Perlecan-Deficient Mutation Impairs Corneal Epithelial Structure

Abstract

Purpose.: To elucidate the role of perlecan (Hspg2), a large multidomain heparan sulfate proteoglycan expressed in the basement membrane, in the structure of the corneal epithelium.

Methods.: A previously developed perlecan-deficient (Hspg2 ^−/−-Tg) mouse model was used. Histologic analysis of their corneas was performed by light and transmission electron microscopy. The localization of perlecan in the corneas of wild-type (WT) mice and Hspg2 ^−/−-Tg mice was examined by immunohistochemistry. The effects of perlecan deficiency on corneal epithelial structure was analyzed with respect to the expression of corneal epithelial proliferation and differentiation markers, such as Ki67, cytokeratin12 (K12), connexin43 (Cx43), Notch1, and Pax6 by immunohistochemistry and real-time polymerase chain reaction (PCR).

Results.: The Hspg2 ^−/−-Tg mice had microphthalmos and a thinner corneal epithelium compared with that of the WT mice. Perlecan was localized in the corneal epithelial basement membrane in the WT mice, but not in the Hspg2 ^−/−-Tg mice. The Hspg2 ^−/−-Tg corneal epithelium exhibited thinner wing cell layers and a decreased number of Ki67-positive cells, but no dead cells, compared with the WT corneal epithelium. Immunohistochemistry and real-time PCR analysis revealed a significantly decreased expression of corneal epithelial differentiation markers such as K12, Cx43, Notch1, and Pax6 in Hspg2 ^−/−-Tg mice, compared with those of the WT mice.

Conclusions.: The findings of this study highlight a strong correlation between the presence of perlecan in the basement membrane and the structure of corneal epithelium and that the perlecan-deficient mutation impairs corneal epithelial structure.

The surface of a mammalian cornea is composed of a nonkeratinized, self-renewing, pluristratified epithelium of ectodermal origin. The corneal epithelium consists of basal, wing, and superficial cells that are separated from the stroma by the basement membrane (BM). Corneal epithelial cells exhibit a dynamic homeostasis, turning over approximately every 7 to 10 days. Many cellular processes, such as proliferation, apoptosis, differentiation, migration, adhesion, and stratification, are essential for the structure of corneal epithelium. 

Perlecan (Hspg2) is a large (>400 kDa), multidomain heparan sulfate proteoglycan (Hspg) expressed in BM. ^{1 –6} The protein core consists of five domains that share homology with other molecules involved in nutrient metabolism, cell proliferation, and adhesion, including laminin, the low-density lipoprotein (LDL) receptor, epithelial growth factor (EGF), and the neural cell adhesion molecule (N-CAM). ^{1 –3} Within the protein core there are numerous sites for O-linked glycosylation, as well as four potential sites for heparan sulfate (HS)/chondroitin sulfate (CS) chain attachment. These chains, which are usually HS, have been shown to be involved in many interactions, including those associated with growth factors, extracellular matrix (ECM) molecules, and neuromuscular junction proteins. ^{1 –3,7} Perlecan regulates cells through a basic mechanism involving the binding of various proteins via the protein core and/or the glycosaminoglycan chains. In vertebrates, perlecan functions in a diverse range of developmental and biological processes, from the development of cartilage to the regulation of wound healing. ^{8 –13} Recent reports from other groups also emphasized a key role for perlecan in regulating cell proliferation and cell survival in different tissues. For example, it has been reported that perlecan HS deficiency induces apoptosis of lens epithelial cells. ¹⁴ Sher et al. ¹⁵ found that perlecan regulates both the survival and terminal differentiation steps of keratinocytes and that it is critical for the formation of normal epidermis. 

