Proliferation and apoptosis in the epithelium of the developing human cornea and conjunctiva

Life Sciences 68 (2001) 2987–3003

Proliferation and apoptosis in the epithelium of the developing human cornea and conjunctiva D.T. Yewa,*, O. Shaa, W.W.Y. Lib, T.T. Lamb, D.E. Lorkec a

Department of Anatomy, The Chinese University of Hong Kong, Shatin, New Territories, Hong Kong, China b Department of Ophthalmology and Visual Sciences, The Chinese University of Hong Kong, Shatin, New Territories, Hong Kong, China c Department of Neuroanatomy, University Clinic Eppendorf, Martinistr. 52, 20246 Hamburg, Germany Received 9 June 2000; accepted 18 December 2000

Abstract To determine the distribution of proliferating and apoptotic cells in the human cornea during prenatal and early postnatal development, we examined sections of the bulbar conjunctiva, the limbus as well as the central and peripheral cornea between 11 weeks of gestation and 6 months after birth. The objective was to localize dividing cells by proliferating cell nuclear antigen-like immunoreactivity (PCNA-LI) and apoptotic cells by terminal transferase-mediated nick-end labeling (TUNEL). Before the 17th gestational week, PCNA-LI was absent in all 4 regions examined, indicating negligible cell proliferation during early development. After 20 weeks, strong PCNA-labeling was observed in all regions examined suggestive of high proliferative activity not only in the limbus and the bulbar conjunctiva, but also in the central and peripheral cornea. This rise in proliferative activity was followed by a steady decline: after 28 weeks, anti-PCNA staining gradually disappeared in the central and peripheral cornea, so that, at 6 months after birth, it was confined to the limbus and the bulbar conjunctiva, resembling the picture described for the adult cornea. TUNEL-positive cells were virtually absent in all 4 regions examined before the 38th gestational week. Apoptotic cells only started to appear at 38 weeks; at this stage, they were confined to the bulbar conjunctival epithelium. At 6 months after birth, TUNELpositive cells were observed in the bulbar conjunctival epithelium and the entire cornea; the limbus, however remained devoid of apoptotic cells throughout the entire prenatal and early postnatal period. The present study for the first time localizes proliferating and apoptotic cells in the epithelium of the developing human cornea. Three stages of development can be distinguished: Minimal proliferation (until 17th week), vigorous proliferation over the entire cornea including the limbus and the bulbar conjunctiva (until 28th week) and gradual decrease in proliferative activity (after 28th week) accompanied by the appearance of apoptotic cells. © 2001 Elsevier Science Inc. All rights reserved. Keywords: Apoptosis; Conjunctiva; Cornea; Development; Human; Limbus; PCNA; Proliferation; TUNEL

* Corresponding author. Tel.: (852)-2609-6899; fax: (852)-2603-5031 E-mail: [email protected] (D.T. Yew) 0024-3205/01/$ – see front matter © 2001 Elsevier Science Inc. All rights reserved. PII: S 0 0 2 4 - 3 2 0 5 ( 0 1 )0 1 0 9 9 -2

2988

D.T. Yew et al. / Life Sciences 68 (2001) 2987–3003

Introduction The identification of proliferating cells in the adult corneal epithelium has been the subject of intensive research given the great clinical importance of cell renewal, particularly during wound healing after corneal injury [for review, see 1, 2]. A large number of studies on the adult rodent [3–6] and human [7, 8] cornea have shown that corneal epithelium undergoes continuous cell renewal and regeneration throughout adult life. Cells desquamated from the corneal surface are replaced by mitotic division of a small contingent of cells called “stem cells” [9] which are located in the limbus of the cornea, the peripheral area between the cornea and the adjacent bulbar conjunctiva. These stem cells undergo unlimited self-replication and have a relatively slow turnover rate. They give rise to “transient amplifying cells” derived from each stem cell mitosis. Transient amplifying cells multiply by undergoing a few additional rounds of cell divisions, but are already committed to cellular differentiation; they are located in the basal epithelia of the limbus and the peripheral cornea [1, 2, 4]. During subsequent maturation, corneal epithelial cells move to the center of the cornea (centripetal migration), where dying cells are lost by desquamation from the surface. In contrast, studies on cellular proliferation in the developing cornea are rare. Until recently, the only techniques available for assessing proliferative activity have been counts of mitotic figures, which is a rather crude technique suffering from many limitations [10], or measurements of the incorporation of thymidine analogues, e.g. BrdU or tritiated thymidine. These techniques can only be performed in viable cells and are therefore not applicable in the developing human fetus. Consequently, recent studies on proliferative activity in the developing cornea have been mainly performed in the avian and rodent eye [11–13], and studies on the developing human cornea have been confined to morphological evaluation [14–16], enzyme histochemistry [17, 18] and immunohistochemical demonstration of differentiationrelated parameters, e.g. keratins [19], actin filaments [20], extracellular matrix maturation [21] or PAX6 expression [22]. With the discovery of the proliferating cell nuclear antigen (PCNA), which can be detected by immunohistochemistry in formaldehyde-fixed specimens [23], the assessment of proliferative activity in post mortem specimens has become possible [10, 24, 25]. PCNA is an auxiliary protein of DNA polymerase associated with the DNA replication fork during the S phase [26]. Its rate of synthesis is correlated with the proliferative rate of cells [27], so that PCNA-like immunoreactivity (PCNA-LI) can be used as a reliable marker for epithelial proliferative activity [4, 5, 8, 28]. Apoptosis, programmed cell death, has been shown to be as important to development as proliferation and differentiation [29]; it represents a major regulatory mechanism in eliminating abundant and unwanted cells during embryonic development [30]. Apoptotic cell death is an active gene-directed process of self destruction characterized by the fragmentation of genomic DNA [for review, see 31–34]. This phenomenon can be detected using the terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling (TUNEL) assay where DNA breaks are labeled [33, 35]. The aim of the present study is to assess the proliferative activity in the bulbar conjunctiva, the limbus and the cornea of immersion-fixed developing human fetuses by means of immunohistochemical detection of PCNA and to determine whether apoptosis plays a role during human corneal development.

