Pigment epithelium‐derived factor peptide promotes limbal stem cell proliferation through hedgehog pathway

Abstract Expansion of limbal epithelial stem cells (LSCs) is crucial for the success of limbal transplantation. Previous studies showed that pigment epithelium‐derived peptide (PEDF) short peptide 44‐mer could effectively expand LSCs and maintain them in a stem‐cell state, but the mechanism remained unclear. In the current study, we found that pharmacological inhibition of Sonic Hedgehog (SHh) activity reduced the LSC holoclone number and suppressed LSC proliferation in response to 44‐mer. In mice subjected to focal limbal injury, 44‐mer facilitated the restoration of the LSC population in damaged limbus, and such effect was impeded by the SHh or ATGL (a PEDF receptor) inhibitor. Furthermore, we showed that 44‐mer increased nuclear translocation of Gli1 and Gli3 in LSCs. Knockdown of Gli1 or Gli3 suppressed the ability of 44‐mer to induce cyclin D1 expression and LSC proliferation. In addition, ATGL inhibitor suppressed the 44‐mer‐induced phosphorylation of STAT3 at Tyr705 in LSC. Both inhibitors for ATGL and STAT3 attenuated 44‐mer‐induced SHh activation and LSC proliferation. In conclusion, our data demonstrate that SHh‐Gli pathway driven by ATGL/STAT3 signalling accounts for the 44‐mer‐mediated LSC proliferation.

Pigment epithelium-derived factor (PEDF) is a 50-kDa secreted glycoprotein with multiple biologic effects on various types of cells. 12 The amino acid positions Val78-Thr121 of human PEDF (termed 44mer) is responsible for neurotrophic and mitogenic activity. [13][14][15] 44mer has been proven to be able to promote LSC proliferation and meanwhile maintain a stem-cell state. 14 Further in vivo studies have confirmed that the 44-mer effectively repopulates LSCs in damaged limbus in rabbits. 16,17 These encouraging results suggest the potential role of 44-mer in treating LSCD or increasing the survival of limbal grafts. However, the mechanism of 44-mer-mediated LSC self-renewal has yet to be elucidated.
Sonic Hedgehog (SHh) signalling pathway is critical for maintaining and supporting stem cell properties in various tissues. [18][19][20][21] Secreted SHh ligands act on responding cells by turning on an intracellular signalling pathway that induces Gli transcription factors 18 . The active stage of the SHh pathway is tightly controlled and regulated by the integrating pathways to keep cellular processes balanced. Corneal wounding induced transient up-regulation of SHh ligands and expression of Gli3 in the limbus, suggesting that LSC behaviour is regulated by the SHh pathway. 22 In this study, we investigated whether SHh signalling pathway could be induced by 44-mer and promote the expansion of LSCs.
We found that 44-mer-induced Gli1 and Gli3 expressions to promote LSC proliferation, and such effect was suppressed by ATGL and STAT3 inhibitor.

| Animals
New Zealand albino rabbits (3.0-3.5 kg, 6 months old) and 3-5month-old Balb/c mice were used. All procedures were approved by the Mackay Memorial Hospital Review Board for animal investigation and were conducted in accordance with the ARVO statement for the Use of Animals in Ophthalmic and Vision Research.

| Limbal stem cell culture
LSCs were isolated from rabbits and used for cell-suspension culture, colony-forming efficiency (CFE) and BrdU labelling assay as described previously. 14 Briefly, the limbal rings were washed in phosphate-buffered saline containing 50 μg/mL gentamicin. After the iris and excessive sclera were removed, the limbal rings were exposed to dispase II (1.2 IU/mL in Hanks' balanced salt solution free of Mg 2+ and Ca 2+ ) at 4°C for 16 hours. The loosened epithelial sheet was harvested with a cell scraper and separated into single cells by treating with 0.5 mL trypsin (0.25% and 0.01% EDTA) for 15 minutes at 37°C with gental shaking. Cells were transferred to 9 mL of 10% FBS/DMEM/F-12 medium and were then collected by centrifugation (400 g for 5 minutes). LSCs were cocultured with MMC-treated NIH-3T3 fibroblast feeder cells located within the transwell (0.4 μm pore, BD Biosciences, Bedford, MA). For passage, near confluent cells were harvested with 0.25% trypsin and then 1 10 5 subcultured cells were cultured in the respective medium described above.

| Colony-forming efficiency
Approximately 1 x 10 3 LSCs were seeded in a 3.8-mm 2 dish and cocultured with MMC-treated NIH-3T3 feeder cells located within the transwell. The medium was changed every 2-3 days. At 10 days, colonies were fixed by 4% paraformaldehyde (room temperature for 1 hour) for immunostaining and crystal violet staining. The CFE (%) was calculated using the following formula: number of colonies formed/number of cells plated × 100%.

