Cell Delivery with Fixed Amniotic Membrane Reconstructs Corneal Epithelium in Rabbits with Limbal Stem Cell Deficiency

Abstract

Purpose.: To explore the feasibility and efficacy of a cell delivery system using amniotic membrane (AM) fixed by a novel biomembrane-fixing device (BMFD) for corneal epithelium reconstruction in rabbits with limbal stem cell deficiency (LSCD).

Methods.: Sixty female rabbits with LSCD were created and randomly assigned to three groups of 20 each: LSCD rabbits without treatment (the control), LSCD rabbits treated with BMFD-fixed AM (BMFD-AM), and rabbits treated with male human limbal epithelial cells delivered with BMFD-fixed AM (BMFD-AM+cells). They were followed up with slit lamp observation and corneal fluorescein staining for 14 days. Cytokeratin K3 and K4 and mucin 5AC were used to evaluate corneal conjunctivalization. Sry gene detection was used to trace the delivered cells.

Results.: The corneal re-epithelialization time was 5.60 ± 0.46 days in the BMFD-AM+cell group, significantly shorter (P < 0.05) than in the LSCD (12.45 ± 0.65 days) and the BMFD-AM (9.25 ± 0.51 days) groups. Conjunctivalization and neovascularization were observed to be severe in the LSCD group and moderate in the BMFD-AM group. The prevention of conjunctivalization in the BMFD-AM+cell group was evidenced by positive K3/K12 and negative MUC5AC and K4 observed on re-epithelialized corneal epithelium. The histologic sections at different time points and positive Sry gene expression indicated that the delivered cells adhered to the wounded corneal surface and proliferated well.

Conclusions.: These findings demonstrate that the BMFD with fixed AM served well as a cell delivery system for the ocular surface. The delivered limbal epithelial cells promoted corneal re-epithelialization and prevented corneas from conjunctivalization and neovascularization in rabbits with experimental LSCD.

Causes of corneal limbal stem cell deficiency (LSCD) include ocular burns, Stevens Johnson syndrome, congenital aniridia, ocular cicatricial pemphigoid, severe microbial infection, and recurrent pterygium. The severe cases have persistent corneal epithelial defect with significant visual impairment, which often causes vision loss. ^1,2 LSCD presents a significant challenge for the ophthalmologist in the clinic. 

Progress has been made in reconstructing the ocular surface of eyes with severe LSCD by using epithelium grafts recently generated by tissue-engineering technology. The transplantation of cultivated epithelia carried by a variety of scaffolds has obtained efficacy in ocular surface reconstruction. The amniotic membrane (AM) has been shown to promote clonogenicity and re-epithelialization and to reduce inflammation, neovascularization, and scar formation. The AM has been commonly used and proven to be one of the most ideal scaffolds. ^{3 –6} However, reduced corneal transparency and visual acuity have been observed after healing. Nishida et al. ^7,8 developed a thermosensitive technology to produce carrier-free cell sheets, which displayed a structure similar to that of the normal epithelium and improved vision recovery after transplantation. 

Single-cell suspension transplantation has been commonly used for clinical cell therapy in many diseases, such as umbilical cord blood cell transplantation in hematologic diseases, ⁹ liver cells transplantation in the hepatic coma of severe hepatitis, ¹⁰ spleen cells transplantation in severe hemophilia A and advanced liver cancer, ¹¹ and pancreatic islet cells in the treatment of diabetes. ¹²  

It is difficult to deliver a cell suspension onto the ocular surface because of the surface's unique anatomic characteristics. The transplanted cells cannot attach to the surface long enough to proliferate. Thus, the single-cell suspension has not been successfully used for treating corneal diseases. In a previous study, we designed a novel biomembrane-fixing device (BMFD) ¹³ to fix the AM on the ocular surface. There is an airtight space between the corneal surface and the BMFD-AM system that allows suspended single cells to be successfully delivered onto the ocular surface. In the present study, we explored the therapeutic efficacy of this novel approach to corneal epithelial reconstruction in experimental LSCD rabbits by delivering cultured human limbal epithelial cells onto the ocular surface with the BMFD-AM system. 

