Regeneration of Corneal Epithelium With Dental Pulp Stem Cells Using a Contact Lens Delivery System

Abstract

Purpose: The corneal epithelium is sloughed off surface of the eye by the action of blinking and is continually replaced by division and maturation of the limbal stem cells (LSCs). In the case of injury or disease, LSCs can be lost or damaged to a point at which the corneal epithelial layer is no longer maintained. leading to LSC deficiencies (LSCDs). When this occurs, the opaque conjunctiva overgrows the anterior surface of the eye, leading to vision impairment or loss. Dental pulp stem cells (DPSCs) are promising candidates as autologous LSC substitutes. In this study, contact lenses (CLs) are used as a novel medical device to deliver DPSCs onto corneal surface to enhance corneal epithelium regeneration.

Methods: Dental pulp stem cells labeled with green fluorescent Qtracker 525 were seeded onto the pretreated CLs, allowed to adhere, then delivered to debrided human corneas. Expression of KRT3, 12, 13, and 19 was investigated by immunostaining, then standard and confocal microscopy.

Results: Dental pulp stem cells were successfully isolated, labeled, and delivered to the corneal surface using CLs. Following removal of CLs, confocal microscopy showed that the DPSCs had migrated onto the cornea. Coexpression of KRT12 and green fluorescent Qtracker 525 confirmed that the DPSCs had transdifferentiated into corneal epithelial progenitors. Delimitation of KRT 19 and green fluorescence provides evidence that Qtracker 525-labeled DPSCs establish a barrier to the invasion of the cornea by conjunctiva.

Conclusions: In this study we show that DPSCs, delivered using CLs, can be used to enhance repair and regeneration of the human corneal epithelium.

The cornea plays a crucial role in vision and is responsible for two-thirds of total refractive power of the eye. Corneal epithelium consists of five to six layers of cells (50 μm) and together with the tear film is the first barrier against outside environment.¹ These epithelial cells maintain transparency and are unique to the cornea. As with other tissues, the cornea has its own stem cell population, the limbal stem cells, which reside within a protective niche, the limbus. The primary function of limbal stem cells is to maintain the integrity and transparency of the cornea.^2–4 

In the event of injury or disease affecting the limbus, corneal epithelial renewal may be impaired or lost due to limbal stem cell deficiency (LSCD). This leads to corneal vascularization, conjunctivalization, scarring, and pain,^3,5,6 which will dramatically decrease quality of life and can lead to blindness. 

Stem cells have presented a great opportunity for tissue regeneration and therapies⁷ and can be obtained from various sources, such as the limbus, skin, bone marrow, oral mucosa, and many others. In 2000, a new source of stem cells was discovered in human adult tooth pulp, these are dental pulp stem cells (DPSCs).⁸ Their discovery holds particular promise for orofacial tissue regeneration^9–11 because they originate from craniofacial tissue derived from neural crest cells that have undergone an epithelial-mesenchymal transition. Dental pulp stem cells are easily accessible by tooth extraction (wisdom, ectopic tooth, or even decayed tooth) or root canal surgery; they can be cryopreserved and biobanked to serve as a source of autologous stem cells for future use.¹² 

Various markers were identified to support the suitability of mesenchymal stem cells for terminal differentiation into target cells. Cytokeratin 3 and 12 are markers of corneal epithelium and also referred to as markers of differentiation.^13,14 In vivo studies^15,16 reported traces of coexpression of cytokeratin 3 and 12 in DPSCs of rabbits and this supports the concept that DPSCs are promising candidates for enhancement of corneal epithelial regeneration. Cytokeratin 13 and 19 are both markers of conjunctival epithelium and as such were used in this study to signify the extent of conjunctivalization.^17–21 

Contact lenses (CLs) are hugely popular, mainstream medical devices used to improve vision. They are also routinely used as bandages to aid corneal healing due to disease or surgery.^22,23 Scleral lenses have been used as carriers of adipose-derived stem cells in a rabbit corneal epithelium repair model²⁴ and soft CLs have been used to deliver expanded epithelial cells to the cornea.²⁵ These characteristics make CLs ideal candidate carriers for the transfer of DPSCs or transdifferentiated DPSCs to an injured cornea. 