| Western blotting
Cell lysis and SDS PAGE were performed as described previously. 23 NE-PER nuclear and cytoplasmic extraction kit (Pierce, Rockford, IL) was used to separate total cell lysate into cytoplasmic and nuclear fractions according to the manufacturer's instructions. Each cellular fraction was then resolved by SDS-polyacrylamide gel electrophoresis, electrotransferred to polyvinylidene difluoride membranes (Millipore, Bedford, MA), and processed for immunoblot analysis.
Proteins of interest were detected using the appropriate IgG-HRP secondary antibody and ECL reagent. X-ray films were scanned on a Model GS-700 imaging densitometer (Bio-Rad Laboratories, Hercules, CA) and analysed with Labworks 4.0 software. Blots from at least three independent experiments were used for quantification.

| Quantitative real-time PCR
Experiments were performed as described previously. 23 The sequences of the PCR primers were rabbit Gli1 (accession number:

| siRNA transfection
LSCs were transfected with 10 nmol/L of siRNA targeting Gli1 or Gli3 genes (listed in

| Partial limbal injury
Mice were anaesthetized by intraperitoneal injection of a mixture of zoletil (6 mg/kg) and xylazine (3 mg/kg). One drop of 0.5% proparacaine hydrochloride (Alcaine; Alcon,Fort Worth, TX) was given before ocular procedures. The epithelium of the inferior 120 degree limbus was removed with a 0.5 mm metal burr (Rumex international Co, Clearwater, FL), 24 0.75 mm into the cornea and 0.75 mm into the conjunctiva in the experimental eye.

| Organ culture
To evaluate the expressions of Gli1 and Gli3 in the nucleus of remaining LSCs after partial limbal injury, murine eyes were enucleated and placed in a 24-well culture plate containing 2 mL LSC culture medium 14 supplemented with 10 µmol/L 44-mer for 2 hours at 37°C.
LSC culture medium supplemented with DMSO served as the control.
Each globe was fixed in 4% paraformaldehyde in 0.1 mol/L phosphate buffer (PH 7.4) for 48 hours at 4°C and then embedded in paraffin.

| Statistical analysis
Results were presented as mean ± SD. The statistic significances of the experimental results were assessed by Student's t test using . A two-tailed P < 0.05 was considered statistically significant. Statistical significance was as follows: P < 0.05 (*), P < 0.01 (**), and P < 0.001 (***).

| SHh signalling promotes LSC proliferation
We first examined the influence of SHh signalling on LSC clonogenicity by exposing LSCs to SHh inhibitors (HPI4 or cyclopamine) for 10 days. The clonogenicity of LSC was significantly suppressed by SHh inhibitors compared to control ( Figure 1A). In addition the BrdU proliferation assay showed that HPI4 and cyclopamine significantly reduced the proliferation of ∆Np63α-positive LSCs compared to control (17.8 ± 2.7% and 21.3 ± 3.2% vs 80.5 ± 7.1%), suggesting that suppression of SHh signalling impaired LSC proliferation ( Figure 1B). To confirm our findings, LSCs were further treated with recombinant human SHh at various concentrations (50, 100 and 200 ng/mL) for 24 hours before 2 hours of BrdU labelling. SHh was found to increase LSC proliferation (1.8-4.0-fold; Figure 1C) and cyclin D1 expression (1.4-2.6-fold; Figure 1D) in a dose-dependent manner.  Figure 3D). HPI4 injection did not show any severe adverse effect, such as weight loss ( Figure 3E). These findings suggest that the 44mer enhances restoration of LSCs in the damaged limbus through SHh signalling pathway.

| Knockdown of Gli3 impairs the mitogenic activity of 44-mer on LSCs
To examine whether Gli1 and Gli3 are required for the 44-mer-mediated LSC proliferation, we used siRNA to silence the two genes.

| ATGL activates STAT3 signalling to promote Gli expression
Given that 44-mer bound to its receptor ATGL to initiates signal transduction, 23  and Gli3 proteins were suppressed to near basal levels ( Figure 6C).

