Pigment Epithelial-Derived Factor Peptide Regenerated Limbus Serves as Regeneration Source for Limbal Regeneration in Rabbit Limbal Deficiency

Abstract

Purpose: To demonstrate that a 44-amino acid peptide from pigment epithelial-derived factor (PEDF) induces the regeneration of limbal excision wound, and the regenerated limbus can act as the regeneration source for new limbal excisional injuries in rabbit model of limbal deficiency.

Methods: Half circumference partial limbal excision was followed by PEDF peptide treatment to achieve limbal wound regeneration. Three months later, a second stage half circumference partial limbal excision removed the remaining native limbal tissue followed by PEDF peptide treatment. The structure and function of the regenerated limbus were analyzed at 3 and 6 months. Conjunctivalization was analyzed by impression cytology. Immunohistochemical analysis was performed with antibodies to corneal epithelium-associated keratin 3 (K3), conjunctival epithelium-associated keratin 13 (K13), ΔNp63α, ABCG2, and BrdU. Extensive limbal excision was performed to examine the regeneration potential of the PEDF peptide.

Results: Total limbal stem cell deficiency occurred with severe inflammation and conjunctivalization of the limbal wound and adjacent cornea in vehicle control eyes. In PEDF peptide treated eyes, the regenerated limbus prevented fibrovascular invasion and goblet cell migration into the corneal surface. Immunohistochemical staining of the regenerated limbus showed a wide distribution of cells expressing ΔNp63α and ABCG2 as in the native limbus. BrdU labeling assay revealed the presence of slow-cycling cells in the basal layer of the regenerated limbus. The PEDF peptide can heal extensive limbal excisional wounds and sustain ocular surface integrity.

Conclusions: The addition of PEDF peptide has the potential to repair limbal excisional wounds with the recovery of normal limbus-like anatomy and function. The PEDF peptide is a potential remedy for extensive limbal injury.

The maintenance of a healthy functional corneal epithelium is provided by a unique subpopulation of stem cells (SCs) located in the basal epithelial layer of the limbus.¹ These limbal epithelial stem cells (LSCs) cycle slowly to generate progenitors that support corneal epithelial turnover and respond to injury.^1,2 Limbal damage deprives the cornea of SC source and leads to limbal stem cell deficiency (LSCD). Clinically, LSCD manifests as ocular surface insufficiency, characterized by recurrent or persistent corneal epithelial defects, conjunctivalization, corneal neovascularization, chronic stromal inflammation, scarring and, ultimately, loss of corneal transparency.³ 

Current therapeutic efforts adopted in the treatment of LSCD have been focusing on replacement of deficient LSCs by transplantation. Limbal tissue autografts, allografts, and cell transplantation (a limbal equivalent followed by ex vivo expansion of LSCs) have been used extensively for clinical therapy with wide success.^4–9 In particular, cell-based therapy has emerged as a new trend for the treatment of LSCD for its advantages compared to tissue transplantation: requirement for minimal amount of donor tissue, ease and safety of surgical techniques, reduced rejection rate, and unnecessary systemic immunosuppression.^5,7,9–12 However, cell-based therapy is costly, requires highly specialized and equipped facility with skilled personnel thus limiting its application outside medical centers. Moreover, the isolation and expansion of LSC is difficult. A major technique limitation to this technology is a lack of consensus about a universal method to isolate bona fide limbal epithelial cells at large or fast enough to avoid treatment delays. 

