Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Stem Cells. Author manuscript; available in PMC 2010 July 19.
Published in final edited form as:
PMCID: PMC2906385

Characterization of Putative Stem Cell Phenotype in Human Limbal Epithelia


This study evaluated proposed molecular markers related to stem cell (SC) properties with the intention of characterizing a putative SC phenotype in human limbal epithelia. Human corneal and limbal tissues were cut in the vertical and horizontal meridians for histology, transmission electron microscopy (TEM), and immunostaining. Semiquantitative reverse transcriptase-polymerase chain reaction (RT-PCR) and in situ hybridization were used to evaluate gene expression. TEM showed that the limbal basal cells were small primitive cells. Immunostaining disclosed that p63, ABCG2 and integrin α9 were primarily expressed by the basal epithelial cells of limbus. Antibodies against integrin β1, epidermal growth factor receptor (EGFR), K19, enolase-α, and CD71 stained the basal cells of the limbus more brightly than the suprabasal epithelia. Integrin α6, nestin, E-cadherin and connexin 43 did not stain the limbal basal cells, but the suprabasal epithelia of the cornea and limbus showed strong immunoreactivity. K3 and involucrin stained only corneal and limbal superficial cells. RT-PCR showed higher levels of p63, ABCG2 and integrin α9 mRNA, but lower levels of K3, K12 and connexin 43 expressed in the limbal epithelia than the corneal epithelia. In situ hybridization showed that p63 transcripts were located in basal layer of the limbal epithelium. This work suggests that the basal epithelial cells of the limbus are p63, ABCG2 and integrin α9 positive, and nestin, E-cadherin, connexin 43, involucrin, K3, and K12 negative, with relatively higher expression of integrin β1, EGFR, K19, and enolase-α. This putative SC phenotype may facilitate the identification and isolation of limbal epithelial SCs.

Keywords: Cornea, Limbus, Epithelium, Stem cells, Stem cell phenotype


Recent data indicate that not only embryonic stem cells (SCs) are pluripotent, but there are also tissue-specific SCs residing in many adult tissues, including blood, liver, brain, skeletal muscle, intestine, skin, and cornea. SCs derived from these adult tissues have the ability to regenerate these tissues, and they offer great therapeutic potential for treating diseased and damaged tissues and serving as gene delivery vehicles [1, 2]. Widely accepted criteria for defining SCs are: A) slow cycling or long cell cycle time during homeostasis in vivo; B) poorly differentiated with primitive cytoplasm; C) high capacity for error-free self renewal, and D) activation to proliferate by wounding or placement in culture [3-7].

The ocular surface is an ideal region to study epithelial SC biology because of the unique spatial arrangement of SCs and transient amplifying cells. The compartmentalization of the corneal epithelial SCs within the limbus provides a valuable opportunity to study the behavior of adult SCs [4, 8, 9]. The supporting data for the limbal location of corneal epithelial SCs include: A) the limbal basal cells lack the corneal epithelial differentiation-associated keratin pair K3 [10] and K12 [11]; B) the limbal basal epithelium contains slow-cycling cells identified as the “label-retaining cells” following pulse-chase labeling of all cells with a DNA precursor, such as [3H]-thymidine or bromodeoxyuridine (BrdU) [12], and the limbal basal epithelium exhibits high proliferative potential in culture [12-15]; C) experimental studies and clinical observations show abnormal corneal epithelial wound healing with conjunctivalization, vascularization, and chronic inflammation when the limbal epithelium is partially [16, 17] or completely defective [18, 19], and D) limbal cells are essential for the long-term maintenance of the central corneal epithelium, and they can be used to reconstitute the entire corneal epithelium in patients with limbal SC deficiencies [9, 20]. Collectively, these data leave little doubt that corneal epithelial SCs reside in the limbus.

To date, no direct methods have been established to identify the corneal SCs because of the lack of specific molecular markers, although a variety of SC-associated markers have been proposed. The major markers proposed for epithelial SCs in ocular or non-ocular tissues in the past decade can be categorized into at least three groups: A) nuclear proteins such as the transcription factor p63; B) cell membrane or transmembrane proteins including integrins (integrin β1, α6, α9), receptors (epidermal growth factor receptor [EGFR], transferrin receptor CD71), and drug resistance transporters (ABCG-2), and C) cytoplasmic proteins such as cytokeratins (CK) (cytokeratin 19), nestin, and α-enolase. In addition, a variety of differentiation markers have also been proposed to distinguish the SCs from differentiated cells. These include cytokeratins K3 and K12, involucrin, intercellular adhesive molecule E-cadherin, and gap junction protein connexin 43, etc. This study was conducted to evaluate currently proposed molecular markers related to SC properties with the intention to characterize a putative SC phenotype in human limbal epithelia. While no single marker can identify adult SCs to date, characterization of a putative SC phenotype may shed light on the understanding of SC features.