In the cornea, perlecan is expressed in the BM of the corneal epithelium. ¹⁶ However, the functions or roles of perlecan in the cornea have yet to be well investigated. Therefore, in the present study, the role of perlecan in the structure of corneal epithelium was investigated by use of perlecan-deficient (Hspg2 ^−/−-Tg) mice. By genetically disrupting perlecan expression in the BM of corneal epithelium, the results of this study revealed that perlecan is essential in the structure of corneal epithelium. To the best of our knowledge, this study is the first to demonstrate the involvement of perlecan in the structure of the corneal epithelium. 

Materials and Methods

Animal Experiments

Some perlecan-deficient (Hspg2 ^−/−) mice die around embryonic day (E)10 due to defects in the myocardial basement membranes, and the mice that survive this stage die perinatally of premature cartilage development. ^12,17 In a previous study, a perlecan transgenic mouse line (Tg, Col2a1-Hspg2 ^Tg/−) that expresses recombinant perlecan in cartilage was created by use of a cartilage-specific Col2a1 promoter/enhancer to reverse the cartilage abnormalities of Hspg2 ^−/− mice. ¹³ Perinatal lethality–rescued mice (Hspg2 ^−/−-Tg, Hspg2 ^−/−; Col2a1-Hspg ^+/Tg) were then created by mating the transgenic mice with heterozygous Hspg2 ^+/− mice. The Hspg2 ^−/−-Tg mice exhibited normal cephalic development, and those mice were then maintained in a mixed genetic background of C57BL/6 and SVJ 129. In this study, we used 8- and 16-week-old Hspg2 ^−/−-Tg mice and Hspg2 ^+/+-Tg mice as well as wild-type (WT) mice, and the eyes of those mice were dissected and prepared for histologic or molecular analysis. All animal experiments in this study were performed in accordance with the guidelines set forth in the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Histologic Analysis

The excised mouse eyes were fixed in 20% formalin in phosphate-buffered saline (PBS) at 4°C overnight and then embedded in paraffin. Next, 3-μm-thick sections of the eyes were mounted on microslides (New Silane; Muto-Glass, Tokyo, Japan). Histologic examination was performed after Harris hematoxylin and eosin (H-E) staining. Histology of the corneas of the 8-week-old Hspg2 ^−/−-Tg and WT mice littermates was then compared by use of light microscopy (AX80; Olympus Corp., Tokyo, Japan). 

Morphometric Measurements

For the morphometric measurements, corneal thicknesses were calculated in 8-week-old Hspg2 ^−/−-Tg and WT mice. Next, 3-μm-thick tissue sections of the cornea stained with H-E staining were viewed by light microscopy (40× magnification) with a computerized image analyzer (KS400; Carl Zeiss AG, Oberkochen, Germany), and measurements were made by use of a calibrated eyepiece graticule. Corneal thickness was measured in the central region of the serial sections of each eye. The mean thickness was then calculated by averaging those measurements. The epithelial, stromal, endothelial, and whole corneal thicknesses were then compared. The ratio of the epithelial cell layer thickness to the full corneal thickness was also calculated. 

Examination by Transmission Electron Microscopy

For the transmission electron microscopy (TEM) examinations, the eyes of 8- and 16-week-old Hspg2 ^−/−-Tg and WT mice were dissected and fixed in cold 2.5% glutaraldehyde with PBS overnight at 4°C and then sectioned into small pieces. Those sections were then postfixed with 2% osmium tetroxide in the same buffer, dehydrated through a series of ethyl alcohol solutions, and embedded in Epon. All sections were examined by use of an electron microscope (H-7100; Hitachi, Tokyo, Japan) at an accelerating voltage of 75 kV. 

