SEARCH

SEARCH BY CITATION

Keywords:

  • Limbus;
  • Cornea;
  • Cell culture;
  • Developmental biology;
  • Differentiation;
  • Fetal stem cells;
  • Proliferation;
  • Tissue-specific stem cells

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. SUMMARY
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES
  11. Supporting Information

The adult cornea harbors stem cells (SCs) in its periphery, in a niche known as the limbus. Over the past 2 decades there has been substantial research into these adult corneal SCs, their limbal niche, and their therapeutic applications. However, few studies have investigated how this niche and its SCs develop in humans. To better characterize this development, human fetal corneas from 8.5- to 22-weeks'-gestation (n = 173), neonatal (n = 2), and adult (n = 10) specimens were obtained. Histological and immunohistochemical assessments were conducted to determine embryological changes and expression of developmental and SC-related genes. Fresh fetal corneas were explanted to propagate corneal progenitors and cells characterized using reverse transcription-polymerase chain reaction, immunohistochemistry, flow cytometry, and colony-forming assays. A novel “ridge-like” structure was identified, circumscribing the fetal cornea, which we hypothesize represents the rudimentary SC niche. Immunohistochemistry disclosed “stem-like” cells across the cornea, becoming confined to this ridge with increasing gestational age. In addition, for the first time, pure long-term cultures of fetal corneal epithelium, which displayed phenotypical and functional properties similar to those of adult limbal SCs, were established. Optimization of culture techniques and purification of this SC population will allow for further investigation of their proliferative ability, with potential research and clinical applications. This study expands our understanding of limbal niche development and opens new avenues for investigation. STEM CELLS 2009;27:2781–2792


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. SUMMARY
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES
  11. Supporting Information

The adult corneal surface consists of a self-renewing, stratified squamous epithelium maintained by quiescent, slow-cycling, epithelial stem cells (SCs) [1]. When primed, these cells undergo asymmetric division whereby one daughter remains to sustain the SC pool, whereas the other is evicted and programmed to differentiate into a transient amplifying cell. These cells acquire excessive proliferative activity as they migrate centripetally to replace terminally differentiated cells sloughed from the ocular surface as a consequence of injury or physiological aging [2]. Adult corneal SCs reside within a specific niche with anatomical and functional dimensions to maintain “stemness” [3, 4]. This niche is located within the limbus, a transitional zone partitioning the cornea from the conjunctiva, and is characterized by stromal invaginations, known as the “Palisades of Vogt” [3, 5, 6]. These papillae-like projections protect SCs from traumatic and environmental insults [7], allow epithelial-mesenchymal interactions, and provide access to chemical signals that diffuse from the rich underlying vascular network. Moreover, the niche provides a greatly increased surface area to concentrate and confine SCs in a narrow annulus that circumscribes the adult peripheral cornea. Recently, additional structures, “limbal crypts”, made up of solid cores of compacted cells have been identified [3], which may be similar to the recently discovered “focal limbal projections” [8].

The human cornea begins to differentiate at approximately 6-weeks' gestation (WG) and its development involves the interaction of the lens vesicle with the overlying surface ectoderm [9]. The lens is thought to inhibit corneal stromal development until it starts to recede, late in embryogenesis [10]. The corneo-scleral junction appears at the end of the embryonic period and is well demarcated by gestational week 11 [9]. However, little is known about the development of the human limbus, including when the palisades are formed, when SC activity first appears and when it localizes to the limbal niche. Immunohistochemical studies on the fetal cornea [11] identified the absence of CK-3 (corneal differentiation marker) in early fetal corneas (8WG), its presence in superficial cells by 12-13WG and by 36WG, positive staining in all suprabasal cells. The authors concluded this expression was inversely related to the distribution of SCs within the developing cornea. Of note, the authors suggested that the composition of the basal cornea and limbus may be similar until the postnatal period, with both possessing an undifferentiated phenotype [11]. Using a SC-specific microarray, we recently identified differential expression of CK15 between the central corneal and limbal regions of fetal (18WG) and adult corneas [12]. Given CK15 is regarded as a marker of the adult limbal phenotype [12, 13], these results suggest segregation of SC activity in the peripheral fetal cornea.

Du and colleagues [14] reported culturing limbal epithelial cells from a human fetal (20WG) cornea; however, their study lacked experimental and methodological details. It is possible that isolated fetal corneal epithelial cells have greater proliferative and survival ability than their adult counterparts, as has been shown in other fetal ocular cells [15] and epidermal keratinocytes [16], rendering them a potentially useful alternative cellular source for therapeutic applications.

In this study we aimed to characterize the anatomical development of the human cornea, with a particular focus on identifying stem, proliferating, and differentiating cells, and finally to isolate, culture, and characterize human fetal corneal epithelium, to assess their SC content and proliferative ability.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. SUMMARY
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES
  11. Supporting Information

Tissue Preparation

Human tissue was obtained in accordance with the tenets of the World Medical Associations Declaration of Helsinki and approved by the UNSW Human Research Ethics Committee. Premortem informed consent for tissue donation for research purposes was obtained. Normal postmortem human fetal whole eyes from 8-11WG (n = 10) and fetal anterior segments from 12-22WG (n = 163) were collected (supporting information Table 1). Fresh fetal corneas (n = 7) for tissue culture were obtained from the Prince of Wales Hospital (Sydney, Australia, http://www.sesiahs.health.nsw. gov.au/jmo/powh/index.asp) and 2 neonatal whole eyes, aged 3 weeks (corrected for prematurity), were obtained from the University of NSW. Normal human adult corneal rims (n = 10) were sourced from the Lions' NSW Eye Bank with consent. Fetal specimens were fixed in 2% paraformaldehyde and processed within 4 hours of death; adult corneal rims were fixed within 12 hours of death. Corneal specimens were bisected vertically or horizontally and embedded on the cut surface.

Histological and Immunostaining Analysis

Hematoxylin and eosin (H&E) sections of all 173 fixed fetal specimens were assessed. Representative corneal specimens were then selected for immunohistochemical analysis from a series of gestational ages, in the range 8-22WG. This included specimens in triplicate from 12WG, 14WG, 17WG, and 20WG and also individual specimens at the outskirts of the range (8WG, 10WG, and 22WG). Four observers (two blinded) systematically scored components of the ocular surface (conjunctiva; limbus; peripheral and central cornea) on H&E-stained sections; they also scored immunoreactivity for selective markers on the developing ocular surface. Sections were photographed using a DP70 camera, mounted on a BX51 microscope (Olympus, Sydney, Australia, http://www.olympus-global.com).