In this study, standard soft CLs were used to deliver in vitro expanded DPSCs to an ex vivo model system comprising human donor corneas in organ culture. 

Methods

Human Corneas

Human corneal tissue was obtained from the Manchester Eye Bank (Manchester, UK). All tissue was consented for research and had been deemed unsuitable for transplantation due to low endothelial counts (< 2200 cells mm²). On arrival at the laboratory, corneas were transferred into 12-well tissue culture plates (Corning Ltd., Ewloe, Flinshire, UK) with culture medium (Dulbecco's modified Eagle medium (DMEM; 4500 mg/L glucose) supplemented with 10% (vol/vol) fetal bovine serum (FBS; Latch International, Uckfield, UK), 2 mM L-glutamine (PAA Laboratories Ltd., Yeovil, Somerset, UK) and 1% (vol/vol), penicillin/streptomycin stock solution (PAA Laboratories) and incubated at 37°C in 5% CO₂. In this study, 20 corneas were used; the details of these are listed in Table 1. The study conformed to the tenets of the Declaration of Helsinki. 

Table 1

 

Human Donor Corneas Used in This Study

 

Dental Pulp Stem Cell Isolation

Ethical approval was obtained to harvest DPSCs from the extracted teeth of consented patients undergoing tooth extraction at Manchester Dental Hospital and Manchester Royal Infirmary (Integrated Research Application System Project ID: 106973 REC Ref: 13/LO/0631). The pulp tissue was transported at 37°C to the laboratory in thermally insulated containers (Labcold Ltd., Basingstoke, Hampshire, UK) for further processing. In this study five cell lines from three patients were used (Table 2). 

Table 2

 

Patients and Cell Lines Used in the Study

 

Consented patients donated intact teeth following tooth extraction under local and/or general anaesthesia. The teeth were debrided using sterile saline and gauze swabs, then fractured within a sterile stainless steel mortar by striking with a sterile stainless steel pestle and mallet (Fig. 1A). The pulp was carefully extracted from the pulp chamber using an endodontic k-file (Pro-Fit 25 mm, ISO Size 30; QED Ltd., Peterborough, UK), endodontic probe, and forceps. The tooth pulp was aseptically sectioned into approximately 1-mm³ fragments and seeded into an organ culture dish (Corning Life Sciences, Corning, NY, USA) with culture medium containing 2.5 μg mL⁻¹ amphotericin B antimycotic (Fungizone Gibco by Life Technologies, Inc., Paisley, UK) and incubated at 37°C in CO₂ 5% (Fig. 1B). 

Figure 1

 

Preparation of the dental pulp, cell culture, and characterization. The tooth was cleaned and washed in sterile saline, fractured, and the pulp extracted using a 25-mm K-file (A). The pulp was dissected into 1-mm³ cubes and seeded into an organ culture dish. Within 48 hours of culture, DPSCs were emerging from the tissue (B). Dental pulp stem cells were induced to differentiate along the adipo-, chondro-, and osteocyte lineages (C–E). The DPSCs were immunostained with fluorescent antibody against the stem cell marker CD44 (F). Dental pulp stem cells were labeled with Qtracker 525 and seeded onto CLs (G, H). Scale bars: 100 μm (C), 200 μm (B, E, F, H), and 400 μm.

Dental Pulp Stem Cell Characterization

Once DPSCs cultures were established, they were differentiated along adipogenic, chondrogenic, and osteogenic lineages to show multipotency.^26–29 After 15 days, lineage-specific histologic staining was performed and cultures were examined by light microscopy (Figs. 1C–E). 

Cell Maintenance

Confluent cell cultures were washed twice with HBSS (Sigma-Aldrich Co. Ltd., Poole, Dorset, UK) and trypsinized (Life Technologies, Inc.) for 5 minutes at 37°C, CO₂ 5%. Trypsin activity was quenched with culture medium and cells recovered by centrifugation (90g). Cell pellets were gently resuspended, and the cells counted and sized using a Millipore Scepter portable cell counter (Merck Millipore, UK Ltd., Watford, UK). Aliquots of 5 × 10⁵ cells were seeded into 75-cm² flask (Corning Ltd.) for expansion; culture medium was changed every 3 days. 