Immunohistochemical Staining

Deparaffinized sections were washed in 100% ethanol and rehydrated with PBS. Antigen retrieval was performed by boiling the sections in 0.01 M citrate buffer (pH 6) for 10 minutes. Next, the slides were washed with PBS and blocked with 4% normal serum (species selected according to the secondary antibody) in PBS and 0.3% bovine serum albumin for 10 minutes at room temperature. The slides were incubated with primary antibody overnight at 4°C (Table 1), washed with PBS, incubated with the secondary antibodies, and counterstained with DAPI (H-1200; Vector Laboratories, Inc., Burlingame, CA). Ki67-positive cells were quantified by capturing the image of individual nuclei from Hspg2 ^−/−-Tg and WT sections processed in parallel and immunostained on the same slides. All sections were viewed with a fluorescence microscope (AxioVision 3.1; Carl Zeiss Meditec, Inc.) and confocal microscopy (TCS-SP5/TIRF; Leica Microsystems AG, Solms, Germany). 

Table 1.

 

Primary Antibodies

Antigen	Class	Dilution	Supplier
Anti-perlecan	Rabbit polyclonal	1/100	Seigaku, Tokyo Japan
Anti-Ki67	Rabbit polyclonal	1/200	Abcam, Cambridge UK
Anti-cytokeratin12	Goat polyclonal	1/200	Santa Cruz Biotech, Santa Cruz, CA
Anti-connexin43	Rabbit polyclonal	1/2000	Abcam
Anti-Notch1	Rabbit polyclonal	1/200	Abcam
Anti-Pax6	Mouse monoclonal	1/200	R&D Systems Minneapolis, MN

Assessment of Cell Death by TUNEL Assay

Deparaffinized sections were rehydrated through graded alcohols and then washed with PBS. The tissue sections were treated with proteinase K (80 μg/mL) for 20 minutes at room temperature. The slides were then washed twice with PBS. Next, the TUNEL assay (TUNEL in Situ Cell Death Detection Kit, fluorescein; Roche Diagnostics GmbH, Mannheim, Germany) was performed according to the manufacturer's instructions. Briefly, the sections were first counterstained with DAPI. The samples were then visualized by fluorescence microscopy, and images were obtained for quantitative analysis. TUNEL-positive cells were then quantified by capturing the image of individual nuclei from Hspg2 ^−/−-Tg and WT sections processed in parallel and immunostained on the same slides. 

Analysis by Real-Time Polymerase Chain Reaction

Total RNA was isolated from the dissected corneas (NucleoSpin RNA II; Macherey-Nagel GmbH, KG, Duren, Germany) according to the manufacturer's instructions. cDNA was generated from 1.0 μg total RNA (ReverTra Ace-α; Toyobo Co., Ltd., Osaka, Japan). Real-time PCR was performed with SYBR green master mix (Fast SYBR Green Master Mix; Applied Biosystems, Inc. [ABI], Foster City, CA) on a commercial system (Prism 7500; ABI). In this study, we did not isolate the RNA from the epithelia but from the whole cornea. Therefore, the PCR analysis for the level of reduced expression of differentiation and developmental regulator molecules in Hspg2 ^−/−-Tg eyes may be semiquantitative, not absolutely quantitative. Primers sequences are listed in Table 2. 

Table 2.

 

Primers Used in RT-PCR

Gene	Primer	Primer Sequence
Ki67	Forward	5′-GCAGGAAGCAACAGATGAGAAGCC-3′
	Reverse	5′-GCTCAGGTGATACATGCCTCCTGC-3′
Active caspase3	Forward	5′-AGGTGGCAACGGAATTCGAGTC-3′
	Reverse	5′-ACACGGGATCTGTTTCTTTGCG-3′
Cytokeratin12	Forward	5′-TCTTCATGCTGGTGGTGTCCTTG-3′
	Reverse	5′-TCAAGAAACCAGGCCTCTGCATC-3′
Connexin43	Forward	5′-TCTTCATGCTGGTGGTGTCCTTG-3′
	Reverse	5′-CGATCCTTAACGCCCTTGAAGAAG-3′
Notch1	Forward	5′-GGAGGACCTCATCAACTCACATGC-3′
	Reverse	5′-CCGTTCTTCAGGAGCACAACAG-3′
Pax6	Forward	5′-AAGGATGTTGAACGGGCAGAC-3′
	Reverse	5′-TGTTGCTGGCAGCCATCTTG-3′
GAPDH	Forward	5′-AAGAGAGGCCCTATCCCAACTC-3′
	Reverse	5′-TTGTGGGTGCAGCGAACTTTATTG-3′

