Clusterin Promotes Corneal Epithelial Cell Growth through Upregulation of Hepatocyte Growth Factor by Mesenchymal Cells In Vitro

Abstract

Purpose.: Although the cornea expresses high levels of clusterin (CLU), the role of CLU in the cornea is poorly understood. This study was performed to investigate the possible role of CLU in corneal epithelial homeostasis.

Methods.: CLU was overexpressed in 3T3 cells by transfection of a vector encoding full-length CLU (Clu-3T3). Colony-forming efficacy (CFE) was compared in mouse corneal cell line (TKE2) and human primary corneal/limbal epithelial cells that were cocultured with Clu-3T3 and mock-3T3. To determine whether feeder cells have a contact effect, cocultures without feeder-epithelium contact were also performed. Neutralizing antibody against CLU was used to assess the effects of secretory CLU in TKE2 cells cocultured with Clu-3T3 cells. The expression of growth factors associated with limbal stem/progenitor cell maintenance and growth were analyzed by RT-PCR and Western blot analysis.

Results.: TKE2 cells cocultured with Clu-3T3 feeders showed higher CFE and were larger in colony size than TKE2 cells cocultured with mock-3T3 feeders. Increased CFE of TKE2 was observed without direct contact with Clu-3T3 cells, which was significantly blocked by treatment with CLU neutralizing antibody. Clu-3T3 cells expressed higher levels of HGF than mock-3T3 cells, which were significantly suppressed with anti-HGF neutralizing antibodies. Collectively, the promotion of colony-forming and cell proliferation by Clu-3T3 cells was partially mediated by the induction of HGF.

Conclusions.: Clusterin indirectly enhances the CFE of corneal/limbal epithelial cells by inducing the production of HGF by feeder cells, suggesting a role in epithelial-mesenchymal interaction.

Clusterin (CLU) is a glycoprotein first isolated from ram rete testis fluid; it is also known as apolipoprotein J, sulfated glycoprotein-2, glycoprotein III, and testosterone-repressed message-2. It is composed of two 35- to 40-kDa subunits (α and β) encoded by a single gene, and it forms a heterodimer stabilized by disulfide bonds. ^{1 –4} It has been implicated in diverse physiological functions, including cell-cell interactions, ⁵ complement inhibition, lipid transportation, cell survival, apoptosis, ⁶ aging, ⁷ and cell protection from cytotoxic stress. ^{8 –10} CLU is ubiquitously expressed in many tissues and body fluids; however, it was also found to be the most abundant gene transcript in the human corneal epithelium. ^11,12  

Despite these initial reports, the function of CLU in the homeostasis and survival of human corneal/limbal epithelial cells is still not clearly understood. Recent studies suggested that CLU is involved in cell development and proliferation. ^13,14 CLU expression is also upregulated by a multitude of factors, including cellular injury and stress and cell growth, differentiation, and aging. ^6,7,15 Given that the corneal epithelium is a nonkeratinized, stratified squamous epithelium constantly undergoing proliferation and migration, we hypothesized that CLU plays a role in the maintenance of corneal epithelial homeostasis. When the corneal epithelium is injured, limbal epithelial stem cells govern the renewal of the corneal epithelium by generating transient-amplifying cells (TACs) that migrate centripetally from the limbus to the corneal basal layer. ^16,17 TACs further proliferate and differentiate into terminally differentiated cells to maintain ocular surface homeostasis. ^18,19 In this study, we investigated the effect of CLU on cultured murine corneal epithelial cells and primary human limbal epithelial cells using a coculture system, with NIH-3T3 feeder cells overexpressing CLU. 

Materials and Methods

Murine Corneal/Limbal Epithelial Cell Line (TKE2) Culture

TKE2 is a murine limbal/corneal epithelium-derived progenitor cell line. ²⁰ TKE2 cells were maintained in defined keratinocyte serum-free medium (KSFM; Gibco-Invitrogen Corp., Carlsbad, CA) supplemented with 10 ng/mL human recombinant epithelial growth factor, 1% penicillin/streptomycin, and growth supplement supplied by the manufacturer until use. Cell cultures were incubated at 37°C, under 95% humidity and 5% CO₂, and culture medium was changed every 3 to 4 days. 