D.T. Yew et al. / Life Sciences 68 (2001) 2987–3003

2989

Methods Human specimens Eyes were dissected bilaterally from ten autopsies obtained through a collaborative collection of the Department of Anatomy, Chinese University of Hong Kong, and Jinan University. These fetuses, five male and five female, were obtained with the permission of the patients and the hospitals concerned; their medical history revealed no eye diseases. Fetal age (conceptual age), estimated by crown rump length [36], was 11, 14, 16, 17, 20, 23, 28, 38 weeks of gestation, 4 and 6 months postnatal. The eyeballs were enucleated and immersion fixed in 4% buffered paraformaldehyde [4]. They were then dissected into halves along the middle coronal plane, and the anterior (corneal) parts were subsequently divided into two halves along the median sagittal plane. Finally, each specimen consisted of one quarter of the eyeball, each containing tissue from the cornea as well as the limbus and the bulbar conjunctiva. Specimens were then dehydrated in graded alcohols, cleared in xylene and embedded in paraffin wax (Histosec, Merck, Germany). 8mm-sections were mounted on gelatin-coated slides and stored for subsequent reactions. Histology Sections were deparaffinized and rehydrated through xylene and graded alcohol. For routine histological evaluation, sections were stained with haematoxylin and eosin. Immunohistochemistry (PCNA staining): To block endogenous peroxidase activity, rehydrated sections were pretreated for 5 min with methanol containing 0.4% H2O2. After 2 rinses in 0.01M phosphate buffered saline (PBS; pH 7.4), sections were incubated in diluted normal blocking serum (Elite Vecta Stain ABC Kit, Mouse IgG) for 30 min and then incubated in (1st) monoclonal mouse proliferating cell nuclear antigen (clone PC10, M 0879, DAKO A/S, Glostrup, Denmark, 1:100) for 1 h; (2nd) biotinylated secondary antibody (Elite Vecta Stain ABC Kit, Vector Laboratories, CA, 1:200) for 1 h; (3rd) ABC Reagent (Elite Vecta Stain ABC Kit, Vector Laboratories, CA) for 1 h. Sections were rinsed 3 times in 0.01M PBS inbetween. The immunocytochemical staining signals were visualized by incubating the sections in 0.05% of the substrate, i.e. 3939-diaminobenzidine tetrahydrochloride (DAB) in 0.01M PBS containing 0.01% H2O2. All stainings included negative controls with omission of the primary antibody, which did not show any immunoreaction. TUNEL: Apoptosis was detected using terminal dUTP nick-end labeling (TUNEL) assays. The ApopTagTM peroxidase kit (Oncor, Gaithersburg, MD) was applied following the manufacturer’s protocol. Briefly, after deparaffination, rehydration, quenching of endogenous peroxidase activity and predigestion with 20 mg/ml proteinase K, the sections were incubated with an equilibration buffer containing dUTP-digoxigenin. Terminal deoxynucleotidyl transferase (TdT) was added to end-label the 39 ends of the DNA fragments; the reaction was stopped by dipping the slides in the stop buffer. Thereafter, anti-digoxigenin antibody conjugated with peroxidase was added, followed by incubation in 0.5 mg/ml DAB. For positive control, retina sections from the rcs rat mutant rat strain with known apoptotic cells were used [37]. For negative control, the TdT enzyme was substituted by distilled water. All the incubations were performed in a humid chamber at room temperature. Finally, the sections were dehydrated, cleared and mounted.