Results

Histologic Analysis of the Hspg2 ^−/−-Tg Eyes

The eyes of 8-week-old Hspg2 ^−/−-Tg mice that were approximately the same body weight and length as WT mice exhibited microphthalmos and a small palpebral fissure (Fig. 1A). Under light microscopy at low magnification, the Hspg2 ^−/−-Tg eyes stained with H-E staining also exhibited microphthalmos (Figs. 1B, 1C). Under high magnification, the Hspg2 ^−/−-Tg eyes showed a thinner corneal epithelium compared with that of the WT eyes (Figs. 1D, 1E). The localization of perlecan was examined by immunostaining with specific antibody for their core protein. Eight-week-old WT and Hspg2 ^−/−-Tg mice were stained with anti-perlecan antibody (Alexa488, green), and the corneal nuclei were stained with DAPI (blue). Perlecan was strongly expressed in the corneal epithelial BM of the WT mice (Fig. 1F). However, the expression of perlecan was not recognized in the corneal epithelium of the Hspg2 ^−/−-Tg mice (Fig. 1G). 

Figure 1.

 

Histologic analysis. Representative macroscopic images of the eye in situ (A). H-E-stained sections (B–E) show the histologic features of whole eyes visualized by light microscopy at low (B, C) and high (D, E) magnifications. The 8-week-old Hspg2 ^−/−-Tg mice had microphthalmos, whereas the WT mice did not (A–C). The corneal epithelium of the 8-week-old Hspg2 ^−/−-Tg mice thinner than that of the WT mice (D, E). Immunohistochemical staining of perlecan in the corneas of the WT and the Hspg2 ^−/−-Tg mice (F, G). Perlecan (Alexa 488, green) was strongly expressed in the corneal epithelial basement membrane of the WT mice (F). However, the expression of perlecan was not recognized in the corneal epithelium of the Hspg2 ^−/−-Tg mice (G). Epi, epithelium; St, stroma. Scale bars: (B, C) 600 μm; (D, E) 40 μm; (F, G) 30 μm.

Analysis by TEM

TEM was performed to further examine the corneal morphology in the Hspg2 ^−/−-Tg and WT mice. The corneal epithelia from WT mice and Hspg2 ^−/−-Tg, 8 weeks (Figs. 2A, 2B) and 16 weeks (Figs. 2C, 2D) of age, were analyzed by TEM. At 8 and 16 weeks of age, the Hspg2 ^−/−-Tg mice showed thinner corneal epithelia compared with the WT mice. Eight-week-old WT mice showed 9 to 10 corneal epithelial layers (Fig. 2A). In contrast, the 8-week-old Hspg2 ^−/−-Tg mice showed thinner undifferentiated wing cell layers compared with the WT mice (Fig. 2B). Corneal wing-cell layers of the 16-week-old Hspg2 ^−/−-Tg mice were thinner and undifferentiated compared with those of the WT mice. As the ages of the mice progressed, the Hspg2 ^−/−-Tg mice showed a thinner corneal epithelium compared with that of the WT mice (Fig. 2A-D). Under high magnification, no significant difference was observed between the Hspg 2 ^−/− Tg mice and WT mice in regard to the structure of superficial cells (Figs. 2E, 2F), basal cells (Figs. 2G, 2H), and epithelial BM (Figs. 2I, 2J). 

Figure 2.