Recent reports have indicated that there are LSC remains on the ocular surface of some of the patients with LSCD. The inadequate number of LSC and suboptimal ocular surface general condition prevent these surviving LSC from maintaining healthy ocular surface. Amniotic membrane (AM) alone has been reported to reconstruct ocular surface possibly by improving the general condition of the ocular surface.^13–17 

Pigment epithelial-derived factor (PEDF) is a highly conserved 50-kDa secreted glycoprotein that is encoded by a single gene and is widely expressed in a variety of cell types with distinct biological effects. Although PEDF is known commonly as a strong antiangiogenic factor, recent studies have accumulated evidence linking it with SC biology. The effects of PEDF in a variety of SCs is mainly in supporting SC survival and maintaining multipotency.^18–22 The 44 amino acid peptide (positions Val78-Thr121) of human PEDF (PEDF peptide, 44-mer) determines its neurotrophic and mitogenic effect on adult neural progenitor cells.^22,23 Recently, the PEDF peptide was demonstrated to increase the numbers of SC related holoclones in limbal progenitor cell culture.²⁴ Moreover, limbal progenitor cell proliferation in response to corneal epithelial wounding was intensely activated by PEDF peptide. In another study, we induced partial limbal deficiency by creating a full thickness half circumferential limbal excision and observed limbal regeneration facilitated by PEDF peptide. The regenerated limbus prevented fibrovascular invasion of the cornea and sustained corneal epithelium integrity even after repeated large epithelial debridement.²⁵ However, since only half of the limbus was excised and regenerated, whether cells from the regenerated limbus is capable of sustaining the ocular surface remains unclear. A definitive proof that the regenerated limbus can support corneal surface cell turnover is the generation of a condition that all the limbus are regenerated limbus. In order to prove that the regenerated limbus can be the sole source of progenitor/SCs for limbal regeneration induced by PEDF peptide, we deprived the study eye of its native limbal tissue by two consecutive half circumference limbal excision followed by PEDF peptide treatment. 

Material and Methods

Chemicals and Antibodies

Dimethyl sulfoxide (DMSO), 5-bromo-2′-deoxyuridine (BrdU), Triton X-100, Hoechst 33258 dye, n-heptanol, and formalin were all from Sigma-Aldrich Corp. (St. Louis, MO, USA). ΔNp63α polyclonal antibody and all the fluorescent dye-conjugated secondary antibodies were purchased from BioLegend (San Diego, CA, USA). Mouse anti-keractin-3 (K3; clone AE5; CBL218) was purchased from Millipore Corporation (Bedford, MA, USA). Mouse anti-keractin-13 (K13; clone AE8: sc-57003) was purchased from Santa Cruz Biotechnology (Heidelberg, Germany). Rat monoclonal anti-ABCG2 antibody (ab24115) was from Abcam (Cambridge, MA, USA). Mouse anti-BrdU antibody (GTX42641) and vimentin polyclonal antibody (GTX100619) were from GeneTex (San Antonio, TX, USA). 

Treatment Preparation

The 44 amino acid peptide of PEDF (Val78-Thr121) was synthesized by GenScript (Piscataway, NJ, USA). The PEDF peptide was reconstituted in DMSO as stock (5 mM) and mixed with 0.2 mg TOBREX eye ointment (Alcon, Fort Worth, TX, USA; containing 0.3% Tobramycin and 0.5% Chlorobutanol) to a concentration 100 μM. Vehicle-containing ointment was prepared by mixing 1 μL DMSO in 0.2 mg TOBREX eye ointment. 

Animals

Twenty-four New Zealand albino rabbits (3.0−3.5 kg, 6 months of age) were used in this study. All animals were handled in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and institutional guidelines for animal investigation. All procedures were performed with animals under general anesthesia induced by intramuscular injection of ketamine hydrochloride (35 mg/kg) and xylazine hydrochloride (5 mg/kg) and topical anesthesia with 0.5% proparacaine hydrochloride (Alcaine; Alcon). The right eye of each animal was surgically injured in the same manner by a single investigator, and the left eye was left untouched. 

Two-Stage Partial Limbal Deficiency Animal Model

Two halves circumferences (180-degree) limbal excision was performed in two stages. After the first half circumference partial limbal excision, a second half circumference partial limbal excision was carried out 3 months later, depriving the rabbit eye of its native limbal tissue. The limbal excision was created surgically by lamellar dissection of the limbal zone, 2 mm into the cornea, 3 mm into the conjunctiva, and 100 to 120 μm in depth.^26,27 Debridement of the corneal epithelium was performed by application of n-heptanol soaking for 30 seconds followed by mechanical scraping.^26,28 Amniotic membrane as a substrate was grafted over the limbal excisional wound and sutured with interrupted 10-0 nylon sutures.²⁵ Also, 10-0 Nylon corneal sutures were used to mark the starting and the ending edges of each half circumference limbal excision to ensure complete removal of the native limbus at the end of the two-staged partial limbal excisions. 

