Fate of Corneal Epithelial Cells Separated from Limbus In Vivo

Abstract

Purpose.: To characterize corneal epithelial cells separated from limbus in vivo by transplantation of a stainless steel ring with or without creating a defect inside the ring.

Methods.: A stainless steel ring (diameter, 8 mm; width, 300 μm; depth, 250 μm) was transplanted into rabbit corneal stroma using 10-0 nylon interrupted suture after cutting to a 250 μm depth by corneal vacuum trephine (diameter, 8.0 mm). Epithelial cells were removed inside the ring, and re-epithelization was evaluated after 1 week. Hematoxylin staining and immunostaining against p63, Ki67, and cytokeratin 3 were performed for phenotypic analysis of corneal epithelia. A corneal epithelial defect was centrally created inside the ring (4, 5, and 6 mm diameter) after transplantation. When re-epithelization was achieved, a central epithelial defect was continuously created until cells were exhausted within the ring. The number of created defects was also analyzed to assess the potential of re-epithelialization.

Results.: Ring-transplanted corneal stroma showed few signs of inflammation, and when epithelium was totally removed from inside the ring, complete epithelial defects were persistent for ≥ 1 month. Corneal sensation was significantly decreased in corneas with the ring (P < 0.05). Immunostaining demonstrated similar expression patterns for p63, Ki67, and cytokeratin3 as the controls. When rings were transplanted into the intact cornea, inside epithelia prevented epithelial defects in vivo for ≤ 6 months.

Conclusions.: Transient-amplifying cells might maintain homeostasis for >1 month when separated from their limbus in vivo. This model will be useful for future stem cell research or wound healing models.

Corneal epithelial homeostasis has been recognized to move according to the XYZ theory ¹ (i.e., stem cells [SCs] self-produce SCs and transient-amplifying cells [TACs], which proliferate several times and differentiate into totally differentiated cells [TDCs]). It is believed that populations of corneal epithelial stem cells exist at the limbus, a narrow area between the cornea and conjunctiva in human ^2,3 and other species. ⁴ Corneal epithelial SCs have common features with other adult somatic stem cells, including small size ⁵ and high nuclear-to-cytoplasmic ratio. ⁶ They also lack expression of differentiation markers such as cytokeratins 3 and 12. ⁷ Although signaling and mechanisms of SC differentiation have not been elucidated yet, the fate of tissue-specific SCs is believed to be at least as long as lifespan (> 1 year), and TACs are believed to have a shorter lifespan (a few weeks or less). However, the lifespan of TACs has not been defined because these cells are not clearly detectable in vitro and in vivo. SCs cycle slowly during homeostasis and therefore retain DNA labels for long periods; however, in the event of injury they can become highly proliferative. ^8,9  

Clinically, when limbus is totally destroyed by severe inflammatory diseases such as alkali burn, Stevens-Johnsons syndrome, or ocular pemphigoid, normal corneal epithelial homeostasis cannot be maintained. ¹⁰ As a result, corneal epithelial defects may appear and conjunctival epithelial cells invade from a limbal location to the central cornea to cover epithelial defects by neovascularization. In human corneal transplants, all corneal epithelial cells are replaced by recipient epithelium after 3 months ¹¹ ; however, limbal epithelial cells in transplanted allografts exist after much longer periods. ¹² These data support the existence of stem cells in limbus, and not in a central location. A recent report demonstrated that corneal epithelial stem cells are distributed throughout the ocular surface in several species. ⁴  

In vitro, SC and TAC function have been evaluated by colony-forming efficiency (CFE) over long periods, with holoclone (SCs), melaclone (TACs), and paraclone (terminally differentiated cells) cells demonstrating particular colony shapes, cellular morphology, and secondary CFE on 3T3 feeder layers. ⁶ However, in vivo, no such assay to evaluate the function of SCs and/or TACs has been established. Because features such as inflammation, blinking, tears, and cytokines will affect the fate of SCs and TACs, functional assays in vivo are supposedly more complex and unstable than in the in vitro model. Here, we developed a novel and simple assay to demonstrate the fate of centrally located TACs and their contribution during wound healing using a stainless steel ring transplantation model. Such stainless steel rings completely separate limbal SCs and peripheral and/or central corneal TACs. 

