Concise Review: Limbal Epithelial Stem Cell Therapy: Controversies and Challenges§


  • Anna R. O'Callaghan,

    Corresponding author
    1. Department of Ocular Biology & Therapeutics, University College London Institute of Ophthalmology, London, United Kingdom
    • Department of Ocular Biology & Therapeutics, UCL Institute of Ophthalmology, 11-43 Bath Street, London EC1V 9EL, United Kingdom
    Search for more papers by this author
    • Telephone: 02076086996; Fax: 02076086887

  • Julie T. Daniels

    1. Department of Ocular Biology & Therapeutics, University College London Institute of Ophthalmology, London, United Kingdom
    Search for more papers by this author

  • Author contributions: A.O.C.: collection and/or assembly of data, data analysis and interpretation, and manuscript writing; J.T.D.: manuscript writing, and final approval of manuscript.

  • Disclosure of potential conflicts of interest is found at the end of this article.

  • §

    First published online in STEM CELLSEXPRESS October 13, 2011.


Limbal epithelial stem cells (LESCs) are a population of stem cells responsible for maintenance and repair of the corneal surface. Injury and disease can result in a deficiency of these stem cells, the vision affecting condition called limbal stem cell deficiency (LSCD) in which the cornea becomes opaque, vascularized, and inflamed. Cultured LESC therapy was first described in 1997;29:19231932–19231932.and LESCs cultured from either patients or donors have been used to successfully treat LSCD. In this review, some of the challenges and controversies associated with cultured LESC therapy will be discussed including alternative stem cell sources. STEM CELLS 2011;29:1923-1932


Limbal epithelial stem cells (LESCs) are located in the limbal region of the cornea (Fig. 1A) and are responsible for the maintenance and repair of the corneal surface. Injury and disease can result in limbal stem cell deficiency (LSCD). Acid and alkali burn, aniridia, Stevens Johnson syndrome, and contact lens-induced keratopathy are all examples of causes of LSCD. In LSCD, the conjunctival epithelium migrates across the limbus causing the cornea to become opaque thus affecting vision. This is accompanied by inflammation, vascularization, and severe discomfort.

Figure 1.

Limbus location and limbal epithelial cells in culture. (A): Location of the limbus. (B): Superior, inferior, nasal, and temporal regions of the cornea/limbus. (C): Colony of limbal epithelial cells (outer edges indicated with arrows) in coculture with 3T3 fibroblasts. (D): Confluent area of cultured limbal epithelial cells. Scale bar = 125 μm.

Cultured LESC therapy has been used to successfully treat LSCD and was originally described by Pellegrini et al. in 1997 [1]. In this original procedure, a small biopsy 1 mm2 was taken from the limbus of the patient's healthy eye. Limbal epithelial cells were cultured from this biopsy to form an epithelial sheet using growth-arrested 3T3 fibroblasts as a feeder layer (Fig. 1 shows photos of limbal epithelial cells cultured in this way). This cultured limbal epithelial cell sheet was transplanted onto the corneal surface after removal of the conjunctiva from the corneal wound bed. A therapeutic contact lens was placed over the transplanted sheet in this procedure but amniotic membrane and fibrin have subsequently been used as carriers for cultured corneal epithelial transplantation [2, 3]. Baylis et al. (2011) have recently reviewed the outcomes of cultured limbal epithelial cell therapy that have been published since 1997 with data from 583 patients. The overall success rate of this procedure at the time of their review was 76% [4].