Materials and Methods

Materials and Reagents

Mouse monoclonal antibodies (mAb) against integrin β1, p63, EGFR, CD71, E-cadherin, and involucrin were purchased from Lab Vision (Fremont, CA; Human ABCG2 mAb was from Calbiochem (San Diego, CA; Cytokeratin 19 (K19) mAb was from DAKO (Carpinteria, CA; AE5 mAb for keratin 3 (K3) was from ICN Pharmaceuticals (Costa Mesa, CA;, and goat polyclonal antibody against enolase α was from Santa Cruz Biotechnology (Santa Cruz, CA; Anti-nestin mAb was from Chemicon International (Temecula, CA; A rabbit antibody against integrin α9 was kindly provided as a gift by Dr. M. A. Stepp, George Washington University Medical Center, Washington DC. Fluorescein Alexa-Fluor 488 conjugated secondary antibodies (goat anti-mouse or anti-rabbit immunoglobulin [IgG], donkey anti-goat IgG) were from Molecular Probes (Eugene, OR; Vectastain Elite Kits were from Vector Laboratories (Burlingame, CA; Anti-human integrin α6 mAb, rabbit polyclonal antibody against Connexin 43, Hoechst 33342, DNA size marker and other reagents came from Sigma (St. Louis, MO; GeneAmp RNA-PCR kit was from Applied Biosystems (Foster City, CA; Riboprobe combination system and restriction endonucleases were from Promega (Madison, WI; [35S] UTP was from Amersham Biosciences (Piscataway, NJ; All plastic ware was from Becton Dickinson (Lincoln Park, NJ;

Human Corneal and Limbal Tissue Preparation

Fresh normal human corneal tissues were obtained from the National Disease Research Interchange (NDRI; Philadelphia, PA) for this study. The corneal and limbal specimens were prepared by cutting the tissues in the vertical meridian from 6 o'clock to 12 o'clock through the central cornea (Fig. 1A) and in the horizontal direction across the superior peripheral cornea and limbus (Fig. 1B). The tissue specimens were embedded in a mixture of 75% (volume [v]) OCT compound (Sakura Finetek USA Inc., Torrance, CA; and 25% (v) Immu-Mount (Thermo-Shandon; Pittsburgh, PA; and frozen in liquid nitrogen. Frozen sections (6-10 μm thick) were used for haematoxylin-eosin staining and for immunostaining. Some corneal limbal tissues were fixed in 10% phosphate buffered formalin for 1 day, then transferred to 70% ethanol, dehydrated, and embedded with paraffin. Paraffin sections (5 μm thick) were then cut for in situ hybridization experiments.

Figure 1
Haematoxylin-eosin staining showing palisades of Vogt limbal architecture

Transmission Electron Microscopy

Normal human corneoscleral tissues, which did not meet the criteria for clinical use and were preserved for less than 24 hours post mortem, were obtained from Lions Eye Bank of Texas (LEBT; Houston, TX) for this study. The specimens, about 1 mm cubes, were excised from central and peripheral cornea and limbus, and fixed for 1 hour in 3% glutaraldehyde buffered to pH 7.2 with 0.01 M PIPES (piperazine-N,N′-bis[2-ethane sulfonic acid]). They were rinsed in buffer and post-fixed in PIPES-buffered osmium tetroxide (pH 7.2) for 1 hour at room temperature, then rinsed in several changes of distilled water and dehydrated through a graded series of ethanol. The dehydrated tissues were incubated in two 45-minute changes of propylene oxide followed by a 1:1 mixture of propylene oxide and Spurr's resin for one and a half hours. The tissue pieces were then incubated in pure resin for one and a half hours, after which they were transferred to fresh resin in block molds and allowed to cure at 60°C overnight. One μm thick sections cut from the hardened blocks were mounted on glass slides, stained with an alcoholic solution of toluidine blue and basic fuchsin, and examined under the light microscope. Areas of interest were trimmed and 60 nm sections were cut and mounted on copper grids (300 mesh). The grids were stained with uranyl acetate and lead citrate and photographed with a Zeiss EM-900 transmission electron microscope (Zeiss; Peabody, MA; Photographs were taken with Kodak 4489 Electron Microscope film (Eastman Kodak; Rochester, NY;

Immunofluorescent Staining

Immunofluorescent staining was performed by a previously reported method [21, 22] to evaluate the expression and location of different molecular markers that have been proposed to identify SCs or differentiated cells. In brief, human corneal and limbal frozen sections were thawed, dehydrated, and fixed in cold methanol (for cytoplasmic and nuclear protein staining) or 2% paraformaldehyde (for all membrane protein staining) at 4°C for 10 minutes. Some sections were left unfixed for integrin α9 antibody staining. Sections were blocked with 5% normal horse serum in phosphate buffered saline (PBS) for 1 hour to decrease nonspecific antibody interactions. Primary mAb against nuclear p63 (clone 4A4, 1:1,000, 1 μg/ml), ABCG2 (clone BXP-21, 1:100, 2.5 μg/ml), EGFR (1:20, 10 μg/ml), integrin β1 (1:200), integrin α6 (1:200, 10 μg/ml), CD71 (1:100, 2 μg/ml), K19 (1:100, 0.4 μg/ml), nestin (1:100, 10 μg/ml), E-cadherin (1:200, 2.5 μg/ml), K3 (AE5) (1:50, 20 μg/ml) or involucrin (1:40, 5 μg/ml), or polyclonal anti-sera, goat against enolase-α (1:100, 2 μg/ml), or rabbit against integrin α9 (1:200) or connexin 43 (1:200, 2.5 μg/ml), were applied and incubated for 1 hour at room temperature. Secondary antibodies, Alexa-Fluor 488 conjugated goat anti-mouse or anti-rabbit IgG or donkey anti-goat IgG (1:300) were then applied and incubated in a dark chamber for 1 hour, followed by counterstaining with Hoechst 33342 DNA binding dye (1 μg/ml in PBS) for 2 minutes. After washing with PBS, Antifade Gel/Mount (Fisher Scientific; Norcross, GA; and a coverslip were applied. Sections were examined and photographed with an epifluorescent microscope, Eclipse 400, (Nikon; Tokyo, Japan; with a digital camera (model DMX 1200, Nikon) with similar exposure times for the cornea and limbus.