Immunohistochemical analysis was performed on tissues using a panel of markers (supporting information Table 2) associated with proliferation (Ki67), differentiation (CK3), apoptosis (caspase-3), and stem-like cells (p63, p63α, CK-15, and ABCG2) in the developing human cornea, as previously described [17, 18]. Heat-assisted antigen retrieval in citrate buffer (pH 6.0) was necessary to unmask CK3, p63, and p63α antigenic sites and immunoreactivity visualized with 3-amino-9-ethylcarbazole (Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com) for all antibodies. Ki67 proliferation index (PI) was calculated using the equation provided below. An unpaired t test was performed at each gestational age to identify differences between the expression of Ki67 in the limbus and that in central cornea.

  • equation image

For immunofluorescence, two fetal corneas (19WG) were incubated with an anti-CK15 antibody (supporting information Table 2), rinsed extensively in phosphate-buffered saline, and then incubated overnight at 4°C with an Alexa 488-conjugated secondary antibody (Molecular Probes Inc., Eugene, OR, http://probes.invitrogen.com). Shallow radial incisions were made in the peripheral cornea, and the specimen was mounted in glycerol, cover-slipped, and viewed with an inverted Olympus FV 1000 laser scanning microscope (Olympus).

Electron Microscopy

To elucidate surface topography, two additional whole fetal corneas (17WG and 18.5WG) were prepared for scanning electron microscopy as previously described [19].

Cell Culture

Human fetal (18WG) and adult corneal epithelial cells were propagated from tissue explants as previously described [20]. In total, five fresh fetal corneas were used for this approach. Establishment of culture was most successful when these corneas were placed epithelium facedown, as whole explants. Adult limbus was identified, cut into small segments, and placed into 6-well plates (CellStar; Greiner Bio-One, Frickenhausen, Germany, http://www.gbo.com/en). Media formulated to support adult corneal progenitors (Cnt50; Millipore, Billerica, MA, http://www.millipore.com) [20, 21] was used and replaced on alternate days, until sufficient epithelial growth was noted (10–12 days), at which time explants were removed and cells subcultured. For enzymatic dissociation fetal corneas (n = 2) were placed in Dispase II (2.4 U/ml; Roche Diagnostics, Indianapolis, IN, http://www.roche-applied-science.com) at 37°C for 90 minutes. Dislodged epithelial sheets were washed, plated out, and cultured as per whole corneal explants. Colony-forming efficiency was assessed by plating disaggregated cells at low density (1–5 × 104 cells per well) into triplicate wells of a 6-well plate. Cells were allowed to flourish for 7 days, after which colonies were counted. A colony consisted of a tight cluster of greater than 12 cells. The difference in colonies formed between fetal and adult cells was analyzed using a 2-way ANOVA with Bonferroni's post-tests for individual comparisons.

Reverse Transcription-Polymerase Chain Reaction

To assess whether cultured fetal corneal epithelial cells expressed adult corneal SC-related genes, reverse transcription-polymerase chain reaction (RT-PCR) was performed. Total RNA was extracted (RNAgents; Promega, Madison, WI, http://www.promega.com) and reverse-transcribed into complementary DNA using Superscript III (Invitrogen, Carlsbad, CA, http://www.invitrogen.com). Transcript levels were assessed by PCR using intron-spanning primers for the putative corneal SC markers ΔNp63α, ABCG2, and CK-15 (supporting information Table 3). Thermal cycling conditions consisted of the following: denaturation at 94°C for 2 minutes, followed by 28–35 cycles each at 94°C for 30 seconds, 60°C for 30 seconds, and 72°C for 30 seconds and a final extension at 72°C for 5 minutes to terminate reactions [18].

Flow Cytometry

Permeabilized cells were immunolabeled for p63, ABCG2, CK15, or isotype control IgG (supporting information Table 4), as detailed previously [18, 19].

Statistical Analysis

Statistical analysis was carried out using GraphPad Prism version 5 software (GraphPad Software Inc., La Jolla, CA, http://graphpad.com). p values of <0.05 were regarded as statistically significant. A Kruskal-Wallis 1-way analysis of variance with Dunn's post-test was applied when considering ordinal data. Unpaired t tests and ANOVA with Bonferroni post-test were used for continuous data, for two or more groups, respectively.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. SUMMARY
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES
  11. Supporting Information

Anatomical Development of the Fetal Cornea

Macroscopically, fetal corneal diameter increased with gestational age, from 2.5 mm at 13WG to 5 mm at 22WG (Fig. 1A). Histologically, at 8.5WG, the cornea was evident only as a continuation of surface ectoderm, although segregation between epithelial and stromal precursors was evident, with epithelial cells polarizing in an orderly manner. The lens was adherent anteriorly to the primitive cornea (Fig. 1B–1D). By 11WG, eyes were significantly larger and structurally more advanced. The eyelids were fused and the lens was retreating from the posterior cornea. The corneal epithelium comprised one to two layers of irregular cells, and the central corneal stroma was more prominent compared with those of the younger specimens (Fig. 1E–1G).

thumbnail image

Figure 1. Human fetal eye development. Fetal corneal size at different ages was compared to an adult eye (A). Corneas were embedded in paraffin, sectioned, stained with hematoxylin and eosin, and photographed. Panels (B–D) and (E–G) display fetal whole eyes from 8.5WG and 11WG, respectively. Sections in (B) and (E) were photographed at ×100, (C) and (F) at ×200, and (C) and (G) at ×1000 (oil immersion). The area encompassed by the rectangle in (E) is magnified in (F) and the area encompassed by the rectangles in (C) and (F) are magnified in (D) and (G), respectively. Abbreviations: AC, developing anterior chamber; C, developing cornea; EL, developing eyelid; L, developing lens; SE, surface ectoderm; WG, weeks' gestation.