2990

D.T. Yew et al. / Life Sciences 68 (2001) 2987–3003

Evaluation: The ocular surface was divided into four regions: bulbar conjunctiva, limbus, central cornea and peripheral cornea. “Bulbar conjunctiva” was the conjunctiva adjacent to the limbus, characterized by goblet cells and underlying blood vessels. “Limbus” was distinguished by the presence of the chamber angle. The cornea was characterized by an underlying Bowman’s membrane; “central cornea” was selected from the midarea of the cornea, “peripheral cornea” from the area adjacent to the limbus [3, 11, 38]. Results At the earliest stages examined (between 11 and 17 gestational weeks), the epithelium of the bulbar conjunctiva was still relatively thin. It consisted of one basal layer of cuboidal cells and only one to two superficial layers of compacted squamous cells (Fig. 1a). The limbus, located at the junction between the cornea and the bulbar conjunctiva, was also thin; its epithelium was made up of a single layer of cuboidal to columnar cells. Close to the cornea, an additional compact layer of squamous cells was found on top of the basal columnar cells

Fig. 1 a. Bulbar conjunctiva of human fetus at 14 gestational weeks lined by an epithelium consisting of 2–3 cell layers. b: Limbus near the cornea at 14 gestational weeks showing a 2-layered epithelium composed of basal cuboidal and a thin layer of superficial squamous cells (arrowheads). c: Cornea of human fetus at 11 gestational weeks consisting of a stroma with numerous keratocytes (arrows) covered by an epithelium of basal cuboidal and extremely flattened superficial cells (arrowheads). (HE stain, 3200)

D.T. Yew et al. / Life Sciences 68 (2001) 2987–3003

2991

Fig. 2. 4 weeks of gestation, PCNA immunoreaction: the epithelia (arrowheads) of the bulbar conjunctiva (a), the limbus (b) and the cornea (c) are devoid of PCNA labeling. (3200)

(Fig. 1b). The one- to two-layered corneal epithelium was composed of cuboidal basal cells and flattened superficial cells (Fig. 1c). The underlying stroma contained numerous keratocytes; the inner border of the cornea was lined by a one-layered cuboidal endothelium (Fig. 1c). At this time, no significant PCNA-LI could be observed in either of these regions: the bulbar conjunctiva (Fig. 2a), as well as the limbus (Fig. 2b) and the cornea (Fig. 2c) were de-

2992

D.T. Yew et al. / Life Sciences 68 (2001) 2987–3003

Fig. 3. 20 weeks of gestation, HE stain: the bulbar conjunctiva (a) shows extensive infoldings (crypts) in its multilayered pseudostratified columnar epithelium; goblet cells (arrow) are now visible. Both the limbus (b) and the cornea (c) show a 2-layered epithelium composed of basal cuboidal and compacted superficial squamous cells. Note the distinct Bowman’s membrane (arrowheads) underlying the epithelium of the cornea. (a, b: 3200; c: 3400)

void of PCNA staining, indicating that cell proliferation is minimal at this early stage of human corneal development. The most obvious developmental change observed in the fetuses between the 20th and the 23rd gestational week was the marked increase in thickness of the bulbar conjunctiva which

D.T. Yew et al. / Life Sciences 68 (2001) 2987–3003

2993

Fig. 4 a, b. 20 weeks of gestation, PCNA immunoreaction: the epithelium of the bulbar conjunctiva (a) shows some PCNA positive cells located both superficially (arrowheads) and deeply (arrow). b: transitional region between limbus (arrows) and bulbar conjunctiva (arrowheads). Note that, in the limbus, PCNA labeled cells are only found in the deeper layer (arrows), whereas in the conjunctiva, staining is also observed in the superficial layer (arrowheads). c, d: Cornea of human fetus at 23 weeks of gestation showing PCNA labeling in the superficial squamous cells (arrowheads) of its center (c), and in both superficial (arrowheads) and deep (arrows) layers of its periphery (d). (a,c 3400; b,d 3600)

consisted of a pseudostratified squamous epithelium of 4 layers with invaginated crypts (Fig. 3a), occasionally, goblet cells could be detected. In contrast, histology of the limbus had hardly changed: it consisted of a single layer of basal columnar cells covered by one layer of flattened squamous cells (Fig. 3b). Like the limbus, the corneal epithelium at this stage was still thin consisting of a basal layer of cuboidal to columnar cells and a superficial layer of flattened squamous cells (Fig. 3c). Underneath, a clear homogenous and amorphic Bowman’s membrane had become recognizable (Fig. 3c). At this stage, PCNA-LI cells were detected in