Proliferation Assay

Dental pulp stem cells were seeded into 24-well plates at 1 × 10⁴ cells per well and cultured for 24 hours. The medium was aspirated and replaced with saline (control; 0.9% [wt/vol] NaCl) or borate buffer saline (BBS; CL storage solution). After 1 hour of exposure at 37°C in 5% CO₂, the cultures were washed with PBS, and 350 μL of 20% (vol/vol) CellTiter 96 AQueous One Solution Reagent (Promega UK, Southampton, UK) in phenol red-free DMEM (Sigma-Aldrich Co. Ltd.) were pipetted into each well followed by incubation for a further 1 hour at 37°C under 5% CO₂. To measure the absorbance of the resulting colorimetric formazan product, 100 μL of solution from each well were pipetted in triplicate into a 96-well plate, and absorbance was recorded at 490 nm using a Biochrom Asys UVM 340 Microplate reader (Biochrom, Cambridge, UK). Data were normalized to the saline control and expressed as a percentage versus saline ± SEM. 

Qtracker Cell Labeling

At an early passage (at latest p2), cells were labeled with the Qtracker 525 Cell Labelling Kit (Life Technologies, Inc.) as per the manufacturer's protocol. Labeled cells were seeded into a 25-cm² flask (Corning Ltd.) and fluorescence was checked after 1 day in culture. Proliferation of the labeled DPSCs was compared with that of the nonlabeled DPSCs as per proliferation protocol described above. Data were normalized to the control, and expressed as a percentage versus nonlabeled cell group ± SEM (Fig. 2). 

Figure 2

 

Proliferation and attachment assays. The proliferation rate of Qtracker-labeled DPSCs versus nonlabeled cells was assessed. The proliferation rate was higher for labeled cells with P value of 0.0436 (A). The effect of BBS versus saline (0.9% NaCl) on proliferation rate was also assessed. Borate-buffered saline was found to decrease proliferation rate of cells, with P < 0.0001 (B). An attachment assay provided evidence to demonstrate that the PBS and FBS washing steps increase cell attachment; P value 0.0383 (C). Fetal bovine serum was found to increase cell attachment and growth (D).

Contact Lens Preparation and DPSCs Seeding

Bausch and Lomb UK Ltd. (Kingston-upon-Thames, Surrey, UK) PureVision2 CLs were incubated in PBS at room temperature for 72 hours with PBS changes every 24 hours. Then CLs were incubated in 100% FBS for 3 hours before cell seeding. Fetal bovine serum was selected in preference to human serum as its use resulted in increased cell attachment (Fig. 2). The PBS wash step was introduced as the borate storage buffer in which the CLs are supplied was found to have a cytotoxic effect (Fig. 2). The CLs were incubated in 24-well plates and handled with fine stainless sterile forceps to prevent distortion. A cell viability assay (Life Technologies, Inc.) was performed after the DPSCs were seeded onto CLs (Supplementary Fig. S1). 

Labeled DPSCs were suspended in fresh culture medium and 1 × 10⁵ cells seeded onto the corneal contacting side of a CL in a maximum of 150 μL of culture medium. A maximum of 200 μL of culture medium was added outside of the corneal contacting side of the lens. Cell attachment and fluorescence were checked microscopically after 24 hours of incubation at 37°C in 5% CO₂. 

Preparation of Human Donor Corneas

Human donor corneas were incubated in 6-well plates with DMEM before CL transfer. The corneas were debrided of epithelium using a scalpel (verified by hematoxylin-eosin staining to confirm that the Bowman's layer was not damaged, data not shown) and the culture medium aspirated. After removal of debris, CLs preseeded with DPSCs were transferred onto corneas and fresh culture medium added to submerge CL margins. The cornea kept its shape and maintained corneal surface tension in the culture dish due to the presence of the ridged scleral rim and an air bubble trapped below the posterior of the cornea (see Supplementary Fig. S2). Following 5 days of incubation at 37°C in 5% CO₂, the CLs were removed and both corneas and CLs were analyzed using laser confocal microscopy (Nikon UK Ltd., Kingston-upon-Thames, Surrey, UK). 