Methods

Transplantation of Stainless Steel Ring in Rabbit Cornea

Japanese white rabbits aged 1 month and older were used in this study. Animals were handled according to guidelines described in the ARVO statement for the Use of Animals in Ophthalmic and Vision Research and approved by the Animal Care and Use Committee of Tokyo Dental College (approval number, 227402). 

Stainless steel rings (diameter, 8 mm; width, 300 μm; depth, 250 μm) were created by cutting off the tip of corneal vacuum trephines (diameter, 8.0 mm; KAI Industries Co., Ltd., Gifu, Japan). Under general anesthesia by intravenous injection of diazepam (0.5 mg/kg) and pentobarbital sodium (30 mg/kg), rings were transplanted into the corneal stroma of the left eye (right eye was control) using 10-0 nylon (Mani, Tochigi, Japan) interrupted suture after cutting to a depth of 250 μm using corneal vacuum trephine (diameter, 8.0 mm; Hessberg-Barron; Jedmed Instrument Company, St. Louis, MO) (Fig. 1A). 

Figure 1.

 

Epithelial physiology with ring transplantation on rabbit cornea. (A) A stainless steel ring (diameter, 8 mm; width, 300 μm; depth, 250 μm) was transplanted into rabbit corneal stroma using interrupted sutures of 10-0 nylon after cutting the cornea to a depth of 250 μm depth by vacuum trephine. (B) Corneal stroma with ring transplantation for 1 week showed little inflammation. (C) When an 8-mm central defect was created after cutting with 8-mm diameter vacuum trephine to a depth of 250 μm, the defect was completely re-epithelized after 4 days. (D) When the epithelium was totally removed inside the ring after transplantation, epithelial defects persisted for ≥1 month (n = 3). (E) When the ring was removed after 1 week, epithelial defects completely healed within 4 days (n = 3).

To demonstrate the long-term effect of such ring transplantation, a rabbit with a ring transplant was maintained until epithelial exhaustion was observed, as determined by enlargement of epithelial defects visualized by fluorescein. 

Creating a Persistent Rabbit Epithelial Defect Model Using Stainless Steel Rings

When epithelial cells inside the ring were removed, re-epithelization was evaluated after 1 week. As a control, epithelial defects were not created after ring transplantation, and defect size was analyzed. Inflammation including conjunctival injection, corneal infiltration, and corneal opacity were also examined to determine whether transplanted rings induced severe inflammation. 

Corneal Sensation and Nerve Terminal in Ring-Transplanted Rabbit

Stainless steel rings were transplanted into rabbit corneas as described above. After 1 week, corneal sensation was measured with a Cochet-Bonnet esthesiometer (n = 3, Handaya Co., Tokyo, Japan). We evaluated the blinking just after touching as a marker of their corneal sensation. Immunostaining against βIII-tubulin was performed to analyze the effect of stainless ring transplantation on nerve terminals of the central cornea. ¹³  

Immunostaining

To characterize epithelial differentiation and proliferation at 1 week after stainless steel ring transplantation, immunostaining of the cornea-specific differentiation marker cytokeratin3, ¹⁴ proliferation marker Ki67, ¹⁵ and epithelial progenitor marker p63 ^16,17 was performed. Immunostaining for βIII-tubulin, a neuron-specific marker, was also performed to examine the effect on neurons after stainless steel tris-buffered saline Tween-20 ring transplantation. Antibodies used are summarized in Table 1. Frozen sections (thickness, 5 μm) were fixed with 2% paraformaldehyde (Wako, Osaka, Japan) for 15 minutes. After blocking with 10% normal donkey serum (Chemicon International, Inc., Temecula, CA) at room temperature (RT) for 1 hour, sections were incubated with cytokeratin K3 (1:100), Ki67 (1:100), and p63 (1:100) at RT for 1.5 hours. After twice washing with phosphate-buffered saline for 5 minutes each wash, sections were incubated with a cyanine 3-conjugated donkey anti-mouse IgG antibody (1:100; Chemicon International, Inc.). After three additional washes with TBST (0.825 mM Tris, 136.9 mM NaCl, 1.34 mM KCl, 0.1% Tween 20 [Sigma, St. Louis, MO]), the sections were incubated with 1 μg/mL 4′,6-diamidino-2-phenylindole (DAPI; Dojindo Laboratories, Tokyo, Japan) at RT for 5 minutes. Finally, sections were washed three times in TBST and coverslipped using an aqueous mounting medium (Fluoromount/Blus; Diagnostic BioSystems, Pleasanton, CA) and images acquired (Axioplan 2 microscope; Zeiss, Göttingen, Germany). 