Download figure to PowerPoint

In older specimens (12-22WG), individual components of the ocular surface, including conjunctiva, cornea, and limbus, were easily identified. The conjunctiva was relatively homogenous over this period, although its thickness increased from two layers between 12WG and 14WG to three layers from 14.5WG to 22WG (Fig. 2A–2C). Goblet cells were obvious from as early as 12WG (Fig. 2A). The epithelium became less densely cellular and began to stratify with increasing age (Fig. 2A–2C). Cellularity of the conjunctival stroma also reduced with age, making way for loose connective tissue (Fig. 2B, 2C). Occasional blood vessels were noted in specimens as early as 12WG (micrograph not shown) and vasculogenesis was a common feature in the conjunctival stroma of specimens >15WG (Fig. 2B, arrow).

thumbnail image

Figure 2. Fetal ocular surface development (12-22WG). Representative photomicrographs illustrate architectural changes in the developing human fetal cornea (A–I). Tissue in these panels was stained with hematoxylin and eosin. Each column represents different regions of the ocular surface from the same specimen. The single-headed arrow in (B) indicates blood vessels in the conjunctival stroma. The double-headed arrow in (I) indicates thickness of Bowman's layer. Panel (J) shows suprabasal cells within the “ridge-like” limbal area, which display a vacuolated appearance, whereas basal epithelial cells (large arrows) are more typical. Superficial cells have a flattened appearance (small arrows). This section was derived from an 18WG fetal cornea and was stained with periodic acid-Schiff (PAS). Inset (J) shows PAS-positive goblet cells in the conjunctiva of the same specimen. Panel (K) shows the ridge-like limbal area of a 20WG fetal cornea, stained with caspase-3 (red immunoreactivity). Original magnification was ×1000 (oil immersion) for panels (A–J) and ×400 for (K). The bars in panel (L) represent the median epithelial thickness across four ocular surface regions, for specimens 8.5-22WG (n = 173). Groups were compared using the Kruskal-Wallis test, with Dunn's post-test. Asterisks denote significance. Abbreviations: ***, <0.0001; CC, central cornea; G, goblet cell; PC, peripheral cornea; WG, weeks' gestation.

Download figure to PowerPoint

The main changes to the cornea over this period (12-22WG) included increasing thickness and decreasing stromal cellularity with increasing age (Fig. 2G–2I), although the cornea remained far more cellular than that in adulthood. The number of epithelial layers increased from two to three in the peripheral cornea over this period, but stayed constant at two layers in the central cornea (Fig. 2G–2I). Epithelial layers were generally more ordered than those in the conjunctiva (compare Fig. 2A–2C with Fig. 2G–2I) and were typically stratified squamous in type. Stromal keratocytes appeared more elongated than those in the conjunctiva and there was no evidence of a vasculogenesis in the fetal cornea.

The development of Bowman's layer remained incompletely defined. As early as 13WG, the corneal epithelium and underlying stroma began to separate (micrograph not shown). By 17WG, a thin Bowman's-like structure had developed (Fig. 2H), and by 20-22WG it was present and spanned the peripheral and central fetal cornea (Fig. 2I, double-headed arrow). Bowman's layer appeared earlier in the peripheral cornea than in the central cornea, and from 19WG onward, the epithelium in the peripheral cornea was at least one layer thicker than that in the central cornea, suggesting that fetal corneal epithelial proliferation is initiated in the periphery and is directed centripetally as occurs in the adult cornea [22].

Anatomical Development of the Fetal Limbus—The Limbal Ridge

The limbus was distinguishable from as early as 10WG, with the looser stroma of the conjunctiva integrating with the dense and ordered corneal stroma (Fig. 2D–2F). Furthermore, this region was one to two layers thicker than the corresponding epithelium in either the conjunctiva or the cornea, at all gestational ages (Fig. 2L). However, the papillae-like structures (Palisades of Vogt) characteristic of the adult limbal SC niche [5, 6] were not identified in any of the 173 fetal specimens. To ensure these folds were not deeper or in an alternative orientation, two fetal corneas (19WG) were serially cut from end to end, but these natural undulations were never observed. In addition, a postnatal specimen (3 weeks) confirmed our observations in fetal tissue, suggesting that limbal palisades do not form until after birth.

Rather than limbal palisades, fetal eyes had a “ridge-like” structure that identified the early fetal limbus. Within this ridge, the limbal epithelium was thicker than that in the surrounding region and comprised cells with a low nuclear-to-cytoplasm ratio and a “vacuolated” appearance (Fig. 2D, 2E, 2J). These cells were confined to a narrow region of the limbus, approximately 8–12 cells wide, and were present in fixed and fresh-frozen specimens. Initially, it was thought their appearance may be consistent with glycogen vacuoles. To test this, sections (14-18WG) were stained with periodic acid-Schiff (PAS). Although the goblet cells that lined the conjunctiva of these specimens were PAS-positive (Fig. 2J, inset), the vacuolated limbal epithelial cells did not exhibit any obvious staining (Fig. 2J).

Given the recognized role of apoptosis in morphogenesis, we postulated that some of the cells within the limbal ridge may be undergoing programmed cell death [23]. Sections from two fetal corneal specimens (14WG, micrograph not shown; 20WG, Fig. 2K) were stained for cleaved (active) caspase-3, revealing an immunoreactive layer of suprabasal cells within the fetal limbal epithelial ridge.

Surface Topography of the Fetal Cornea

Scanning electron microscopy and confocal microscopy were employed to visualize the topography of the corneal surface. From the histological sections, it was hypothesized that a limbal ridge would be obvious. Evaluation of photographs taken at the time of dissection indicated that the epithelial ridge was visible beyond the pigmented iris (Fig. 3A, arrow), an observation that was subsequently confirmed by confocal microscopy. Two whole fetal corneas (18WG and 19WG) were immunolabeled with CK-15, a known adult limbal epithelial progenitor cell marker [12, 13]. Immunolabeling for CK15 showed that cells within the ridge in whole-mounted fetal corneas were CK15-positive and that the region containing the CK15-immunoreactive cells was approximately 100-μm wide (Fig. 3B, 3C). Scanning electron microscopy revealed a clear demarcation zone (Fig. 3D, arrows) between the smooth cornea and the contoured conjunctiva. Closer inspection of the limbal junction showed a slightly elevated region, 80- to 100-μm wide, in keeping with the histological findings (Fig. 3E, arrows). Cells within this ridge differed morphologically from those in the cornea or conjunctiva, being more elongated, tightly packed with distinct rounded or indented profiles (Fig. 3F). Cells within the limbal ridge and throughout the cornea also featured abundant apical microvilli (Fig. 3F).

thumbnail image

Figure 3. The limbal “ridge” in the developing fetal ocular surface. The limbal ridge is visible just with macrophotography (A). The area encompassed by the rectangle in panel A represents the area shown in (B), which is a photomicrograph of a 19WG fetal cornea, immunolabeled with CK15, highlighting and confirming the ridge-like structure. The dashed lines with the scissor symbol represent the region where the cornea was cut to flat-mount it. The arrowheads in panel (B) identify occasional superficial CK15+ cells in the peripheral cornea. The dashed box in panel (B) is magnified in (C), which shows specific cytoplasmic staining for CK15 within the ridge. Original magnifications: ×100 for (B) and ×400 for (C). Panels (D–F) are scanning electron micrographs from an 18.5WG human fetal cornea. The dashed boxes in (D) and (E) are enlarged in (E) and (F), respectively. Arrows in (D, E) point to the limbal ridge. Asterisks in (F) identify limbal epithelial cells with collapsed cytoplasm. The double-headed arrow in (F) indicates the thickness of the limbal ridge. Original magnifications: ×16 (D), ×65 (E), and ×900 (F). Abbreviations: CC, central cornea; Cj, conjunctiva; L, limbus; PC, peripheral cornea; WG, weeks' gestation.