2994

D.T. Yew et al. / Life Sciences 68 (2001) 2987–3003

both the superficial and the deep layers of the bulbar conjunctival epithelium (Fig. 4a), whereas staining was confined to the basal layer of the limbus (Fig. 4b). In the corneal epithelium, PCNA positive cells were only present in the superficial epithelial cell layer of the center (Fig. 4c); in contrast, they were detected in both the superficial and deeper layers of the peripheral cornea (Fig. 4d). PCNA staining was not only detected in the nuclei of these cells but also frequently in their cytoplasm. Between 28 and 38 gestational weeks, the limbus contained two to three cell layers and was hard to delineate from the bulbar conjunctiva as a separate region (Fig. 5a). The threelayered corneal epithelium of fetuses of these ages was similar in appearance to the limbus; it was made up of one row of basal cuboidal cells underlying two layers of squamous cells (Fig. 5b). Some of the epithelial cells in the superficial layer had lightly stained cytoplasm (Fig. 5b). At this stage, the number of PCNA-LI cells had increased in the bulbar conjunctiva; staining was again seen in the superficial and in the deeper layers of the epithelium (Fig. 6a). Often, not only the nucleus, but also the cytoplasm was labeled with PCNA activities. In the limbus, PCNA-LI cells were only present in the basal layer of the epithelium (Fig. 6b); as before, either nuclei or cytoplasm were labeled. In the corneal epithelium of the center, PCNA positive cells were still present in the surface layer, showing cytoplasmic labeling (Fig. 6c); in contrast, the peripheral cornea was devoid of anti-PCNA staining (Fig. 6d). At 6 months postnatal, the epithelium of the bulbar conjunctiva was usually made up of 5– 6 layers (Fig. 7a). The limbus was clearly delineated at this older age, having an epithelium of only one layer (Fig. 7b). The corneal epithelium consisted of about 4 cell layers (Fig. 7c). PCNA staining in the bulbar conjunctiva was again observed in both the deep and superficial layers, it was detected in both the nuclei and the apical cytoplasmic regions (Figs. 8a). In contrast, PCNA activity was confined to cell nuclei of the epithelial cells of the limbus (Fig. 8b). In the corneal epithelium staining was no longer detectable (Fig. 8c). During initial prenatal development, i.e. before the 38th gestational week, virtually no positive TUNEL staining was observed in the bulbar conjunctiva, the limbus, the central and the pe-

Fig. 5. 28 weeks of gestation, HE stain. a: Transitional zone between bulbar conjunctiva (arrowheads) and limbus (asterisks). Note that the epithelium of the limbus is still composed of 2–3 cell layers, the shape of the superficial cells being squamous. b: Cornea: Note that the most superficial layer of squamous cells in the peripheral cornea (asterisk) shows a dark staining, whereas in the central region (arrowheads) of the cornea, superficial cells have a lightly stained cytoplasm. (3140)

D.T. Yew et al. / Life Sciences 68 (2001) 2987–3003

2995

Fig. 6. a: Bulbar conjunctiva of human fetus, 28 weeks of gestation, showing PCNA positive cells in superficial (arrowheads) and basal (arrows) layers. b: PCNA-positive cells in the limbus at 28 weeks of gestation. Note that they are confined to the basal layer (arrow). c,d: PCNA reaction in the cornea at 38 weeks of gestation, showing positive cells (arrowheads) in the central (c), but not in the peripheral cornea (d). (3400)

ripheral cornea, indicating that there was insignificant apoptosis (Fig. 9a). At 38 gestational weeks, some apoptotic cells were observed in the bulbar conjunctival epithelium (Fig. 9b), whereas the entire corneal epithelium and the limbus were still devoid of positive TUNEL staining (Fig. 9c). At 6 months postnatal, an increase in apoptotic cells was observed in the bulbar conjunctival epithelium. At this stage, TUNEL-positive cells were also seen in the central and the peripheral parts of the cornea, both superficially and in the deeper layers (Fig. 10a). Their density was, however, lower than in the conjunctival epithelium. The limbus remained devoid of TUNEL-positive cells throughout the entire prenatal and early postnatal period (Fig. 10b). Discussion The present study has been undertaken in order to localize proliferating cells in the human cornea during prenatal and early postnatal development. The proliferative activity of the bul-

2996

D.T. Yew et al. / Life Sciences 68 (2001) 2987–3003

Fig. 7. 6 months postnatal, HE stain: the bulbar conjunctiva (a) consists of a 5–6 layered pseudostratified epithelium, the limbus (b) shows a single layer of columnar cells, and the corneal epithelium (c) is composed of about 4 layers, the superficial layer being squamous. (3400)

bar conjunctival, limbal and corneal epithelium has been visualized by PCNA-LI. Our results show that in the oldest specimens examined, i.e. the human cornea at 6 months after birth, PCNA activity is restricted to the basal layer of the limbus. This observation indicates that the distribution of proliferating cells in the early postnatal human cornea is already similar to that of the adult. Numerous studies reveal that cell proliferation in the adult cornea is confined to its limbal region [1–4, 9]. In this transitional zone between the cornea and the conjunctiva, stem cells are located undergoing mitotic division. Their function is to replace

