Adhesion, Migration, and Proliferation of Cultured Human Corneal Endothelial Cells by Laminin-5

Abstract

Purpose.: To investigate the expression of laminin-5 (LM5) and its receptors by human corneal endothelial cells (HCECs) and whether recombinant human LM5 influences adhesion, proliferation, and migration of cultured HCECs.

Methods.: The expression of LM5 and its receptors was examined in human donor corneas by immunohistochemistry, reverse transcription–polymerase chain reaction, and flow cytometry. HCECs cultured under serum-free conditions were used for analysis of the biological effects of LM5. Changes in HCEC adhesion and proliferation due to LM5 were evaluated by counting the number of cells. HCEC migration was assessed by quantifying the percentage of wound closure in the wound-healing assay with an image-processing and -analysis software program.

Results.: Adult HCECs expressed the LM5 receptor α3β1 integrin, but not LM5 itself. Significantly more cells became adherent to recombinant LM5 (1.0 μg/mL)-coated dishes than to uncoated dishes in the cell adhesion assay. The proliferation of cultured HCECs was moderately promoted by LM5 (1.0 μg/mL) and soluble LM5 (20 ng/mL and 50 ng/mL) in the cell proliferation assay. A significantly higher percentage of wound closure was obtained with medium containing soluble LM5 than with control medium in the wound-healing assay.

Conclusions.: HCECs express the LM5 receptor α3β1 integrin. Recombinant LM5 promotes adhesion, migration, and moderate proliferation of cultured HCECs. It may be a critical factor in promoting HCEC culture and may contribute to the practical use of tissue-engineered HCECs.

Corneal endothelial cells (CECs) form a single layer lining Descemet's membrane and are essential for the maintenance of corneal hydration, thickness, and transparency. Severe CEC loss due to intraocular surgery, glaucoma, trauma, or congenital diseases can cause corneal stromal edema and opacity, leading to impaired visual acuity. Full-thickness corneal transplantation and Descemet's stripping with automated endothelial keratoplasty (DSAEK) ^{1 –3} are procedures that require fresh human corneas and are applied clinically for CEC loss, but there is a worldwide shortage of donors. Although human CEC (HCECs) are maintained in a nonreplicative state in vivo, ^{4 –7} adult HCECs have precursors largely committed to the CEC lineage ^8,9 and retain their proliferative capacity in vitro. ^10,11 Cultured HCECs have been proposed for the treatment of CEC loss, and they are expected to become useful for transplantation in the future. ^{12 –16} However, a very efficient culture technique is critical for regenerative medicine with HCECs. The optimum culture condition for HCECs that allows cell adhesion, proliferation, and migration has not yet been established. ^{17 –20}  

Laminin-5 (LM5) is composed of three chains, designated α3, β3, and γ2, that are encoded by three different genes (LAMA3, LAMB3, and LAMC2, respectively). ^21,22 LM5 is expressed in the basement membrane of stratified and transitional epithelium, where it co-localizes with the anchoring filaments. Some investigators have proposed that LM5 mediates epithelial cell adhesion through α3β1 integrin in focal adhesions and α6β4 integrin in hemidesmosomes. Deficiency or abnormality of LM5 is associated with epithelial cell fragility and epidermolysis bullosa. ^23,24  

We have previously reported that LM5 is synthesized by human corneal epithelial cells and is secreted into the extracellular matrix of the basement membrane, and we have also shown that LM5 has a crucial role in the adhesion and migration of corneal epithelial cells in vitro. ^25,26 Furthermore, highly adhesive LM5 is a key factor in the differentiation of corneal epithelial progenitors from differentiated cells. ²⁷ Moreover, LM5 has a role in the differentiation of human mesenchymal stem cells. ²⁸  

Regarding the relation between CECs and LM5, Descemet's membrane in infants and cell aggregates from donor HCECs show LM5 expression, ^20,29 but the actions of LM5 in CECs are still unknown. In this study, we investigated the effects of recombinant human LM5 on HCECs. We discuss a possible role for LM5 in regenerative medicine with the use of HCECs. 