Download figure to PowerPoint

Expression of Proliferation and Differentiation Markers in the Developing Cornea

Ki67, a widely accepted marker of proliferation [24], was used to assess the spatial and temporal distribution of proliferating cells within the developing cornea. Ki67 expression was identified in cells across the entire cornea in specimens 8-14WG (Fig. 4A, 4B). However, in older specimens (17-20WG), Ki67 immunoreactivity was reduced in the central cornea, and concentrated in the limbus. This staining was semiquantitatively expressed as a PI [25], shown in Figure 4I. Although initially the expression was similar in the limbus and central cornea, from 14WG the limbal epithelial PI was consistently higher than that in the central corneal epithelium; however, this did not reach significance until 20WG (Fig. 4I; p < .05). By 22WG, Ki67 immunoreactivity was predominantly localized to the basal layer of the limbal epithelium (Fig. 4E, arrows), with only occasional Ki67-immunoreactive cells detected in the peripheral and central cornea (Fig. 4F). In adult corneas, Ki67 stained basal and suprabasal layers of the limbal epithelium (Fig. 4G, arrows), with weak to absent staining in the periphery (micrograph not shown) and no immunoreactivity in the central corneal epithelium (Fig. 4H).

thumbnail image

Figure 4. Proliferation in the corneal epithelium. Photomicrographs illustrating expression of Ki67 protein in the developing human fetal cornea at 14WG (A, B), 20WG (C, D), and 22WG (E, F) compared to an adult cornea (G, H). Column 1 shows the limbal region and column 2 the central corneal region from the same respective specimen. Positive immunoreactivity is denoted by red nuclear/cytoplasmic staining. Arrows in (E) and (G) point to basal limbal epithelium. All sections were lightly counterstained with hematoxylin. Sections in the first column (A, C, E, G) were photographed at ×400 (original) magnification, whereas sections in the second column (B, D, F, H) were photographed at ×1000 (oil immersion). Panel I shows the semiquantification of the proportion of Ki67 reactive cells in the specimen. A PI of basal fetal limbal and central corneal epithelium was calculated for each immunolabeled specimen and was compared graphically. Data represent mean ± SEM. n = 3 for each age, except for 22WG, where n = 1. Unpaired t tests were performed, comparing the limbal and central corneal regions at each gestational age. Asterisks denote significance. Abbreviations: *, p < .05; ***, p < .0001; PI, proliferation index; WG, weeks' gestation.

Download figure to PowerPoint

Differentiated corneal cells were identified using CK3, an accepted marker of advanced corneal epithelial differentiation [11, 26, 27]. CK3 immunoreactivity, indicating the presence of differentiated corneal cells, was essentially absent from corneas between 8.5WG and 10WG (micrograph not shown). From 12WG to 17WG there was positive staining in scattered superficial cells across the entire corneal epithelium (data not shown). By 20WG, intense immunoreactivity was noted in the superficial layer of the corneal and limbal epithelium, comparable to the adult corneal epithelium (micrographs not shown).

Expression of Stem Cell-Associated Markers in the Developing Cornea

A panel of putative adult corneal SC markers was used, to look for expression of SC-associated markers in the developing fetal cornea. Given its essential role in epithelial proliferation and development [28], the expression of p63 was first assessed. Expression of p63 was not detected in specimens aged between 8WG and 12WG (micrographs not shown). However, from 12WG to 22WG, p63 immunoreactivity was present in clusters of cells in the basal epithelium across the entire corneal surface (Fig. 5A–5C, small arrows). Since the p63-α isoform is considered to be more specific for adult limbal epithelial stem cells [29, 30], fetal corneal specimens were also stained for this variant, where its expression appeared similar to that of the pan-p63 marker (Fig. 5D–5F). This is in contrast to the expression in adult specimens, where p63α appeared more selective than the pan-p63 antibody (compare Fig. 5G–5I and Fig. 5J–5L).

thumbnail image

Figure 5. Expression of p63 in fetal and adult human corneas. Representative photomicrographs illustrating the expression of the p63 protein in the developing human fetal cornea, aged 17 weeks [(A–C) and (D–F): pan and α-isoform, respectively], in comparison to that of an adult cornea [(G–I) and (J–L): pan and α-isoform, respectively]. Immunoreactivity is denoted by the red nuclear staining. Sections were not counterstained. Each row represents different regions of the ocular surface from the same specimen. Rows 1 and 2 are from the same fetal specimen, whereas rows 3 and 4 are from the same adult specimen. The small arrows in (A–F) highlight immunoreactive nuclei. The large arrows in (D–F) point to corneal nerves. The hatched line in each panel is a guide to denote the basement membrane. Original magnification: ×1000 (oil immersion). Abbreviation: WG, weeks' gestation.

Download figure to PowerPoint

CK15, an intermediate filament protein, which identifies the limbal phenotype in the adult [13] and fetal cornea [12], was next examined. In early development (8-14WG), CK15 was expressed across the ocular surface (Fig. 6A, 6B). From 17WG onward, expression was increasingly restricted to the limbal epithelium (Fig. 6C, 6D) and, by 22WG, to the limbal basal cells alone (micrographs not shown), similar to the expression patterns found in adult specimens (Fig. 6E, 6F). Figure 6G summarizes these results schematically. Expression of ABCG2, a well-recognized putative adult limbal SC marker [31], closely resembled the CK15 results (micrographs not shown).

thumbnail image

Figure 6. Expression of CK15 in the developing human fetal and adult cornea. CK15 expression was identified in the developing fetal (A–D) and adult (E, F) cornea. Positive immunoreactivity is denoted by red cytoplasmic staining. Fetal corneas incubated with an isotype control (IgG1) displayed no reactivity (inset, C). All sections were lightly counterstained with hematoxylin. Original magnification: ×400. Panel (G) shows a schematic summary of the CK15 results. Black-shaded areas are CK15+ and correspond to the stained epithelium developed by the 3-amino-9-ethylcarbazole (red) chromogen. Panel (H) shows a similar scheme, but for CK3 expression. The schemes are based on an original diagram by Rodrigues and colleagues [11]. Abbreviation: WG, weeks' gestation.