Immunohistochemical Staining

Immunohistochemical staining was performed by a previously reported method [23] to evaluate certain markers including p63, ABCG2, and integrin α9. After fixing with cold methanol (for p63), or 2% paraformaldehyde (for ABCG2), or non-fixation (for integrin α9), the sections were treated with 0.3% H2O2 in PBS containing 0.5% horse serum to quench the endogenous peroxidase activity and then incubated with 5% horse serum to block the non-specific sites. ABCG2 (1:100) or p63 (1:1,000) mAb, or rabbit anti-integrin α9 (1:200) was applied and incubated for 1 hour at room temperature, followed by incubation with biotinylated anti-mouse or anti-rabbit IgG secondary antibody, using a Vectastain Elite ABC Kit (Vector Laboratories) according to the manufacturer's protocol. The samples were finally incubated with 3,3′-diaminobenzidine (DAB) peroxidase substrate to give a brown stain and counterstained with hematoxylin. After washing with PBS and mounted, the sections were examined and photographed with an epifluorescent microscope.

Total RNA Extraction and RT-PCR

The cornea and limbus were separated by an 8-mm trephine from human corneoscleral tissues preserved for less than 36 hours, and the epithelia was scraped and collected into 4 M guanidium solution. Total RNA was isolated by acid guanidium thiocyanate-phenol-chloroform extraction using our previously described method [21]. The RNA was quantified by its absorption at 260 nm and stored at -80°C before use. With a housekeeping gene, glyceraldehyde-3-phosphate dehydrogenase (GAPDH), as internal control, the mRNA expression of different molecular markers by corneal and limbal epithelia were analyzed by semiquantitative reverse transcriptase-polymerase chain reaction (RT-PCR) as described in our previous reports [21, 24]. Briefly, first-strand cDNAs were synthesized from 0.5 μg of total RNA with murine leukemia virus reverse transcriptase. PCR amplification of the first-strand cDNAs was performed with specific primer pairs, designed from published human gene sequences (Table 1) for different markers in a GeneAmp PCR System 9700 (Applied Biosystems). Semi-quantitative RT-PCR was established by terminating reactions at intervals of 20, 24, 28, 32, 36, and 40 cycles for each primer pair to ensure that the PCR products formed were within the linear portion of the amplification curve. The fidelity of the RT-PCR products was verified by comparing their size to the expected cDNA bands and by sequencing the PCR products.

Table 1
Human primer sequences used for semi-quantitative RT-PCR

Preparation of Riboprobe and In Situ Hybridization

The 2.0 kb p63α probes were generated from plasmid p63α-bluescript KS by SmaI digestion and Sp6 RNA polymerase for the antisense probe, and by XbaI digestion and T7 RNA polymerase for the sense probe. The linearized plasmids were gel-isolated and used as templates for antisense and sense [35S]-UTP labeled riboprobes. The transcription mixture (30 μl) included 1 μg of linearized template cDNA, 3 mM of ATP, GTP, CTP and [35S]-UTP, 10 mM dithiothreitol (DTT), RNase Inhibitor (1 unit/μl of transcription mix), and polymerase (0.7 unit/μl of transcription mix). Transcription was performed for at least 2 hours at 37°C. The template cDNAs were digested by RNase-free DNase (2 μl at 1 unit/ul, 15 min at 37°C). The riboprobes were then precipitated and resuspended in 50 ml of diethyl pyrocarbonate (DEPC)-treated water.

The 5 μm paraffin sections were dewaxed and rehydrated. After digestion by proteinase K and post-fixation in 10% formaldehyde-PBS, the sections were prehybridized for 1 hour at 60°C in the hybridization mix (50% formamide, 10× salts, 0.05M DTT, 500 μg/ml poly ribo A, 50 ug/ml tRNA and 10% Dextran sulfate). The probes were denatured for 5 min at 100°C and added to the hybridization mix (140,000 cpm/μl). The hybridization reaction was carried out at 60°C for overnight. After incubation, the sections were stringently washed and digested with RNase A. Sections were then exposed to photographic emulsion for 10-14 days before developing and haematoxylin counterstaining. Hybridization signals were obtained with a digital capture system.


A New Look at Limbal Structure and Ultrastructure

Previously reported studies [9, 25, 26] evaluated limbal structure in radially oriented tissue sections (Fig. 1A). In order to completely evaluate the complex architecture of the limbal palisades of Vogt, we also evaluated cross-sections of the superior limbus (Fig. 1B). The palisades of Vogt in the superior limbus are structured with papilla-like columns, with smaller densely packed basal cells, larger less densely packed cells within the columns, and flattened squamous cells in the apical layers. There are also blood vessels, nerves, and connective tissue between the epithelial columns (Fig. 1B).