Table 1.

 

Sources of Primary Antibodies

Antigens	Clone	Dilution	Source
p63	4A4	1:50	Santa Cruz*
Cytokeratin 3	AE5	1:100	Progen†
Ki67	MIB-1	1:100	Dako‡
βIII-tubulin	MAB1195	1:100	R&D Systems§

 All antibodies were mouse monoclonal category.

*  Santa Cruz, CA.

†  Glostrup, Denmark.

‡  Heidelberg, Germany.

§  Minneapolis, MN.

Colony-Forming Efficiency

Corneal epithelial cells inside the ring were removed from the stroma after 1 week and 1 month after treatment (Dispase II; Roche Diagnostics, Indianapolis, IN) at 37 °C for 1 hour. Isolated cells were treated with trypsin/EDTA for 5 minutes and seeded on mitomycin C-treated 3T3 fibroblast feeder layers seeded at a density of 20 cells per cm² for 2 weeks. Thereafter, culture dishes were visualized using rhodamine B (Wako) and scanned for comparison. 

Analysis of Defect Size and Cell Exhaustion

Corneal epithelial defects were centrally created inside the ring (diameter, 4, 5, and 6 mm) using a handy router (Konan, Osaka, Japan) just after ring transplantation. When re-epithelization was achieved completely inside the ring, central epithelial defects were continuously created within the ring until epithelial cells were exhausted (total epithelial defect was observed inside the ring). Epithelial defects inside the ring were visualized by fluorescein staining and subsequently photographed. Images were captured using a specialized software program (ImageJ; developed by Wayne Rasband, National Institutes of Health, Bethesda, MD; available at http://rsb.info.nih.gov/ij/index.html.) and defect size was measured. 

Statistical Analysis

The significance of differences between groups was determined by Student's t-test, with P < 0.05 considered statistically significant. 

Results

Consistent Total Epithelial Defect Created Inside the Ring

Corneal epithelial cells inside the ring were completely removed after ring transplantation. Little inflammation was evident in corneal stromas subject to ring transplantation for 1 week, and no conjunctival or sclera infection (Fig. 1B) was observed. When epithelial defects were created inside the cutting edge without the ring, defects completely healed within 4 days (Fig. 1C). When the epithelium was totally removed from inside the ring, epithelial defects persisted for ≥ 1 month (n = 3; Fig. 1D). When the ring was removed after 1 week, epithelial defects completely healed within 4 days (n = 3; Fig. 1E). 

Long-term analysis of ring-transplanted corneas showed that epithelial defects were maintained inside the ring. Unexpectedly, corneal epithelial layers were maintained for up to 6 months without creating defects (Fig. 2A). Corneal sections showed no epithelial ingrowth underneath the ring, and one or two layers with fragile epithelial cells inside the ring (Fig. 2B), which may demonstrate exhaustion of corneal epithelial cells inside the ring. 

Figure 2.

 

Corneal sensation and nerve terminals between the ring and corneas. (A) When the ring was transplanted, corneal sensation was decreased significantly compared with normal cornea (n = 5; *P < 0.01). In normal corneas, βIII-tubulin was observed (B) in the stroma and (C) outside the ring, but much less staining was observed inside the ring (C). (D) A higher magnification image. Ring−, without ring; Ring+, with ring.

Corneal Sensation after Ring Transplantation

One week after stainless steel ring transplantation into the cornea, corneal sensation was significantly decreased (n = 5; P < 0.001; Fig. 2A). At 3 weeks, although βIII tubulin was expressed in the epithelium and stroma of normal corneas and outside stainless steel rings transplanted in corneas (Figs. 2B, 2C, 2D); very little staining was observed inside the stainless steel rings (Figs. 2C, 2D). 

