expression of integrins was assessed by the Human Cell Surface Marker Screening Panel (BD Biosciences, Franklin Lakes, NJ, USA) according to manufacturer's protocol. HCECs at passages 2 through 5 were used for these experiments. Briefly, cultured HCECs were detached by treatment with Accutase (BD Biosciences) at 37°C, washed twice with PBS, passed through a BD Falcon 70-μm cell strainer (BD Biosciences), incubated in OptiMEM with the addition of 100 units/mL DNase for 15 minutes at room temperature, and resuspended in BD Pharmingen Stain Buffer (BD Biosciences) containing 5 mM EDTA. Cells were incubated with primary antibodies (integrin α1, integrin α2, integrin α3, integrin α6, integrin β1, mouse isotope IgG, and rat isotype IgG; BD Biosciences) at the dilution indicated by manufacturer's protocol for 30 minutes on ice. The cells were washed with BD Pharmingen Stain Buffer containing 5 mM EDTA and then incubated with AlexaFluor 647 conjugated goat anti-mouse IgG or goat anti-rat IgG (1:200 dilution, BD Biosciences) for 30 minutes. Cells were then washed with BD Pharmingen Stain Buffer containing 5 mM EDTA, fixed in BD Cytofix Fixation Buffer (BD Biosciences) for 10 minutes, and analyzed by flow cytometry using the BD FACSCant (BD Biosciences) and CellQuest Pro software (BD Biosciences). 

Immunoblotting

The CECs were seeded, with or without neutralizing integrin α3β1 and α6β1 antibodies, on a culture plate coated with laminin -511 or -211. A noncoated plate was used as a control. Three hours after seeding, the CECs were washed with ice-cold PBS, and then lysed with ice-cold radio immunoprecipitation assay (RIPA) buffer (Bio-Rad Laboratories, Hercules, CA, USA) buffer containing phosphatase inhibitor cocktail 2 and protease inhibitor cocktail. The lysates were centrifuged, and the supernatant representing total proteins was collected. An equal amount of protein was fractionated by SDS-PAGE, and polyvinylidene fluoride (PVDF) membranes were blocked with 3% nonfat dry milk, followed by an overnight incubation at 4°C with the following primary antibodies: phosphorylated focal adhesion kinase (FAK; 1:1000; Cell Signaling Technology, Inc., Danvers, MA, USA), FAK (1:1000; Cell Signaling Technology, Inc.), and GAPDH (1:3000; Abcam, Cambridge, UK). The blots were washed and incubated with horseradish peroxidase-conjugated secondary antibodies (1:5000; Cell Signaling Technology, Inc.). The blots were developed with luminal for enhanced chemiluminescence using the ECL Advanced Western Blotting Detection Kit (GE Healthcare, Piscataway, NJ, USA), and documented using an LAS4000S (Fuji Film, Tokyo, Japan) cooled charge-coupled-device camera gel documentation system. The relative density of the immunoblot bands was determined by Image J (National Institutes of Health, Bethesda, MD, USA) software. Relative fold differences were compared with the control values. 

Statistical Analysis

The statistical significance (P-value) of differences between mean values of the two-sample comparison was determined with the Student's t-test. The statistical significance in the comparison of multiple sample sets was analyzed with Dunnett's multiple-comparisons test. Results are expressed as mean ± SEM. 

Results

Expression of Laminin Isoforms in Corneal Endothelium and Descemet's Membrane

Expression patterns of laminin isotypes vary in different tissues; therefore, we analyzed the expression of specific laminin chains in human corneal endothelium and the Descemet's membrane at the mRNA and protein levels. Gene expression of laminin chains (α1, α2, α3, α4, α5, β1, β2, β3, β4, γ1, γ2, and γ3) was analyzed by RT-PCR in the corneal endothelium obtained from three independent donor corneas (Table 1). Laminin α5, β1, β2, β3, γ1, and γ2 chains were expressed in the corneal endothelium, while α1, α2, α3, α4, β4, and γ3 were not expressed (Fig. 1A). 

Figure 1

 

Laminin-511 and -521 are expressed in corneal endothelium and Descemet's membrane. (A) Expression of genes of laminin chains α1, α2, α3, α4, α5, β1, β2, β3, β4, γ1, γ2, and γ3 in the corneal endothelium were assessed by RT-PCR. (B) Expression of laminin α2, α5, β1, β2, β3, and γ1 chains in donor Descemet's membrane were stained by immunohistochemistry. Scale bar: 50 μm. The experiments were performed in triplicate.