Download figure to PowerPoint

Culture and Functional Characterization of Human Fetal Corneal Epithelial Cells

Fetal corneal epithelial cells were cultured and propagated to over 10 generations (Fig. 7A–7C) and retained a small, undifferentiated, and tightly packed morphology (Fig. 7B, 7C; arrows). Over this period, a minority of cells became larger, resembling terminally differentiated cells (Fig. 7B, 7C; asterisks). Morphologically, fetal cells resembled adult limbal epithelial cells cultured in the same media (Fig. 7D).

thumbnail image

Figure 7. Cultured fetal corneal epithelial cells. Panels (A–C) are phase-contrast photomicrographs of fetal corneal epithelial cells. Panel (A) shows cells migrating from a tissue explant after 3 days. Panels (B) and (C) show fetal cells at passages 1 and 7 respectively in comparison to adult limbal epithelial cells at passage 2 (D). Original magnification: ×40 (A) or ×200 (B–D). White arrows in (B–D) identify small proliferating cells, whereas asterisks indicate large terminally differentiated cells. Panel (E) shows a comparison of colony-forming activity between cultured fetal and adult corneal epithelial cells (blue and red bars, respectively). Here, cells at passage 4 were plated in triplicate at four different densities (x-axis) and the colony formation assessed after 7 days in culture, by counting the number of colonies that had formed in each well (y-axis). A colony was defined as a cluster of ≥12 cells. After 12 days, the culture was arrested and the colonies were stained with hematoxylin (E, inset). Groups were compared using a 2-way ANOVA with Bonferroni's post-test for individual comparisons. Reverse transcription-polymerase chain reaction analysis (F) for SC-related gene transcripts was performed on total RNA extracted from cultured fetal corneal epithelial cells at passage 2 (lanes 1 and 2) and cultured adult limbal cells at passage 3 (lanes 3 and 4). Lanes 2 and 4 represent control reactions where the RNA was not reverse-transcribed. A 50 bp molecular weight ladder and a no-template control were included in adjacent lanes (not shown). Flow cytometric analysis for the same markers was performed on cultured fetal and adult cells. Adult limbal epithelial cells from passage 4 (G–J) and fetal corneal epithelial cells from passages 3, 6, and 9 [(K–N), (O–R) and (S–V), respectively] were assessed. Purple histograms represent the stained population, whereas the unfilled histograms set the background reference with an isotype-matched antibody. Positive staining (%) was defined by a gate that included <5% of the background population.

Download figure to PowerPoint

A known property of SCs is their ability to form distinct colonies when seeded at low densities. To test this, epithelial cells were seeded low density and a linear relationship was observed between the number of cells seeded and the number of colonies formed, consistent with the behavior of stem cells (Fig. 7E). Fetal cells formed more colonies compared to their adult counterparts (2-way ANOVA, p < .01). Colony-forming efficiency was calculated to be 0.37% in the fetal corneal epithelial cells and 0.28% in the adult cells.

Expression of Adult SC Transcripts in Cultured Fetal Corneal Epithelial Cells

To assess whether the cultured fetal corneal epithelial cells expressed transcripts for putative adult SC-related genes, RT-PCR was performed and demonstrated a single amplicon at the expected size for ΔNp63α, ABCG2, and CK-15, respectively, for both cultured fetal corneal epithelial and adult limbal epithelial cells (Fig. 7F).

Flow Cytometric Analysis of Cultured Corneal Epithelial Cells

Protein expression assessed by flow cytometry (Fig. 7G–7V) showed a significant right shift in the cell populations expressing p63 (pan), ABCG2, and CK-15 (purple histograms), in both fetal and adult populations, whereas only a small shift was identified for p63α. The percentage of positive cells for each marker was comparable between fetal and adult cell types (compare Fig. 7G–7J with Fig. 7K–7V). Interestingly, later generation fetal corneal epithelial cells continued to express similar levels of SC markers, suggesting they were maintained in a “stem-like” state under these conditions (Fig. 7O–7R, 7S–7V).

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. SUMMARY
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES
  11. Supporting Information

This study aimed to determine the structural and functional development of the human fetal cornea, focusing on when and how limbal SCs and their niche assemble. Our results provide compelling new in vivo and in vitro evidence of stem-like cells in the developing cornea. With increasing gestational age, these cells appeared to become confined to a novel ridge-like microstructure, which we propose is the rudimentary SC niche.

The fetal period encompassing eyelid closure is one of considerable growth (Fig. 1A) and consolidation for the human ocular surface (Figs. 1, 2). These events cannot occur without exquisitely regulated physiological programs, including proliferation (Fig. 4), differentiation (Fig. 6), apoptosis (Fig. 2K), angiogenesis, and matrix remodeling. The onset of fetal corneal epithelial proliferation and the point in time when the epithelial layers form during gestation is a topic of debate [10, 32]. In the current investigation, we examined epithelial layers in 173 fetal corneas, in the range 8.5-22WG. In doing so, we identified a bilayered epithelium during the developmental period studied. Notably, the first signs of a multilayered (three to four layers) epithelium appeared in our oldest fetal specimen (22WG). This increase was most obvious in the peripheral cornea, suggesting proliferation may commence from the limbal zone [10]. Interestingly, this proliferative surge coincides with ocular exposure to amniotic fluid [10]. In view of its corneal wound-healing properties [33] and mitogenic activity [34], it is tempting to speculate that amniotic fluid may be a trigger for corneal maturation.

The immaturity of the fetal corneal epithelium was evident from the selective suprabasal expression of CK3, essentially undetectable in the youngest fetal specimens but widespread with increasing gestational age (Fig. 6H) [11]. Also noted was a corneal epithelium that was undergoing rapid expansion, as indicated by the Ki67 PI (Fig. 4). Notably, the percentage of Ki67-expressing cells declined with increasing gestational age, and in older fetal specimens expression was mainly confined to the limbal zone (Fig. 4E, 4I). This trend was broadly similar to that previously reported by Yew and colleagues [32] using the proliferating cell nuclear antigen.