Materials and Methods

This study was conducted in accordance with the Declaration of Helsinki. Corneas from donors 55 to 76 years of age were obtained from the Northwest Lions Foundation. Some were cut into sections for immunohistochemistry, reverse transcription–polymerase chain reaction (RT-PCR), and flow cytometry, and others were separated into cells and cultured for cell adhesion, proliferation, and wound-healing assays. 

Immunohistochemistry

Corneas were bisected, placed in OCT compound (Miles, Elkhart, IN), frozen in liquid nitrogen, and stored at −80°C. Unfixed 5-μm cryostat sections were collected on glass slides (Superfrost Plus; Fisher Scientific, Pittsburgh, PA) and fixed in methanol for 10 minutes. After the specimens were rinsed with phosphate-buffered saline (PBS), they were incubated for 30 minutes in PBS with 1% bovine serum albumin (BSA, Sigma-Aldrich, St. Louis, MO) and 0.1% Triton X-100 (Rohm & Haas, Philadelphia, PA) to prevent nonspecific staining. After two 5-minute rinses with PBS, the sections were incubated for 1 hour at room temperature with mouse anti-integrin-α_{3, -}α₆, -β₁, and -β₄ antibodies (1:100; Chemicon, Temecula, CA) and mouse anti-laminin-α3, -β3, and -γ2 antibodies (1:100; Santa Cruz Biotechnology, Santa Cruz, CA). Mouse IgG (1:1000; Sigma-Aldrich) was used as the control. After three washes with PBS (−), the sections were incubated at room temperature for 0.5 hour with fluorescein-labeled goat anti-mouse IgG (Alexa Fluor 488, 1:200; Molecular Probes, Eugene, OR). After another three washes, the sections were mounted with antifade mounting medium containing Hoechst 33342, to simultaneously counterstain nuclear DNA. Images were obtained with a fluorescence microscope (Nikon, Tokyo, Japan) and were saved to a personal computer. 

Flow Cytometry

To detect the expression of integrin-α₃, -α₆, -β₁, and -β₄ in endothelial cells, we used flow cytometry (FACScan; BD Biosciences, Franklin Lakes, NJ). After they were washed with buffer (FACS buffer; phosphate-buffered saline [pH 7.4], 0.5% bovine serum albumin, and 0.02% sodium azide), 1 × 10⁶ cells were treated with Fc-block (BD Biosciences, San Jose, CA) for 15 minutes and then incubated with mouse anti-integrin-α₃, -α₆, -β₁, and -β₄ antibodies (1:100; Chemicon) or with isotype control mouse IgG (BD PharMingen, San Jose, CA), for 1 hour at room temperature. Next, the cells were washed twice with the buffer and incubated with fluorescein isothiocyanate-conjugated anti-mouse IgG (BD PharMingen) for 30 minutes. Finally, the cells were washed three times with the buffer and analyzed. A kit containing propidium iodide (BD Biosciences) was used for staining, to gate out dead cells according to the manufacturer's instructions, and the data were analyzed (Cell Quest software; BD Biosciences). 

Reverse Transcription–Polymerase Chain Reaction

Total RNA was extracted (Trizol reagent; Life Technologies, Rockville, MD). RNA quantity was determined by an optical density measurement at 260 nm. cDNA was synthesized from 1.0 mg of total RNA using oligo (dT) primers with reverse transcriptase (SuperScript; Life Technologies). The cDNA samples were subjected to PCR with specific primers for LAMA3 (α3 chain), LAMB3 (β3 chain), LAMC2 (γ2 chain), and GAPDH (Table 1). The GAPDH served as an internal standard for sample normalization. The conditions for amplification were LAMA3, LAMB3, LAMC2, and GAPDH, 94°C for 45 seconds, 55°C for 45 seconds, and 72°C for 1 minute. Each PCR was performed in a 50-μL reaction volume containing 1 μL of 10 mM dNTPs, 5 μL of 10× buffer, 1 μL Taq polymerase, and 2 μL of 10 mM sense and antisense primers. Normalized samples were amplified in a linear range established by using serial cDNA dilutions and varying the number of cycles. Amplified products were separated by electrophoresis in 3% agarose gels after staining with ethidium bromide. 