D.T. Yew et al. / Life Sciences 68 (2001) 2987–3003

2997

Fig. 8. 6 months postnatal: in the bulbar conjunctiva (a), positive PCNA reaction is observed in the nuclei (arrows) and the apical cytoplasmic region (arrowheads) of some cells. In the limbal epithelium (b) staining is mainly found in the nuclei of the basal layer (arrows). The cornea (c) shows no significant PCNA reaction. (a,b: 3200; c: 3400)

cells that are continuously lost from the corneal surface. During this process of cell renewal, postmitotic cells are displaced into the suprabasal layers (vertical movement), migrate from the periphery to the center, becoming increasingly more mature during this centripetal movement, and are finally desquamated [2, 6, 19]. In addition, we have demonstrated that the distribution of proliferative activity during prenatal development is strikingly different from this adult and early postnatal picture. At the

2998

D.T. Yew et al. / Life Sciences 68 (2001) 2987–3003

Fig. 9. TUNEL reaction for the detection of apoptotic cells. a: Cornea of human fetus, 11 weeks of gestation, devoid of TUNEL staining. b, c: 38 weeks of gestation: Whereas labeled nuclei (arrows) are detected in the bulbar conjunctiva (b), the cornea (c) still does not exhibit any TUNEL reaction. (a,c: 3200; b:3250)

earliest stage examined, the human cornea almost completely lacks PCNA-LI, indicating negligible cell proliferation between 11 and 17 gestational weeks. Virtual absence of mitotic cell division during early corneal development has also been observed in the rodent eye [13]. At this early stage of development, the human central corneal, limbal and bulbar conjunctival epithelium are each composed of only one to two cell layers. This histological aspect is very similar to that of the rodent cornea during the first postnatal week [11]. BrdU-labeling studies in the mouse have shown that the proliferative rate of the corneal epithelium is also very low at this stage of development, with only very few cells being labeled which are randomly distributed in the basal cell layer of the entire cornea [13]. Our older specimens show that between 20 and 23 gestational weeks, PCNA-LI appears in all regions examined. This indicates a sharp increase in proliferative activity involving all parts of the cornea, both central and peripheral, including the limbus and the adjacent bulbar conjunctiva. A similar sharp rise in the proliferative rate has also been observed in the murine cornea during the first postnatal week using BrdU-labeling [13]. As in the human cornea,

D.T. Yew et al. / Life Sciences 68 (2001) 2987–3003

2999

Fig. 10. Human fetus, 6 months postnatal: The cornea (a) shows positive TUNEL nuclei (arrows) in superficial and deep layers of epithelium. In contrast, the limbus (b) remains devoid of TUNEL staining. (3200)

cells with high mitotic activity are evenly distributed throughout the murine cornea during this period of most vigorous cell proliferation; they are, however, confined to its basal layer. In contrast, we have also found PCNA-LI cells in the superficial layers of the central and peripheral cornea of the human fetus aged 20–23 gestational weeks. This positive PCNA staining observed in the superficial layers may indicate that cell division in the developing human cornea is not confined to the basal cell layer. This interpretation seems, however, unlikely to us, because basal lamina attachment is generally essential for epithelial cell proliferation [39, 40]. After displacement into suprabasal layers, cells tend to become postmitotic and lose their capacity for cell division [2]. It is therefore more probable that the labeled cells observed in the superficial layers represent postmitotic cells shortly after their last mitotic division, which are still stained due to the specific metabolism of PCNA. PCNA is a ubiquitous and tightly regulated cell cycle specific nuclear protein belonging to the replication complex permitting DNA synthesis [26, 41]; its expression is an obligatory event in G1 to S transition. Two forms of PCNA have been described: a soluble form lost on organic solvent fixation and not involved in replication and an insoluble form tightly associated with ongoing DNA-synthesis