The dogma that stem cells of the cornea reside solely in the limbus has been challenged by Majo et al. (2008). In their controversial paper, they showed that the central cornea of the mouse contains oligopotent stem cells [5]. In addition, the authors suggested that the central mammalian corneal epithelium also contains stem cells. However, their human data did not appear to support this theory [6] suggesting that there may be species-specific differences in the location of corneal stem cells. Although it is widely accepted that stem cells of the cornea reside in the limbus, several other studies have suggested that this is not the only location of cells which can regenerate the corneal epithelium [7, 8]. A study by Dua et al. (2009) looked at eight eyes all of which had a central area of normal corneal epithelium despite being LESC deficient. These normal areas remained unchanged over a mean follow-up of 60 months. The authors proposed that either some LESC niches remained that were not clinically visible but contributing to the maintenance of the healthy central corneal epithelium or that basal transient amplifying cells of the surviving central epithelium were capable of maintaining the central cornea for long periods of time. They suggested that the latter explanation was more likely. Dua et al. (2009) ask an important question of what would happen if these healthy islands of central cornea are damaged or destroyed as a result of injury or disease. If stem cells are present, corneal epithelial regeneration should occur—however, this remains to be answered. Chang et al. (2008) also demonstrated that the human central corneal epithelium was capable of corneal epithelial regeneration following ablation of the limbus, at least in the first 12 hours following wounding. In this study, laser ablation was performed to either leave a central island of corneal epithelium separated from the peripheral cornea and limbus or in addition removing the limbus. Epithelial recovery was observed whether or not the limbus was present.

Several different putative stem cell niche structures have been observed in the limbus. Our group has shown that putative stem cells exist in structures termed limbal crypts and focal stromal projections, which are preferentially located in the superior and inferior regions of the limbus [9]. Other studies have also suggested that stem cells may be more abundant in the superior and inferior limbus [10, 11]. The superior, inferior, nasal, and temporal regions of the limbus are indicated in Figure 1B. Dua et al. have proposed that stem cells reside within structures called limbal epithelial crypts. These limbal epithelial crypts were variably distributed along the limbus, typically clustered together, and most abundant in the nasal region [12]. As LESCs are not uniformly distributed throughout the limbus, the location of a limbal biopsy for use in cultured LESC therapy is very important. For LESC therapy, a biopsy needs to be taken from a stem cell-rich region of the limbus. It is important to take a small biopsy so as not to deplete the population of stem cells in the donor eye and cause this eye to also become LESC deficient. There is currently no definitive marker for LESCs and this presents a challenge for identifying and sorting stem cells from the limbal biopsy taken for use in LESC therapy. In the absence of a LESC marker, a combination of positive expression of stem cell-associated markers, for example, ABCG2 and negative expression of differentiation markers, for example, CK3 are used to identify putative LESCs. Features of stem cells such as small cell size and high nuclear to cytoplasmic ratio can also be used to aid the identification of putative LESCs. These markers have been reviewed elsewhere [13–15].

The lack of a definite LESC marker also contributes to the challenge of how to assess graft quality and the likelihood of a successful outcome prior to transplantation. Rama et al. (2010) have found that the percentage of cells which stain intensely for ΔNp63α (referred to as p63-bright cells) can be used to predict the clinical outcome of a LESC graft. Cultures containing more than 3% p63-bright cells were found to be associated with successful transplantation in 78% of patients whereas cultures containing 3% or less p63-bright cells were associated with only 11% of successful transplants [16]. This group now only use cultures containing more than 3% p63-bright cells to prepare grafts. However, an adequate number of stem cells in the graft is not the only factor required for a successful clinical outcome, and Pellegrini et al. (2011) have suggested that in failed transplants stem cells may be lost during engraftment due to a hostile in vivo microenvironment. Strict inclusion and exclusion criteria for patients as well as suitable preparation of the ocular surface prior to transplantation may therefore also be necessary for success [17].

Another challenge is how the outcomes of cultured stem cell therapy for the treatment of LSCD can be assessed clinically, and how these results can be compared both between and within different studies conducted by different groups. How do you measure success when some patients will not regain total vision? Normal vision was only restored by LESC therapy in patients with undamaged corneal stroma, and corrective surgical procedures such as penetrating keratoplasty are required to improve visual acuity where stromal scarring is present [16]. A universal grading system for clinical outcomes would make it easier for studies to be compared and the best treatment method to be identified.

The lack of a definite LESC marker contributes to the difficulty of assessing the extent to which a patient has LSCD. Initial patient diagnosis is important not only for selecting the best treatment for the patient but also for assessing the clinical outcome. It is therefore important to be able to accurately distinguish between total and partial LSCD. The lack of a clear and definitive method for determining if a patient is totally LESC deficient makes it difficult to compare outcomes as the starting position may not be the same.