Primary Human Corneal/Limbal Epithelial Cell Culture

Human primary corneal/limbal epithelial cells (HLECs) were isolated from the limbus of eye bank corneas after the central corneal buttons were used for transplantation. Iris, endothelium, and conjunctiva were surgically removed from corneal limbus, and the limbus was treated with 2.5 U/mL neutral protease (Dispase II; Roche, Basel, Switzerland) in F12/Dulbecco's modified Eagle's medium (DMEM) at 4°C overnight. The epithelium was separated from the stroma with a cell scraper and was dispersed in 0.05% trypsin EDTA at 37°C for 30 minutes. HLECs were suspended in defined KSFM. Unless indicated otherwise, 5 × 10⁴ HLECs were seeded in 25-cm² flasks. The cultures were incubated at 37°C, under 95% humidity and 5% CO₂, and the medium was changed every 3 days. 

Establishment of a 3T3 and STO Cell Line Stably Expressing Clusterin

pCAGIPuro plasmid was a kind gift from Hitoshi Niwa (Laboratory for Pluripotent Cell Studies, Riken Center for Developmental Biology, Kobe, Japan). The plasmid, with full-length CLU cDNA containing a COOH-terminal HA, was transfected into NIH-3T3 cells and STO cells using reagent (LipofectAMINE 2000; Life Technologies, Inc., Carlsbad, CA) (Clu-3T3, mock-3T3, Clu-STO, mock-STO). The selection of clones was carried out for 2 weeks with 10 μg/mL puromycin (Life Technologies, Inc.) and were maintained in DMEM (Gibco-Invitrogen Corp., Carlsbad, CA) supplemented with 10% FCS and 1% penicillin/streptomycin. No predictable morphologic changes owing to CLU transfection were observed in either cell line. 

Colony-Forming Efficiency

To evaluate the proliferative potential of cell colonies, Clu or mock-3T3 and STO cells were used in a colony-forming efficiency (CFE) assay, as previously described. ^21,22 Clu or mock-transfected cells in DMEM containing 10% FCS were plated at a density of 3 × 10⁵ cells in six-well culture plates (Iwaki, Naperville, IL). The next day, transfected cells were treated with mitomycin C (MMC; Nakarai Tesque, Kyoto, Japan) (4 μg/mL) for 2.5 hours at 37°C, and then single cells were seeded at 300 cells/well in a mixture of equal parts defined KSFM and supplemental hormonal epithelial medium ²³ containing 5% FCS. For separate cultures, corneal/limbal epithelial cells were incubated with cell culture inserts (Transwell; Corning Inc., Corning, NY) to inhibit contact between epithelial and feeder cells. CFE was calculated by the percentage of colonies at day 14 generated by the number of epithelial cells plated in the well. Colony size (mm²) and number of colonies (n = 3) were quantified using ImageJ software (developed by Wayne Rasband, National Institutes of Health, Bethesda, MD; available at http://rsb.info.nih.gov/ij/index.html). Growth capacity was evaluated on day 14 when cultured cells were stained with rhodamine B (Wako, Osaka, Japan) for 30 minutes. 

Western Blot Analysis

Western blot analysis was performed using standard Western blotting methods. Clu-3T3, mock-3T3 cells, and TKE2 cells were dissociated with lysis buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1% Nonidet P-40 [Calbiochem, Darmstadt, Germany]) and were homogenized. Culture supernatants of Clu-3T3 or mock-3T3 cells were concentrated five times with spin column (Microcon; Millipore, Bedford, MA). Protein concentration of the supernatant was measured using a BCA protein assay kit (Pierce, Rockford, IL). All samples were then diluted in 2× sample buffer (100 mM Tris-HCl [pH 6.8], 4% SDS [Invitrogen], 20% glycerol [Wako], and 12% 2-mercaptoethanol [Wako]) and were boiled. Ten micrograms of each sample were loaded on a 10% Tris-HCl gels (Ready Gels J; Bio-Rad Laboratories Inc., Hercules, CA) and were transferred onto polyvinylidene difluoride membranes (Millipore, Billerica, MA). Membranes were reacted with antibodies against CLU (M-18; Santa Cruz Biotechnology, Santa Cruz, CA), hepatocyte growth factor (HGF) (D-19; Santa Cruz Biotechnology), and β-actin (mabcam8226; Abcam, Cambridge, MA) for 60 minutes at room temperature. After three washes in TBST, donkey biotinylated anti-rabbit IgG (Zymed, San Francisco, CA) was added for 30 minutes at room temperature. The protein bands were visualized by enhanced chemiluminescence (GE Healthcare, Piscataway, NJ). 