After surgery, animals were sorted randomly into the PEDF peptide treatment group (n = 8) or sham-operated vehicle control group (n = 4). Two rabbits that neither were surgically wounded nor received any treatment were used as normal controls. Immediately after each limbal excision, the wounded eye was treated with PEDF peptide-containing ointment or vehicle-containing ointment once a day for 2 weeks. The follow-up period was 6 months. In order to detect label retaining cells (LRCs) in the regenerated limbus, animals were pulse labeled intraperitoneally with BrdU (Sigma-Aldrich Corp., Steinheim, Germany; 50 mg/kg body weight in PBS) once a week for 2 weeks immediately after the second half circumference partial limbal excision. Two rabbits were pulse labeled with BrdU immediately after the first half circumference partial limbal excision to detect LRC in the native limbus as positive controls. In all animals, 3 months after BrdU labelling, limbal tissue was harvested for identification of BrdU-labeled cells with anti-BrdU antibody (GeneTex). 

Extensive Partial Limbal Deficiency Animal Model

Two hundred seventy-degree limbal excision was carried out by performing two interrupted 135-degree limbal excision. The limbal excision, corneal epithelium debridement, and AM graft were performed as described above. Animals were sorted randomly into the PEDF peptide treatment group (n = 4) or sham-operated vehicle control group (n = 4). Two rabbits were used as normal controls. Immediately after the surgery, the wounded eye was treated with PEDF peptide-containing or vehicle-containing ointment once a day for 2 weeks. The follow-up period was 3 months. 

Immunofluorescence

Deparaffinized tissue sections or 4% paraformaldehyde-fixed LSCs were blocked with 10% goat serum and 5% BSA in PBS containing 0.1% Tween-20 for 1 hour. Staining was performed using primary antibodies against ΔNp63α (1:150 dilution; BioLegend), K3 (1:250 dilution; Millipore Corporation), K13 (1:250 dilution; Santa Cruz Biotechnology), ABCG2 (1:100 dilution; Abcam), vimentin (1:100 dilution; GeneTex), and BrdU (1:100 dilution; GeneTex) at 37°C for 2 hours, followed by incubation with the appropriate rhodamine- or FITC-conjugated donkey IgG (1:500 dilution) for 1 hour at room temperature. Images were captured using a Zeiss epifluorescence microscope with a CCD camera and photographs taken using the Zeiss Axiovision version 3.1 software (Carl Zeiss MicroImaging GmbH, Germany). 

Statistical Analysis

Statistical analyses were performed using GraphPad Software (San Diego, CA, USA). The results are expressed as mean ± SEM of (n) independent experiments. Comparisons were analyzed using the Student's t-test, Wilcoxon nonparametric paired test, and 1-way analysis of variance (ANOVA). Differences are considered significant when P < 0.05. 

Results

Ocular Surface Can Be Sustained by an Entirely Regenerated Limbus

Recently, we reported the regeneration of limbal tissue on a half circumference limbal excision wound with the assist of the PEDF peptide. Progenitor cells were generated from such regenerated limbus to repopulate ocular surface. However, it is unclear if regenerated limbus can sustain the whole ocular surface in the absence of native limbal tissue. To generate an eye with only regenerative limbal tissue, a two-staged half circumference limbal excision was performed as described in Methods to remove the native limbal tissue (Fig. 1). Surgical results indicated that in vehicle control eyes (n = 4), total LSCD occurred with severe inflammation and 360-degree conjunctivalization of the limbal wound and adjacent cornea. Repeated corneal epithelial debridement led to a persistent epithelial defect and corneal opacity. In eyes treated with the PEDF peptide after each half circumference limbal excision (n = 6), limbus was regenerated over the 360-degrees of the limbal excisional wound without conjunctivalization. The PEDF peptide treated limbus withstood repeated extensive scrapping of the corneal epithelium, performed monthly for 6 months (Fig. 1). By removing the remaining native limbus, we demonstrated that regenerated limbal tissue is capable of maintaining the ocular surface, even after repeated scrapping of the epithelium. 