Cryosectioning

All corneas were fixed with 4% (wt/vol) paraformaldehyde (PFA; Sigma-Aldrich Co. Ltd.) for 24 hours at 4°C. After washing with PBS-sucrose, the corneas were placed in KP-CryoCompound (optimal cutting temperature [OCT]) blocks (Klinipath BV, Duiven, The Netherlands) and frozen on the surface of liquid nitrogen. Frozen tissue/OCT blocks were stored at −40°C. 

Tissue sectioning was performed at −18°C to −20°C using a Bright cryostat (Bright Instruments Company Ltd, Huntingdon, UK). Ten-micron-thick tissue sections were placed onto glass slides (SuperFrost Plus, VWR, Lutterworth, Leicestershire, UK) and dried overnight at 37°C. 

Immunostaining

Corneal tissue sections, whole corneas and CLs were fixed with 4% (wt/vol) PFA in 0.1 M PBS for 24 hours (tissue and tissue sections) or 20 minutes (CLs). Corneas were washed in PBS-sucrose for 72 hours, changing the solution every 24 hours. After blocking with normal donkey serum (1:100), samples were incubated overnight at 4°C with antibodies against cytokeratins 3 (mouse), 12 (mouse), and 19 (rabbit), all 1:400 for CLs and 1:800 (cytokeratin 13, rabbit 1:100) for corneas (all Abcam, Cambridge, UK). Samples were rinsed with PBS and secondary donkey anti-rabbit or donkey anti-mouse antibodies Alexa Fluor (1:500, Invitrogen) were added for 2 hours at room temperature in the dark. The CLs and tissue sections were mounted with Vectashield with 4′,6-diamidino-2-phenylindole (DAPI) (Vector Laboratories, Peterborough, UK). Images were captured using a fluorescence microscope (Olympus BX60; Olympus, Southend-on-Sea, UK) and laser confocal microscope (Nikon C1). Images and videos were edited using ImageJ (http://imagej.nih.gov/ij/; provided in the public domain by the National Institutes of Health, Bethesda, MD, USA), Pixelmator (3.1v; Pixelmator team, Vilnius, Lithuania), and Hype Pro (Version 3.0.3 [466] Professional Edition; San Francisco, CA, USA). 

Data Analysis

All experiments were repeated at least three times with experimental and technical triplicates. Data were analyzed using GraphPad Prism version 6.0c for MAC OS X (GraphPad Software, La Jolla, CA, USA). One-way ANOVA followed by t-test was performed to compare samples. A P value less than 0.05 was considered statistically significant. The data shown in Figures 2A and 2B were normalized to the control, and expressed as a percentage versus the nonlabeled cell group (Fig. 2A) and the saline control (Fig. 2B) ± SEM. ImageJ was used to calculate the area of cell attachment (Fig. 2C) and the cell count (Fig. 2D). 

Results

In this study, DPSCs were isolated using an explant method. A confluent primary culture was obtained after approximately 15 days (Figs. 1A, 1B). Plasticity of DPSCs was confirmed following adipo-, chondro-, and osteogenic-differentiation (Figs. 1C–E). Further DPSC characterization was performed for CD29, CD34, CD45, and CD90 markers using FACS analysis and RT-PCR for CD44 and CD90 expression (Supplementary Fig. S3). Dental pulp stem cells were successfully labeled with Qtracker and transferred onto CL (Figs. 1G, 1H). Counterintuitively, labeled DPSCs had significantly increased proliferation rate compared with nonlabeled control DPSCs (Fig. 2A). Bausch and Lomb PureVision2 CLs were pretreated before the DPSCs were attached; this involved washing the CLs with PBS for 72 hours to eliminate the cytotoxic effect of the antimicrobial borate-based CL storage solution (Figs. 2B, 2C). The CLs were then incubated for 3 hours with 100% FBS to increase cell attachment (Fig. 2D). 