Another question that remains to be answered is how does cultured LESC therapy actually work? Is it replacing stem cells, or somehow causing the patients' own stem cells to become functional/active again, for example, by providing growth factors. Current evidence regarding what happens to the transplanted cells appears to be contradictory. Daya et al. [18] showed by DNA analysis that no ex vivo donor stem cell DNA was present beyond 9 months. In contrast, Djalilian et al. [19] showed that there was evidence for the persistence of donor epithelial cells up to 3.5 years following limbal allograft tissue transplantation. The mode of action may depend on whether the patient has total or partial LESC deficiency. The patient may have LESCs present that are unable to properly function due to the unsuitable environment created by damage or disease. Once a suitable microenvironment has been restored by cultured LESC therapy, these existing LESCs may be able to resume normal function.

Sex-mismatching and DNA fingerprinting can be used to study the survival of allogenic-transplanted cells [20–23]. Similarly xenotransplantation studies, for example, human cells transplanted onto an animal disease model allow the origin of the remaining cells following transplantation to be determined. For example, in a study by Du et al. (2003), cultured human limbal cells were transplanted onto a rabbit model of LSCD and after 40 days the majority of repopulated epithelial cells were observed to express anti-human nuclear antibody [24]. However, tracking of autologous transplanted cells is more difficult as these methods cannot be used, and transplanted cells would have to be labeled to track them in vivo—it would have to be proved that such tagging would not adversely affect the function or characteristics of these transplanted cells and not pose a risk to the patient. Cell tracking studies will provide insight into what happens to autologous transplanted cells following therapy, but a safe and enduring label has not yet been optimized for this use.

Other challenges for cultured LESC therapy and the use of alternative sources of cultured cells for the treatment of LCSD include regulatory issues and variability both between donors and within the culture process. From a regulatory standpoint, cell therapies should ideally be animal product free. At present, the gold standard culture method for limbal epithelial cells for use in the clinic uses mouse 3T3 fibroblasts as a feeder layer and bovine serum. At present, the 3T3 coculture system seems to be optimal for the expansion of LESCs [13] and is allowed for clinical use until a suitable alternative is found. It has been suggested that this concern over the use of the 3T3s as a feeder layer is exaggerated; as very few 3T3s remain on the final graft, these cells are growth arrested and therefore could not engraft, and there is no evidence that these cells have had a harmful effect from their use in graft production [25]. However, the risk of murine viral transmission to human cells in vitro and whether this would be important for patient health is unknown. Because of donor variability, the production time for a cultured cell sheet for transplantation varies causing scheduling issues for patient surgery. Investigation into whether grafts could be stored prior to use, or whether this would adversely affect their function and viability would be useful to alleviate such issues. Amniotic membrane is often used in LESC therapy and this also varies from donor to donor and also within a donor.

Depending on the cause of the LESC deficiency, the patient may receive an autologous or allogeneic-cultured LESC graft. Ideally, the patients' own cells would be used as this would reduce the risk of rejection and eliminate the need for immunosuppression. However, it is not always possible to use the patients' own LESCs for treatment as both eyes may be affected (bilateral damage). Other sources of stem cells from the patient may be a viable alternative and are currently being investigated.


Cultured oral mucosal epithelial stem cells have already been used for the treatment of LESC deficiency in patients, a procedure called COMET (cultivated oral mucosal epithelial transplantation). COMET has been shown to be successful in restoring the corneal surface with follow-up times of up to 35 months [36–43]. A summary of the reported studies in human patients to date is shown in Table 1. Oral mucosa is an easily accessible source of stem cells and heals quickly with minimal scarring. Peripheral neovascularization is commonly seen in COMET, and longer follow-up times will show whether this vascularization remains only peripherally or becomes more extensive. It has been suggested that antiangiogenic therapy for corneal neovascularization could be used in combination with COMET to improve the therapeutic outcome [49].

Table 1. Alternative sources of cultured stem cells for the treatment of LSCD
inline image

Animal models are useful for testing the safety and efficacy of cultured cell therapies prior to use of these treatments in human patients. Cultured LESC therapy for the treatment of LSCD has been shown to be successful in both rabbits and humans, as has COMET. The rabbit cornea is the most commonly used model for LSCD, probably because it is of a similar size to the human cornea and thus useful in terms of developing the surgical techniques required for transfer of epithelial cell sheets onto the patients' eye. LSCD can be achieved in these animal models by chemical or mechanical means, or a combination of both. Surgical removal of the limbus and corneal epithelium, and NaOH-induced chemical burn are examples of methods used to cause LSCD. Although models exist for LSCD caused by chemical burn injury, no specific animal models exist for Stevens Johnson syndrome. This presents a challenge for developing therapies for this cause of LSCD.