Transmission electron microscopy (TEM) shows that the basal cells of central corneal epithelium are columnar cells with a low nucleus/cytoplasm (N/C) ratio (Fig. 2A). The nucleus has loose chromatin and a pronounced nucleolus with a number of coiled DNA (heterochromatin). The cytoplasm of these cells contains a large number of ribosomes and tonofilaments. The basal epithelial cells are connected to the Bowman's membrane by hemidesmosomes (Fig. 2B). In contrast, the basal cells of limbal epithelium are smaller with a larger N/C ratio (Fig. 2E). Their nucleus has more euchromatin as open DNA and barely detectable nucleolus. In contrast to the central cornea, the cytoplasm of the limbal basal cells contains more tonofilaments, and there is no underlying Bowman's membrane. The limbal basal cells have basal invaginations through the basement membrane to connect with the underlying matrix, which contains collagen, fibrils, dilated capillaries, and some macrophages (Fig. 2F). This structure provides an ideal environment for cell nutrition. The peripheral corneal structure is between the central cornea and limbus. Their basal epithelial cells have an intermediate size between corneal and limbal epithelial cells, and the nuclei have less coiled DNA than corneal cells. They interdigitate with the basement membrane by basal infoldings (Figs. 2C, 2D).

Figure 2
Transmission electron microscopic images of the central cornea (A, B), peripheral cornea (C, D) and limbus (E, F)

Immunostaining of SC-Associated Markers in the Cornea and Limbus

Nuclear protein p63 was recently proposed as an SC marker to identify epidermal SCs as well as limbal epithelial SCs [27]. Using the mAb clone 4A4, which reacts with all p63 isoforms [28], the p63 protein was immunodetected primarily in the nuclei of limbal epithelial basal layer, but not in most limbal suprabasal and corneal epithelial cells (Fig. 3). Immunohistochemical staining with the same antibody revealed more clearly that within limbal basal layer, patches of p63 negative cells were interspersed with more numerous p63 positive cells (Fig. 4). ABCG2, a member of the ATP binding cassette (ABC) transporters, has been proposed as a universal marker for SCs [29]. Using mAb clone BXP-21 [30], the ABCG2 transporter protein was primarily immunodetected in the cell membrane and cytoplasm of certain limbal basal epithelial cells, but not in most limbal suprabasal cells and corneal epithelial cells (Fig. 3). There were ABCG2 negative cells interspersed with the ABCG2 positive cells in the basal layer (Fig. 4). Interestingly, integrin α9 was also immunodetected at the cell membranes and cytoplasm of certain limbal basal epithelial cells but not in limbal suprabasal and cornea cells (Figs. 3, ,4).4). This is a similar expression pattern to p63 and ABCG2.

Figure 3
Immunofluorescent staining of proposed SC-associated markers, p63, ABCG2, integrin α9, integrin β1, EGFR, K19, enolase-α, CD71, and integrin α6 on frozen sections of limbus (left panels) and cornea (right panels)
Figure 4
Immunohistochemical staining for p63, ABCG2, and integrin α9 on limbal frozen sections

Integrin β1 and α6 were previously proposed as putative SC markers for epidermal keratinocytes [31, 32]. Our results showed that integrin β1 was abundantly expressed by the cell membranes of corneal and limbal epithelia with a much higher level of expression by the limbal basal cells (Fig. 3). In contrast, integrin α6 antibody strongly stained the cell membranes of suprabasal layers of corneal and limbal epithelia. It did not stain the basal cell layers of the limbal epithelia (Fig. 3). The EGFR antibody stained the cell membranes of limbal basal epithelia much stronger than the suprabasal cells, and it also strongly stained most corneal epithelial cells (Fig. 3), a pattern similar to integrin β1. Transferrin receptor CD71 mAb strongly stained the cell membranes of certain basal cells in the limbus and most corneal epithelial cells (Fig. 3). Cytokeratin 19 (K19), a proposed marker for skin hair follicle SCs [33], was expressed in the cytoplasm of basal epithelial cells more strongly than the suprabasal cells in limbus. K19 was also expressed by most corneal epithelial cells. Enolase-α was largely expressed by limbal basal epithelial cells, although some limbal suprabasal and corneal cells were stained (Fig. 3). The immunostaining patterns of these proposed SC-associated markers by corneal and limbal epithelial cells are summarized in Table 2.

Table 2
Localization of sc-associated markers

Immunostaining of Differentiation Associated Markers in the Cornea and Limbus

Cytokeratin 3 (K3), a corneal specific marker [10], was strongly expressed by all corneal epithelia and the superficial limbal epithelial cells. Involucrin antibody strongly stained only superficial epithelial cells in the cornea and limbus. Neither antibody stained the basal layer of the limbal epithelium (Fig. 5).

Figure 5
Immunofluorescent staining for differentiation associated markers, K3, involucrin, connexin 43, E-cadherin, and nestin on frozen sections of limbus (left panels) and cornea (right panels)

The absence of gap junction components, such as connexin 43 and 50, was recognized as a feature of SCs [34]. The connexin 43 antibody stained membranes of suprabasal epithelia, but not the basal layer of the limbus (Fig. 5). E-cadherin had a similar expression pattern to connexin 43. This antibody only stained suprabasal epithelia of the cornea and limbus, but not the basal layer of the limbal epithelia (Fig. 5). Interestingly, nestin, a neuron SC marker, was expressed in the cytoplasm of superficial corneal and limbal epithelia. The limbal basal cells were totally nestin negative (Fig. 5). These immunostaining patterns suggest that connexin 43, E-cadherin, and nestin may serve as differentiation markers for corneal epithelia. The localization of these differentiation associated markers in corneal and limbal epithelia is summarized in Table 3.