RT-PCR

Total cellular RNA was isolated from Clu-3T3 or mock-3T3 cells with purification mini kits (RNeasy; Qiagen, Hilden, Germany), according to the manufacturer's specifications. One microgram of total RNA was converted to cDNA (iScript cDNA Synthesis kit; BioRad) and was subsequently used for reverse transcription-polymerase chain reaction (RT-PCR) analysis. PCR was performed in a thermal cycler as follows: 94°C for 5 minutes, followed by 30 amplification cycles (94°C, 30 seconds; 60°C, 30 seconds; 72°C, 1 minute). Primer sets for mouse glyceraldehyde-3-phosphate dehydrogenase (gapdh) (sense, 5′-TGACGTGCCGCCTGGAGAAA-3′; antisense, 5′-AGTGTAGCCCAAGATGCCCTTCAG-3′), mouse hgf (sense, 5′-CACCTCCTCCTGCTTCATGT-3′; antisense, 5′-CACCTGTTGGCACACTCATC-3′), mouse kgf (sense, 5′-TTGACAAACGAGGCAAAGTG-3′; antisense, 5′-TTGCAATCCTCATTGCATTC-3′), mouse igf1 (sense, 5′-TGGATGCTCTTCAGTTCGTG-3′; antisense, 5′-GGGAGGCTCCTCCTACATTC-3′), mouse fgf2 (sense, 5′-AGCGGCTCTACTGCAAGAAC-3′; antisense, 5′-CCGTTTTGGATCCGAGTTTA-3′), and mouse egf (sense, 5′-TCCCAGCGAGAAAGACTGAT-3′; antisense, 5′-TTGGCCATTTCAATCACAGA-3′) were synthesized at Operon Biotechnologies (Tokyo, Japan). Equal amounts of PCR-amplified products were visualized by ethidium bromide. The mRNA expression levels for each gene were normalized to the gapdh gene. 

Quantitative Real-Time PCR

Total RNA was extracted (RNeasy; Qiagen) and digested (DNase I; Qiagen) according to the manufacturer's instructions. Single-strand cDNA was synthesized with reverse transcriptase (SuperScript II; Invitrogen). Semiquantitative real-time PCR was performed with a double-stranded DNA-binding dye (SYBR Green I; Applied Biosystems, Foster City, CA) using a sequence detection system (ABI Prism 7700; Applied Biosystems). Expression levels of mRNA were normalized by the median expression of a housekeeping gene (gapdh). Copy number was expressed as the number of transcripts per nanogram of total RNA. Primer sequences for mouse gapdh and hgf are described here. 

ELISA for Mouse Clusterin

Cells were incubated for 48 hours in 96-well plates. Supernatants of Clu-3T3 and mock-3T3 cells were stored at −70°C until further ELISA assay. Mouse CLU concentrations were measured by an enzyme-linked immunosorbent assay kit (Life Diagnostic, West Chester, PA). The detection limit of the kit was approximately 3.9 ng/mL. 

Immunohistochemistry for Mouse Clusterin

Frozen sections of normal mouse cornea were fixed for 10 minutes in cold acetone. Four-well chamber slides of TKE2 were fixed for 10 minutes in 100% methanol. These were permeabilized cell membranes with 1% nonionic surfactant (Triton X-100; Sigma, St. Louis, MO) for 10 minutes at room temperature. Frozen sections and slides were blocked by incubation with 10% normal goat serum (Chemicon International Inc., Temecula, CA) for 30 minutes at room temperature. Antibodies to CLU (1:40) (H-330; Santa Cruz Biotechnology) were applied and incubated for 60 minutes at room temperature, followed by incubation with Cy3-conjugated secondary antibody. Isotype antibodies were used as negative controls. After three washes with TBST (0.825 mM Tris, 136.9 mM NaCl, 1.34 mM KCl, 0.1% Tween 20), the sections were incubated with 1 μg/mL 4′,6-diamidino-2-phenylindole (DAPI; Dojindo Laboratories, Tokyo, Japan) at room temperature for 5 minutes. Finally, sections were washed three times in TBST and coverslipped using an aqueous mounting medium (PermaFluor; Beckman Coulter, Marseille, France). 