Figure 1

 

The PEDF peptide (44-mer) regenerated limbus supplies epithelial cells for sustenance of the ocular surface. Diagram of the strategy to achieve total limbal regeneration by two-staged total native limbal excision and evidence of the long-term maintenance of corneal epithelium. Left panels: White and black dotted lines demarcate the areas corresponding to the first and second limbal regeneration, respectively, at 6-months follow-up. Three months after the first 180-degree limbal excision, the second-step removal of the remaining half intact native limbus was performed and left to heal for another 3 months. Treatment with the 44-mer or vehicle containing ointment was performed immediately after each limbal excision, once a day for 2 weeks. Limbal function was monitored by photography of limbal vascular invasion after the first limbal excision (asterisks) and by corneal epithelial wound healing stained with topical fluorescein after the second limbal excision. Representative images are from three independent experiments.

Structure Characterization of the Limbal Tissue Regenerated Secondarily From a Regenerated Limbus

With the assistance of the PEDF peptide, the two-staged half circumference limbal excision surgery generates a secondary regenerated limbus using primary regenerated limbus as a source. The secondary regenerated limbus is functional as evident from the prevention of corneal conjunctivalization (Fig. 1). Structurally, the regenerated tissue obtained from both first (primary) and second (secondary) half circumference limbal excision displayed multilayered epithelium and contained a regular, packed basal layer and the same thickness of stratified epithelium as a native limbus (Fig. 2A). Notably, goblet cells (Fig. 2A, indicated by arrows) were restricted to the conjunctival epithelium and not on the corneal epithelium, indicating the prevention of conjunctival invasion. Immunostaining of the healing limbal tissue revealed that both regenerated tissues contained cells positively stained for ΔNp63α and ABCG2 at roughly the same density and pattern as the limbus of normal controls, indicating the presence of limbal progenitor cells in the regenerated limbus. The regenerated limbus contained some K3 positive keratinocytes in the superficial layer that was not observed in the normal controls. Keratin 13 immunostaining was used as negative control for conjunctival cells (Fig. 2B). 

Figure 2

 

The 44-mer-induced regenerative epithelium sustains the integrity of limbal-like epithelium after two-staged total native limbal excision. (A) Representative photomicrographs of H&E stained regenerated limbal-like epithelium at 6 months after the first half circumference limbal excision (first limbal regeneration) and at 3 months after the second half circumference limbal excision (second limbal regeneration). Arrows denote goblet cells. Representative results from three separate experiments are shown. Original magnification: ×200. (B) Immunofluorescence analysis of the distribution of ΔNp63α-positive/K3-negative LSCs, ABCG2-positive LSCs, corneal-like K3-positive, and conjunctival epithelium-associated K13-positive cells in the normal control and the regenerative limbal tissue. Nuclei were visualized with Hoechst 33258 staining. Original magnification: ×400.

These structure analyses showed that the secondarily regenerated limbus that is generated in the presence of only primarily regenerated limbus have similar structure as primarily regenerated limbus and native limbus. These data indicated that the limbal regenerating potential of the primarily regenerated limbus is similar to the native limbal tissue. 