Table 3
Localization of differentiation associated markers

Gene Expression of Molecular Markers in Corneal and Limbal Epithelia

With the housekeeping gene GAPDH as an internal control, semi-quantitative RT-PCR disclosed a differential expression pattern of SC associated markers and differentiation markers by human corneal and limbal epithelia. The expressions of ΔNp63, ABCG2, and integrin α9 transcripts were markedly higher by limbal epithelia than corneal epithelia (Fig. 6). Although some bands of ΔNp63 and integrin α9 were clearly visible in one corneal sample (Fig. 6), this may have resulted from sample overlapping of the peripheral cornea and limbus when they were separated by a trephine. There was no significant difference in integrin β1, EGFR, K19, and enolase-α mRNA between limbal and corneal epithelia (data not shown). The mRNA of corneal specific differentiation markers, K3 and K12, were abundantly expressed by the corneal epithelia but expression was weak or undetectable in the limbal epithelia (Fig. 6). Expression of connexin 43 mRNA was also barely detectable in limbal epithelia, but was abundant in corneal epithelia (Fig. 6).

Figure 6
Semi-quantitative RT-PCR for SC-associated markers, p63 (440 bp), ABCG2 (379bp), and integrin α9 (123 bp), and differentiation associated markers, K3 (145bp), K12 (150 bp), and connexin 43 (154 bp) expressed by corneal (C1, C2, C3) and limbal ...

In Situ Hybridization of p63 Riboprobe in Corneal and Limbal Epithelia

To confirm the p63 expression pattern, in situ hybridization with [35S]-labeled sense and antisense p63α RNA riboprobes was performed on cornea and limbal tissues. The radioautograms showed that strong p63 mRNA signals were only in the basal layer of limbal epithelia, but not in the suprabasal limbal and entire corneal epithelia (Figs. 7C, 7D), when the sections were hybridized with antisense riboprobes. In contrast, there was no signal in control sections hybridized with p63α sense riboprobes (Figs. 7A, 7B).

Figure 7
In situ hybridization of human limbal paraffin sections with [35S]-labeled p63 α riboprobes


The Limbal Structure and Corneal Epithelial Stem Cell Niche

In this study, the limbal structure was analyzed in tangential cross-sections cut through the superior limbus. These sections provided a look at the papilla-like epithelial columns in the limbus (Fig. 1B). The orientation of these sections provides a clear demonstration of the differential expression pattern by the basal and suprabasal limbal epithelial cells. The column architecture of the limbus has many advantages for the SCs. It provides a greater surface area for the large quantity of SCs that are required for self-renewal. The undulating architecture also provides SCs with different numbers of overlying epithelial cells, which allows more efficient response to different depths of corneal epithelial wounding. With stroma tissue inserted between limbal papilla-like columns, the basal layer of limbal epithelia is adjacent to a rich vascular network that provides ample nutrients and other supportive factors. TEM provides morphological evidence that limbal basal cells are cuboidal primitive cells, distinguishable from the large columnar cells in basal layer of the corneal epithelia (Fig. 2).

Basal Limbal Epithelial Cells Uniquely Express p63, ABCG2 and Integrin α9

Recently, the nuclear transcription factor p63, a member of the p53 family, was proposed to be a marker of corneal epithelial SCs. p63 is highly expressed in the basal cells of many human epithelial tissues, and the truncated dominant-negative ΔNp63 isoform is the predominant species in these cells [28, 35, 36]. It was reported that p63 knockout mice lack stratified epithelia and contain clusters of terminally differentiated keratinocytes on the exposed dermis [28], and that p63 expression is associated with proliferative potential in human keratinocytes [27, 37]. Our findings are consistent with these previous reports. Nuclear p63 was strongly expressed in certain limbal basal cells, evidenced by immunostaining (Figs. 3, ,4),4), RT-PCR (Fig. 6), and in situ hybridization (Fig. 7). The presence of p63 in these basal cells may be an indication of their high proliferative potential.

ABCG2, a member of the ABC transporters, formally known as breast cancer resistance protein 1 (BCRP1), has been identified as a molecular determinant for bone marrow SCs, and has been proposed as a universal SC marker [29, 38]. Our results demonstrated for the first time that the ABCG2 gene is expressed primarily by the limbal basal cells, evidenced by immunostaining (Figs. 3, ,4)4) and RT-PCR (Fig. 6). ABCG2 expression may be a common feature of SCs as a mechanism to prevent them from damage by drugs and toxins. Further studies are necessary to explore the potential that ABCG2 could serve as a marker for corneal epithelial SCs.

Integrin α9 was reported to localize to the basal cells of the epidermis, conjunctiva, and limbus after birth and into adulthood in the developing ocular surface of mice [39]. Our results showed that integrin α9 expression was limited to certain basal cells of the limbal epithelia, evidenced by immunostaining (Figs. 3, ,4)4) and RT-PCR (Fig. 6). Cell-matrix adhesion is largely mediated by integrins, the major components of stable adhesions to the basement membrane in hemi-adhesion junctions. The higher expression of integrin α9 and β1 (see below) may indicate the strong adhesion of limbal basal cells to specific extracellular matrix ligands and may explain the resistance of limbus to shear forces. Further studies are needed to determine if integrin α9 and β1 are specific cell surface markers of the limbal SCs.