The labeled DPSCs were transferred to the corneal contacting side of the prepared CLs. Within 24 hours, cells were attached to the CLs and produced extracellular matrix, cellular clusters, and spheroids. Appropriate cell numbers and volume of medium were critical to prevent the CL from drying or DPSCs escaping corneal contacting side area. Once DPSCs were attached to corneal contacting side of the CLs (after at least 24 hours of incubation) CLs were placed cell- side down onto the debrided surface of human donor corneas. The CLs were carefully removed after 5 days of incubation, and cell migration and corneal epithelial marker expression analyzed. 

Following incubation of the cell-loaded CLs on the denuded corneas, both the corneas and CLs were fixed separately and immunostained for cytokeratins 3, 12, 13, and 19 and examined by confocal and light microscopy. Most of the green fluorescent DPSCs were seen to be attached to the debrided corneal surface (Fig. 3), with some DPSCs remaining attached to the CL (Figs. 4A–C). Immunostaining for the cytokeratins showed that the green Qtracker 525–labeled DPSCs on corneas expressed cytokeratin 3 and 12 (red), resulting in an orange coloration of the cells (Figs. 4B, 4E). At the periphery of corneas, the DPSCs appeared to restrict the intrusion of the cytokeratin 19 (a conjunctival cell marker)–positive cells onto the central debrided cornea (Figs. 4C, 4F). Cytokeratin 3 was not expressed by the green DPSCs remaining on CLs (Figs. 4A, 4D). 

Figure 3

 

Laser confocal microscopy of human donor cornea after DPSC-coated CLs removal. Four different corneas (A, E, and I are the same cornea) were cultured with DPSCs preseeded CLs. Once the CLs were removed, the corneas were imaged using a Z-stack laser confocal microscope Nikon C1 with 488-nm laser wavelength. (A–D) Superimposition of sections, (E–H) sectional view, and (I–L) volumetric view in xyz orientation. DPSCs were labeled with Qtracker 525 and are bright green. Large numbers of cells can be seen attached to the corneal surface. X and y were identical (521 μm) for all four corneas and 1273 μm. Z was 400 μm for (A), 384 μm for (B), 416 μm for (C), and 344 μm for (D). Scale bar: 300 μm.

Figure 4

 

Light fluorescent microscopy and laser confocal microscopy of CLs and corneas immunostained for cytokeratin 3, 12, and 19. Fluorescent microscopy (A–C) of CLs preseeded with DPSC after they were removed from the corneal surface and laser confocal microscopy (D–F) of corneas after the removal of the CLs; both were immunostained for KRT 3 (A, D), KRT 12 (B, E), and KRT 19 (C, F). Small islands of DPSCs were attached to CLs after removal from the corneal surface; KRT 3 was expressed by the green DPSCs on the cornea but not on the CLs (A, D), but there is clear expression of KRT 12 on both CLs (B) and the cornea (E, orange color, highlighted by arrows in [B]). At the periphery of the cornea, the green DPSCs (white crosses) were bordering the KRT 19+ cells and restricted them from migration toward the center of the cornea (C). Scale bar: 300 μm.

The corneas were sectioned and immunostained for cytokeratins 3, 12, 13, and 19 (Fig. 5). It can be seen that both cytokeratin 3 and 12 were expressed by the green Qtracker 525–labeled DPSCs (Figs. 5A, 5B), supporting the concept that DPSCs differentiated into corneal epithelial progenitor-like cells once attached to Bowman's membrane. Cytokeratin 19–positive cells and green Qtracker 525–labeled DPSCs converged at the periphery of the cornea, suggesting that the transdifferentiated DPSCs restrict the intrusion of KRT 19–positive conjunctival epithelium (Fig. 5C), whereas the control cornea (not treated with DPSCs) together with KRT 13 showed complete conjunctivalization (Figs. 5D, 5E). 