The PEDF Peptide-Regenerated Limbus Contains Slow Cycling Cells

The slow cycling cells, detected by BrdU label retention, are presumed to be dormant SCs. Slow cycling cells were identified in the limbus.^1,2,29 We reasoned that during limbal regeneration, some of the SCs moved into the excision wound would become dormant. In vivo BrdU labelling was performed to identify slow cycling cell in the regenerated limbus. Pigment epithelial-derived factor peptide (n = 6) or vehicle (n = 4) treatment and BrdU pulse-labeling were performed immediately after 180-degree limbal excision and chased for 3 months as described in Methods. BrdU-labeled cells could indeed be detected in the primarily regenerated limbus of experimental eyes (Fig. 3). Although the density of labeled cells in the regenerated limbus is slightly lower than in the native limbus (20.5 ± 4.5% vs. 26.0 ± 2.7%), the existence of slow cycling cell in the regenerated limbus provided another similarity to native limbal tissue (Figs. 3A and 3B). 

Figure 3

 

Identification of LRCs in the native (A) and regenerated limbus (B). Diagrams summarize the limbal excision area and the location of tissue sampling for immunostaining studies. Ocular surface sections were double-stained with BrdU to identify LRCs (red) and ΔNp63α to identify limbal epithelium (green). BrdU-positive cells, identified as nuclei stained blue with Hoechst 33258, were counted using the Image-Pro Plus 4.5.1 computer program. Percentages of BrdU-positive LRCs from representative photographs are shown (n = 4). Right panels show the study design of BrdU pulse chase of the 44-mer induced limbal regeneration model in the rabbit.

We investigated the effect of 44-mer on the structure of the regenerated limbus 14 days after first and second half circumference limbal excision (Fig. 4). Hematoxylin and eosin (H&E) stain of the limbal wound revealed that the limbal epithelium was completely removed during the surgery. A layer of epithelium, underneath the AM, was regenerated at day 14 after limbal injury (Fig. 4A). Immunofluorescence analysis indicated that ΔNp63α and ABCG2-positive cells were identified on the regenerated limbal tissue treated with 44-mer, which were not observed in limbal wounds treated with vehicle ointment. Keratin 13-positive cells were observed only in the vehicle-treated wounds. Keratin 3-positive corneal cells were mainly distributed at the top layer of the regenerated limbal epithelium (Fig. 4B). Taken together, 14 days after half circumference limbal excision and 44-mer treatment, limbal progenitor cells were regenerated in the limbal excisional wound. 

Figure 4

 

Structural and immunofluorescence analysis of the regenerated limbus 14 days after first and second half circumference limbal excision. (A) H&E staining. Arrows denote healed limbal wounds at day 14. Original magnification: ×200. (B) Immunofluorescence analysis of LSC markers (ΔNp63α and ABCG2), corneal epithelium-associated K3, and conjunctival epithelium-associated K13. Data represent two independent experiments and two rabbits per group.

The PEDF Peptide Regenerates Healthy Limbus and Maintains Ocular Surface Integrity After Extensive Limbal Excision

The limbus has limited regeneration potential. Extensive limbal damage leads to ocular surface failure. This also limits the amount of donor limbal tissue that can be harvested without compromising the ocular surface. To test the limbal healing potential of the PEDF peptide, an extensive 270-degree limbal excision was carried out (Fig. 5). In vehicle controls (n = 4), prominent inflammation and fibrovascular membrane encroachment were observed over both the remaining intact limbus and the limbal excisional wound (indicated by black and white asterisks, respectively). Repeated debridement cause delayed wound healing and exacerbation of limbal and corneal vascularization were seen in vehicle control eyes. In pigment epithelial-derived factor peptide treated eyes (n = 4), the wounded area healed with recovery of normal limbal anatomy. Repeated debridement of the corneal epithelium resulted in a slight delay of epithelial recovery at day 3 and total recovery at day 7 after surgery in PEDF peptide treated eyes. Impression cytology confirmed ingrowth of goblet-cell containing epithelium in the limbal and peripheral corneal surface of vehicle-treated eyes. These results indicate that the PEDF peptide can regenerate the limbus and maintain ocular surface integrity after extensive limbal excision. 

