The Successful Culture and Autologous Transplantation of Rabbit Oral Mucosal Epithelial Cells on Amniotic Membrane

Abstract

purpose. To determine the feasibility of using human amniotic membrane (AM) as a substrate for culturing oral epithelial cells and to investigate the possibility of using autologous cultivated oral epithelial cells in ocular surface reconstruction.

methods. An ocular surface injury was created in one eye of each of eight adult albino rabbits by a lamellar keratectomy, and a conjunctival excision was performed, including and extending 5 mm outside the limbus. Oral mucosal biopsy specimens were obtained from these eight adult albino rabbits and cultivated for 3 weeks on a denuded AM carrier. The cultivated epithelium was examined by electron microscopy (EM) and immunohistochemically labeled for several keratins. At 3 to 4 weeks after the ocular surface injury, the conjunctivalized corneal surfaces of the eight rabbits were surgically reconstructed by transplanting the autologous cultivated oral epithelial cells on the AM carrier.

results. The cultivated oral epithelial sheet had four to five layers of stratified, well-differentiated cells. EM revealed that the epithelial cells were very similar in appearance to those of normal corneal epithelium, had numerous desmosomal junctions, and were attached to a basement membrane with hemidesmosomes. Immunohistochemistry confirmed the presence of the keratin pair 4 and 13 and keratin-3 in the cultivated oral epithelial cells. Corneas that were grafted with the cultivated oral epithelial cells on an AM carrier were clear and were all epithelialized 10 days after surgery.

conclusions. Cultures of oral epithelial cells can be generated to confluence on AM expanded ex vivo from biopsy-derived oral mucosal tissue. Autologous transplantation was performed with these cultivated oral epithelial cells onto the ocular surfaces of keratectomized rabbit eyes. Autologous transplantation of cultivated oral epithelium is a feasible method for ocular surface reconstruction. The long-term outcome of such transplantation is not yet clear, and its feasibility in clinical use should be evaluated further.

The normal ocular surface is covered with highly specialized corneal and conjunctival epithelia, which are formed by two phenotypically different types of epithelial cells. ^{1 2 3 4 5} The conjunctival epithelium is well vascularized and consists of loosely organized cell layers populated by mucin-secreting goblet cells. The cornea is covered by nonkeratinized, stratified epithelium that is responsible for maintaining ocular surface integrity and is essential for vision. Corneal epithelial cells are derived from stem cells located in the limbal epithelium. ^{6 7} In severe ocular surface diseases, such as Stevens-Johnson syndrome (SJS) and ocular cicatricial pemphigoid (OCP), in which limbal epithelial cells are destroyed, the neighboring conjunctival epithelial cells invariably cover the corneal surface resulting in chronic inflammation, stromal scarring, and neovascularization, which severely affects visual acuity. ^{2 8 9}  

Many attempts have been made to establish a surgical treatment for severe ocular surface diseases. ^{10 11 12 13 14} In such diseases, because the prognosis after penetrating keratoplasty is poor, the alternative surgical treatment of corneal epithelial transplantation (limbal transplantation or keratoepithelioplasty), in conjunction with amniotic membrane (AM) transplantation, has been developed to improve the outcome of ocular surface reconstruction. ^{15 16 17 18} The most recently developed treatment for these diseases involves the transplantation of cultivated corneal epithelial stem cell for reconstructing the ocular surface after damage caused by corneal epithelial stem cell deficiency. ^{19 20 21} Our group developed a corneal limbal epithelial culture system, using AM as a carrier, in an animal model. ^{22 23} We have since adopted this system for clinical use in severe ocular surface disorders and have successfully achieved ocular surface reconstruction. ^{24 25} However, despite the success of these surgical procedures, several problems remain. First, transplantation of corneal epithelial cells from donors (allografts) requires sufficient donor material; second, it carries a risk of rejection. Consequently, after surgery, an intensive, long-term course of immunosuppressants must be prescribed to prevent postoperative inflammation and allograft rejection, markedly reducing the quality of life of these patients. In some cases damaged corneas have sometimes been successfully reconstructed by transplanting autologous limbal epithelial cells (autografts). ^{19 20} However, this is not possible in severe ocular surface diseases, such as SJS and OCP, because they are usually bilateral, and therefore autologous corneal epithelial transplantation is not an option in these cases. 

This article describes an attempt to overcome the problems of allogeneic transplantation by using oral epithelial cells as a substitute for corneal epithelial cells. We cultured oral epithelial cells on AM expanded ex vivo from biopsy-derived oral mucosal tissues of eight rabbits. We then transplanted the autologous cultivated oral epithelium onto the ocular surfaces of keratectomized rabbits and evaluated the survival of the tissue. This study is a first step toward assessing the use of autologous transplantation of mucosal epithelial cells of nonocular surface origin. 