Prior to this study, a comprehensive examination of the developing human limbus was lacking. Despite careful tissue orientation and microscopic examination, neither palisades nor other limbal SC structures that shelter adult corneal SCs [3, 6] were identified in fetal or neonatal specimens, suggesting that this anatomical structure develops postnatally. Although the literature on this event is sparse, our theory is supported by Rodrigues' observations [11], whose study contains a schema depicting the absence of palisades until after birth. One study identified limbal palisades in a neonatal eye; however, the age of that specimen was not disclosed [35]. In support of our observations, Yeung and colleagues [36] were also unable to identify limbal epithelial crypts in a 4-month-old infant cornea. This data is consistent with reports suggesting that the human cornea continues to develop until 6 months after birth [37], paralleling animal studies showing centripetal streaming of progeny cells in the cornea to be incomplete until the 10th postnatal week [38]. An incompletely matured corneal SC niche in the early postnatal period may partially explain the relatively late onset of corneal anomalies in congenital limbal stem cell deficiencies (LSCD), such as aniridia [39, 40].

A major finding of this study is the discovery of a previously undocumented topography for the fetal ocular surface epithelium: a ridge-like elevation that circumscribes and demarcates the developing cornea. Although we [20] and others [32] have previously published photos that documented this structure, it went unnoticed and hence its significance in corneal development and relevance to corneal health was not discussed. We propose that the fetal limbal ridge is analogous to other epithelial SC niches such as the crypts of Lieberkuhn in the small intestine, known to harbor epithelial SCs [41] and the hair follicle bulge, a SC-containing structure that becomes pronounced during development [42]. Interestingly, although the limbal ridge initially presents as a focal swelling between 12WG and 20WG, this subsides and becomes inconspicuous from 22WG to adulthood, perhaps as a result of physical stresses enforced by ocular and eyelid motion. When the ocular surface is no longer protected by sealed eyelids, the developmental process may be initiated to ensure early progenitors are protected throughout life, hence the development of deeper undulating and projecting niches to protect and house corneal SC during adulthood. It was interesting that the apoptosis marker, caspase-3, localized to a discrete row of cells within the limbal ridge (Fig. 2E), implying that a fine-tuned balance is struck between cell proliferation and programmed cell death within this structure, a feature seen in other ocular structures, such as the lens [23]. However, in addition to proapoptotosis, antiapoptotic functions (proliferation and differentiation) have been ascribed to the caspases including caspase-3 [43], which may be highly relevant to the developing corneal epithelial SC.

In this study, we provide convincing evidence that SC-like activity is present within this newly identified microstructure. Progenitor cells are initially distributed across the entire ocular surface, eventually becoming confined to the limbal ridge (Fig. 6A–6G). Our results are analogous to those reported in the developing human hair follicle SC niche, where SCs are present throughout the peripheral epithelium within the embryo, but subsequently the SC niche becomes progressively limited to the upper part of the outer root sheath, known as the “bulge” [44, 45]. Moreover, our results are supported by similar findings in the developing rat cornea, where stem-like cells are initially found throughout the basal layer of the corneal and limbal epithelium. As the cornea matures, these cells become sequestered in the limbus alone [46]. The trigger for this consolidated localization is incompletely understood but seems to commence before eyelid opening, at 25-26WG [10]. This is supported by studies in rats, where epidermal growth factor receptor expression by epithelial SCs becomes localized to the basal limbus 12 days prior to eyelid opening [47]. Therefore, although eyelid opening is a recognized period for corneal development [48], these results suggest that, for SC activity, genetic and signaling programs must be in place before the actual event. In the current study, the one exception to the general SC marker expression trend was the pattern of p63 staining, where equivalent levels were found in the fetal limbal and central corneal epithelium at all ages (Fig. 5) [49]. This widespread expression in the central cornea may relate to its more generalized function in epithelial development [28, 50].

Given the presence of SC activity in the developing fetal cornea, we determined whether these cells could be propagated. For the first time, we successfully cultured primary fetal corneal epithelial cells to over 10 generations. Of the methods employed, the explant technique was most successful, possibly because it provided sufficient time for progenitors to emigrate while soluble supportive factors diffused from the tissue during the primary expansion period [51]. Although enzymatic dissociation was also explored, this proved ineffective as cultures became overwhelmed with fibroblasts (data not shown), possibly as a result of the thin epithelium and poorly developed Bowman's layer (Fig. 2). Future studies will establish whether these cultured fibroblasts may be used as feeders to support clonogenic expansion of undifferentiated corneal epithelial progenitors in a human-human system to avoid potential xenogeneic contamination from animal-derived support cells such as fetal murine 3T3 fibroblasts [52].

The cultured fetal and adult cells both stained positive for a number of SC-related markers, in both their messenger and protein form. However, it seems unlikely that the very high percentages seen for ABCG2 and CK15 in the flow cytometry results represent the true percentage of SCs in the culture, and likely reflects some degree of in vitro upregulation of these markers, which has been described before in the adult limbal SC literature [53].

Interestingly, adult-derived corneal epithelial cells generally reached confluence earlier than fetal counterparts (data not shown). This difference may indicate that fetal cells were maintained in a relatively quiescent and undifferentiated state, consistent with our flow cytometry findings, where the proportion of cells expressing SC-related antigens was similar between passages 3 and 6 and was maintained up to passage 9. One of the challenges of culturing adult limbal SCs to date has been their tendency to differentiate and become senescent [20, 54]. Evidence from other organ systems indicates fetal cells have greater proliferative capacity than their adult counterparts. For example, bone marrow-derived fetal SCs are superior candidates for biomedical engineering compared to their perinatal or adult counterparts [55], whereas in the skin, cultured neonatal keratinocytes form a significantly higher number of clonogenic cells [16]. Although the clonogenic potential of fetal corneal SCs was not significantly different than that of adult corneal SCs, it is possible that optimizing our culture conditions (e.g., nurturing with human fetal cornea fibroblasts) may prolong this activity.

Although further in vitro optimization is clearly required, with the cell culture work in this study representing only very preliminary assays, our fetal corneal SC cultures may offer considerable promise for toxicological studies, biomaterial testing, and incorporation into corneal scaffolds for transplantation [56]. Although there would be significant ethical and legal hurdles to overcome, other clinical applications may eventually include their potential use as an alternative cellular source to treat bilateral LSCD [17], a blinding disorder that remains one of the most difficult to treat in ophthalmic medicine [8], complicated further by the current and projected worldwide shortage of corneal donor tissue [57].