Basal Limbal Epithelial Cells Express Higher Levels of Integrin β1, EGFR, K19, and Enolase-α

Integrins β1 and α6 were previously proposed as putative SC markers for epidermal keratinocytes. Integrin β1-enriched human epidermal basal cells from both keratinocyte culture and foreskin biopsies were demonstrated to have a higher colony-forming efficiency than unfractionated cells [31, 40, 41]. Murine epidermal keratinocytes with high levels of integrin α6 and low to undetectable expression of the transferrin receptor (CD71), termed α6briCD71dim cells, were reported to possess characteristics of SCs [42, 43]. Our results showed that integrin β1 mAb more strongly stained the basal cells than suprabasal cells of the limbus, but it also stained all corneal epithelial cells (Fig. 3). Integrin α6 mAb stained the suprabasal limbal and corneal epithelial cells but not the basal limbal cells (Fig. 3). Transferrin receptor (CD71) antibody stained the basal cells of the limbus more strongly than the suprabasal cells, but it also stained most corneal epithelium.

It has been reported that basal limbal epithelial cells express higher levels of EGFR than the more mature and differentiated suprabasal limbal epithelial cells [22, 44]. This notion was supported by our results that the EGFR antibody more strongly stained the basal layers of limbal epithelium than suprabasal epithelia (Fig. 3). The presence of high levels of EGF receptors might allow these cells to be rapidly stimulated by growth factors to undergo cell division during development and following wounding.

As a member of the cytokeratin family of intermediate filaments, K19 has been suggested as a marker for the epidermal SCs in skin hair follicles. K19 was expressed in the hair follicle and was absent from the interfollicular epidermis in hairy sites. K19 was also noted in the slow cycling [3H]-thymidine-label-retaining cells by double-labeling experiments [33]. In this study, we observed that K19 was expressed at a higher level in the basal than the suprabasal layers of the limbal epithelium, but it was also expressed in all layers of the corneal epithelium. A cytoplasmic glycolytic enzyme, enolase-α, was proposed as a corneal epithelial SC marker [45, 46]. Enolase-α was localized to the limbal basal cells and the basal cells of other stratified epithelia. Our data showed that enolase-α was located in the cytoplasm of limbal basal epithelial cells as well as some basal corneal cells.

Basal Limbal Epithelial Cells Express Low Levels of Nestin, E-Cadherin, Connexin 43, Involucrin, K3, and K12

The intermediate filament nestin was previously proposed as a neuron SC marker [47]. Our immunostaining data revealed for the first time that nestin was strongly expressed in the cytoplasm of superficial cells of human corneal and limbal epithelia, but not in the basal limbal cells (Fig 5). Further studies are necessary to evaluate the functional role of nestin in the limbal epithelia.

Cell-to-cell communication plays an important role in cellular development and differentiation. Gap junctional communication is formed by a family of related amphipathic polypeptides called connexins. These intercellular communicating channels allow direct passive diffusion of low molecular weight solutes between neighboring cells. E-cadherin is a transmembrane Ca2+-dependent homophilic adhesion receptor that plays an important role in cell-cell adhesion. E-cadherin mediated cell-cell contact results in cell activation and an increase in key signaling molecules that are involved in cell proliferation and survival [48, 49]. Our study showed that the connexin 43 and E-cadherin mAb stained only the superficial corneal and limbal epithelia. These findings suggest that connexin 43 and E-cadherin are expressed by differentiated epithelial cells, and the absence of these intercellular communication molecules in the basal limbus may be an inherent feature of SCs, reflecting the need for SCs to maintain the uniqueness of their own intracellular environment [34, 48]. With more abundant integrins (β1 and α9) and lack of connexin 43 and E-cadherin, the limbal SCs in vivo are likely to strongly adhere to the extracellular matrix, thus maintain their niche, and yet are less adhesive to one another, enabling individual SCs to be rapidly mobilized and exit from their niche for self-renewal.

K3 and K12 are well known as corneal specific markers [10, 11, 50]. Consistently, our results showed that the basal limbal cells were K3 negative (Fig. 5). K3 and K12 transcripts were barely detected in limbal epithelia but were strongly expressed by the corneal epithelia (Fig. 6). Involucrin, another differentiation marker [51, 52], was also negative in the basal epithelial layer of the limbus.

Application of the Putative Stem Cell Phenotype of Human Limbal Epithelia

Our findings demonstrate that the basal epithelial cells of the limbus are small active primitive cells, that are p63, ABCG2 and integrin α9 positive, and nestin, E-cadherin, connexin 43, involucrin, K3, and K12 negative. They have a relatively higher expression of integrin β1, EGFR, K19, and enolase-α. This expression pattern may represent a unique phenotype of the putative SCs in the basal layer of the limbal epithelia, distinguishing them from suprabasal limbal and corneal epithelial cells. While no single marker for the corneal epithelial SCs has been identified to date, the characterization of this putative SC phenotype in our study provides greater understanding of limbal SC features, and may practically be used in future studies to identify and isolate the corneal epithelial SCs. This panel of SC-associated and differentiation-associated molecular markers can be utilized to distinguish corneal epithelial cells at different phases of their life cycle. The unique phenotype may be useful for further characterization of putative SCs in culture. This anatomical localization of gene expression can be used to study the local environmental factors at the limbus, such as different type collagens, extracellular matrix proteins or growth factors, which regulate gene expression and promote stemness. Using this putative SC phenotype, isolation or enrichment for limbal SC could be possibly achieved by adherence to collagen IV or other extracellular matrix based on their higher expression of certain integrins such as α9 and β1, or by fluorescence activated cell sorting with flow cytometry based on other cell surface markers such as ABCG2 and EGFR. Partially enriched basal limbal epithelial cells could be further utilized to search for truly unique markers for limbal epithelial SCs.


We thank Dr. Mary Ann Stepp for kindly providing the integrin α9 antibody.