Materials and Methods

Preparation of AM

With proper informed consent in accordance with the tenets of the Declaration of Helsinki for research involving human subjects and on approval by the Institutional Review Board of the Kyoto Prefectural University of Medicine, human AMs were obtained at the time of Cesarean section. Under sterile conditions, the membranes were washed with sterile phosphate-buffered saline (PBS) containing antibiotics (5 mL of 0.5% levofloxacin) and stored at −80°C in Dulbecco’s modified Eagle’s medium (GibcoBRL, Rockville, MD) and glycerol (Wako Pure Chemical Industries, Osaka, Japan) at the ratio of 1:1 (vol/vol). Immediately before use, the AM was thawed, washed three times with sterile PBS containing antibiotics, and cut into pieces approximately 4 cm × 4 cm. For the oral epithelial cultures, membranes were then deprived of their amniotic epithelial cells by incubation with 0.02% EDTA (Nacalai Tesqu Co., Kyoto, Japan) at 37°C for 2 hours to loosen cellular adhesion, followed by gentle scraping with a cell scraper (Nunc International, Naperville, IL). 

Primary Cultures of Oral Epithelial Cells

We cultured the rabbit oral epithelial cells by using a previously reported culture system for corneal epithelial stem cells, with several modifications. Oral epithelial cells were cocultured with mitomycin C (MMC)-inactivated 3T3 fibroblasts. Briefly, confluent 3T3 fibroblasts were incubated with 4 μg/mL MMC for 2 hours at 37°C under 5% CO₂ to inactivate their proliferative activity. They were then rinsed with PBS to remove MMC, trypsinized, and plated onto plastic dishes at a density of 2 × 10⁴ cells/cm². Denuded AMs were spread, epithelial basement membrane side up, on the bottom of culture plate inserts (Corning, Inc., Corning, NY), and these inserts were placed in dishes containing treated 3T3 fibroblasts. We also used an air-lifting technique to promote oral epithelial differentiation and the epithelial barrier function. ²⁶  

Oral mucosal biopsy specimens, each 4 to 6 mm² in size, were taken from eight adult albino rabbits (2–2.5kg) with anesthesia induced by intramuscular injection of xylazine hydrochloride (5 mg/mL) and ketamine hydrochloride (50 mg/mL). Animals were treated in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and with the experimental procedure approved by the Committee for Animal Research at Kyoto Prefectural University of Medicine. Submucosal connective tissues were removed with scissors to the extent possible. The resultant samples were cut into small explants that were immersed three times in PBS solution containing antibiotics: 50 IU/mL penicillin-streptomycin and 5 μg/mL amphotericin B for 10 minutes at room temperature. These explants were then incubated at 37°C for 1 hour with 1.2 IU dispase, as previously described, ²⁷ and treated with 0.25% trypsin-EDTA solution for 30 minutes at room temperature to separate the cells. Enzyme activity was stopped by washing with culture medium comprising DMEM and Ham’s F12 (1:1 mixture) with 10% fetal bovine serum (FBS), insulin (5 μg/mL), cholera toxin (0.1 nmol/L), human recombinant epidermal growth factor (10 ng/mL), and penicillin-streptomycin (50 IU/mL). The cell suspension was filtered through a cell-dissociation sieve (Sigma, St. Louis, MO) to remove unsatisfactory segments, yielding a suspension of purified oral mucosal cells. The suspension was centrifuged twice for 5 minutes at 1000 rpm, and the resultant cell pellet was resuspended in culture medium. The oral epithelial cells (1 × 10⁵ cells/mL) were then seeded onto denuded AM spread on the bottom of culture inserts and cocultured with MMC-inactivated 3T3 fibroblasts. The culture was submerged in medium for 2 weeks and then exposed to air by lowering the medium level (air-lifting) for 1 week. Cultures were incubated at 37°C in a 5% CO₂-95% air incubator for up to 21 days, and the medium was changed every day. 

Ocular-Surface Injury

To simulate the condition found in the stem cell deficiencies, ²² an ocular-surface injury was created in one eye of each of the eight adult albino rabbits by excising all the conjunctival tissue within 5 mm of the limbus and performing a superficial keratectomy of the entire corneal surface, including the limbal epithelial cells. Antibiotic eye drops (0.5% levofloxacin) and intramuscular gentamicin (1 mg/kg) were administered after surgery. 