 

TEM of the corneal epithelium. TEM images show ultrastructural features of corneal epithelium from WT (A, C, E, G, I) and Hspg2 ^−/−-Tg (B, D, F, H, J) mice. Corneal epithelia of the 8- and 16-week-old Hspg2 ^−/−-Tg mice (B, D) were thinner and had thinner wing cell layers compared with those of the WT mice (A, C). As the ages of the mice progressed, the corneal epithelium of the 16-week-old Hspg2 ^−/−-Tg mice became thinner and the wing cell layer was undifferentiated compared with that of the WT mice (D). Under high magnification, no significant difference was observed between the Hspg2 ^−/−-Tg mice and WT mice in regard to the structure of the superficial cells (E, F), basal cells (G, H), and epithelial basement membrane (I, J). Epi, epithelium; St, stroma; mv, microvilli; bc, basal cell, BM, corneal basement membrane. Scale bar: (A–D) 5 μm; (E-H) 2 μm; (I, J) 0.5 μm.

In the corneal stromal layer, the keratocytes were localized between stromal lamellae, with no significant difference found between the 8-week-old Hspg2 ^−/−-Tg and WT mice (Figs. 3A, 3B). Under high magnification, cross-sections of the collagen fibers from the Hspg2 ^−/−-Tg and WT mice demonstrated parallel bundles of a regular diameter (Figs. 3C, 3D). Under low magnification, no significant difference was observed between the Hspg2 ^−/−-Tg and WT mice as to the thickness of the endothelial layers (Figs. 3E, 3F). Under high magnification, Descemet's membrane was found to be composed of electron-dense material in both the Hspg2 ^−/−-Tg and WT mice (Figs. 3G, 3H). The corneal endothelium was found to have some desmosomes and gap junctions, with no significant difference found between the Hspg2 ^−/−-Tg and WT mice (Figs. 3I, J). 

Figure 3.

 

Electron microscopy of corneal stroma and endothelium. Keratocytes (A, B). Collagen fibers cut in cross-section (C, D). Stroma and corneal endothelium observed under low magnification (E, F). Descemet's membrane observed under high magnification (G, H). Corneal endothelium observed under high magnification (I, J). No significant differences were found between the 8-week-old Hspg2 ^−/−-Tg mice and WT mice in regard to the construction of the stroma and endothelium. K, keratocytes; St, stroma; Des, Descemet's membrane; End, endothelium. Scale bar: (A, B) 2 μm; (C, D, I, J) 0.5 μm; (E, F) 5 μm; (G, H) 1 μm.

Corneal Thickness Morphometry

The thickness of the corneal epithelium was examined in the 8-week-old mice, as that is the age at which the development of the corneal epithelium is complete. Histologic examination of those mice revealed that the corneal epithelial thickness was markedly thinned in the Hspg2 ^−/−-Tg mice. The corneal thickness of the central region was then calculated (Fig. 4A). The thickness of the central whole corneal cell layers was found to be 25.6% thinner in the Hspg2 ^−/−-Tg mice (on average, 85.12 μm thick compared with 114.53 μm in the WT mice; n = 6; P = 0.0411). The thickness of the central epithelial cell layers was found to be 45.5% thinner in the Hspg2 ^−/−-Tg mice (on average, 18.51 μm thick compared with 33.94 μm in the WT mice; n = 6; P = 0.0022). The average thicknesses of the central corneal stromal layers and endothelial layers were not significantly different between the Hspg2 ^−/−-Tg and WT mice. Because of the microphthalmos of the eyes of the Hspg2 ^−/−-Tg mice, we calculated the comparison of the ratio of the epithelial cell layer thickness to the full central corneal thickness in the central region (Fig. 4B). The ratio of the central epithelial cell layer thickness to the central whole corneal thickness was found to be significant lower in the Hspg2 ^−/−-Tg mice, 22.6% compared with 29.4% in the WT mice (n = 6, P = 0.0043). These findings suggest that the corneal epithelial cell layer in the Hspg2 ^−/−-Tg mouse is thinner regardless of the microphthalmos. 

Figure 4.