Methods

Experimental LSCD Animal Model

Sixty female white New Zealand rabbits weighing 2 to 2.5 kg were purchased from Guangdong Medical Laboratory Animal Center. All described procedures were approved by the Institutional Animal Care and Use Committee and were conducted in compliance with the Guide for the Care and Use of Laboratory Animals of China and the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

The rabbits were anesthetized by intramuscular injection of ketamine (50 mg/kg) and chlorpromazine (25 mg/kg), together with 0.4% oxybuprocaine hydrochloride eye drops (Santen Pharmaceutical Co., Ltd. Japan) applied three times at 5-minute intervals. The limbal and central corneal epithelia were completely removed with a surgical blade, including the 0.1- to 0.2-mm-thick corneal layer of the 2-mm outer edge of the cornea, to create the LSCD model. 

Sixty LSCD rabbits were randomly divided into three groups of 20 each. Control group LSCD rabbits without special treatment received only topically applied KSFM four to six times a day. The rabbits in The BMFD-AM group were treated by AM fixed with BMFD (BMFD-AM) in addition to topical KSFM. The rabbits in the BMFD-AM+cell group were treated with male human limbal epithelial cells delivered by BMFD-fixed AM. The AM was replaced every three days. All the cases were followed up until completely corneal re-epithelialization. 

The BMFD-AM Device

The BMFD used for rabbits was refined from the PMMA ring, as our previous report. ¹³ The size of the BMFD ring, 0.5 to 1 mm thick with a 16- to 18-mm inner diameter and a 18- to 20-mm outer diameter, was redesigned in accordance with the rabbit's conjunctival sac. There is a gap in the middle of the ring that makes it easy to insert a blunt needle through it. The rings were sterilized with ethylene oxide before use. The cryopreserved human AM, prepared as described elsewhere, ¹⁴ was thawed and washed with PBS before use. 

A piece of preserved AM with a size two times that of the BMFD was flattened onto a sterile glass dish with the epithelial side down. A BMFD ring with the contoured side up was placed on and wrapped by the AM, to form the BMFD-AM device (Fig. 1). The BMFD-AM was inserted downward into the lower fornix by opening the lower eyelid and then into the upper fornix by raising the upper eyelid, so that it was fully inserted into the conjunctival sac. After the membrane's position was adjusted, the excess was cut off along a circle 5 mm from the outer edge of the BMFD ring, and the BMFD-AM was thus fixed on the ocular surface. 

Figure 1.

 

BMFD-AM system. (A) AM fixed by BMFD. (B) A cross-section illustrating a cell suspension delivered by the BMFD-AM system.

Cell Delivery

A spontaneously derived human corneal epithelial cell line (SDHCEC1) ¹⁵ was generated from a 29-year-old, healthy male donor. SDHCEC1 cells express corneal-type keratins K3/K12 but not the conjunctival markers K4 and mucin 5AC. It possesses high proliferative capacity and progenitor cell phenotype with expression of K14, K19, β1-integrin, and vimentin. ¹⁵ The frozen SDHCEC1 cells were thawed and cultured with 1× KSFM, containing l-glutamine (Invitrogen) to obtain the concentrated single-cell suspension solution (×10⁷/mL). The well-suspended single-cell solution (0.5 mL) was injected into the space between the AM and the ocular surface of the LSCD eyes in the BMFD-AM+cell group with a 1-mL syringe with a blunt needle and was allowed to adhere to the wounded corneal surface for 15 to 20 minutes. This procedure was repeated three or four times, and 0.15% cyclosporine eye drops were applied to the treated eyes in this group two times a day during the second week. 

Ocular Surface Evaluation

The physical signs on the ocular surface were observed by slit lamp examination of the rabbits' corneas once a day. A fluorescein sodium solution was used to check the integrity of the corneal epithelium. The corneal surface was also examined for clarity and neovascularization by slit lamp, and the condition was evaluated and scored according to Table 1. The time until total re-epithelization (in days) and the adverse reactions were recorded. 

Table 1.