Statistical Analysis

Differences between two paired groups were analyzed by the paired Student's t-test and considered significant at P < 0.05. Values are expressed as the mean ± SD. 

Results

Clusterin Expression in Clu-3T3 Cells

Western blot analysis showed higher levels of intracellular CLU (mature form; pre-sCLU and precursor holoprotein form; pre-CLU) expression levels in the Clu-3T3 cells than mock-3T3 cells (Fig. 1A). In the culture supernatant, expression of the secretory form, sCLU, was higher as well in Clu-3T3 cells (Fig. 1B). This trend was also found in Clu-STO and mock-STO cells (data not shown). Furthermore, the concentration of sCLU in the condensed supernatant of Clu-3T3 cells was detected by ELISA (approximately 100 ng/mL) but was under detection levels in mock-3T3 cells (Fig. 1C). Western blot analysis (Fig. 1D) and immunocytochemistry (Fig. 1E) confirmed that TKE2 corneal cells expressed CLU. Immunohistochemistry of the mouse cornea also confirmed the expression of CLU in vivo (Fig. 1F). 

Figure 1.

 

Clusterin expression in Clu-3T3 and TKE2 cells. (A) Western blot analysis of cell lysate shows the expression of sCLU precursor form (pre sCLU) and CLU precursor form (pre CLU) in Clu-3T3 cells but not in mock-3T3 cells (control). (B) Western blot analysis of culture supernatant shows that sCLU was detected more in Clu-3T3 than in mock-3T3. (C) ELISA of secreted CLU protein concentration in culture supernatant of Clu-3T3 and mock-3T3 cells (n = 3; mean ± SD; *P < 0.01). (D) Western blot and (E) immunocytochemistry show the expression of CLU by TKE2 corneal epithelial cells. (F) CLU expression (red) in the mouse corneal epithelium shown by immunohistochemistry.

CFE of TKE2 and Corneal/Limbal Cells Using Clu-3T3 Feeder Cells

To observe the effects of CLU on the clonal growth of corneal epithelial cells, colony-forming assays were performed in mouse corneal epithelial cells (TKE2 cells) cocultured with mitomycin C–treated Clu-3T3 and mock-3T3 feeder cells for 14 days. TKE2 cells cultured with CLU-3T3 cells showed a higher colony-forming ability than did those in mock-3T3 cells, with a 50% increase in CFE (Figs. 2A, 2B). Larger colonies were obtained with Clu-3T3 cells than with mock-3T3 cells. To test the effect of colony formation in response to culture supernatants produced by Clu-3T3 cells, we performed the colony-forming assay using culture inserts to separate epithelial cells from feeder cells (Fig. 2C). CFE of TKE2 cultured separately from Clu-3T3 cells was slightly lower than cocultured cells with feeder contact; however, significantly higher colony-forming capacity was observed compared with mock-3T3 feeder cells (Fig. 2D). 

Figure 2.

 

Effect of Clu-3T3 cells on TKE2 colony formation. (A) Murine corneal epithelial cells, clone TKE2, were cocultured with mitomycin C–treated Clu-3T3 and mock-3T3 cells at a density of 300 cells/well of six-well plates. After 14 days, colonies were stained with rhodamine B to show larger colony sizes with Clu-3T3 compared with mock-3T3 cells. (B) CFE on Clu-3T3 was significantly higher than mock-3T3 feeder cells (n = 3; mean ± SD; *P < 0.01). (C) TKE2 cells cultured with mitomycin C–treated Clu-3T3 and mock-3T3 cells using transwell cultures to separate epithelial cells from feeder cells. (D) CFE was also significantly higher with Clu-3T3 cells than with mock-3T3 cells (n = 3; mean ± SD; *P < 0.05).