This study was supported by NIH Grants, EY014553 (D.Q.L.) and EY11915 (S.C.P.), National Eye Institute, Bethesda, MD, a grant from Lions Eye Bank of Texas, an unrestricted grant from Research to Prevent Blindness, a post-doctoral research fellowship from Fight For Sight, the Oshman Foundation, and the William Stamps Farish Fund.


For reprints contact: moc.sserPdeMahplA@stnirpeR


1. Hall PA, Watt FM. Stem cells: the generation and maintenance of cellular diversity. Development. 1989;106:619–633. [PubMed]
2. Potten CS, Loeffler M. Stem cells: attributes, cycles, spirals, pitfalls and uncertainties. Lessons for and from the crypt. Development. 1990;110:1001–1020. [PubMed]
3. Blau HM, Brazelton TR, Weimann JM. The evolving concept of a stem cell: entity or function? Cell. 2001;105:829–841. [PubMed]
4. Watt FM, Hogan BL. Out of Eden: stem cells and their niches. Science. 2000;287:1427–1430. [PubMed]
5. Lavker RM, Sun TT. Epidermal stem cells: properties, markers, and location. Proc Natl Acad Sci USA. 2000;97:13473–13475. [PubMed]
6. Watt FM. Epidermal stem cells as targets for gene transfer. Hum Gene Ther. 2000;11:2261–2266. [PubMed]
7. Cotsarelis G, Kaur P, Dhouailly D, et al. Epithelial stem cells in the skin: definition, markers, localization and functions. Exp Dermatol. 1999;8:80–88. [PubMed]
8. Dua HS, Azuara-Blanco A. Limbal stem cells of the corneal epithelium. Surv Ophthalmol. 2000;44:415–425. [PubMed]
9. Tseng SC. Concept and application of limbal stem cells. Eye. 1989;3:141–157. [PubMed]
10. Schermer A, Galvin S, Sun TT. Differentiation-related expression of a major 64K corneal keratin in vivo and in culture suggests limbal location of corneal epithelial stem cells. J Cell Biol. 1986;103:49–62. [PMC free article] [PubMed]
11. Kurpakus MA, Maniaci MT, Esco M. Expression of keratins K12, K4, and K14 during development of ocular surface epithelium. Curr Eye Res. 1994;13:805–814. [PubMed]
12. Cotsarelis G, Cheng SZ, Dong G, et al. Existence of slow-cycling limbal epithelial basal cells that can be preferentially stimulated to proliferate: implications on epithelial stem cells. Cell. 1989;57:201–209. [PubMed]
13. Ebato B, Friend J, Thoft RA. Comparison of limbal and peripheral human corneal epithelium in tissue culture. Invest Ophthalmol Vis Sci. 1988;29:1533–1537. [PubMed]
14. Kruse FE, Tseng SC. A tumor promoter-resistant subpopulation of progenitor cells is larger in limbal epithelium than in corneal epithelium. Invest Ophthalmol Vis Sci. 1993;34:2501–2511. [PubMed]
15. Pellegrini G, Golisano O, Paterna P, et al. Location and clonal analysis of stem cells and their differentiated progeny in the human ocular surface. J Cell Biol. 1999;145:769–782. [PMC free article] [PubMed]
16. Chen JJ, Tseng SC. Corneal epithelial wound healing in partial limbal deficiency. Invest Ophthalmol Vis Sci. 1990;31:1301–1314. [PubMed]
17. Chen JJ, Tseng SC. Abnormal corneal epithelial wound healing in partial-thickness removal of limbal epithelium. Invest Ophthalmol Vis Sci. 1991;32:2219–2233. [PubMed]
18. Huang AJ, Tseng SC. Corneal epithelial wound healing in the absence of limbal epithelium. Invest Ophthalmol Vis Sci. 1991;32:96–105. [PubMed]
19. Jenkins C, Tuft S, Liu C, et al. Limbal transplantation in the management of chronic contact-lens-associated epitheliopathy. Eye. 1993;7:629–633. [PubMed]
20. Tsai RJ, Li LM, Chen JK. Reconstruction of damaged corneas by transplantation of autologous limbal epithelial cells. N Engl J Med. 2000;343:86–93. [PubMed]
21. Li DQ, Tseng SC. Three patterns of cytokine expression potentially involved in epithelial-fibroblast interactions of human ocular surface. J Cell Physiol. 1995;163:61–79. [PubMed]
22. Liu Z, Carvajal M, Carraway CA, et al. Expression of the receptor tyrosine kinases, epidermal growth factor receptor, ErbB2, and ErbB3, in human ocular surface epithelia. Cornea. 2001;20:81–85. [PubMed]
23. Yoshino K, Tseng SC, Pflugfelder SC. Substrate modulation of morphology, growth, and tear protein production by cultured human lacrimal gland epithelial cells. Exp Cell Res. 1995;220:138–151. [PubMed]
24. Li DQ, Lokeshwar BL, Solomon A, et al. Regulation of MMP-9 production by human corneal epithelial cells. Exp Eye Res. 2001;73:449–459. [PubMed]
25. Goldberg MF, Bron AJ. Limbal palisades of Vogt. Trans Am Ophthalmol Soc. 1982;80:155–171. [PMC free article] [PubMed]
26. Gipson IK. The epithelial basement membrane zone of the limbus. Eye. 1989;3:132–140. [PubMed]
27. Pellegrini G, Dellambra E, Golisano O, et al. p63 identifies keratinocyte stem cells. Proc Natl Acad Sci USA. 2001;98:3156–3161. [PubMed]