 

Score Criteria for Ocular Surface Evaluation

Score	Corneal Opacity	Fluorescein Stained Area	Neovascularization
0	None	None	None
1	Mild turbidity, iris texture can be seen	≤1/4 quadrant	2 mm within limbus
2	Moderate turbidity, iris texture unclear	1/4 < area ≤1/2	Present in the peripheral cornea, ≤1/2 quadrant
3	Severe turbidity, pupil seen faintly	1/2 < area ≤3/4	Present in the peripheral cornea, >1/2 quadrant
4	Severe turbidity, pupil not visible	>3/4	New blood vessels growing in the entire cornea

Histology and Immunohistochemistry Staining

Rabbit corneas from each experimental group were collected and embedded in OCT compound (Tissue Tek; Sakura Fintek, Torrance, CA) for sectioning. Cryostat sections (5 μm) were mounted on slides for hematoxylin and eosin (HE) staining and immunofluorescent staining. Immunodetection of mouse anti-human K3/K12 (Chemicon, Temecula CA), mouse anti-human K4 (Wuhan Boshide Bioengineering Co, Ltd., Wuhan, China), and mouse anti-human monoclonal antibody mucin 5AC (Jitai Biological Scientific, Co., Shanghai, China) were performed by immunofluorescent staining with secondary antibodies, including FITC goat anti-mouse antibody IgG (Sigma-Aldrich, Inc., St. Louis, MO). The nuclei were counterstained with 0.5 g/mL Hoechst 33342 dye (Sigma-Aldrich. Inc.), and the specimens were observed with a confocal laser scanning microscope (Carl Zeiss Meditec, Oberkochen, Germany). 

Sry Gene Detection

The DNA of all the samples was extracted by the chelex 100 method ¹⁶ and quantified by quantitative real-time PCR (Quantifier Y Human Male DNA Quantification Kit; cat. no. 4343906; Applied Biosystems, Inc. [ABI], Foster City, CA). A fluorescence multiplex amplification system (Power Plex Y; Promega, Madison, WI) was used for amplification of the PCR fragments. The fragments underwent capillary electrophoresis by automatic genetic analyzer and were genotyped (model 3100 with Genescan and Genotype software; ABI). 

Data Analysis

Statistical analysis was performed by paired-samples t-test (SPSS 11.5; SPSS, Chicago, IL). Results of comparisons are expressed as the mean ± SE; P < 0.05 was considered statistically significant. 

Results

The BMFD designed for this study was refined from the PMMA ring, as we reported in our previous study. ¹³ There is an airtight space between the cornea surface and the BMFD-AM device. A gap in the middle of the ring makes it easy to insert a blunt needle through which the delivered cells are injected. In the present study, we observed that the cultured human corneal epithelial cells could be successfully delivered onto the ocular surface with this novel device for reconstruction of the corneal epithelium in LSCD rabbits. 

The Delivered Corneal Epithelial Cells Promoted Corneal Re-epithelization in the Experimental LSCD Rabbits

To evaluate the effect of cell delivery on corneal re-epithelialization, we developed an experimental LSCD model in the rabbit. In the LSCD group without treatment, the conjunctival cells grew onto the wounded cornea with goblet cells, collagen fibers, and new blood vessel hyperplasia, as well as inflammatory cell infiltration, as observed by histopathologic examination and the positive conjunctival epithelial markers MUC5AC and K4. 

On day 1, at 1 hour after surgery, a layer of the delivered cells covered the damaged corneal surface in the BMFD-AM+cell group. These corneas were negative to fluorescein staining or displayed only a few fluorescein dots (fluorescein score, 1.55 ± 1.50; Fig. 2). In contrast, positive fluorescein staining was observed on the entire cornea in the other two groups (score, 3.75 ± 0.55 in the LSCD group; 3.65 ± 0.49 in the BMFD-AM group). The difference was significant (P < 0.05) between the BMFD-AM+cell group and each of the other two groups, but there was no significant difference between the other two groups (P > 0.05, Fig. 2). 

Three days later, in the BMFD-AM+cell group several cases (15%, 3/20) had re-epithelialization of the entire damaged cornea, and other cases had only a partially positive corneal fluorescein staining (score, 1.95 ± 1.32). However, much higher fluorescein staining scores were observed in the LSCD (score, 2.85 ± 0.75) and BMFD-AM (score, 2.45 ± 0.69) groups. There was a significant difference (P < 0.05) between the BMFD-AM+cell group and the LSCD group (Fig. 2). 

Figure 2.

 

Corneal fluorescein staining photography and statistical analysis. (A) Representative images of corneal fluorescein staining in the three groups at days 1, 3, 7, and 14 after surgery. (B) Average scores of corneal fluorescein staining in the three groups at the examination time points. Data are presented as the mean ± SE; n = 20 in each group (*P < 0.05).