Autologous Oral Epithelial Transplantation

At 3 to 4 weeks after the ocular surface injury, the conjunctivalized ocular surfaces of the eight rabbits were surgically reconstructed by transplanting autologous oral epithelial cells cultivated on AM. In all cases, the damaged corneal surface, including the 5-mm zone of adjacent conjunctival tissue, was carefully excised under anesthesia. All animals in which oral epithelial cells had been placed in culture 3 to 4 weeks earlier received autologous cultivated oral epithelial cells on AM. All received the oral epithelial sheet as a 12-mm diameter disc of AM. The sheets were sutured to the keratectomized corneal surface with 10-0 nylon sutures and covered with a soft contact lens (58% water content) held in place by four peripheral anchoring sutures (Fig. 1) . After surgery, topical antibiotics (0.5% levofloxacin) and steroids (0.1% betamethasone) were applied three times daily. For our experimental controls, four eyes received no transplant, four eyes received acellular AM transplants onto keratectomized corneas, and four eyes received the oral tissue transplant directly onto the keratectomized corneal-limbal zone. 

Immunohistochemistry

Immunohistochemical studies of several keratins in the oral epithelial sheet were performed using our previously described method. ^{28 29} Normal rabbit cornea, conjunctiva, and oral samples were also examined for purposes of comparison. Briefly, cryostat sections (7 μm thick) were placed on gelatin-coated slides, air dried, and rehydrated in PBS at room temperature for 15 minutes. To block nonspecific binding, the tissues were incubated with 1% bovine serum albumin (BSA) at room temperature for 30 minutes. Subsequently, the sections were incubated at room temperature for 1 hour with the primary antibody (Table 1) , and then washed three times in PBS containing 0.15% Triton X-100 (PBST) for 15 minutes. Control incubations consisted of incubation with the appropriate normal mouse and rabbit IgG (Dako, Kyoto, Japan) at the same concentration as the primary antibody and also the omission of the primary antibody for each specimen. After staining with the primary antibody, the sections were then incubated at room temperature for 1 hour with appropriate secondary antibodies, fluorescein (FITC)-conjugated donkey anti-mouse IgG (Jackson ImmunoResearch, West Grove, PA) and fluorescein (FITC)-conjugated donkey anti-rabbit IgG (Vector Laboratories, Burlingame, CA). After several washings with PBS, the sections were coverslipped using antifading mounting medium containing propidium iodide (Vectashield; Vector Laboratories) and examined by confocal microscopy (Fluoview; Olympus, Tokyo, Japan). 

Electron Microscopy

Rabbit oral mucosal epithelial cells cultured on denuded AM were examined by scanning electron (SEM) and transmission electron microscopy (TEM). Normal rabbit cornea, conjunctiva, and oral samples were also examined for comparison. Specimens were fixed in 2.5% glutaraldehyde in 0.1 M PBS, washed three times for 15 minutes in PBS, and postfixed for 2 hours in 2% aqueous osmium tetroxide. They were washed three more times in PBS before being passed through a graded ethanol series (50%, 70%, 80%, 90%, 95%, and 100%). For SEM preparation, specimens were transferred to hexamethyldisilazane (TAAB Laboratories Equipment, Ltd., Aldermastron, UK) for 10 minutes and allowed to air dry. When dry, specimens were mounted on aluminum stubs and sputter coated with gold before examination in a digital scanning electron microscope (JSM 5600; JEOL, London, UK). For TEM, the specimens were embedded in epoxy resin (Agar 100; Agar Scientific, Ltd., Stansted, UK). Ultrathin (70 nm) sections were collected on copper grids and stained for 1 hour with uranyl acetate and 1% phosphotungstic acid and then for 20 minutes with Reynolds’ lead citrate before examination by transmission electron microscope (JEM 1010; JEOL). 

Results

Cultivation of Oral Epithelial Cells

Epithelial cells from the oral mucosa began to form colonies on the denuded AM within 3 days. After 10 days in culture, a confluent primary culture of oral epithelial cells had been established that covered the whole AM (Fig. 2A) . At 3 weeks, the cultivated oral epithelial cells showed four to five layers of stratification, were well differentiated (Fig. 2B) , and appeared very similar to both normal corneal epithelium (Fig. 2C) and the corneal epithelium cultivated using our established technique (Fig. 2D) . ^{22 23 24 25 26 27}  

Immunohistochemistry of Keratins

The patterns of expression of the keratins in the cultivated oral epithelium were investigated with immunohistochemistry. Negative control sections, incubated with normal mouse and rabbit IgG, and omission of the primary antibody exhibited no discernible specific immunoreactivity over the entire region, except that there was slight nonspecific immunoreactivity for keratin-12 in the subepithelial regions of the cornea, conjunctiva, and oral mucosa. However, there was no nonspecific immunoreactivity in the epithelial regions. The immunoreactivity observed in each specimen was compared with the control. Immunohistochemistry showed the presence of keratins (4/13 and 3) in the cultivated oral epithelial cells (Figs. 3A3 3A4 3A5) . The keratin-4 and -13 pair were expressed in the superficial and intermediate layers, with no discernible immunostaining in the basal cell layers. In contrast, keratin-1, keratin-10, and cornea-specific keratin-12 were not expressed in any layers of the oral epithelial sheets (Figs. 3A1 3A2 3A6) . Cornea-specific keratin-3 was expressed in all epithelial layers (Fig. 3A5) . 