PEDF has been shown to induce phosphorylation of STAT3 in
LSCs, and PEDF-ATGL signalling is relied on STAT3 signalling. 14,23 To determine whether the effect of 44-mer on SHh activation is dependent on ATGL/STAT3 signalling, pharmacological inhibitors for ATGL and STAT3 were evaluated. Our results showed that Atglistatin suppressed the phosphorylation of STAT3 at Tyr705 in response to 44-mer ( Figure 7A). With the 44-mer treatment for 3 hours, the mRNA levels of Gli1 and Gli3 were significantly up-regulated by 2.5-and 2.4-fold, respectively, compared to control ( Figure 7B). STAT3 inhibitors suppressed the 44-mer-induced Gli1 and Gli3 mRNA and protein expressions to near basal levels ( Figure 7B,C). Collectively, pharmacological inhibition of ATGL/ STAT3 signalling attenuated the ability of the 44-mer to up-regulate Gli1 and Gli3 expressions.

| D ISCUSS I ON
Damage to the human corneal limbus may lead to permanent dysfunction of the stem cells. Currently, there is no available therapeutics to slow down the progress of LSCD. We have previously demonstrated that LSCs can be expanded by 44-mer both in vitro and in vivo. 14,16,17 In this present study, we have further shown that ATGL-STAT3-SHh signalling pathway is essentially involved in 44mer-mediated LSC expansion. Myc, Bcl2, Bmi1 and Snail, which further regulate cell behaviour. 18,26 The role of SHh in the LSC population has remained unknown until our current study showing that intrinsic SHh is crucial for LSC self-renewal. In addition, 44-mer behaves as an activation signal to interact with SHh/Gli signalling, resulting in modulation of Gli protein expression to promote LSC proliferation. Our study proved that 44-mer up-regulates Gli1 and Gli3 in LSCs. We also found that knockdown of Gli3 reduced Gli1 mRNA and protein levels in 44-mer-treated LSCs, suggesting that Gli1 gene is probably a direct transcriptional target of Gli3. 27 Other signalling molecules, which were reported to be positive regulators of Gli function, include EGF, PDGF, FGF and IGF. 18,21,28 Although EGF, and FGF promote LSC proliferation, 29 the interaction between EGF, FGF and SHh-Gli in LSCs is unknown.
STAT3 is a cytoplasmic protein with Src Homology-2 domains that act as signal messengers and transcription factors, participating in cellular responses to cytokines and growth factors. PEDF was found to cause a striking activation of STAT3 in myoblasts, hepatocytes and LSCs; pharmacological inhibition of STAT3 blocks PEDF function in these cells. 14,23,30 Accordingly, in the present study, we (C) LSCs were exposed to the 44mer for 6 h after pretreated with STAT3 inhibitor. Gli proteins were detected by Western blot analysis. Blotting of β-actin serves as a loading control found that SHh signalling activated by the 44-mer was regulated by STAT3 activity as well. These results suggest that cell signalling in response to 44-mer appears to rely essentially on STAT3 activity.
Other well-studied signalling systems dependent on STAT3 include G-CSF receptor signalling, HGF and IL-10. 31 Apart from promoting SHh signalling activation, the mechanism by which 44-mer increases LSC population may also involve the direct binding of STAT3 to the △Np63α promoter. 32 Crosstalk between SHh and STAT3 has been studied in gastric metaplasia, lung adenocarcinoma and skin tumours. All these studies suggest that the interaction between STAT3 and SHh was indirectly mediated through up-regulation of third-party factors, such as IL-6, IL1β, TNFα, IL-11 and TIF1. 26,33,34 The precise pathway by which the STAT3 and SHh pathway interacts in LSCs requires further studies.
Various membrane receptors for PEDF have been reported in different cells. [35][36][37] It was found that ATGL is a putative receptor for PEDF in rabbit corneal epithelial cells. 38 In our study, either direct inhibition of SMO by cyclopamine, 26 or inhibition of downstream of SMO by HPI4, 39 abolished the 44-mer-induced LSC expansion. A recent report indicated that SHh activates phospholipase A2 to release arachnoid acids, which promotes SMO ciliary accumulation and signalling. 40 Because PEDF can regulate triglyceride metabolism in hepatocytes through ATGL, 41 one may speculate that PEDF augments the function of SMO to enhance the expression of downstream Gli proteins.
The restoration of LSC population in the limbal wound promoted by the 44-mer is probably because of the migration of LSCs, either from lateral cell displacement or the central zone. 16 The target genes of the SHh pathway responsible for epithelial-mesenchymal transition 26 may contribute to LSC migration. On the other hand, a scenario of one-for-one replacement from neighbouring LSCs, stimulated by death of LSCs, 42 is another possible mechanism in which symmetrical division could be promoted by the 44-mer. Our study progresses the LSC biology field by providing clues on the molecular programs that orchestrate LSC activity. Exogenous cues, such as PEDF, which modify the signals that determine stem-cell fate decisions, have the translational potential for the treatment of LSCD.

ACK N OWLED G EM ENTS
We thank Chu-Ping Ho for assistance with animal experiments, Dr.
Erh-Hsuan Lin for technique support of immunostaining, and Dr. Tim J Harrison and Dr. Yihe Chen for polishing the English in this article.

DATA AVA I L A B I L I T Y S TAT E M E N T
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.