Figure 5

 

The PEDF peptide (44-mer) regenerated limbus can sustain the ocular surface after extensive limbal excision. External photographs of the limbal area 2 months after 270-degree lamellar excision and 2 weeks after corneal epithelial wounding challenge. The 44-mer and vehicle treatment are described in Methods. Dotted lines demarcate the area of limbal resection. Black and white asterisks indicate the nonresection and resection areas, respectively. Corresponding high magnification photographs are presented. Corneal epithelial wounding challenges were performed and the eyes were stained with fluorescein. Impression cytology confirmed ingrowth of epithelium containing goblet-cells in the limbal and peripheral corneal surface of vehicle-treated eyes. Representative results from three separate experiments are shown.

Discussion

Structural lesions of the limbus do not heal spontaneously and represent a serious clinical challenge for the reconstruction of the ocular surface. In this study, we performed two partial limbal excisions to reproduce extensive limbal structural lesion in rabbits. In this animal model, we aim a partial removal of LSCs to test the capacity of PEDF peptide to regenerate the excised limbal zone. We demonstrated that the PEDF peptide treatment regenerated a limbus-like structure at the excised limbal zone with anatomical and functional similarities to neighboring normal limbus, which was consistent with our previous study.²⁵ However, in this study, we performed a second stage partial limbal excision of the remaining normal limbus, aiming the removal of the limbal zone that served as the native regeneration source for the excised limbal zone. Surprisingly, after the PEDF peptide treatment, the limbus-like structure at the recovered limbal zone acted as a regeneration source for the newly excised limbal zone with the recovery of a functional limbus. 

Slow-cell cycle and prolonged retaining of DNA label are still considered to be the two most reliable characteristics of LSC. The presence of BrdU-labelled slow cycling cells in the PEDF peptide regenerated limbus confirmed the existence of putative SCs at the recovered limbal zone. Since we excised the full thickness of the limbal epithelium and a layer of stroma underneath to exclude the possibility of limbal regeneration from residual LSCs in the wound,³⁰ we trust that the mechanisms driving inter quadrant repopulation of limbal progenitor/SCs induced by the PEDF peptide might involve enhancement of proliferation, migration, and plasticity of residual viable cells with regenerative properties. 

The PEDF and its 44-mer peptide have already been demonstrated to have mitogenic effect on adult limbal progenitor cells in vitro and in vivo.^24,25 We demonstrated that the PEDF peptide regenerated limbus contained limbal progenitor cells characterized by staining for molecular markers ΔNp63α and ABCG2.²⁵ In this study, we demonstrated a widespread presence of slow cycling LRCs that repopulated the excised limbal zones in the PEDF peptide treated eyes. Although the limbal excision was performed in a two stage manner, the large amount of limbal progenitor/SCs removed surgically was far from the capacity of migration or dedifferentiation of residual cells with regenerative properties. This result suggested that residual LSCs must have underwent both symmetrical and asymmetrical division to replenish the LSC population within the niche.^31,32 

The replenishment of excised limbal zones with slow cycling LRCs in the PEDF peptide treated eyes also indicated that migration of limbal progenitor/SCs occurred during the process of limbal regeneration. The transference of cells with regenerative properties from untreated to treated limbal zones occurred most probably by migration via the central cornea zone or via lateral cell displacement. Lateral cell displacement may be a function of increased cell motility/migration, whereas transference through the center of the cornea may be a function of enhanced proliferation. Moreover, transference of SCs may occur via the ability of early transient amplifying cells present in the proximity of the excised zone to migrate outward and recover their SC phenotype in situ by dedifferentiation³³ or conjunctival transdifferentiation.^28,34 Current advancements in genetic lineage tracing of stem and progenitor cells provided direct evidence for limbal/corneal progenitor cell survival and migration during corneal renewal under steady-state and wound healing conditions.³⁵ It holds great promise for progressing the field of LSC biology by providing clues as to the molecular programs that govern SC activity within the limbus and development of therapeutic opportunities directed at modifying signals that regulate LSC function in situ such as the PEDF peptide induced limbal regeneration demonstrated in our study. 