3000

D.T. Yew et al. / Life Sciences 68 (2001) 2987–3003

[41]. The monoclonal antibody employed in the present study (clone PC10) has been shown to recognize the insoluble form, to be highly specific for PCNA and to bind to it with a particularly high affinity [23]. Synthesis of PCNA increases significantly during S phase, with peak synthesis in early S phase [42], and correlates with the proliferative state of the cell [27]. For that reason, anti-PCNA immunoreactivity is employed as a reliable marker of cell proliferation. There is, however indication, that due to the relatively long half-life of PCNA, anti-PCNA staining can persist for a short period after the last mitosis [4, 10]. The half-life of PCNA is about 20 h, and PCNA has been detected by immunological methods in cells that have recently left the cell cycle [24]. It is therefore conceivable that the PCNA-LI cells observed in the superficial layers of the cornea represent postmitotic cells shortly after their last division, which have detached from the basement membrane and moved to the corneal surface. During this period of rapid corneal proliferation, PNCA-LI has not been confined to the nucleus, but has also occasionally been observed in the cytoplasm. Being closely related to DNA replication, PCNA is a nuclear protein [23]. There are, however, several reports that PCNA is present in the cytoplasm [43–45]. Applying the same monoclonal antibody as in our investigation, cytoplasmic staining has also been described in the limbus of the adult rodent cornea [4]. This study includes a very thorough assessment of different fixation procedures upon PCNA epitope preservation, thereby virtually excluding fixation artifact as a cause of cytoplasmic labeling. Before assuming its function in the nucleus, PCNA is synthesized in the cytoplasm and is then transferred into the nucleus. Consequently, during liver regeneration, PCNA is first found in the cytoplasm and later detected in the nuclei [44]. In a recent biochemical analysis of PCNA expression in normal and regenerating rat livers, Tanno and Taguchi [46] describe 3 different forms of PCNA, two of which have been detected in the cytoplasm. They hypothesize that, depending on the presence of a nuclear localization signal, PCNA is either transferred into the nucleoplasm or remains in the cytoplasm. The present results indicate that the cytoplasmic type may be detectable by immunochemistry in the proliferative compartment of the adult cornea [4], but also, as shown in our study, in rapidly dividing corneal and bulbar epithelial cells of the fetal human. During later developmental stages (fetuses older than 28 gestational weeks), PCNA labeling decreases in the human cornea; it first disappears in the peripheral cornea and later-on also in the central cornea, so that staining is confined to the limbus and the bulbar conjunctiva in our oldest specimens. This indicates that after a limited period during which cell divisions are found in the entire cornea, proliferative activity becomes restricted to only the limbus of the cornea and the bulbar conjunctiva. A similar decline in proliferative activity has been observed during the development of the murine cornea: after a period of high proliferative rate during which BrdU incorporation is observed in the whole cornea, the proliferative rate diminishes [13]. This decrease in the proliferative rate in the cornea is closely related to ongoing differentiation of the corneal epithelium: Whereas a 64-kDa corneal keratin, regarded as a marker for an advanced stage of corneal epithelial differentiation, is virtually absent in the human cornea at 12 weeks of age, it is expressed by the suprabasal layers of the entire cornea at 36 weeks of age [19]. Only in the limbal-conjunctival junction, containing mitotically active stem cells, 64-kDa corneal keratin expression remains low. In contrast, type III collagen is only found in the immature epithelial basement membrane of the human cornea, but can no longer be detected after 27 weeks of gestation [21].

D.T. Yew et al. / Life Sciences 68 (2001) 2987–3003

3001

In addition, our study shows that apoptosis only plays a minor role during early human corneal development. Neither during the initial development (before the 17th gestational week), characterized by negligible proliferative activity, nor during the subsequent phase of vigorous cell proliferation (until the 28th gestational week), TUNEL-positive cells are observed. Only during the final developmental stages examined, few apoptotic cells can be detected, first in the bulbar conjunctiva (after 38 gestational weeks), later (postnatally) also in the cornea. Paucity of apoptosis during early corneal development is also indicated by Western blotting analyses of caspase 3 expression currently performed in our laboratory. Preliminary results suggest that caspase 3 expression in the developing cornea is very low before gestational week 21 and gradually increases thereafter. The distribution of TUNEL-positive cells in the 6-month-old human cornea slightly differs from that of the adult. Whereas at 6 months after birth, TUNEL-positive cells are located both in the superficial and the deep layers of the cornea, they are confined to the surface of the cornea in the adult rodent [47, 48] and human [49]. Both in the adult and in the developing cornea, the limbus is devoid of apoptotic cells. The fact that the limbus hosting high proliferative activity is devoid of apoptotic cells both in the adult and in the developing cornea and that apoptotic cells only appear in the developing cornea after the period of most vigorous proliferative activity indicates that apoptosis is inversely related to cell proliferation in the human cornea. Conclusion The present study for the first time assesses the localization of proliferating and apoptotic cells in the embryonic human cornea. The results indicate that development of the human cornea proceeds in 3 stages. During the first stage (until 17th week), cell proliferation and apoptosis are minimal. The 2nd stage (until 28th week) is characterized by a high proliferative activity, which can be observed in all parts of the cornea; apoptotic cell death remains negligible. Throughout this stage, the cornea mainly grows in diameter [50, 51]; the number of its cell layers only increases from 2 to 3 [present results, 15]. In the 3rd stage (after 28th week), cell proliferation decreases and becomes confined to the limbus, which is the location of stem cells in the adult cornea. During this stage, which is also characterized by ongoing differentiation of the corneal epithelium [19, 21], the cornea grows both in diameter [50, 51] and in thickness, with the number of its cell layers increasing to about five [present results, 15]. In this final stage, apoptotic cells start to appear, first in the conjunctiva and later also in the cornea.

References 1. Tseng SCG. Concept and application of limbal stem cells. Eye 1989;3:141–57. 2. Kruse FE. Stem cells and corneal epithelial regeneration. Eye 1994;8:170–83. 3. Cotsarelis G, Cheng S-Z, Dong G, Sun T-T, Lavker RM. Existence of slow-cycling limbal epithelial basal cells that can be preferentially stimulated to proliferate: implications on epithelial stem cells. Cell 1989; 57:201–9. 4. Gan L, Van Setten G, Seregard S, Fagerholm P. Proliferating cell nuclear antigen colocalization with corneal epithelial stem cells and involvement in physiological cell turnover. Acta Ophthalmologica Scandinavica 1995;73:491–5.