Table 1.

 

Primers Used for DNA Amplification of LM-5

Products	Sequence (5′-3′)	Size (bp)
LAMA3 (α3 chain)	F: GGGATGCCTCCAGCAGTGAG	547
	R: GTGCATTCATCATCACATTCT
LAMB3 (β3 chain)	F: TGAGGTTCAGCAGGTACTGC	663
	R: TAACTGTCCCATTGGCTCAG
LAMC2 (γ2 chain)	F: CTGAGTATGGGCAATGCCAC	452
	R: GCTCTGGTATCAACCTTCTG
GAPDH	F: GCACCGTCAAGGCTGAGAAC	120
	R: TGGTGAAGACGCCAGTGGA

HCEC Culture

Primary culture of HCECs was performed as described previously. ¹¹ All primary cultures and serial passaging of HCECs was performed in growth medium that consisted of low-glucose Dulbecco's modified Eagle's medium (DMEM) supplemented with 15% fetal bovine serum, 2.5 mg/L amphotericin B (Fungizone; Invitrogen-Gibco, Grand Island, NY), 2.5 mg/L doxycycline, and 2 ng/mL basic fibroblast growth factor (bFGF; Sigma, St. Louis, MO). Small explants of the endothelial layer, including Descemet's membrane, were removed with sterile surgical forceps. The explants from each cornea and were placed endothelial cell-side down into four 35-mm culture dishes coated with bovine extracellular matrix, which was obtained by incubating bovine CECs on culture dishes, as described elsewhere. ¹¹ The dishes were then carefully placed in the incubator. After 3 days, the medium was exchanged and then replaced every other day thereafter. When the proliferating cells had reached a sufficient density, passaging was done at ratios ranging from 1:1 to 1:4. Subsequent passaging was performed by the same method, but at a ratio of 1:16. To prevent interference by serum factors, cells from the third passage were incubated in serum-free medium (i.e., DMEM supplemented with B27; Invitrogen-Gibco), 2.5 mg/L amphotericin B, and 2.5 mg/L doxycycline. 

Adhesion Assay

Recombinant human LM5 (1.0 μg/mL; Oriental Yeast Co. Japan), human type IV collagen (10 μg/mL), and human fibronectin (10 μg/mL; BD Biosciences, Bedford, MA) were used to coat 12-well culture dishes, after which third-passage HCECs were incubated at a density of 6000 cells/cm² in serum-free medium. After 1 hour, the plate was washed twice with PBS, and the number and morphology of the adherent cells were determined. For cell counting, 0.05% trypsin/EDTA was added for 5 minutes at 37°C, and dissociation into a single-cell suspension was achieved by pipetting. The cells were subsequently resuspended in the basal medium. The viability of isolated HCECs was >90%, as shown by trypan blue staining (Wako Pure Chemical Industries, Osaka, Japan). The number of cells was then determined (Coulter counter; Beckman-Coulter, Hialeah, FL). 

Proliferation Assay

HCECs from the third passage were incubated at a density of 6000 cells/mm² in serum-free medium, and each plate was washed twice with PBS after 1 hour. The serum-free medium was subsequently exchanged every other day, and the number and morphology of adherent cells were compared between cultures on LM5-coated dishes and the controls after 1, 3, and 7 days. 

Because the initial number of adherent cells was different between uncoated and LM5-coated dishes due to promotion of cell adhesion by LM5, we next added soluble LM5 to the culture medium to obtain an equal number of cells initially and evaluated proliferation only. HCECs were incubated in 12-well dishes coated with human type IV collagen (10 μg/mL), which promotes cell adhesion to dishes in the serum-free culture condition. On the next day, after the dishes were washed twice with PBS, culturing was continued in serum-free medium with soluble LM5 at 20, 50, or 100 ng/mL. The medium was exchanged every other day, and the number and morphology of adherent cells were compared with control cultures after 1, 3, and 7 days. The cells were counted as for the adhesion assay (Coulter counter; Beckman Coulter). 