We then used immunohistochemistry to evaluate the presence of laminin chains expressed by endothelial cells in the Descemet's membrane. Immunostaining for laminin α5, β1, β2, and γ1 chains was evident as a linear staining pattern along the endothelial face of the Descemet's membrane (Fig. 1B), suggesting that laminin-511 and -521 represent the major laminin forms in the adult human Descemet's membrane. 

Effect of Laminin-511 and -521 on Cell Adhesion and Proliferation of HCECs

The detection of laminin-511 and -521 as the predominant laminin forms in the Descemet's membrane led to further evaluation of the suitability of these laminins as a substrate for in vitro expansion of HCECs. Primary HCECs seeded on laminin-511 and -521 formed a monolayer of cells with a hexagonal phenotype after 48 hours of culture. In contrast, control HCECs seeded on laminin-211 coated or uncoated culture plates formed patchy colonies rather than a confluent monolayer (Fig. 2A). The effect of laminin-511 and -521 on cell adhesion was evaluated by determining the numbers of adherent HCECs 24 hours after seeding. Incubation on laminin-511 and -521 coated plates increased the number of adherent cells 1.5-fold compared to uncoated controls, while laminin-211 had no effect on the number of adherent cells (Fig. 2B). The effect of laminins on HCEC proliferation was further assessed by measuring the incorporation of BrdU into the newly synthesized DNA. HCECs cultured on laminin-511 and -521 showed 2.6- to 3.2-fold increases in BrdU incorporation compared to the uncoated control. Of interest, HCECs cultured on laminin-211 demonstrated almost same proliferative potential when compared to HCECs cultured on uncoated control plates (Fig. 2C). In addition, laminin-511 and -521 increased percentage of Ki67-positive HCECs compared to the uncoated control and laminin-211 (Figs. 2D, 2E). 

Figure 2

 

Laminin-511 and -521 enhance cell adhesion and proliferation of HCECs. (A) Phase-contrast images of HCECs cultured on plates coated with laminin-511 and -521 (Descemet's membrane type laminins) and laminin-211 (non-Descemet's membrane type laminin). Scale bar: 100 μm. (B) The effect of laminins on cell adhesion was tested by seeing CECs at a density of 5.0 × 10³ cells/well, and the number of adhering CECs was determined with a CellTiter-Glo luminescent cell viability assay after 24 hours. The experiments were performed in triplicate. *P < 0.01. (C) The stimulating potency of laminins on cell proliferation was evaluated by seeding CECs at a density of 5.0 × 10³ cells/well. The proliferation of the CECs was evaluated by a BrdU incorporation assay after 24 hours of incubation with 10 μM BrdU. The experiment was performed in duplicate. *P < 0.01. (D, E) CECs were seeded at a density of 1.25 × 10⁴ cells/cm² for 72 hours, and then expression of Ki67 was assessed by immunofluorescence staining. Ki67 positive cells were plotted as a graph. The experiment was performed in duplicate. *P < 0.01.

Cell Density and Functional Phenotype of HCECs Cultured on Laminin-511 and -521

Cultured primary HCECs formed a contact inhibited monolayer of cells after 30 days of cultivation under all experimental conditions (Fig. 3A). A cell density of 2200 to 2400 cells/mm² was achieved when HCECs were cultured on laminin-511 or -521, whereas the density was only 1100 cells/mm² in uncoated control plates. Of interest, an increase in HCEC cell density similar to that seen with laminin-511 and -521 was observed following culture on laminin-211 (Fig. 3B). 

Figure 3

 

Laminin-511 and -521 maintain cell density and functional phenotype during cell culture. (A, B) CECs obtained from donor corneas were divided into four groups and seeded on a noncoated plate, or on plates coated with laminin-511, -521, or -211. Images show representative phase contrast images of CECs after 110 days cultivation. Graphs indicate the average cell densities of CECs. Scale bar: 100 μm. *P < 0.01. (C) The functional phenotype was assessed by staining for ZO-1 and Na⁺/K⁺-ATPase at the plasma membrane in the CEC culture. Scale bar: 50 μm. The experiments were performed in duplicate.