Figure 5

 

Keratin 3, 12, 13, and 19 immunostaining of corneal sections after DPSC-seeded CLs were removed and control cornea that had not been treated with DPSC-seeded CLs. It can be clearly seen that KRT 3 (A) and KRT 12 (B) red fluorescence overlapping with green fluorescence of DPSCs at the center of the cornea, and thus fluorescence is orange, although KRT 19 stain (C) is limited by green DPSCs at the periphery of the cornea. As a demonstrative control (D, E), cornea with no DPSCs underwent conjunctivalization within 5 days and thus conjunctival epithelium reaches the center of the cornea without interruptions. The bottom row shows DAPI stain corresponding to the top row (A and a, B and b, C and c). Scale bar: 200 μm.

Discussion

Corneal epithelial injury leading to a reduction in vision can dramatically affect a patient's quality of life, and often necessitates surgical intervention for restoration of vision.³⁰ Corneal epithelial cell loss can cause permanent vision impairment and often requires corneal grafting. Corneas are the most frequently transplanted tissue worldwide (>68,000 tissues supplied for keratoplasty in 2012 in the United States)³¹ and as a result of its immune-privileged status, the procedure is very successful. However, surgery will be successful only if the recipient has a functional stem cell population. This requirement limits what can be done for those patients with damaged limbal stem cell niches and who have developed LSCD. To address this, recent studies have looked at restoring the stem cell population in a hope to reestablish a functional limbus and then undertake central corneal transplantation if indicated. 

Amniotic membrane transplantation (AMT) has become a popular therapy for ocular surface damage since 1998 in the United Kingdom.³² Studies have reported that AMT can be used to treat partial LSCD on its own or to treat total LSCD by limbal allografting.² Drug therapy can be combined with AMT; these include the use of antivirals and steroid therapies to arrest inflammation and treat ocular herpes.³³ However, donor variation of the amniotic membrane, infection transmission, contradictory actions of proinflammatory cytokines (IL-6 and IL-8), and uncertainty about the exact mechanism of action of the amniotic membrane led researchers to generate a defined synthetic membrane.³⁴ The use of either amniotic or synthetic membranes requires suturing; therefore, the development of a nonsurgical alternative would be beneficial. 

Contact lenses have become more and more common not only to correct vision or aesthetics, but also to administer various therapeutic effects by drug-release potency²² or cell therapy.^24,25 For the past decade, it has been demonstrated that CLs provide an adequate reservoir for drugs and/or cells to adjust the environment of the cornea. It has been shown by DiGirolamo et al.²⁵ that soft CLs can be used as a carrier for autologous epithelial cells to aid ocular rehabilitation in patients with unilateral LSCD, although this method is impossible with bilateral LSCD. Scleral lenses have been used by Espandar et al.²⁴ to correct the inflammatory environment in the event of alkali burn in rabbits. Their study showed that scleral lenses could be used as a carrier for adipose-derived stem cell enhanced regeneration, but there were limitations as described by authors. 

Nowadays, there are many procedures available to treat corneal disease or injury and restore at least some functional vision.^2,35,36 In this study, we have demonstrated the possibility of using soft CLs as a stem cell carrier to enhance corneal epithelium regeneration. This novel method of SC delivery could provide a simple, nonsurgical alternative to current surgically based methods. The proposed protocol would simplify transplantation, thereby reducing theater time and associated costs, and could be adapted for use in the third world where corneal damage due to disease is prevalent. 

This study is focused to use CLs as a carrier for DPSCs to induce regeneration of corneal epithelial. Dental pulp stem cells can be easily accessed with tooth extraction or minimally invasive pulpectomy, without the risks associated with limbal grafting or harvesting epithelial/limbal cells from a healthy eye. The use of CLs allows stem cells to be transferred onto the cornea, and as such it is perfect for clinical application. Without CLs, suspended DPSCs would not survive/attach to the cornea due to corneal shape, tearing, and blinking. Previously, Karaöz et al.¹⁴ showed that DPSCs have the properties of epithelial stem cells. Their study demonstrated expression of several markers similar to those of bone marrow stem cells (BMSCs), and it was stated that DPSCs possess higher proliferation rate and stronger neural and epithelial stem cell properties, including the expression of corneal epithelial specific cytokeratin 12, than BMSCs. In this study, we demonstrated the possibility of using soft CLs as carriers to deliver DPSCs to enhance corneal epithelium regeneration. We showed that no additional external chemical stimulus was needed to induce the DPSCs to transdifferentiate to corneal epithelial-like cells. Once the cells had transferred from the CL to the corneal surface, they began to express the corneal-specific markers cytokeratin 3 and 12. In addition to the expression of these markers, the DPSCs were able to prevent conjunctival cells growing in the central cornea. 