Immunohistochemical examination of the rabbit normal corneal, conjunctival, and oral epithelial cells showed that keratin-1 and -10 were not expressed in any layers of the corneal (Figs. 3B1 3B2) , conjunctival (Figs. 3C1 3C2) , and oral epithelial cells (Figs. 3D1 3D2) . Keratin-4 was expressed in the superficial layer (Figs. 3B3) , and keratin-3 and -12 were expressed in all epithelial layers of the corneal epithelium (Figs. 3B5 3B6) . Keratin-4 and -13 were expressed in the superficial and intermediate layers of the conjunctiva (Figs. 3C3 3C4) . These keratins were expressed in all epithelial layers of oral epithelium (Figs. 3D3 3D4) , and, finally, cornea-specific keratin-3 was expressed in the oral epithelial cells (Figs. 3D5) . No keratin-12 immunoreactivity was found in conjunctival (Figs. 3C6) and oral epithelium (Figs. 3D6) . 

Ultrastructural Features of the Epithelium

SEM examination of the cultivated oral epithelial cells revealed a continuous layer of flat, squamous, polygonal epithelial cells (Fig. 4A) . These cells appeared healthy and well formed with distinct cell boundaries and ranged from 12 to 90 μm in size, the average exposed cell surface diameter was 28 ± 10.29 μm (SD). Some of the oral epithelial cells appeared to be in the process of desquamating (Fig. 4B) . They were closely attached to each other with tightly opposed cell junctions and distinct cell boundaries (Fig. 4C) . The apical surface of the cells was covered with microvilli (Fig. 4D) . 

TEM examination of the oral epithelial culture sheet showed that the cells produced five to six layers of well-stratified epithelium (Fig. 5A) , appeared healthy, and were differentiated into basal columnar cells, suprabasal cuboid wing cells, and flat squamous superficial cells (Figs. 5A 5B) . The epithelial cells in the basal cell layers were columnar (Fig. 5C) , and recently divided cells provided evidence of rapid proliferation in the basal cell layers of the epithelial sheet (Fig. 5D) . The basal epithelial cells adhered well to the AM substrate with hemidesmosome attachments and produced basement membrane material (Fig. 6A) . In all cell layers, the epithelial cells were comparatively closely attached to neighboring cells by numerous desmosomal junctions (Fig. 6B) , and in the superficial cell layer what appeared to be tight junctions were evident between neighboring cells (Fig. 6C) . The apical surface of the most superficial cells was covered with a glycocalyx-like material (Fig. 6D) . 

SEM examination of the normal corneal, conjunctival, and oral cells showed that the surface morphology of the corneal epithelial cells was different from that of the other cell types. The corneal epithelial cells revealed a continuous layer of flat squamous polygonal epithelial cells with an average size of 28.4 ± 10.0 μm with distinct surface microvilli (Fig. 7A) . The conjunctiva consisted of smaller polygonal epithelial cells with an average size of 12.3 ± 2.3 μm mixed with numerous goblet cells (Fig. 7C) . The oral mucosal samples (Fig. 7E) cells had an average size of 27.5 ± 8.4 μm, and their apical surfaces had long parallel ridge-like structures quite different from the microvilli on the corneal epithelial cells. 

TEM examination of the normal corneal, conjunctival, and oral cells showed that the corneal epithelial cells could be easily distinguished from the other cell types by their ultrastructure. The corneal epithelial cells consisted of five to six layers of well-stratified epithelium (Fig. 7B) , appeared healthy, and were differentiated into basal columnar cells, suprabasal cuboid wing cells, and flat squamous superficial cells. The rabbit conjunctiva could be easily identified by the presence of numerous goblet cells (Fig. 7D) . The oral samples (Fig. 7F) had many more cell layers than were present in the normal corneal epithelial cell layer. 

Autologous Transplanted Epithelium

At 3 to 4 weeks after the ocular surface injury, conjunctival epithelium completely covered the damaged corneal surface in all eight rabbits with considerable neovascularization and subconjunctival inflammation evident (Figs. 8A1 8B1 8C1 8D1) . The extent of injury was similar in all animals. After the removal of conjunctivalized tissue, we reconstructed the ocular surface with a cultivated oral epithelial sheet on AM (day 0). No signs of infection, bleeding, or sheet detachment were observed. In the early stages (day 2) after transplantation, the eight eyes that had received 12-mm diameter discs of AM containing autologous cultivated oral epithelial cells all possessed an epithelialized area (Figs. 8A2 8B2 8C2 8D2) . Most of the area covered with cultivated oral epithelial cells was not stained by fluorescein and was separated from the outer rim of healing conjunctiva by an annular epithelial defect, which stained with fluorescein. At 10 days after transplantation, the area covered by the epithelium had expanded outward and was connected with healing conjunctival epithelium in some areas (Figs. 8A3 8B3 8C3 8D3) . Moreover, the corneal surfaces of all eyes were clear and smooth, and the entire corneal surfaces were completely covered with transplanted autologous oral epithelium (Figs. 8A4 8B4 8C4 8D4) . The control animals that had received no transplant (Fig. 8E) or had received acellular AM (Fig. 8F) did not show any evidence of epithelialization at day 10 (Figs. 8E3 8F3) . The control animals that had undergone direct transplantation of oral tissue showed considerable neovascularization and inflammation, resulting in the failure of ocular surface reconstruction (Fig. 8G) . 