Culture on laminin-511 or -521 stimulated persistent expression and subcellular localization of Na⁺/K⁺-ATPase and ZO-1 at the plasma membrane. However, HCECs cultured on uncoated control plates showed partial loss of Na⁺/K⁺-ATPase and ZO-1 expression during culture (Fig. 3C). These data suggest that laminins present in the Descemet's membrane maintain the cell density and a functional cellular phenotype in vitro. 

Effect of Laminin E8 Fragments on Cell Culture of HCECs

We therefore examined the effect of laminin E8 fragments (E8s) on cell density and cellular phenotype of cultured HCECs. Laminin E8s are truncated proteins composed only of the α, β, and γ chain C-terminal regions, which include the active integrin binding site.²⁹ Culture on plates coated with either laminin-511 or laminin-511-E8s resulted in a 1.3-fold increase in cell number compared to uncoated control plates 24 hours after cell seeding (Fig. 4A). Phase contrast images of confluent monolayers of primary HCECs derived from three independent donor corneas showed that HCECs cultured on laminin E8s assumed a hexagonal contact-inhibited phenotype while HCECs cultured on the FNC coating mix showed less hexagonality and had a greater size variation (Fig. 4B). The average cell density was significantly higher for HCECs cultured on laminin E8s than on FNC (2397.1 ± 149.0 cells/mm², 1203.3 ± 209.7 cells/mm², respectively; Fig. 4C). 

Figure 4

 

Laminin E8 fragments enable HCEC culture while maintaining cell density. (A) The effect of laminin E8s was evaluated by seeding CECs at a density of 5.0 × 10³ cells/well, and determining the number of adhered CECs with a CellTiter-Glo luminescent cell viability assay after 24 hours. Culture plates were coated with laminin-521 or laminin-511-E8s. The experiments were performed in triplicate. *P < 0.01. (B, C) CECs obtained from three independent donors were cultured on FNC coating mix (a widely used coating reagent) or laminin 511-E8s. Phase contrast images depict CECs after 110 days cultivation. Graphs show the average cell densities of the CECs. Scale bar: 100 μm. *P < 0.01.

Involvement of Integrin α3β1 and α6β1 in Mediating Binding to Laminins

Laminin isoforms bind to receptors expressed on the cell surface, including integrin receptors and nonintegrin receptors such as syndecans and dystroglycans. Various integrins are known to bind to laminins, so we first evaluated the expression of integrin chains in human corneal endothelium at the mRNA and protein levels. PCR analysis of donor corneal endothelium confirmed the expression of integrins α1, α2, α3, α4, α5, α6, α7, α8, α9, α10, α11, αV, αIIb, β1, β5, and β8 (Fig. 5A). We next assessed the protein expression of these integrins on the cell surface of cultured HCECs. Flow cytometry revealed expression of α1, α2, α3, α6, and β1 integrin chains by the HCECs (Fig. 5B). The involvement of integrin α3β1 and α6β1 in activation of FAK was evaluated. Phosphorylation of FAK was enhanced in the HCECs cultured on the laminin-511 in comparison to the HCECs cultured on uncoated plate and laminin-211. However, functional blocking of α3β1 and α6β1 integrins by neutralizing antibodies suppressed phosphorylation of FAK of the HCECs cultured on laminin-511 (Fig. 5C). We then evaluated the effect of α3β1 integrin and α6β1 integrin on cell adhesion. α3β1 and α6β1 integrin neutralizing antibodies suppressed the adhesion of HCECs after 24 hours of seeding, even in the presence of laminin-511 (Fig. 5D). These findings indicate that laminins regulate HCEC adhesion through binding to integrin α3β1 and α6β1. 

Figure 5

 

Interaction between laminins and HCECs is facilitated by integrin α3β1 and α6β1. (A) The expression pattern of integrin α and β chains in donor corneal endothelium was evaluated by PCR. (B) The expression of integrins detected by PCR was evaluated on the cell surface of cultured HCECs by flow cytometry. Three independent cultured CECs were used for the experiments. (C) The involvement of integrin α3β1 and α6β1 in activation of FAK was evaluated by functional blocking of with neutralizing antibodies. The CECs were seeded, with or without neutralizing integrin α3β1 and α6β1 antibodies, on a culture plate coated with laminin 511 or 211. A noncoated plate was used as a control. Phosphorylation of FAK and total FAK were evaluated by Western blotting after 3 hours. The experiments were performed in duplicate. (D) The involvement of integrin α3β1 and α6β1 in laminin-CECs interaction was evaluated by functional blocking of with neutralizing antibodies. The CECs were seeded at a density of 5.0 × 10³ cells/well with or without neutralizing integrin α3β1 and α6β1 antibodies, and the number of adhered CECs was determined by performing a CellTiter-Glo luminescent cell viability assay after 24 hours. The experiments were performed in triplicate. *P < 0.01.