Wound-Healing Assay

HCECs from the third passage were suspended in DMEM with 15% fetal bovine serum and incubated in 12-well tissue culture dishes (1 × 10⁶ cells/well) until confluence was reached. The medium was removed and serum-free medium containing 50 ng/mL of soluble LM5 was added to the wells 12 hours before wound creation. A 1000-μL pipette tip was used to create wounds at intervals of approximately 2 mm. Immediately after wounding, the medium was exchanged for fresh medium containing 50 ng/mL of soluble LM5, to remove cellular debris. Wounds were examined under a microscope, and images were saved to a personal computer at 0, 24, 48, and 72 hours after wound creation. The wound closure area was quantified with an image-processing and analysis software program (Image J 1.42q; developed by Wayne Rasband, National Institutes of Health, Bethesda, MD; available at http://rsb.info.nih.gov/ij/index.html). 

Statistical Analysis

Statistical comparisons were performed with the unpaired Student's t-test for two groups. One-way analysis of variance and Scheffé's multiple comparison test (Excel; Microsoft, Redmond, WA) were used for comparisons among three or more groups. Analyses were performed with the same software, and the level of significance was set at P < 0.05. 

Results

Expression of LM5 Subunits and Receptors in the Human Cornea

Expression of LM5 subunits (α3, β3, and γ2) and their receptors (integrin α3β1 and α6β4) was investigated in the human cornea. Although the three LM5 subunits were expressed in the basement membrane of the corneal epithelium, these subunits were not detected in the corneal endothelium by immunohistochemistry (Fig. 1A). The mRNAs of the subunits were detected in both the human corneal epithelium and the HCECs by RT-PCR (Fig. 1B). 

Figure 1.

 

Expression of LM5 subunits and their receptors in the human cornea. (A) LM5 subunits (α3, β3, and γ2) were expressed in the basement membrane of the corneal epithelium, but not in the CECs. (B) Expression of LM5 subunits by RT-PCR. LM5 subunits were detected in the HCECs by RT-PCR. (C) Expression of a LM5 receptor (α3β1 integrin) was revealed in the corneal endothelium by immunohistochemistry. (D) Flow cytometry also showed α3β1 integrin expression by HCECs. Ep, epithelium; BM, Bowman's membrane; EC, endothelial cell.

One of the LM5 receptors (α3β1 integrin) was detected in the human corneal endothelium by immunohistochemistry (Fig. 1C). α3β1 integrin expression was also detected by flow cytometry of dissociated HCECs (Fig. 1D). The other LM5 receptor (α6β4 integrin), which is found in hemidesmosomes, was not detected in the CECs, although both α3β1 and α6β4 integrins were expressed in the basement membrane of the corneal epithelium (data not shown). 

Effect of LM5 on HCEC Adhesion

To determine whether LM5 promotes HCEC adhesion, we conducted a cell-adhesion assay in serum-free medium. Figure 2 shows representative photographs of adherent HCECs on an uncoated control dish (Fig. 2A) and on a dish coated with 1.0 μg/mL LM5 (Fig. 2B). LM5 significantly promoted the attachment of HCECs to the dish after 1 hour of incubation. Two thirds of the seeded cells were adherent to the LM5-coated dishes, which was 10-fold more than the cells adhering to the control dishes (P = 0.006; Fig. 2E). No morphologic differences were observed in the cells adherent to both types of dish. 

Figure 2.

 

Adhesion assay with several adhesive factors. Adherent cells are shown on uncoated (A), LM5-coated (B), fibronectin-coated (C), and human type IV collagen-coated (D) dishes by phase-contrast microscopy. The number of adherent cells after 1 hour of culture is shown (E). Similar findings were obtained with repeated experiments. FN, fibronectin, HC IV, human type IV collagen (n = 3, mean ± SD). NS, not significant; *P < 0.05. Scale bar, 100 μm.