Corneal Epithelial Characterization in Ring Transplantation Model

Immunostaining revealed that cytokeratin K3 and Ki67 staining was similar compared with normal corneas when the rings were transplanted after 1 week (Figs. 3A, 3B, 3C). Cytokeratin K3 was strongly expressed in whole layers except limbus, and Ki67 showed patchy expression in basal cell layers. These expression patterns were similar to those in corneas without ring transplantation, suggesting that ring transplantation did not affect proliferation and differentiation of cells outside the ring. 

Figure 3.

 

Immunostaining against cytokeratin K3 and Ki67. One week after ring transplantation, hematoxylin and eosin staining showed no infiltration of inflammatory cells in the corneal stroma. The expression pattern of cytokeratin K3 and Ki67 at 1 week after ring transplantation was similar to that observed in the normal cornea.

Epithelial Wound Healing in Ring Transplantation Model

After creating 4-mm central corneal defects, these were healed within 3 days. When 5-mm central corneal defects were created within the ring, they were healed within 5 days. However, when 6-mm defects were created, the majority did not heal, and finally all epithelial cells were exhausted inside the ring (Fig. 4A). After the defects were healed, similar-sized defects were repeatedly created. After creating the defect by scraping several times, epithelial cells were exhausted (n = 6; P < 0.01; Fig. 4B). We then evaluated the relationship between the size of the created defect and total re-epithelialization area. Re-epithelialization ability of the residual epithelial cells inside the ring was significantly greater for the 4-mm defects than the 5- and 6-mm defects (Fig. 4C). 

Figure 4.

 

Epithelial defect size and fate of TACs. Central corneal defects (4-, 5-, and 6-mm) were created within the ring. After the defects had healed, defects of exactly the same size were created repeatedly. Although 4-mm defects usually healed within 3 days (A), after scraping several times, epithelial cells were exhausted. (B) Defects of 5- and 6-mm did not heal before, or after scraping three times, and the amount of healing after scraping was significantly more in wounds of 4 mm than in 5- and 6-mm wounds (n = 6; P < 0.01). (C) The ability to resurface the total wound healing area of residual cells inside the ring was greater in 4-mm defects compared with the 5- and 6-mm defects.

Colony-Forming Efficiency Inside the Ring

Before and 1 and 5 weeks after transplantation of the 8-mm ring into the rabbit cornea, corneal epithelial cells inside the ring were collected using dispase and trypsin/EDTA treatment, then seeded on 3T3 feeder layers for 14 days at a cell density of 100 cells per cm². At weeks one and five, significant decreases in CFE were observed inside the ring compared with before ring transplantation (n = 4; P < 0.01; Fig. 5). Although there were no significant changes between weeks one and five, colony size tended to be smaller at 5 weeks. 

Figure 5.

 

Colony-forming efficiency inside the ring. Before and 1 and 5 weeks after the 8-mm ring was transplanted into the rabbit cornea, corneal epithelial cells inside the ring were collected using dispase and trypsin/EDTA treatment, then seeded on 3T3 feeder layers for 14 days (n = 4). A significant decrease of CFE was observed inside the ring (1 and 5 weeks). No significant changes were observed between weeks one and five.

Epithelial Exhaustion after Ring Transplantation for Long-Term

Although one animal was examined for this purpose, corneal epithelial cells were maintained for 6 months within the transplanted ring, despite discontinuity with limbal stem cells (Fig. 6). Cross-sections of long-term ring-transplanted corneas showed no migration within the ring from the limbal side. Epithelial cells inside the ring showed two or three layers in most locations with desquamation, which also supported that limbal or peripheral corneal epithelial cells did not migrate inside the ring. 

Figure 6.

 

Epithelium maintained for ≤6 months within the ring. Epithelial cells were maintained for <6 months within the transplanted ring despite discontinuity with limbal basal cells. (A) Cross-sections of long-term ring-transplanted cornea showed no migration within the ring from the limbal side, but infiltration of inflammatory cells. (B) Six months after ring transplantation, epithelial cells inside the ring eventually began to be exhausted. Figures present with clock hours. Black-lined boxes show the high magnification of the ring location to demonstrate the separation between the inside and outside of the corneal epithelium.