Discussion

Corneal endothelial dysfunction is a major disorder requiring corneal transplantation.³⁰ Modern corneal transplantation techniques such as Descemet's stripping automated endothelial keratoplasty (DSAEK) and Descemet's membrane endothelial keratoplasty (DMEK) are widely performed to replace damaged recipient corneal endothelium with healthy donor corneal endothelium.^31,32 The success of these techniques confirms that replacement of the corneal endothelium, rather than a full thickness corneal transplant, is sufficient for treating corneal endothelial dysfunction. It also raises the possibility of using tissue engineering therapy for treating corneal endothelial dysfunction by regenerating corneal endothelium alone rather than full thickness corneal tissue. Tissue engineering would have the distinct advantages of alleviating associated problems such as shortage of donors, primary graft failure, and technical difficulties in transplantation in severe cases.⁴ 

We previously reported that the corneal endothelium in rabbit and monkey corneal endothelial dysfunction models was regenerated by cell-based therapy without the need for a carrier.¹³ Rho kinase inhibitor enhanced the adhesion of CECs onto the substrate,³³ so we injected cultured CECs into the anterior chamber together with a Rho kinase inhibitor in order to modulate cell adhesion properties. However, a serious technical bottleneck for tissue engineering therapy for corneal endothelial dysfunction is the ability to provide sufficient numbers of cultured CECs.¹⁴ No protocol for culturing CECs for clinical use has been established, although several research groups, including ours, have been pursuing culture methods.^{14–16,34,35} For instance, we demonstrated that conditioned medium obtained from GMP-grade human BM-MSCs potentiated CEC proliferation.¹⁶ We also reported that the inhibition of transforming growth factor-β signaling counteracts the fibroblastic change under culture conditions.¹⁵ These recently updated techniques have enabled efficient in vitro expansion of human CECs for clinical use, and we have recently obtained approval of the Japanese Ministry of Health, Labour, and Welfare and have initiated clinical research into cell-based therapy for corneal endothelial dysfunction patients at the Kyoto Prefectural University of Medicine (Clinical trial registration was obtained from UMIN000012534; http://www.umin.ac.jp/english/). 

Typical culture substrates for CECs have included extracellular matrix (ECM) derived from bovine CECs³⁶ and FNC Coating Mix (Athena Environmental Sciences, Inc.).³⁷ We recently reported that a pericellular matrix prepared from human decidua-derived mesenchymal cells (PCM-DM) provides a xeno-free matrix substrate and enables efficient cell culture,³⁸ while avoiding the risk of contamination by xenogenic pathogens and immunogens from animal-derived matrixes. Interestingly, we noted that the use of PCM-DM enhanced the adhesion, migration, proliferation, and survival of CECs while maintaining cellular functions and cell density through integrin interactions.³⁸ 

Integrins are receptors that sense changes in the extracellular environment and transmit that information.²⁶ We were therefore motivated to examine the effect of constituents of the Descemet's membrane extracellular matrix on CEC culture, since the Descemet's membrane provides the in vivo extracellular environment for CECs. The Descemet's membrane is formed by CECs that secrete the ECM, and is composed of type IV collagen, type VIII collagen, fibronectin, and laminin.³⁹ Its average thickness at birth is 3 μm and it continuously increases in thickness throughout life.³⁹ 

Laminins are the most potent molecules that determine cellular fate by binding to integrins,^19,22 while collagen and fibronectin have less effect on the adhesion of CECs when compared to FNC Coating Mix.³⁸ We therefore hypothesized that laminins present in the Descemet's membrane are pivotal for CEC culture. The expression of laminin α1, α4, α5, β1, and γ1 chains has been indicated in human fetal Descemet's membrane,²⁵ but the expression profile of laminin isoforms and their biological role in the corneal endothelium is unclear. Therefore, we examined the expression of laminin isoforms expressed in the Descemet's membrane and CECs and showed that laminin α5, β1, β2, β3, γ1, and γ2 chains were expressed in the corneal endothelium at the mRNA level. Laminin-511 and -521 were indicated as the isoforms expressed in the Descemet's membrane. 