At day 7 after the surgery, half of the corneas (10/20) in the BMFD-AM+cell group displayed re-epithelialization (fluorescein score, 0.55 ± 0.69) of the entire damaged cornea, significantly better (P < 0.05) than in the other two groups (score, 1.45 ± 0.76 in LSCD group and 1.65 ± 0.81 in the BMFD-AM group; Fig. 2). 

At day 14 after surgery, almost all corneas in the three groups were re-epithelialized (score, 0.65 ± 0.67 in the LSCD group, 0.45 ± 0.51 in the BMFD-AM group, 0.30 ± 0.57 in the BMFD-AM+cell group), except in three cases in the LSCD group. There was no significant difference (P > 0.05) between all the groups (Fig. 2). 

In the BMFD-AM+cell group, all the wounded corneas were re-epithelialized in 4 to 9 (5.60 ± 0.46) days, which was much faster (P < 0.05) than in the other two groups, 6 to 14 (9.25 ± 0.51) days in the BMFD-AM group, and 9 to 21 (12.45 ± 0.65) days in the LSCD group. 

The Delivered Corneal Epithelial Cells Prevented Conjunctivalization and Neovascularization of the Re-epithelialized Corneas

The therapeutic effect was further evaluated based on transparency and neovascularization when the corneas were totally re-epithelialized. In the BMFD-AM+cell group, most of the corneas were smooth and transparent, and iris vessels were observed underneath the cornea. Neovascularization and inflammation were barely detected on the corneal surface of these animals. The scores of corneal opacity and neovascularization in the BMFD-AM+cell group were much less than in the BMFD-AM and LSCD groups, and the differences were statistically significant (P < 0.05). There was no significant difference between the latter two groups (Table 2). 

Table 2.

 

Scores of Corneal Opacity and Neovascularization

Group	Opacity Score	Neovascularization Score
LSCD	2.85 ± 0.19 a	2.65 ± 0.17 a
BMFD-AM	1.70 ± 0.21 a	2.40 ± 0.23 a
BMFD-AM+Cell	0.45 ± 0.11 b	0.75 ± 0.16 b

 Data are shown as the mean ± SE; P < 0.05 was considered significant.

a, b  : the values with the same letter indicate that these values are not significantly different (P > 0.05).

Histopathologic examination showed that the structure of the re-epithelialized cornea in the BMFD-AM+cell group was similar to normal. Immunohistochemical staining showed K3/K12 expression, whereas MUC5AC and K4 were negative (Fig. 3). This phenotype indicates that the cells covering the damaged corneas were corneal epithelial cells. In contrast, conjunctivalization, collagen fibrosis, neovascularization, and inflammatory cell infiltration, as well as positive fluorescent staining of MUC5AC and K4 on the corneas (Fig. 3) were observed in the BMFD-AM and LSCD groups without cell therapy. 

Figure 3.

 

The delivered corneal epithelial cells suppressed conjunctiva epithelialization and neovascularization on corneas with LSCD. In the BMFD-AM+cell group, the re-epithelialized corneas were smooth and transparent, with iris vessels seen beneath the cornea, as observed by slit lamp at day14. The slit lamp images are the representative ones. The images by HE staining showed the normal morphology of the reconstructed corneal epithelium without conjunctivalization and neovascularization. Immunohistochemistry staining showed positive K3/K12 and negative MUC5AC and K4 expressed on the stratified corneal epithelium. In the LSCD and BMFD-AM groups, the corneal surface was covered by a conjunctival epithelium with goblet cells, new blood vessel hyperplasia, inflammatory cell infiltration, and positive staining for MUC5AC and K4.

Successful Delivery of Male Human Corneal Epithelial Cells to the Damaged Female Rabbit Cornea Was Evaluated by Histopathologic Examination and Sry Gene PCR Detection

Histologic examination showed that all epithelial cells were removed from the corneal and limbal surfaces with positive corneal fluorescein staining in the experimental LSCD models (Fig. 4B). In the BMFD-AM+cell group, corneal fluorescein staining was almost negative at 1 hour after cell delivery (Fig. 2; day 1). Histologic examination disclosed that there were scattered cells or a layer covering the damaged corneal surface as early as 1 hour after the cells were delivered (Fig. 4C). Three days later, the corneal fluorescein staining was nearly negative (Fig. 2, day 3). Histologic examination disclosed that the delivered epithelial cells maintained adherence to and formed a monolayer on the damaged corneal surface (Fig. 4D). The delivered cells had proliferated and formed multiple-layered epithelial sheets when examined on days 7 and 14 (Fig. 2). These results suggest that the delivered cells adhered to, survived on, and proliferated well on the damaged corneal surface. 