Discussion

This study revealed novel insights into the pathophysiology of TACs in vivo, a process of corneal epithelial dynamics without normal intake of limbal basal cells, which was recognized as a reservoir of stem cells. We established a model of these cells' function in vivo using stainless steel ring transplantation. These ring devices divided the epithelial layer, and prevented epithelial movement and proliferation from outside the ring. Almost all the corneal nerve was cut by trephine, and the stainless steel ring did not affect the conjunctival injection, irritation, corneal stromal infiltration, and opacity. Corneal sensation was significantly decreased with decrease of stromal βIII-tubulin staining inside the ring. This concise model successfully separated limbal location and the central cornea, and cornea and conjunctiva in vivo, rendering it an ideal model to study regulation of adult somatic SCs. 

Demonstrating clear separation of epithelial layers inside and outside the ring, and a lack of epithelial migration from outside the ring was the first aim of this study. We showed that complete epithelial defects created inside the ring lasted > 1 week, and that CFE of epithelial cells inside the ring was decreased at 1 and 5 weeks. Although epithelial cells inside the ring showed two or three layers with desquamation, cells maintained epithelial layers for approximately 6 months after transplantation. When the ring was removed after 1 week, corneal epithelial cells migrated inside and re-epithelization was complete in < 1 week, demonstrating that the corneal stromal bed was a good environment for epithelial incursion after cutting the stroma and ring transplantation. Although CFE was almost the same at 1 and 5 weeks, each colony size was larger at 1 week, suggesting that the number of TACs was similar but the proliferative potential (one definition of “stemness”) was higher at 1 week (Fig. 5). For a long-term ring transplant model, we did not expect that epithelial cells inside the ring would exist for > 6 months. This finding not only indicated a lack of invasion of epithelial cells inside the ring but also supported existence of early TACs inside the ring. However, because only one animal was examined, further study is required to determine epithelial homeostasis inside the ring over the long-term. 

A recent report demonstrated that SCs were not only located in the limbal basal area but also in peripheral and central locations, suggesting that SCs are enriched in limbal locations. ⁴ That corneal epithelial stem cells are enriched in limbal locations is supported by several reports in humans. ^{18 –21} Some clinical cases have shown that the cornea maintains an intact epithelium even when the limbus is totally invaded by conjunctiva. ²² However, these could be cases with abnormal corneal epithelial homeostasis, which support the hypothesis that SCs were not only present in limbal locations. In our study, central corneal epithelial cells in vivo were separate from limbal epithelial cells and could maintain their homeostasis for > 6 months, suggesting that the lifespan of early TAC was longer than expected, or the existence of SCs was centrally located. In human, many clinical reports suggest that corneal epithelial SCs were located in their limbus. This discrepancy could be explained partially by the report ⁴ that the location of corneal epithelial SCs was different in different species. Further studies to characterize these cells are necessary. 

Inside the ring, the 4-, 5-, and 6-mm defect model demonstrated that 5- and 6-mm defects were not significantly different with respect to time required for complete wound healing and re-epithelization capacity. Epithelial cells inside the ring with 4-mm defects had greater re-epithelization capacity and shorter time for healing than 5- and 6-mm defects. This result suggests that either inner peripheral cells (between 4 and 5 mm) have more potential than outer cells (between 5 and 6 mm) or the size of the defect itself was a factor affecting cellular proliferative and/or re-epithelization capacity. We believe that the second hypothesis is more likely because cellular potential should not be different in similar locations. Indeed, our data suggested that a balance of proliferative capacity and surface area of the defect was involved. 

In summary, our novel stainless steel ring transplantation model provided insights into the fate of TACs inside the ring when epithelial defects were created. More than 6 months after ring transplantation, two or three epithelial layers were maintained inside the ring. This device can be used for future studies of TAC homeostasis and/or in vivo studies of regeneration. 

Footnotes

 Supported by Grant-in-Aid for Young Scientists (B) (18791301) from the Ministry of Education, Culture, Sports, Science and Technology, and Grant-in-Aid for Scientific Research (H18-regeneration-young-002) from the Ministry of Health, Labour and Welfare, Japan.

Footnotes

 Disclosure: T. Kawakita, None; K. Higa, None; S. Shimmura, None; M. Tomita, None; K. Tsubota, None; J. Shimazaki, None