Several integrins (α2β1, α3β1, α5β1, and α6β1) are present in human CECs,⁴⁰ and expression of αvβ3 integrin was reported in rat CECs during postnatal corneal maturation.⁴¹ The αv family integrins were shown to modulate bovine CEC adhesion to collagen.⁴² However, evidence for a relationship between CECs and integrins is mostly indirect rather than direct and not much is known regarding integrin expression.⁴³ In the present study, we demonstrated that the α1, 2, 3, 6, and β1 subunits are expressed at the mRNA level and the protein is localized at the cell surface of the CECs. The neutralizing antibody experiments indicated that the α3β1 and α6β1 integrins are the ones that bind to CECs and transduce the signals. Reports that laminin globular modules of the laminin α5 chain are recognized by α3β1, α6β1, α7β1, and α6β4 integrins in various cell types provide further support for this assertion.^26,44 The laminin α5 chain contains functional RGD sequences and mediates adhesion and migration.⁴⁵ Further studies, such as binding assays, are needed to confirm the direct interaction of these integrins and laminins in CECs. 

FAK is a ubiquitously expressed cytoplasmic tyrosine kinase that regulates signals initiated by integrin-mediated ECM attachment.^46,47 In the present study, we showed that FAK in the CECs was activated by laminin-511, through α3β1 and α6β1 integrins, but not by laminin-211. Coincidently, we also showed that laminin-511, but not laminin-211, enhances cell adhesion and proliferation, resulting in a higher cell density. These data suggest that the integrins expressed by CECs recognized specific laminin isoforms. Therefore, the signal that regulates cell fate may be triggered by FAK activation. FAK mediates cell spreading, adhesion, and migration by balancing the activation of Rho GTPases.^48,49 Though we reported that inhibiting the Rho-Rho kinase pathway enhances proliferation and adhesion of CECs,³³ detailed downstream signaling pathway of FAK in CECs should be studied further. The saturation density of the cultured CECs is also important, but suboptimal seeding density could result in a decrease in cell saturation density and a loss of proliferative potential.⁵⁰ Therefore, one possible explanation is that enhancement of HCEC adhesion by laminin-511 and -521 maintained the proliferative potential, which in turn enabled a higher saturation density, when compared to the HCECs cultured on uncoated control or laminin-211 coated surfaces in this study. 

A recent report indicated that deposition of collagen types III and XVI, agrin, TGFBI, and clusterin is altered in the Descemet's membrane of late-onset Fuchs' corneal endothelial dystrophy.⁵¹ We demonstrated that ECM, which is not expressed in normal Descemet's membrane, does not activate FAK and does not support cell proliferation and adhesion. Consequently, the CECs of Fuchs' corneal endothelial dystrophy may possibly sense the alteration in ECM by integrins and transduce a pathological signal from the ECM. However, this possibility requires further investigation. 

The clinical application of laminin-511 and -521 still presents technical difficulties because laminins are large heterotrimeric proteins and require three independent vectors, coding the α, β, and γ chains, for production of a recombinant protein.²⁹ However, the recent report by Miyazaki et al.²⁹ that recombinant E8s of laminin isoforms support human embryonic stem cells suggests that laminin fragments may be a viable alternative to the native protein for CEC culture. E8s retain the integrin binding site that possess full binding capability to α6β1 integrin, and then E8s are a functionally minimal form of laminin.²⁹ Recombinant E8s also do not present the risk of contamination with xenogenic pathogens and immunogens, so they offer a safe culture alternative. 

In conclusion, we demonstrated that laminin-511 and -521 were the laminin isoforms present in the Descemet's membrane, and that these laminins modulate the adhesion and proliferation of CECs. Binding of these laminin isoforms to integrin α6β1 and integrin α3β1 may transduce functional signals from the microenvironment to the CECs. Lastly, the inclusion of laminin E8s may generate an ideal xeno-free defined substrate for the culture of CECs for clinical applications. 

Acknowledgments

The authors thank Lydia Sorokin for providing antibodies, Michio Hagiya for his invaluable advice about BM-MSCs, and Monty Montoya and Bernardino Iliakis (SightLife) for providing donor corneas. 

Disclosure: N. Okumura, P; K. Kakutani, None; R. Numata, None; M. Nakahara, None; U. Schlötzer-Schrehardt, P; F. Kruse, P; S. Kinoshita, P; N. Koizumi, P