Figure 4.

 

The transplanted cells traced by histopathologic examination and the Sry gene. (A–D) Histopathologic examination (HE staining) before and after cell delivery. (A) Normal rabbit cornea; (B) mechanically de-epithelialized cornea before cells delivery; (C) delivered cells covering the de-epithelialized corneal surface at 1 hour after cell injection; and (D) the delivered cells forming a monolayer on the de-epithelialized corneal surface at day 3. (I, II) The results of Sry gene analysis. (I) Negative expression of the Sry gene by normal female rabbit cornea; (II) positive expression of the Sry gene by the recipient female rabbit corneas with male human cell injection at day 14 in the BMFD-AM+cell group.

When the corneas re-epithelialized completely on day 14 after cell delivery in the BMFD-AM+cell group, they were harvested to have Sry gene PCR detection. Sry gene expression was positive in the corneas with male cell delivery (Fig. 4II) but negative in the recipient corneas from female rabbits (Fig. 4I). This finding demonstrates that the epithelial cells covering the damaged cornea originated from the transplanted male human limbal epithelial cells. 

Discussion

Single-cell suspension injection is mainly used in cell therapy for the tissues or organs with rich blood vessels. It is not suitable for the mucous or skin surface reconstruction because the transplanted cell suspension did not adhere for long and survive. The cornea is transparent and free of blood vessels, which makes it more difficult to implement single-cell suspension injection. How to make the transplanted cells attach and survive on the ocular surface is a challenging task for ocular surface reconstruction by using single-cell suspension injection. 

Many researchers have tried to solve this problem. Gallazzi et al. (IOVS 2004;45:ARVO E-Abstract 1432) found that the topical application of BM-derived stem cells enhances the repair of corneal injuries. Homma et al. ¹⁷ successfully reconstructed the damaged corneas by transplantation of the suspended solution of ES cell–derived epithelial progenitor cells with a temporary tube on the cornea. Wang et al. (IOVS 2004;45:ARVO E-Abstract 3938) injected GFP-labeled bone marrow stoma cells into periocular tissue and observed the transformation of bone marrow stoma cells into corneal cells that reconstructed the damaged corneas. Reinshagen et al. ¹⁸ successfully reconstructed the LSCD rabbit's corneal epithelium by injecting the suspended mesenchymal stem cell solution under the sutured AM. 

In a previous study, we designed a novel BMFD ¹³ to fix an AM on the ocular surface. Different from a commercially available product (ProKera; Bio-Tissue, Miami, FL), our BMFD device is specially designed in size, material, and AM fixation, according to a silicone impression of the eye's anterior segment, ¹³ and matches the fornix quite well. There is an airtight space between the corneal surface and BMFD-AM system that allows suspended cells to be well delivered onto the ocular surface. In this study, we are the first to use this device to deliver a single-cell suspension to reconstruct the ocular surface. 

The rabbits behaved well and tolerated the device. There was no dislocation of the BMFD in this study or in the previous one. Although the AM patch fell off of the BMFD in 50% of the rabbit eyes within 2 to 3 days, ¹³ because of movement of the eye and eyelids, it did not affect the efficacy of the device. We changed the AM at these times to prolong its effects. ¹³ Furthermore, we have used this device, specifically made for humans, in treating more than 50 patients with different ocular surface diseases, including ocular burn. Most patients tolerated it very well with no complaint of discomfort (manuscript in preparation). 

In addition to being part of a cell delivery system, the AM fixed by the BMFD served as a biological dressing to protect the transplanted cells. AM has the ability to maintain stem or progenitor cell characteristics. ¹⁹ It can help to repair the tissue defect and restore its function by supplying appropriate levels of cytokines and a variety of useful components. ²⁰ AM could protect the reconstructed ocular surface, accelerate the speed of corneal re-epithelialization, ²¹ and promote conjunctival stem cell transdifferentiation by reducing inflammation and supplying the essential factors. The nutrient solution could be dropped and penetrate the semipermeable AM membrane to meet the needs of the transplanted cells. These are the advantages of the BMFD-AM that make it superior to other cell delivery methods. 