Histologic examination of transplanted sheets at 10 days after surgery revealed that the sheets adhered well to the host corneal stroma with no evidence of subepithelial cell infiltration or stromal edema. Superficial cells of the transplanted sheets had nuclei, indicating that they were indeed nonkeratinized mucosal epithelial cells (Fig. 9) . 

Discussion

Severe ocular surface diseases such as SJS and OCP are some of the most challenging problems that the clinician faces today. Conventional management is generally unsatisfactory, and the long-term ocular consequences of these conditions are devastating. In the past 10 years, surgical reconstruction of the ocular surface has been greatly advanced by the introduction of limbal epithelial transplantation, AM transplantation, and cultivated corneal epithelial transplantation. ^{12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27} We have since adopted these procedures for clinical use in severe ocular surface diseases and have achieved successful ocular surface reconstruction. ^{24 25} However, we normally used grafts transplanted from donors (allografts), putting the recipients at risk of rejection and necessitating the use of prolonged immunosuppression. Although it has been reported that transplantation of the patient’s own limbal epithelial cells (autograft) is a simple and effective method of reconstructing the corneal surface with unilateral severe ocular surface disease, ^{19 20} we normally cannot perform this procedure because diseases such as SJS and OCP involved complete loss of corneal epithelial stem cells and conjunctival keratinization in both eyes. The problem of allograft rejection is the main reason that we decided to develop a new method of autologous mucosal epithelium transplantation for corneal surface reconstruction. 

In the field of oral surgery, many attempts have been made to reconstruct the oral cavity by skin grafting, oral mucosal grafting, and transplantation of cultured oral mucosal sheets. ^{30 31 32 33 34 35} Epithelial cells isolated from the oral mucosa are generally thought to be at a lower stage of differentiation than skin keratinocytes and offer the following potential advantages: short cell turnover time with resultant short culture time requirement and long-term maintenance under culture conditions without keratinization. ^{36 37} Moreover, the oral mucosa is an ideal location for tissue biopsy, because the resultant scar is inconspicuous. These characteristics suggest that oral epithelial cells could be an ideal substitute for corneal epithelial cells for use in ocular surface reconstruction. 

Cultivated corneal epithelial transplantation has long been a dream of ophthalmologists. In recent years, we have successfully developed a surgical system using cultivated corneal epithelial transplantation. ^{22 23 24 25 26 27} Our system for culturing is unique in several important respects, which include the use of denuded AM as a carrier and a 3T3 fibroblast layer to assist epithelial cell growth. ²³ We have also discovered that air-lifting is crucial for the correct formation of epithelial tight junctions and epithelial cell stratification. ²⁷ We have adopted this system for culturing oral epithelial cells (with some modifications) and have successfully generated confluent cultures of cells, on devitalized human AM, which had been expanded ex vivo from biopsy-derived rabbit oral mucosal tissue. The superficial cells of the cultivated oral cell layer had nuclei, indicating that they were indeed nonkeratinized mucosal epithelium. 

Cytokeratins play an important structural and protective role in maintaining the integrity of the epithelium of the anterior segment of the eye. ^{38 39 40 41} In vivo, cytokeratin filament systems are composed of type 1 (neutral-basic) and type 2 (acidic) obligate heterodimers that exist as specific pairs. ⁴¹ Defined subsets of individual cytokeratin pairs are characteristically expressed, depending on epithelial cell tissue type and level of differentiation. ^{42 43 44} We used immunohistochemistry in this study to demonstrate that the keratin-1 and -10 pair, which is involved in the physiological keratinization process in the epidermis, is not expressed in any layers of the cultivated oral epithelial sheet or in corneal, conjunctival, and oral epithelial cells. We also found that the keratin-4 and -13 pair, which is observed in nonkeratinized, stratified epithelia, is expressed in the superficial and intermediate layers of the cultivated oral epithelial cells, with no discernible immunostaining in the basal cell layers. These keratins were also observed in the conjunctival and oral epithelial cells. These results lead us to believe that the oral epithelial cells cultivated on AM have the characteristics of nonkeratinized mucosa, not of keratinized mucosa. We also observed the general absence of other keratinization-related proteins, such as involucrin and filaggrin (data not shown), which supports our conclusions. Immunohistochemical examination revealed no cornea-specific expression of keratin-12 in any layers of the cultivated oral epithelial sheets or in conjunctival and oral epithelial cells, whereas cornea-specific keratin-3 was expressed in all epithelial layers of the cultivated oral epithelial sheet and oral mucosa. Keratin-3 reportedly, is a reliable marker for corneal differentiation ⁴⁵ and is positive for epithelial cells of the cornea, nose, and some oral mucosa, ^{46 47} and our results are consistent with this. Although the oral epithelial cells cultivated on AM could not become corneal epithelial cells, we suggest that they have the potential ability to become cornealike epithelial cells under our culture conditions. 