Other cultured cell types that have been proposed for the treatment of LSCD include the following: hair follicle stem cells, embryonic stem cells, conjunctival epithelial cells, dental pulp stem cells, umbilical cord lining stem cells, and bone marrow-derived mesenchymal stem cells. These have all been cultured and transplanted into animal models of LSCD with promising results [27-29, 31, 32, 34, 44, 45, 48]. A summary of these studies is shown in Table 1. These cultured cells were generally applied to the recipient eye as an intact cell sheet (either on amniotic membrane, fibrin, or alone) [26-31, 44, 48]. However, bone marrow-derived stem cells have also been administered via systemic injection and injection under transplanted amniotic membrane [32, 33], and embryonic stem cells were applied as a cell suspension [45]. Injection of mesenchymal stem cells under amniotic membrane did not provide an improved corneal surface compared to controls with conjunctivalization and vascularization of the cornea occurring in treated eyes [33]. In contrast, Ma et al. (2006) showed that the ocular surface was repaired when mesenchymal stem cells were transplanted as an intact sheet. In a study by Homma et al. (2004), embryonic stem cells were applied to the surface of the mouse eye as a cell suspension and allowed to attach for an hour. The cornea was completely re-epithelialized within 24 hours of cell application.

Of the limbus alternatives already mentioned (and in addition to oral mucosa), hair follicle, conjunctiva, dental pulp, and bone marrow are all feasible sources of autologous stem cells that could be harvested from patients with LSCD, some of these sources being more easily accessible with less invasive stem cell harvest than others. Nasal epithelium is another autologous source of stem cells which although has not yet been used in cultured stem cell therapy for the treatment of LSCD has shown promise when applied directly to the ocular surface of patients [50]. Advantages of using nasal oral mucosal epithelium are that it is easily accessible, expresses CK3, and contains goblet cells which secrete mucin important for tear film stabilization [50].

Neovascularization was common in COMET and transplantation of other cell types for the treatment of LSCD, and transplanted cells were found to typically express markers related to their tissue of origin. A number of studies have aimed to direct these alternative stem cell sources into a more corneal phenotype for improved functionality on the corneal surface following transplantation. Conditions mimicking the in vivo LESC niche have been used to induce a more corneal phenotype in hair follicle stem cells, embryonic stem cells, and marrow mesenchymal stem cells. Limbal/corneal fibroblasts and extracellular matrix found in the limbus have been used to do this [45–48]. In vivo, limbal fibroblasts are closely associated with LESCs and appear to be an important component of the limbal stem cell niche [51]. Blazejewska et al. (2009) found that pax 6 (involved in eye development) and CK12, which are both expressed in the cornea, were upregulated when mouse hair follicle stem cells were cultured on laminin V with conditioned medium from limbal fibroblasts. Mesenchymal stem cells were found to express CK12 when cocultured with corneal stromal cells in a transwell system [48]. Embryonic stem cells cultured on collagen IV and collagen IV with conditioned medium from limbal fibroblasts were also found to express CK12 [45, 46]. Mesenchymal stem cells have also been shown to express CK3 when cocultured with limbal epithelial cells in a transwell system or when cultured with supernatant medium from limbal epithelial cell culture [35].


Cultured LESC therapy has been shown to be successful in restoring the corneal surface in patients with LSCD. However, there are many challenges associated with cultured LESC therapy, and many questions remain to be answered including how this therapy works and how the outcomes can be best assessed clinically. In cases of bilateral LSCD, alternative sources of stem cells are required to enable autologous therapy. Oral mucosal epithelial cells have already been used in patients for the treatment of LSCD and further work to improve COMET and further research other alternative sources of stem cells for the treatment of LSCD is required.


This work was supported by Fight for Sight (A.O.C.) and NIHR BRC for Ophthalmology at the UCL Institute of Ophthalmology and Moorfields Eye Hospital (J.T.D.).


The authors indicate no potential conflicts of interest.