SUMMARY

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. SUMMARY
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES
  11. Supporting Information

This study offers fresh insight into the way SC activity and the SC niche of the cornea develops. It has identified a new fetal structure, which we have termed “the limbal ridge”, which appears to be the rudimentary form of the limbus, the adult corneal SC niche. Parallels can be made to the development of other epithelial SC niches, including the skin and intestinal niches. It is notable that this study demonstrates that the palisades, or involutions, that characterize the adult limbus appear not to present until some time after birth and this concept of an immature SC niche at birth may explain why certain congenital limbal SC deficiencies have a relatively late onset of expression. In addition, our work suggests that undifferentiated, progenitor cells are initially widespread across the cornea, but localize to the fetal limbal ridge before the time of eyelid opening. Furthermore, initial cell culture work has suggested that these fetal progenitor cells can be cultured and optimization of culture techniques and purification of this SC population may allow for further in vitro studies and investigation of potential clinical applications.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. SUMMARY
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES
  11. Supporting Information

The authors would like to thank Julie Sinuks (Clinical Anatomy and Surgical Skills Unit, UNSW) for assisting in the procurement of neonatal eyes, Murray Smith (School of Medical Sciences, UNSW) for the critical appraisal of the manuscript, and Jenny Norman (Electron Microscopy Unit, UNSW) for technical assistance with scanning electron miscropy.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. SUMMARY
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES
  11. Supporting Information
  • 1
    Cotsarelis G, Cheng S-Z, 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: 201209.
  • 2
    Daniels JT, Dart JKG, Tuft SJ et al. Corneal stem cells in review. Wound Repair Regen 2001; 9: 483494.
  • 3
    Dua HS, Shanmuganathan VA, Powell-Richards AO et al. Limbal epithelial crypts: A novel anatomical structure and a putative limbal stem cell niche. Br J Ophthalmol 2005; 89: 529532.
  • 4
    Li W, Hayashida Y, Chen Y-T et al. Niche regulation of corneal epithelial stem cells at the limbus. Cell Res 2007; 17: 2636.
  • 5
    Davanger M, Evensen A. Role of the pericorneal papillary structure in renewal of corneal epithelium. Nature 1971; 229: 560561.
  • 6
    Townsend WM. The limbal palisades of Vogt. Trans Am Ophthalmol Soc 1991; 89: 721756.
  • 7
    Chen Z, de Paiva CS, Luo L et al. Characterization of putative stem cell phenotype in human limbal epithelia. Stem Cells 2004; 22: 355366.
  • 8
    Shortt AJ, Secker GA, Munro PM et al. Characterization of the limbal epithelial stem cell niche: Novel imaging techniques permit in vivo observation and targeted biopsy of limbal epithelial stem cells. Stem Cells 2007; 25: 14021409.
  • 9
    Duke-Elder S, Cook C. Normal and abnormal development. Part 1: Embryology. In: Duke-ElderS, ed. System of Ophthalmology (Vol. III). London: Henry Kimpton, 1963; 137171.
  • 10
    Sevel D, Isaacs R. A re-evaluation of corneal development. Trans Am Ophthalmol Soc 1988; 86: 178207.
  • 11
    Rodrigues M, Ben-Zvi A, Krachmer J et al. Suprabasal expression of a 64-kilodalton keratin (no. 3) in developing human corneal epithelium. Differentiation 1987; 34: 6067.
  • 12
    Figueira EC, Di Girolamo N, Coroneo MT et al. The phenotype of limbal epithelial stem cells. Invest Ophthalmol Vis Sci 2007; 48: 144156.
  • 13
    Yoshida S, Shimmura S, Kawakita T et al. Cytokeratin 15 can be used to identify the limbal phenotype in normal and diseased ocular surfaces. Invest Ophthalmol Vis Sci 2006; 47: 47804786.
  • 14
    Du Y, Chen J, Funderburgh JL et al. Functional reconstruction of rabbit corneal epithelium by human limbal cells cultured on amniotic membrane. Mol Vis 2003; 9: 635643.
  • 15
    Lin S, Lee O-T, Minasi P et al. Isolation, culture and characterisation of human fetal trabecular meshwork cells. Curr Eye Res 2007; 32: 4350.
  • 16
    Barrandon Y, Green H. Three clonal types of keratinocyte with difference capacities for multiplication. Proc Natl Acad Sci USA 1987; 84: 23022306.
  • 17
    Di Girolamo N, Bosch M, Zamora K et al. A contact lens-based technique for expansion and transplantation of autologous epithelial progenitors for ocular surface reconstruction. Transplantation 2009; 87: 15711578.
  • 18
    Ng J, Coroneo MT, Wakefield D et al. Ultraviolet radiation and the role of matrix metalloproteinases in the pathogenesis of ocular surface squamous neoplasia. Invest Ophthalmol Vis Sci 2008; 49: 52955306.
  • 19
    Di Girolamo N, Chui J, Wakefield D et al. Cultured human ocular surface epithelium on therapeutic contact lenses. Br J Ophthalmol 2007; 91: 459464.
  • 20
    Di Girolamo N, Sarris M, Chui J et al. Localization of the low-affinity nerve growth factor receptor p75 in human limbal epithelial cells. J Cell Mol Med 2008; 12: 27992811.
  • 21
    Qi H, Li D-Q, Shine HD et al. Nerve growth factor and its receptor TrkA serve as potential markers for human corneal epithelial progenitor cells. Exp Eye Res 2008; 86: 3440.
  • 22
    Thoft RA, Friend J. The X, Y, Z hypothesis of corneal epithelial maintenance. Invest Ophthalmol Vis Sci 1983; 24: 14421443.
  • 23
    Bozani´c D, Tafra R, Saraga-Babic M. Role of apoptosis and mitosis during human eye development. Eur J Cell Biol 2003; 82: 421429.
  • 24
    Scholzen T, Gerdes J. The Ki-67 protein: From the known and the unknown. J Cell Physiol 2000; 182: 311322.
  • 25
    Barnard NJ, Hall PA, Lemoine NR et al. Proliferative index in breast carcinoma determined in situ by Ki67 immunostaining and its relationship to clinical and pathological variables. J Pathol 1987; 152: 287295.