Human corneal/limbal epithelial cells (HLECs) formed colonies by coculturing with 3T3 cells in the same fashion, although the CFE was relatively low level compared with TKE2 (Fig. 3A). Once again, Clu-3T3 cells showed a higher CFE in HLECs compared with mock-3T3, showing that the phenomenon was observed irrespective of animal species (Fig. 3B). We examined the reproducibility of results using a different line of feeder cells. As shown in Figures 2C and 2D, STO cells produced colonies that were slightly smaller than 3T3 cells under the same conditions, but significant differences in CFE were observed with Clu-STO cells as with Clu-3T3 cells (Figs. 2C, 2D). Because of the stability and reproducibility of TKE2 combined with 3T3 cells, all subsequent experiments were performed with this combination of cells. 

Figure 3.

 

CFE of different combinations of feeder and epithelial cells. (A) Primary human corneal/limbal epithelial cells were cocultured with mitomycin C–treated Clu-3T3 and mock-3T3 cells. After 14 days, colonies were stained with rhodamine B. (B) CFE of HLE cells was significantly higher in Clu-3T3 feeder cells than in mock control (n = 3; mean ± SD; *P < 0.01). (C, D) Increase in CFE of TKE2 cells was observed using Clu-STO compared with mock-STO cells (n = 3; mean ± SD; *P < 0.05).

Effect of CLU Neutralizing Antibody on CFE of Corneal/Limbal Epithelial Cells

The data presented thus far suggest that stable overexpression of CLU was sufficient to cause the increased CFE of mouse and human corneal/limbal epithelial cells for both 3T3 and STO cells. We further investigated whether enhanced CFE was mediated through overproduction of sCLU. TKE2 cells were cocultured with Clu-3T3 or mock-3T3 cells in the presence of CLU neutralizing antibody or control antibody for 14 days to observe CFE. As shown in Figure 4, the CFE of TKE2 on Clu-3T3 cells was significantly reduced from 35.0% ± 2.6% to 22.2% ± 3.4% with CLU neutralizing antibody treatment. Colony size with CLU neutralizing antibody treatment was smaller than isotype control. CFE was not affected by isotype control antibody treatment. 

Figure 4.

 

Effect of anti–mouse CLU neutralizing antibody on TKE2 colony formation. TKE2 cells were cocultured with mitomycin C–treated Clu-3T3 and mock-3T3 cells and treated with CLU antibody (0.5 μg/mL), which binds to sCLU. After 14 days, colonies were stained with rhodamine B (left), and CFE was evaluated (right). CFE on Clu-3T3 (solid bars) and mock-3T3 (empty bars) was inhibited by the antibody, with statistical significance in the Clu-3T3 group (n = 3; mean ± SD; *P < 0.05).

Expression of Growth Factors by Clu-3T3 Cells

Given that neutralizing antibodies against CLU suppressed the CFE of corneal/limbal epithelial cells, we sought to investigate whether sCLU directly promotes CFE and cell proliferation in corneal/limbal epithelial cells. Interestingly, we found that several different concentrations (0.1–3 μg/mL) of mouse recombinant CLU had no effect on the CFE and cell proliferation of TKE2 cells (data not shown). Therefore, we hypothesized that sCLU acts on 3T3 feeder cells in an autocrine fashion, which, in turn, produces factors that indirectly regulate epithelial proliferation and CFE. We examined the expression of growth factors associated with limbal stem/progenitor cell maintenance and growth (EGF, FGF2, HGF, IGF1, KGF) and compared the expression of these growth factors in CLU-3T3 to those in mock-3T3 by RT-PCR. As a result, we found that higher levels of hgf were expressed in Clu-3T3 cells than in mock-3T3 cells, whereas kgf and igf-1 expression was reduced (Fig. 5A). The increased expression of hgf was confirmed by quantitative real-time PCR. Higher levels of HGF protein were also detected by Western blot analysis of cell lysates (Fig. 5C). These results suggested that the promotion of CFE and cell proliferation by CLU-transfected cells is partially mediated by the secretion of HGF by feeder cells. 

Figure 5.