The essential clinical signs of LSCD include conjunctivalization of the cornea with associated goblet cells, vascularization, fibrovascular pannus, persistent epithelial defects, and scarring. ²² Of these, conjunctivalization of the cornea is the hallmark. K4 is generally accepted as one of the markers of conjunctival epithelium based on previous publications ^{23 –28} and our personal observation in a previous study. ²⁹ We used K4 and MUC5AC, a goblet cell marker, as well as neovascularization, to evaluate corneal conjunctivalization in the LSCD rabbits. Besides the histology and immunohistochemical staining, we scored the degree of conjunctivalization, the area of epithelial defect, and the severity of vascularization and statistically analyzed it to verify the establishment of the LSCD animal model and the significant differences between each group. We recognized that a small amount of K3/K12-positive cells were detected on the BMFD-AM corneas, which may indicate that our LSCD model was not completely successful. However, the LSCD model we established was sufficient to show LSCD, evidenced by significant vascularization and conjunctivalization with positive expression of MUC5AC and K4 on corneas in the BMFD-AM and LSCD groups without cell therapy. This observation had been reported by other researchers. ^14,18,30,31 The results showed that the transplanted cells formed a stratified epithelial layer similar to the normal structure, which expressed the corneal epithelial cell–specific markers K3/K12, but not the conjunctival epithelial and goblet cell markers K4 and MUC5AC (Fig. 4B). 

Restoration of a clear and transparent cornea is associated with fast corneal epithelialization with normal phenotype and complete wound remodeling. Our study showed that the delivered cells served effectively in the reconstruction of the corneal and limbal surfaces. Thus, the fast corneal re-epithelialization restored corneal transparency and suppressed the migration of the conjunctival epithelial cells to the corneal surface and corneal neovascularization. 

The purpose of this study was to verify that the device can be used for cell delivery and that the delivered cells can survive and proliferate on the ocular surface. As Y chromosome tracing is one of the most reliable methods and the Sry gene detection kit is currently available for humans but not for rabbits, we thus used male human corneal epithelial cells as the donor cells and female rabbits as the recipients to trace the implanted cells. The SDHCEC1 cell line, which has been shown to possess high proliferative capacity and progenitor cell phenotype with expression of CK14, CK19, β1-integrin, and vimentin, ¹⁵ is suitable for use in reconstructive corneal tissue engineering. To avoid rapid immune rejection, 0.15% cyclosporine eye drops were applied to the eyes in the BMFD-AM+cell group. Further study of autologous transplantation using rabbit cells is necessary for long-term observation. The results showed that the male human cells were successfully transplanted to the surface of epithelium-deficient rabbit cornea and formed a monolayer, evidenced by Sry gene detection by PCR in the genomic DNA extracted from the recipient corneas of the female rabbits. 

In conclusion, our findings demonstrate that the delivery of single-cell suspension through BMFD-AM is feasible and effective and needs no carrier or scaffold. Cell transplantation with this tool for ocular surface reconstruction may make it easy to evaluate the effects of various cells seeded in vivo and to investigate the mechanism of stem cell differentiation and transdifferentiation in the microenvironment in vivo. It could have clinical use in autologous stem cell transplantation for the treatment of patients with LSCD or persistent epithelial defect. 

Footnotes

 Supported by Grant 2006AA02A133 (WZC) from the National High Technology Research and Development Program of China (863 Project) and Grant QN-08 from the Fundamental Research Funds of State Key Lab of China.

Footnotes

 Disclosure: P. Wan, None; X. Wang None; P. Ma, None; N. Gao, None; J. Ge, None; Y. Mu, None; Z. Wang, None

The authors thank Tong Fei (Forensic Medicine Department, Sun Yat-sen University, Guangzhou, China) for assistance with Sry gene detection, and De-Quan Li, (Ocular Surface Center, Cullen Eye Institute, Department of Ophthalmology, Baylor College of Medicine, Houston, TX) for his work in revising the manuscript.