  • 26
    Moll R, Franke WW, Schiller DL et al. The catalog of human cytokeratins: Patterns of expression in normal epithelia, tumours and cultured cells. Cell 1982; 31: 1124.
  • 27
    Kasper M, Moll R, Stosiek P et al. Patterns of cytokeratin and vimentin expression in the human eye. Histochemistry 1988; 89: 369377.
  • 28
    Yang A, Schweitzer R, Sun D et al. p63 is essential for regenerative proliferation in limb, craniofacial and epithelial development. Nature 1999; 398: 714718.
  • 29
    Di Iorio E, Barbaro V, Ruzza A et al. Isoforms of DeltaNp63 and the migration of ocular limbal cells in human corneal regeneration. Proc Natl Acad Sci USA 2005; 102: 95239528.
  • 30
    Kawasaki S, Tanioka H, Yamasaki K et al. Expression and tissue distribution of p63 isoforms in human ocular surface epithelia. Exp Eye Res 2006; 82: 293299.
  • 31
    Watanabe K, Nishida K, Yamato M et al. Human limbal epithelium contains side population cells expressing the ATP-binding cassette transporter ABCG2. FEBS Lett 2004; 565: 610.
  • 32
    Yew DT, Sha O, Li WWY et al. Proliferation and apoptosis in the epithelium of the developing human cornea and conjunctiva. Life Sci 2001; 68: 29873003.
  • 33
    Castro-Combs J, Noguera G, Cano M et al. Corneal wound healing is modulated by topical application of amniotic fluid in an ex vivo organ culture model. Exp Eye Res 2008; 87: 5663.
  • 34
    Blahovec J, Kostecka Z, Lacroix MC et al. Mitogenic activity of high molecular weight forms of insulin-like growth factor-II in amniotic fluid. J Endocrinol 2001; 169: 563572.
  • 35
    Maseruka H, Ridgway A, Tullo A et al. Developmental changes in patterns of expression of tenascin-C variants in the human cornea. Invest Ophthalmol Vis Sci 2000; 41: 41014107.
  • 36
    Yeung AM, Tint NL, Kulkarni BB et al. Infant limbus: An immunohistochemical study. Exp Eye Res 2009; 88: 11611164.
  • 37
    Lesueur L, Arne JL, Mignon-Conte M et al. Structural and ultrastructural changes in the developmental process of premature infants' and children's corneas. Cornea 1994; 13: 331338.
  • 38
    Collinson JM, Chanas SA, Hill RE et al. Corneal development, limbal stem cell function, and corneal epithelial cell migration in the Pax6+/− mouse. Invest Ophthalmol Vis Sci 2004; 45: 11011108.
  • 39
    Sugar J. Chapter 7: congenital stem cell deficiency. In: HollandEJ, MannisMJ, eds. Ocular Surface Disease: Medical and Surgical Management. New York: Springer, 2002: 9399.
  • 40
    Lee H, Khan R, O'Keefe M. Aniridia: Current pathology and management. Acta Ophthalmol 2008; 86: 708715.
  • 41
    Blache P, van de Wetering M, Duluc I et al. Sox9 is an intestine crypt transcription factor, is regulated by the Wnt pathway, and represses the CDX2 and MUC2 genes. J Cell Biol 2004; 166: 3747.
  • 42
    Paus R, Muller-Rover S, van der Veen C et al. A comprehensive guide for the recognition and classification of distinct stages of hair follicle morphogenesis. J Invest Dermatol 1999; 113: 523532.
  • 43
    Lamkanfi M, Festjens N, Declercq W et al. Caspases in cell survival, proliferation and differentiation. Cell Death Differ 2007; 14: 4455.
  • 44
    Pearton DJ, Ferraris C, Dhouailly D. Transdifferentiation of corneal epithelium: Evidence for a linkage between the segregation of epidermal stem cells and the induction of hair follicles during embryogenesis. Int J Dev Biol 2004; 48: 197201.
  • 45
    Akiyama M, Smith LT, Shimizu H. Changing patterns of localization of putative stem cells in developing human hair follicles. J Invest Dermatol 2000; 114: 321327.
  • 46
    Chung E-H, Bukusoglu G, Zieske JD. Localization of corneal epithelial stem cells in the developing rat. Invest Ophthalmol Vis Sci 1992; 33: 21992206.
  • 47
    Zieske JD, Wasson M. Regional variation in distribution of EGF receptor in developing and adult corneal epithelium. J Cell Sci 1993; 106: 145152.
  • 48
    Zieske JD. Corneal development associated with eyelid opening. Int J Dev Biol 2004; 48: 903911.
  • 49
    Lyngholm M, Høyer PE, Vorum H et al. Immunohistochemical markers for corneal stem cells in the early developing human eye. Exp Eye Res 2008; 87: 115121.
  • 50
    Laurikkala J, Mikkola ML, James M et al. p63 regulates multiple signalling pathways required for ectodermal organogenesis and differentiation. Development 2006; 133: 15531563.
  • 51
    Liu J, Song G, Wang Z et al. Establishment of a corneal epithelial cell line spontaneously derived from human limbal cells. Exp Eye Res 2007; 84: 599609.
  • 52
    Green H. The birth of therapy with cultured cells. Bioessays 2008; 30: 897903.
  • 53
    Chang CY, Green CR, McGhee CN et al. Acute wound healing in the human central corneal epihtelium appears to be independent of limbal stem cell influence. Invest Ophthalmol Vis Sci 2008; 49: 52795286.
  • 54
    Li W, Hayashida Y, He H et al. The fate of limbal epithelial progenitor cells during explant culture on intact amniotic membrane. Invest Ophthalmol Vis Sci 2007; 48: 605613.
  • 55
    Zhang ZY, Teoh SH, Chong MS et al. Superior osteogenic capacity for bone tissue engineering of fetal compared with perinatal and adult mesenchymal stem cells. Stem Cells 2009; 27: 126137.
  • 56
    Vrana NE, Builles N, Justin V et al. Development of a reconstituted cornea from collagen-chondroitin sulfate foams and human cell cultures. Invest Ophthalmol Vis Sci 2008; 49: 53255321.
  • 57
    Whitcher JP, Srinivasan M, Upadhyay MP. Corneal blindness: A global perspective. Bull World Health Org 2001; 79: 214221.

Supporting Information

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. SUMMARY
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES
  11. Supporting Information

Additional supporting information available online.

FilenameFormatSizeDescription
STEM_209_sm_SuppInfo.doc117KSupporting Information

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.