 

Growth factor expression by Clu-3T3 cells and CFE of HLE cells. (A) Total RNA isolated from Clu-3T3 and mock-3T3 cells were subjected to RT-PCR analysis of mRNAs for hgf, kgf, egf, fgf2, igf1, and gapdh (internal standard). Most growth factors were downregulated in Clu-3T3 cells, excluding hgf. (B) Higher expression of hgf mRNA by Clu-3T3 cells was confirmed by quantitative real-time PCR. The data were corrected by gapdh expression (n = 3; mean ± SD *P < 0.05). (C) Western blot of cell lysates also shows the secretion of HGF protein by Clu-3T3 cells. (D) Anti–HGF neutralizing antibodies suppressed the CFE of HLE cells cocultured with Clu-3T3 cells (n = 3; mean ± SD; *P < 0.05).

Discussion

Homeostasis and wound healing of the corneal epithelium are orchestrated by multiple factors, including growth factors, that modulate cell migration and proliferation, cell death, and stem cell maintenance. ^24,25 Among several growth factors identified, HGF is predominantly expressed by fibroblasts and stimulates corneal epithelial cell proliferation through the activation of their cognate receptors. ^26,27 HGF has mitogenic and morphogenic activities in various cell types, ^28,29 and its receptor, c-Met, is highly expressed on corneal epithelial cells. ^30,31 In this study, we demonstrated that CLU-transfected 3T3 cells can upregulate HGF, which then enhances the proliferation of corneal/limbal epithelial cells. Because CLU is expressed predominantly by corneal epithelial cells in vivo, ^11,12 our data suggest that CLU and HGF may be paracrine components of epithelial-mesenchymal interaction during epithelial wound healing. 

We also considered the possibility of enhanced cellular motility using a scratch migration assay, but we found no difference in results using 3T3 and CLU-3T3 supernatants (data not shown). Other factors, such as inhibition of apoptosis by CLU, ³² or HGF ³³ may also be involved in the increased CFE observed in our experiments. Several reports have demonstrated conflicting roles of CLU in cell proliferation. CLU inhibited the epidermal growth factor-stimulated proliferation of prostate cancer cells, ³⁴ and transient overexpression of CLU also decreased proliferative activity in SV40-immortalized prostate epithelial cells. ³⁵ On the other hand, the overexpression of secretory CLU resulted in significant enhancement of cell proliferation in both MIN6 insulinoma cells and primary pancreatic duct cells. ^13,36 In this study, we found that several different concentrations (0.1–3.0 μg/mL) of mouse recombinant CLU had no direct effects on the CFE or cell proliferation of TKE2 cells (data not shown). CLU promoted cell proliferation indirectly through HGF produced by Clu-3T3. Therefore, CLU seems to have different functions depending on cell type, especially when paracrine interaction with other cells is involved. The indirect effect shown in this study may explain the different responses of cells against CLU. 

sCLU has been reported to be a ligand of the LRP-2/megalin receptors. ³⁷ However, we did not detect LRP-2 receptor transcripts in the TKE2 and 3T3 cells (data not shown), despite the suppressive effect of CFE using CLU neutralizing antibody in the coculture experiment (Fig. 4). CLU has been shown to bind with a wide range of soluble ligands, ^4,38 acting as a chaperone for stressed proteins. ³⁹ Therefore, the upregulation of HGF by sCLU may be triggered by some other mechanism that does not involve the LRP-2/megalin receptors. Further studies are required to elucidate this point. 

Our results do not explain all the effects of CLU on epithelial cell proliferation because treatment with anti-HGF neutralizing antibody did not completely inhibit the CFE of TKE2 cells. This raises the possibility that CLU may modulate the expression of the other cell adhesion molecules and soluble growth factors in addition to HGF. CLU may also play a role in cellular protection against environmental stress, which may explain the high levels of CLU expressed in the cornea, one of the tissues most heavily exposed to oxygen and ultraviolet radiation. Additional studies are needed to reveal the full scope of CLU function in corneal homeostasis. 

Footnotes

 Supported by a Grant-in-Aid for Scientific Research (H18-tissue engineering-young-002) from the Ministry of Health and Welfare, Japan.

Footnotes

 Disclosure: N. Okada, None; T. Kawakita, None; K. Mishima, None; I. Saito, None; H. Miyashita, None; S. Yoshida, None; S. Shimmura, None; K. Tsubota, None

Footnotes

 Presented in part at the annual meeting of Japan Corneal Conference, Uruyasu, Japan, February 28 to March 1, 2008.