To assess the potency of LM5, we compared its effect with fibronectin and human type IV collagen. As shown in Figure 2, both LM5 and type IV collagen promoted a higher number of adherent cells (Figs. 2B–D). No morphologic differences were observed among the adherent cells of each group (Figs. 2A–D). The LM5-coated dishes had approximately 1.5 times more adherent cells than the fibronectin-coated dishes (Fig. 2E, P < 0.05), but there was no significant difference between the dishes coated with LM5 and type IV collagen (Fig. 2E). An additive effect of LM5 and type IV collagen on cell adhesion was not detected (data not shown). 

Effect of LM5 on HCEC Proliferation

HCECs were cultured for 7 days, to determine whether LM5 would affect their proliferation. HCECs grown on human collagen type IV–precoated dishes did not proliferate during the observation period. In contrast, the number of adherent cells on the LM5-coated dishes increased by 1.5 times after 7 days of culture (P = 0.007; Fig. 3A). As the initial number of adherent cells was different between uncoated and coated dishes because adhesion was promoted by LM5, we next added soluble LM5 to the culture medium to obtain an equal initial number of adherent cells and evaluated cell proliferation only. HCECs were incubated with or without soluble LM5, and the cells were counted after 14 days. When the HCECs were incubated with soluble LM5 (20 or 50 ng/mL), a significantly higher number of cells was obtained than in the control cultures without soluble LM5 (P = 0.01 and 0.003, respectively; Fig. 3B), indicating that LM5 promoted HCEC proliferation. However, the number of HCECs decreased after incubation with 100 ng/mL of LM5 (P = 0.006; data not shown). 

Figure 3.

 

HCEC proliferation assay, with or without soluble LM5. (A) The number of adherent cells was counted on uncoated and LM5-coated culture dishes after 1, 3, and 7 days of culture. Adherent cells proliferated on LM5-coated, but not on uncoated, dishes, and a significantly higher number of cells were obtained after 7 days of culture compared with 1 day. (B) The number of HCECs in medium containing soluble LM5 was significantly higher than in the control medium without soluble LM5 after 14 days of culture. Similar findings were obtained with repeated experiments (n = 3, mean ± SD). NS, not significant; *P < 0.05; **P < 0.01.

Effect of LM5 on Wound Healing

Sharp wound margins and no HCEC migration were seen in the control cultures (Fig. 4A) and in the cultures with soluble LM5 (Fig. 4B) immediately after wound creation. By 72 hours after wounding, only a small number of HCECs had migrated into the wound in the control cultures (Fig. 4C). In contrast, the wound margins were obscured, and more than half of the wound area was covered by migrating HCECs in the cultures containing LM5 (Fig. 4D). As shown in Figure 4E, the percentage of wound closure quantified by image analysis was significantly higher in the cultures containing soluble LM5 than in the control cultures at 48 and 72 hours after wound creation. 

Figure 4.

 

Wound assay in serum-free medium with or without LM5. Wound healing is shown for cultured HCECs in control medium (A, 0 hr; C, 72 hours) and soluble LM5-containing medium (B, 0 hours; D, 72 hours) under phase-contrast microscopy. Few cells migrated into the wound in the control culture without LM5 (C). In contrast, the wound was covered by cells in the culture containing LM5 (D). (E) The percentage of wound closure was determined with image-analysis software. HCECs cultured with soluble LM5 showed a twofold increase of the percentage of wound closure compared with control HCECs cultured without LM5 after 48 and 72 hours. Similar findings were obtained with repeated experiments (n = 3, mean ± SD). NS, not significant; *P < 0.01.