Our electron microscopy results from this study are of particular interest. SEM examination revealed that rabbit oral epithelial cells, after 3 weeks in culture, appeared healthy and well formed with tightly opposed cell junctions. Some superficial cells appeared to be undergoing the process of desquamation, as would be expected in a healthy stratified epithelial sheet. The cultivated oral cells were similar in size and appearance to rabbit corneal epithelial cells, ²² but were quite different from rabbit conjunctiva and oral mucosa. TEM confirmed that the cultivated oral epithelial sheet was very similar in appearance to that of the corneal epithelium and very different from the conjunctiva and oral mucosa. Similar to the corneal epithelium, it had four to five layers of stratified cells that were differentiated into columnar, wing, and squamous cells. We observed the evidence of rapid proliferation in the basal columnar cells. Both our SEM and TEM results show clearly that our oral mucosal cells, cultivated on AM, resembled normal corneal epithelial cells more closely than any other cell type. 

An important question regarding the growth of oral epithelial cells on AM is how the basal cells attach to the underlying AM and whether a normal barrier function develops in the superficial cells. We believe that these are key points for the successful transplantation of the cultivated oral epithelial sheet. The normal corneal epithelial cells have specialized junctions on their cell surfaces to ensure firm adhesion to neighboring cells and the extracellular matrix below. ⁴⁸ Desmosomal junctions are present between the cell-to-cell surfaces and give the cell sheet structural integrity. ^{48 49 50} Hemidesmosomes present on the basal cell surfaces serve to attach the basal cells to the basement membrane and, as is true of desmosomes, are also linked to the intermediate filament scaffold. ^{48 49 50} Tight junctions are present between the superficial cells, form an impermeable barrier, and are essential for normal functioning of the corneal epithelium. ^{51 52 53} Our TEM results show that cultivated oral epithelium is attached to a basement membrane with hemidesmosomal junctions. Adjacent cells in the cultivated oral sheet are also joined with numerous desmosomal junctions, and what appear to be tight junctions are evident between the most superficial cell layers. From these results, we believe that oral epithelial cells cultivated on AM have junctional specializations similar to those of in vivo corneal epithelial cells. These findings encouraged us to perform the transplantation of cultivated oral epithelial cells on AM. 

Another key point for success in performing cultivated oral epithelial sheet transplantation is how the most superficial cells contact the tear-ocular surface interface. The anterior surface of the normal corneal epithelium has numerous folds in the anterior epithelial cell membranes in the form of microvilli and microplications, together with a glycocalyx layer. ^{54 55} The role of these surface irregularities may be to increase cell surface area and therefore to aid in intra- and extracellular movement of nutritional and waste products across the cell membrane, in addition to stabilizing the corneal tear film that is essential for good vision. Examination by SEM revealed that the apical surface of the cultivated oral epithelial cells is covered with numerous microvilli, almost identical with those found on corneal epithelial cells. We also found evidence of a cell surface glycocalyx, similar in appearance to the glycocalyx present on the surface of corneal epithelial cells. The apical surface of the corneal epithelium is normally covered by the tear film, which has a variety components, including the mucus layer on the epithelial surface. ^{56 57} Mucus is a highly hydrated gel, the most prominent components of which are glycoproteins termed mucins. The mucin MUC1 is the principally membrane-associated mucin and has been identified in the glycocalyx of many epithelium-lined organs including stomach, esophagus, oral mucosa, and ocular surface. ^{58 59} Our finding of a glycocalyx layer on the cultivated oral epithelium led us to consider the interesting hypothesis that this glycocalyx might act as a substitute for the glycocalyx on the ocular surface epithelium. 