Significant advances have been made in recent years regarding the limbal niche that supports LSCs. Niche fibroblasts seem to provide soluble factors to sustain the SC phenotype of LSCs.^36,37 Recombination studies showed that limbal stromal cells govern the phenotype of LSCs.³⁸ Our study provides evidence that LSCs from the intact native limbus migrate into the excision wound and regenerate the limbus. In addition, we observed a rapid accumulation of vimentin-positive fibroblasts in the vicinity of the regenerating epithelium (unpublished data). It is unclear whether PEDF also activates and mobilizes stromal fibroblasts. Pigment epithelial-derived factor may temporarily assume the function of growth factor secreted by fibroblasts and support the epithelium during the establishment of the new niche. After lamellar limbal excision, fibroblasts in the remaining stroma may sustain LSCs migrating into the wound. The regeneration of healthy stromal fibroblasts remains a critical issue for the regeneration of the limbus. More extended studies and further structural analysis of the regenerated limbus with more specific markers are needed to provide proofs of the existence of LSCs. The antiangiogenic effect of the PEDF peptide may also play a role in the regeneration of the excised limbus, providing the space and time for the recovery of the limbus before neovascularization fosters conjunctival colonization and the pathology becomes unsurmountable. 

The capacity of the PEDF peptide to promote limbal regeneration brings new insights in regenerative medicine for ocular surface reconstruction. Current therapy for LSCD have shifted from tissue transplantation to cell therapy. However, it still focuses on the transplantation of LSCs as mainstay therapeutic modality. We pursued a therapeutic modality for LSCD by ex vivo and/or in vivo expansion of viable LSCs. As long as there is viable residual LSCs, in vivo expansion of LSCs can be used in a variety of ocular surface diseases, that includes not only diseases associated with LSC deficiency but also when there is LSC dysfunction or distress such as chronic inflammatory/autoimmune diseases (e.g., Stevens Johnson Syndrome, cicatricial pemphigoid, radiation, and iatrogenic). Extensive damage to the limbus such as in chemical and thermal burns will have enhanced success rate of limbal tissue and cell-based transplantation, and accelerate corneal wound healing. Moreover, PEDF peptide can be used to decrease the need for surgical treatment and/or recurrence rate of pinguecula and pterygium by replenishing focal areas of LSCD. 

In summary, we demonstrated that the PEDF peptide promoted limbal regeneration in rabbit model of limbal deficiency. The regenerated limbus showed anatomical and functional features of a healthy limbus. It contained limbal progenitor cells characterized by staining for markers ΔNp63α and ABCG2. It provided corneal epithelial cells during corneal wound healing. It sustained the capacity of limbal barrier function by blocking conjunctivalization and corneal vascularization when challenged with serial corneal epithelial debridement. Moreover, the PEDF peptide regenerated limbus served as a niche of LSCs demonstrated by its replenishment with slow cycling LRCs after two-stage limbal excision with total excision of native limbus. The regenerated limbus could not only maintain a healthy ocular surface but also functioned as a source of LSCs with capacity of regenerating a new limbus at the site of limbal damage. To our knowledge, this is the first report that implies the regeneration of SC niche by a growth/trophic factor after wounding in the eye. It may be an innovative concept in regenerative medicine and, in the future, the possibility to assist or replace LSC transplantation inferred from our observation may also benefit patient therapy. 

Acknowledgments

The authors thank Chu-Ping Ho and Chin-Min Wang for assistance with animal experiments, and Ju-Yun Wu for assistance with cell culture experiments. 

Supported by grants from the National Science Council, Taiwan (NSC 101-2314-B-195-006-MY3) and Mackay Memorial Hospital (MMH-E-101-006). The authors alone are responsible for the content and writing of the paper. 

Disclosure: S.-I. Yeh, None; T.-C. Ho, None; S.-L. Chen, None; C.-P. Chen, None; H.-C. Cheng, None; Y.-W. Lan, None; J.-W. Hsieh, None; C.-T. Wang, None; Y.-P. Tsao, None