3002

D.T. Yew et al. / Life Sciences 68 (2001) 2987–3003

5. Gan L, Fagerholm P, Kim H-J. Effect of leukocytes on corneal cellular proliferation and wound healing. Investigative Ophthalmology and Visual Science 1999;40(3):575–81. 6. Beebe DC, Masters BR. Cell lineage and the differentiation of corneal epithelial cells. Investigative Ophthalmology and Visual Science 1996;37(9):1815–25. 7. Matsuda A, Tagawa Y, Matsuda H. Cytokeratin and proliferative cell nuclear antigen expression in superior limbic keratoconjunctivitis. Current Eye Research 1996;15:1033–8. 8. Gan L, Fagerholm P, Ekenbark S. Expression of proliferating cell nuclear antigen in corneas kept in long term culture. Acta Ophthalmologica Scandinavica 1998;76:308–13. 9. Zieske JD, Bukusoglu G, Yankauckas MA. Characterization of a potential marker of corneal epithelial stem cells. Investigative Ophthalmology and Visual Science 1992;33(1):143–52. 10. Yu CC-W, Woods AL, Levison DA. The assessment of cellular proliferation by immunohistochemistry: a review of currently available methods and their applications. Histochemical Journal 1992;24:121–31. 11. Chung E-H, Bukusoglu G, Zieske JD. Localization of corneal epithelial stem cells in the developing rat. Investigative Ophthalmology and Visual Science 1992;33(7):2199–206. 12. Smith RS, Hawes NL, Kuhlmann SD, Heckenlively JR, Chang B, Roderick TH, Sundberg JP. Corn1: A mouse model for corneal surface disease and neovascularization. Investigative Ophthalmology and Visual Science 1996;37(2):397–404. 13. Tseng H, Matsuzaki K, Lavker RM. Basonuclin in murine corneal and lens epithelia correlates with cellular maturation and proliferative ability. Differentiation 1999;65:221–7. 14. Wulle KG, Richter J. Electron microscopy of the early embryonic development of the human corneal epithelium. Albrecht Von Graefes Archiv fuer klinische und experimentelle Ophthalmologie 1978;209(1):39–49. 15. Sevel D, Isaacs R. A re-evaluation of corneal development. Transactions of the American Ophthalmological Society 1988;86:178–207. 16. Lesueur L, Arne JL, Mignon-Conte M, Malecaze F. Structural and ultrastructural changes in the developmental process of premature infants’ and children’s corneas. Cornea 1994;13(4):331–8. 17. Pospisilova E, Lichnovsky V, Lojda Z. Enzymatic equipment of the developing human eye. Acta Universitatis Palackianae Olomucensis Facultatis Medicae 1992;134:23–6. 18. Coupland SE, Penfold PL, Billson FA. Histochemical survey of the anterior segment of the normal human foetal and adult eye. Graefe’s Archive for Clinical and Experimental Ophthalmology 1993;231:533–40. 19. Rodrigues M, Ben-Zvi A, Krachmer J, Schermer A, Sun T-T. Suprabasal expression of a 64-kilodalton keratin (no.3) in developing human corneal epithelium. Differentiation 1987;34:60–7. 20. Rodrigues MM, Krachmer J, Rajagopalan S, Ben-Zvi A. Actin filament localization in developing and pathologic human corneas. Cornea 1987;6(3):190–6. 21. Ben-Zvi A, Rodrigues MM, Krachmer JH, Fujikawa LS. Immunohistochemical characterization of extracellular matrix in the developing human cornea. Current Eye Research 1986;5(2):105–17. 22. Nishina S, Kohsaka S, Yamaguchi Y, Handa H, Kawakami A, Fujisawa H, Azuma N. PAX6 expression in the developing human eye. British Journal of Ophthalmology 1999;83:723–7. 23. Waseem NH, Lane DP. Monoclonal antibody analysis of the proliferating cell nuclear antigen (PCNA). Journal of Cell Science 1990;96:121–9. 24. Linden MD, Torres FX, Kubus J, Zarbo RJ. Clinical application of morphologic and immunocytochemical assessments of cell proliferation. American Journal of Clinical Pathology 1992;97(5):S4–13. 25. Schipper DL, Wagenmans MJM, Peters WHM, Wagener DJT. Significance of cell proliferation measurement in gastric cancer. European Journal of Cancer 1998;34(6):781–90. 26. Krude T. Chromatin replication: finding the right connection. Current Biology 1999;9:R394–6. 27. Celis JE, Madsen P, Celis A, Nielsen HV, Gesser B. Cyclin (PCNA, auxiliary protein of DNA polymerase d) is a central component of the pathway(s) leading to DNA replication and cell division. Federation of European Biochemical Societies Letters 1987;220(1):1–7. 28. Yew DT, Lam TK, Tsang D, Au YK, Li WWY, Tso MOM. Changes of cytochemical markers in the conjunctival and corneal epithelium after corneal debridement. Cellular and Molecular Neurobiology 2000;20(4):465–82. 29. Wilson SE, Li Q, Weng J, Barry-Lane PA, Jester JV, Liang Q, Wordinger RJ. The Fas-Fas ligand system and other modulators of apoptosis in the cornea. Investigative Ophthalmology and Visual Science 1996;37(8): 1582–92.