After the successful culture of rabbit oral epithelial cells on AM, we tried to reconstruct the damaged corneal surfaces by transplantation of autologous cultivated oral epithelial cells, to test the viability of using these cells as a substitute for cultivated corneal epithelial cells. Two days after the autologous surgical transplantation of oral epithelial cells, most of the corneal surfaces on which cultivated oral epithelium had been placed were free of epithelial defects. This indicated the complete survival of the transplanted oral epithelial cells. Moreover, the transplanted grafts were surrounded by a conjunctival epithelial defect at 360°, suggesting that there was no contamination of host conjunctival epithelium. Ten days after transplantation, the area covered by the oral epithelium had expanded outward and was in contact with healing conjunctival epithelium in some areas. This suggests that, in the early days after oral epithelial transplantation, the oral epithelial cells on AM survive and spread onto the adjacent keratectomized cornea. We examined sections of these corneas, by the periodic acid-Schiff (PAS) reaction, which is reactive to conjunctival goblet cells, and confirmed that there was no contamination of the host conjunctival epithelium (data not shown). The corneal surfaces of all eyes were clear and smooth, and the entire corneal surfaces were completely covered with transplanted autologous oral epithelium. In addition, all rabbits could follow moving objects. We also observed, as expected, that the control animals that received no transplants or transplants of acellular AM onto a keratectomized cornea did not show any evidence of epithelialization at day 10. Finally, we clearly showed that direct transplantation of oral mucosal tissue cannot be used for ocular surface reconstruction. 

To the best of our knowledge, our study is the first to demonstrate the survival, on keratectomized corneas, of autologous epithelial cells derived from oral biopsy tissue and grown on human AM. We believe that this approach to treating severely damaged eyes with limbal stem cell deficiencies has great potential. However, development of this technique is at an early stage, and there are many questions that must be resolved, including questions about the longevity and mobility of the autologously transplanted oral epithelial cells on the host eye and whether the transplants contain stem cells or progenitor cells. It is very important to determine whether the oral mucosal cell type has the appropriate characteristics to act as a substitute for the corneal epithelium. We must also determine in which cases cultivated oral epithelial transplantation should be used, the best time to perform the grafts, and whether this procedure is superior to existing surgical techniques. We are currently investigating the characteristic of cultivated oral epithelial cells and comparing the gene expression profile of oral epithelial cells with that of corneal epithelial cells. We hope that this will help to answer some of these questions. 

In conclusion, we have successfully generated on AM confluent cultures of oral epithelial cells expanded ex vivo from biopsy-derived oral mucosal tissues. We have successfully performed autologous transplantation of these cells onto keratectomized rabbit corneas. Finally, we believe that autologous transplantation of cultivated oral epithelium is a feasible method for ocular surface reconstruction, although, because the long-term outcome of such transplantation is not yet clear, its feasibility for clinical use should be evaluated further. 

 Supported in part by Grants-in-Aid for Scientific Research from the Japanese Ministry of Health, Labor, and Welfare and the Japanese Ministry of Education, Culture, Sports, Science and Technology and by research grants from the Kyoto Foundation for the Promotion of Medical Science, the Intramural Research Fund of Kyoto Prefectural University of Medicine, the Wellcome Trust, and the TFC Frost Trust.

 Submitted for publication February 26, 2002; revised July 18, 2002; accepted July 23, 2002.

 Commercial relationships policy: N.

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

The precise placement of the AM, the sutures, and the soft contact lens is shown. The cultivated oral epithelial sheet was sutured at the corneal-limbal interface with 10-0 nylon sutures (eight sutures were always used). The soft contact lens was sutured to the bare sclera with 10-0 nylon sutures (four were always used). These sutures were not secured to the cultivated oral epithelial sheet.

Table 1.

 

Primary Antibodies and Source

Antibodies	Category	Dilution	Source
Cytokeratin-1	Mouse monoclonal	×20	YLEM srl
Cytokeratin-10	Mouse monoclonal	×100	Biomeda Corp., Foster City, CA
Cytokeratin-3	Mouse monoclonal	×50	Progen Biotechnik GMBH, Heidelberg, Germany
Cytokeratin-12	Rabbit polyclonal	×200	Provided by Kurpakus M, et al. 60
Cytokeratin-4	Mouse monoclonal	×10	ICN Pharmaceuticals, Inc., Costa Mesa, CA
Cytokeratin-13	Mouse monoclonal	×1	American Research Products, Inc., Kensington, MD

Figure 2.

 

A confluent primary culture of oral epithelial cells taken from the oral tissue after 10 days in culture (A). Light micrographs showing cross-sections of the cultivated oral epithelial cells on AM (B), normal corneal epithelial cells (C), and cultivated corneal epithelial cells on AM (D), stained with hematoxylin and eosin. The cultivated oral epithelial sheet had four to five layers of stratified, well-differentiated cells and appeared very similar to both normal and cultivated corneal epithelium. Original magnification: (A) ×100, (B, C, D) ×400.

Figure 3.