Discussion

The basement membrane protein LM5 is a potent cell adhesion molecule that promotes both cellular adhesion and migration. LM5 also supports various biological activities ^30,31 related to cell survival, antiapoptosis, ³² wound angiogenesis, keratinocyte-to-endothelial cell cross-talk, ³³ and cell proliferation. ³⁴ Conversely, LM5 and its receptor (α3β1) suppress cell migration and wound healing in the skin. ³⁵ The present study demonstrated that LM5 enhanced the adhesion, proliferation, and migration of cultured adult HCECs. We used serum- and growth factor–free culture medium when investigating the biological effects of LM5 on cultured HCECs, because bovine serum contains various unknown factors, and integrin α3β1 is a receptor not only for LM5, but also for fibronectin, epiligrin, thrombospondin, CSPG4, and CD151. Therefore, we concluded that serum-free medium is essential for prevention of the biological effects of these ligands. Among the biological activities detected in this study, the proliferative effect of LM5 on HCECs was only moderate. To test whether LM5 could prolong cell survival by suppressing apoptosis, we used flow cytometry to count the annexin V-positive cells in HCEC cultures, with and without LM5. LM5 did not increase the number of annexin V-positive apoptotic cells (Yamaguchi M, unpublished observation, 2010), suggesting that LM5 promotes cell proliferation without affecting cell survival. 

LM5 is expressed abundantly in the corneal limbal basal epithelium where progenitor cells are located and isolated undifferentiated progenitor cells from corneal limbal epithelium also produce LM5. ²⁷ In addition to LM5 expression in HCECs from infant basement membranes ²⁹ and in HCEC aggregates, ²⁰ HCEC precursors isolated by the sphere-forming assay expressed LM5 (Yamaguchi M, unpublished observation, 2009). The sphere-forming assay is used to isolate HCEC precursors with longer telomeres, higher telomerase activity, and younger progeny than the original cultured HCECs. ³⁶ We found that adult HCECs express a receptor for LM5 (integrin α3β1), but LM5 detected by highly sensitive RT-PCR was undetectable immunohistochemically. This discrepancy may be explained by the facts that most differentiated HCECs does not express LM5, and a small number of HCEC precursors can produce LM5. These results suggest that immature or undifferentiated HCECs express LM5, whereas LM5 is suppressed during development or differentiation. It is unclear whether undifferentiated cells simply produce LM5 or whether LM5 determines the characteristics of these cells. However, LM5 may have the effect of maintaining undifferentiated HCECs during culture. Moreover, our findings suggest that the functional system involving a receptor and ligand is conserved even in adult HCECs and that the α3β1 receptor for LM5 may be the main functional receptor related to its biological effects on these cells. 

DSAEK has become a common procedure for the management of corneal endothelial disorders. In comparison with conventional penetrating keratoplasty, DSAEK allows faster visual rehabilitation with minimal astigmatism or refractive changes and also preserves the structural integrity of the eye. ^{37 –43} With respect to the clinical application of CEC sheet transplantation instead of DSEAK, various transplantation techniques using cultured CECs have been reported. ^12,13,15,44 Cultured CECs are an attractive source for regenerative medicine because there are not enough donor corneas anywhere in the world, while cultured CECs could be used to create a large number of corneal sheets for transplantation. Collagen type IV also promoted cell adhesion to dishes, but had no effect on cell proliferation on collagen type IV–precoated dishes, as shown in Figure 3, whereas a small amount of soluble LM5 on collagen type IV–precoated dishes induced an increase in the number of cells. Our data showing the effects of LM5 on cultured HCECs suggest that promotion of cell adhesion, migration, and moderate proliferation by LM5 contributes to the clinical application of tissue-engineered HCECs. 

In summary, adult HCECs expressed the LM5 receptor α3β1 integrin, but not LM5 itself. Exogenous LM5 promoted the adhesion, migration, and proliferation of cultured HCECs. Our findings suggest that LM5 could play a role in the more efficient culture of HCECs for regenerative medicine by promoting both cell adhesion and migration. 

Footnotes

 Supported in part by a grant from the Ministry of Education, Culture, Sports, Science, and Technology of Japan.

Footnotes

 Disclosure: M. Yamaguchi, None; N. Ebihara, None; N. Shima, None; M. Kimoto, None; T. Funaki, None; S. Yokoo, None; A. Murakami, None; S. Yamagami, None