D.T. Yew et al. / Life Sciences 68 (2001) 2987–3003

3003

30. Verheij M, Bartelink H. Radiation-induced apoptosis. Cell and Tissue Research 2000;301(1):133–42. 31. Wilson SE. Stimulus-specific and cell type-specific cascades: emerging principles relating to control of apoptosis in the eye. Experimental Eye Research 1999;69:255–66. 32. Huppertz B, Frank H-G, Kaufmann P. The apoptosis cascade – morphological and immunohistochemical methods for its visualization. Anatomy and Embryology 1999;200:1–18. 33. Stadelmann C, Lassmann H. Detection of apoptosis in tissue sections. Cell and Tissue Research 2000;301(1):19–31. 34. Hacker G. The morphology of apoptosis. Cell and Tissue Research 2000;301(1):5–17. 35. Wickert H, Zaar K, Grauer A, John M, Zimmermann M, Gillardon F. Differential induction of proto-oncogene expression and cell death in ocular tissues following ultraviolet irradiation of the rat eye. British Journal of Ophthalmology 1999;83:225–30. 36. Dong X, Zhu J, Guan Y, Tse ZS, Di SK, Luo ZB. An observation of the relationship between sitting height and age of 383 fetuses. Journal of Jinan University 1988;4:70–5. 37. Papermaster DS, Windle J. Death at the early age: apoptosis in inherited retinal degeneration. Investigative Ophthalmology and Visual Science 1995;36:977–83. 38. Wei Z-G, Sun T-T, Lavker RM. Rabbit conjunctival and corneal epithelial cells belong to two separate lineages. Investigative Ophthalmology and Visual Science 1996;37(4):523–33. 39. Alberts B, Bray D, Lewis J, Raff M, Roberts K, Watson JD. Molecular Biology of the Cell 3rd Ed. New York: Garland Publishing Incorporate, 1994. 40. Fuchs E, Dowling J, Segre J, Lo SH, Yu Q-C. Integrators of epidermal growth and differentiation: distinct functions for b1 and b4 integrins. Current Opinion in Genetics and Development 1997;7(5):672–82. 41. Bravo R, Macdonald-Bravo H. Existence of two populations of cyclin/proliferating cell nuclear antigen during the cell cycle: association with DNA replication sites. The Journal of Cell Biology 1987;105:1549–54. 42. Morris GF, Mathews MB. Regulation of proliferating cell nuclear antigen during the cell cycle. The Journal of Biological Chemistry 1989;264(23):13856–64. 43. Bravo R, Celis JE. Changes in the nuclear distribution of cyclin (PCNA) during S-phase are not triggered by post-translational modifications that are expected to moderately affect its charge. Federation of European Biochemical Societies Letters 1985;182(2):435–40. 44. Moriuchi T. Proliferating cell nuclear antigen (PCNA): a nuclear protein engaged in eukaryotic DNA replication for one billion years. Medical Science Research 1990;18:911–5. 45. Nakopoulou L, Janinis J, Panagos G, Comin G, Davaris P. The immunohistochemical expression of proliferating cell nuclear antigen (PCNA/Cyclin) in malignant and benign epithelial ovarian neoplasms and correlation with prognosis. European Journal of Cancer 1993;29A(11):1599–601. 46. Tanno M, Taguchi T. Proliferating cell nuclear antigen in normal and regenerating rat livers. Experimental and Molecular Pathology 1999;67:192–200. 47. Ren H, Wilson G. Apoptosis in the corneal epithelium. Investigative Ophthalmology and Visual Science 1996;37(6):1017–25. 48. Gao J, Gelber-Schwalb TA, Addeo JV, Stern ME. Apoptosis in the rabbit cornea after photorefractive keratectomy. Cornea 1997;16(2):200–8. 49. Komuro A, Hodge DO, Gores GJ, Bourne WM. Cell death during corneal storage at 48C. Investigative Ophthalmology and Visual Science 1999;40(12):2827–32. 50. Harayama K, Amemiya T, Nishimura H. Development of the cornea during fetal life: comparison of corneal and bulbar diameter. The Anatomical Record 1980;198:531–5. 51. Denis D, Burguière O, Burillon C. A biometric study of the eye, orbit, and face in 205 normal human fetuses. Investigative Ophthalmology and Visual Science 1998;39(12):2232–8.