The human donor corneas used in this study had been in organ culture in the Manchester Eye Bank for several weeks before use. Corneas can be released for research only after they have been accessed and found to have an endothelial count too low for transplantation, or there are medical contraindications.³⁷ During this period, the maintenance medium is not refreshed, consequently epithelium is not fully maintained and becomes thin and in some cases conjunctivalization of the cornea has occurred. Once in the laboratory and transferred into fresh culture medium, the corneal limbal stem cells are revived and produce a new epithelium,³⁸ which has to be debrided before the DPSCs can be transferred to the surface of the Bowman's membrane. In future studies, we plan to denude the limbus to determine whether it is the cells in the healthy limbus that produce the factors responsible for the induction of transdifferentiation of the DPSCs into corneal epithelial-like cells. However, it may be the unique extracellular matrix of the Bowman's layer that induces the cells to take on an epithelial-like phenotype. If it proves to be that a healthy limbus is needed for DPSC transdifferentiation, we will investigate the possibility of transdifferentiating the DPSCs into corneal epithelium first and then transfer this onto denuded corneas to enhance repair. This may be achievable by noncontact coculture of the DPSCs with a healthy donor limbal explant; once the cells have become corneal epithelial-like they can be transferred to a cornea with no functional limbus. 

It is unclear whether DPSCs need to be isolated using an enzymatic or explant method. Spath et al.³⁹ showed enhanced differentiation abilities of DPSCs isolated by the explant method compared with the enzymatic digestion method, whereas Kerkis and Kaplan⁴⁰ showed no difference between the two methods. The necessity for preselection of isolated stem cells is another area open to question. Using FACS or MACS (Miltenyi Biotec Ltd., Surrey, UK) a homogeneous culture of CD105+/CD44+/CD34+/CD45-, and so forth stem cells can be established, but the transdifferentiation success of the selected stem cell lines differs depending on the method of isolation and selection.^41–43 To date there is no single, definitive DPSC marker, therefore in this study a heterogeneous dental pulp–derived cell population was used. We have analyzed the DPSC populations used for expression of CD29, CD34, CD45, and CD90 using FACS analysis. In agreement with the International Society of Cellular Therapy (ISCT) the DPSCs attached to tissue culture plastic and were successfully differentiated along the adipo-, chondro-, and osteogenic lineage. Thus, according to ISCT, we had isolated the appropriate populations of DPSCs for the purpose of this study.⁴⁴ Although the explant protocol proved to be successful in this study, in the future we plan to use FACS-sorted DPSCs to ensure that homogeneous cell populations are used, which may further enhance corneal epithelial regeneration. 

In summary, it was demonstrated that following tooth extraction stem cells can be easily isolated from the pulp, cryopreserved, and used for corneal epithelial regeneration. Soft CLs were shown to be effective carriers for the purpose of transferring DPSCs onto the corneal surface of the human donor corneas. In the DPSC-loaded CL/human donor cornea ex vivo model, the DPSCs were stimulated to transdifferentiate into corneal epithelial progenitors expressing both KRT 3 and 12. Additionally, the transdifferentiated DPSCs provided a barrier between corneal and conjunctival epithelia and prevented conjunctivalization of the cornea. Further studies using human corneas lacking a healthy limbus are needed to determine whether DPSCs will transdifferentiate in a scenario similar to that of LSCDs seen in the clinical situation. 

Acknowledgments

Disclosure: E. Kushnerev, None; S.G. Shawcross, None; S. Sothirachagan, None; F. Carley, None; A. Brahma, None; J.M. Yates, None; M.C. Hillarby, None