28. Yang A, Schweitzer R, Sun D, et al. p63 is essential for regenerative proliferation in limb, craniofacial and epithelial development. Nature. 1999;398:714–718. [PubMed]
29. Zhou S, Schuetz JD, Bunting KD, et al. The ABC transporter Bcrp1/ABCG2 is expressed in a wide variety of stem cells and is a molecular determinant of the side-population phenotype. Nat Med. 2001;7:1028–1034. [PubMed]
30. Maliepaard M, Scheffer GL, Faneyte IF, et al. Subcellular localization and distribution of the breast cancer resistance protein transporter in normal human tissues. Cancer Res. 2001;61:3458–3464. [PubMed]
31. Jones PH, Watt FM. Separation of human epidermal stem cells from transit amplifying cells on the basis of differences in integrin function and expression. Cell. 1993;73:713–724. [PubMed]
32. Li A, Simmons PJ, Kaur P. Identification and isolation of candidate human keratinocyte stem cells based on cell surface phenotype. Proc Natl Acad Sci USA. 1998;95:3902–3907. [PubMed]
33. Michel M, Torok N, Godbout MJ, et al. Keratin 19 as a biochemical marker of skin stem cells in vivo and in vitro: keratin 19 expressing cells are differentially localized in function of anatomic sites, and their number varies with donor age and culture stage. J Cell Sci. 1996;109:1017–1028. [PubMed]
34. Matic M, Petrov IN, Chen S, et al. Stem cells of the corneal epithelium lack connexins and metabolite transfer capacity. Differentiation. 1997;61:251–260. [PubMed]
35. Yang A, Kaghad M, Wang Y, et al. p63, a p53 homolog at 3q27-29, encodes multiple products with transactivating, death-inducing, and dominant-negative activities. Mol Cell. 1998;2:305–316. [PubMed]
36. Barbareschi M, Pecciarini L, Cangi MG, et al. p63, a p53 homologue, is a selective nuclear marker of myoepithelial cells of the human breast. Am J Surg Pathol. 2001;25:1054–1060. [PubMed]
37. Parsa R, Yang A, McKeon F, et al. Association of p63 with proliferative potential in normal and neoplastic human keratinocytes. J Invest Dermatol. 1999;113:1099–1105. [PubMed]
38. Kim M, Turnquist H, Jackson J, et al. The multidrug resistance transporter ABCG2 (breast cancer resistance protein 1) effluxes Hoechst 33342 and is overexpressed in hematopoietic stem cells. Clin Cancer Res. 2002;8:22–28. [PubMed]
39. Stepp MA, Zhu L, Sheppard D, et al. Localized distribution of alpha 9 integrin in the cornea and changes in expression during corneal epithelial cell differentiation. J Histochem Cytochem. 1995;43:353–362. [PubMed]
40. Jones PH, Harper S, Watt FM. Stem cell patterning and fate in human epidermis. Cell. 1995;80:83–93. [PubMed]
41. Watt FM. Epidermal stem cells: markers, patterning and the control of stem cell fate. Philos Trans R Soc Lond B Biol Sci. 1998;353:831–837. [PMC free article] [PubMed]
42. Kaur P, Li A. Adhesive properties of human basal epidermal cells: an analysis of keratinocyte stem cells, transit amplifying cells, and postmitotic differentiating cells. J Invest Dermatol. 2000;114:413–420. [PubMed]
43. Tani H, Morris RJ, Kaur P. Enrichment for murine keratinocyte stem cells based on cell surface phenotype. Proc Natl Acad Sci USA. 2000;97:10960–10965. [PubMed]
44. Zieske JD, Wasson M. Regional variation in distribution of EGF receptor in developing and adult corneal epithelium. J Cell Sci. 1993;106:145–152. [PubMed]
45. Zieske JD, Bukusoglu G, Yankauckas MA. Characterization of a potential marker of corneal epithelial stem cells. Invest Ophthalmol Vis Sci. 1992;33:143–152. [PubMed]
46. Chung EH, DeGregorio PG, Wasson M, et al. Epithelial regeneration after limbus-to-limbus debridement. Expression of alpha-enolase in stem and transient amplifying cells. Invest Ophthalmol Vis Sci. 1995;36:1336–1343. [PubMed]
47. Lendahl U, Zimmerman LB, McKay RD. CNS stem cells express a new class of intermediate filament protein. Cell. 1990;60:585–595. [PubMed]
48. Scott RA, Lauweryns B, Snead DM, et al. E-cadherin distribution and epithelial basement membrane characteristics of the normal human conjunctiva and cornea. Eye. 1997;11:607–612. [PubMed]
49. Gumbiner BM. Regulation of cadherin adhesive activity. J Cell Biol. 2000;148:399–404. [PMC free article] [PubMed]
50. Liu CY, Zhu G, Converse R, et al. Characterization and chromosomal localization of the cornea-specific murine keratin gene Krt1.12. J Biol Chem. 1994;269:24627–24636. [PubMed]
51. Watt FM, Green H. Involucrin synthesis is correlated with cell size in human epidermal cultures. J Cell Biol. 1981;90:738–742. [PMC free article] [PubMed]
52. Banks-Schlegel S, Green H. Involucrin synthesis and tissue assembly by keratinocytes in natural and cultured human epithelia. J Cell Biol. 1981;90:732–737. [PMC free article] [PubMed]