 

Representative immunohistochemical staining of keratins-1 (A1–D1), -10 (A2–D2), -4 (A3–D3), -13 (A4–D4), -3 (A5–D5), and -12 (A6–D6) in cultivated oral epithelial cells on AM and on normal corneal, conjunctival, and oral mucosa. Keratin-1 and -10 were not expressed in any layers of the cultivated oral epithelial cells (A1, A2) or in the normal corneal (B1, B2), conjunctival (C1, C2), or oral epithelial cells (D1, D2). In contrast, keratin-4 and -13 were expressed in the superficial and intermediate layers of the cultivated oral epithelial cells (A3, A4) and conjunctival epithelial cells (C3, C4). These keratins were expressed in all epithelial layers of oral mucosa (D3, D4), and only keratin-4 was expressed in the superficial layer of the corneal epithelial cells (B3). Keratin-3 was expressed in all epithelial layers of the cultivated oral sheet (A5) and the corneal (B5) and oral mucosa (D5), whereas cornea-specific keratin-12 immunostaining was not found in the cultivated oral epithelial cells (A6) or in normal conjunctival (C6) and oral epithelial cells (D6), but was found in normal corneal epithelial cells (B6). Scale bar: (A–C) 100 μm; (D) 200 μm.

Figure 4.

 

Scanning electron micrographs of cultivated mucosal epithelial cells on denuded AM at low magnification (A, B) and at high magnification (C, D). The cells appeared healthy and well formed with distinct cell boundaries (A). In places, desquamating cells were also found (B). The cells appeared to be in good condition and were closely attached to each other with tightly fitting cell junctions (C). The apical surface of the cells was covered with microvilli (D).

Figure 5.

 

Transmission electron micrographs of rabbit cultivated oral epithelium on AM. The culture formed five to six layers of healthy, well-stratified epithelial cells (A, B). The epithelial cells in the basal cells layers were columnar (C). There was some evidence of basal cell proliferation (D). Scale bars, 2 μm.

Figure 6.

 

High-magnification transmission electron micrographs of rabbit cultivated oral epithelium on AM. The cells were attached to the basement membrane with hemidesmosomal junctions (arrows), and basement membrane extracellular matrix was evident (A). Adjacent cells were joined with numerous desmosomal junctions (arrowheads) (B). Apparent tight junctions (⋆) were occasionally evident between the most superficial cell layers (C). A glycocalyx layer was observed in the apical surfaces (D). Scale bars: (A, C) 500 nm; (B, D) 200 nm.

Figure 7.

 

Scanning electron micrographs of normal rabbit corneal (A), conjunctival (C), and oral epithelial cell layers (E), showing the surface morphology of these cells. The superficial normal rabbit corneal epithelial cell layer was the most similar morphologically to the superficial oral mucosal epithelial cell layer cultivated on AM, as shown in Figure 4 . Transmission electron micrographs of normal rabbit corneal (B), conjunctival (D), and oral epithelial cell layers (F) showing the ultrastructure of these cell layers. The normal rabbit corneal epithelial cell layer clearly had the ultrastructure most similar to that of the oral mucosal epithelial cell layer cultivated on AM, as shown in Figures 5 and 6 . Scale bars: (B, D, F) 2 μm.

Figure 8.

 

Representative slit lamp photographs of eyes of four rabbits taken before transplantation (A1–D1), 48 hours after transplantation with fluorescein (A2–D2), 10 days after transplantation with fluorescein (A3–D3), and without fluorescein (A4–D4). Before transplantation, all eyes showed total limbal stem cell destruction (A1–D1). Forty-eighthours after surgery, most of the corneal surfaces were covered with transplanted cultivated oral epithelial cells, which showed no fluorescein staining (A2–D2). The epithelialized, non-fluorescein-stained area wasclearly separated from the surrounding conjunctival epithelium by an annular epithelial defect. Ten days after surgery, the central epithelializedarea had spread outward (A3–D3). The epithelium at this time covered almost the entire corneal surface and in some areas was in contact with the inner part of the healing conjunctival epithelium. At this time, the corneal surface of all rabbits was covered with clear oral epithelium (A4–D4). Three representative slit lamp photograph series of a control eye that received no transplant (E1–E4), and then a transplant of acellular AM (F1–F4), and finally a transplant of oral mucosal tissue onto the keratectomized cornea in the limbal-corneal area (G1–G4). The time course of photographs 1 through 4 is the same as for those in (A) through (D).

Figure 9.

 

Light micrograph of cultivated oral epithelium 10 days after transplantation. The transplanted grafts adhered well to the host corneal stroma with no evidence of subepithelial cell infiltration or stromal edema. Superficial cells of the transplanted grafts had nuclei, suggesting that they were nonkeratinized mucosal epithelial cells. (⋆) Amniotic stroma. Original magnification, ×100.

The authors thank Michelle A. Kurpakus from the Department of Anatomy and Cell Biology, Wayne State University School of Medicine, Detroit, Michigan, for providing the polyclonal antibody to cytokeratin-12.