 

Morphometry of corneal thickness. Comparison of central corneal thickness. (A) The central corneal epithelial cell layer of the 8-week-old Hspg2 ^−/−-Tg mice was significantly thinner than that of the 8-week-old WT mice (n = 6; P = 0.0022). The full central corneal thickness was significantly thinner in the 8-week-old Hspg2 ^−/−-Tg mice compared with that in the 8-week-old WT mice (n = 6; P = 0.0411). Comparison of the ratio of the epithelial cell layer thickness to the full corneal thickness measured at the central cornea (B). The ratio of epithelial cell layer thickness to full corneal thickness was significantly lower in the Hspg2 ^−/−-Tg mice than in the WT mice (n = 6; P = 0.0043; Mann-Whitney U test: *P < 0.05, **P < 0.01, ***P < 0.0001).

Proliferation and Cell Death in Hspg2 ^−/−-Tg Corneal Epithelium

We posited that the findings of thinner corneal epithelium in the 8-week-old Hspg2 ^−/−-Tg mice could be the result of a decrease in cell proliferation or an increase in cell death. To discern between these two possibilities, immunostaining was performed to investigate the number of Ki67-positive (Figs. 5A1, A2) and TUNEL-positive (Figs. 5D1, 5D2) cells. The Ki67 antigen was designated as a marker for cell proliferation, and the number of Ki67-positive cells was scored across the entire section of the corneal epithelium. The average ratio of Ki67-positive cells to basal cells was 12% per section in the Hspg2 ^−/−-Tg epithelium, compared with 21% in the WT epithelium (n = 6; P = 0.0087; Fig. 5B). Real-time PCR for Ki67 showed a 67% decrease in RNA levels in the Hspg2 ^−/−-Tg epithelium (n = 5; P = 0.0159; Fig. 5C). 

Figure 5.

 

Proliferation and cell death in the Hspg2 ^−/−-Tg corneal epithelium. Immunohistochemistry showed a decreased number of cells containing Ki67 (Alexa488, green, white arrow) in the 8-week-old Hspg2 ^−/−-Tg versus WT corneal epithelium (DAPI, blue; A1, A2). The percentage of Ki67-positive cells in the corneal epithelium showed a 9.0% decrease in the Hspg2 ^−/−-Tg mice (±SEM, n = 6, P = 0.0087; B). Quantification of RNA levels for Ki67 in the corneal epithelium (±SEM; n = 5; P = 0.0159; C). In the superficial corneal cells, there was almost no TUNEL-positive staining (D1, D2). The percentage of TUNEL-positive cells in the corneal epithelium (±SEM, n = 6; E). Quantification of RNA levels for active caspase3 in the corneal epithelium (±SEM; n = 5; F; Mann-Whitney U test: *P < 0.05, **P < 0.01, ***P < 0.0001). Scale bar: 50 μm. Low magnification: A1, D1; high magnification: A2, D2.

TUNEL assay assessment of cell death revealed a very small number of TUNEL-positive cells (<0.3%) per corneal section in both the Hspg2 ^−/−-Tg and WT epithelium (Fig. 5E), and there was no increase in RNA levels of active caspase3 of the apoptosis marker in both Hspg2 ^−/−-Tg and WT corneal epithelium (Fig. 5F), thus indicating that the loss of perlecan did not lead to a significant change in the rate of apoptosis. Therefore, the likely cause of the thinning of the 8-week-old Hspg2 ^−/−-Tg corneal epithelium was determined to be reduced cell proliferation. 

Effect of Perlecan Deficiency on the Expression of Markers of Corneal Epithelial Differentiation

The expression of cytokeratin12 (K12), a corneal differentiation marker, in the 8-week-old Hspg2 ^−/−-Tg mice was significantly decreased compared with that in the WT mice examined by immunohistochemistry (Figs. 6A1, A2). Real-time PCR for K12 in the Hspg2 ^−/−-Tg epithelium showed a 54% decrease in RNA levels compared with the WT epithelium (n = 5; P = 0.4698; Fig. 6E). Connexin43 (Cx43), a gap junction protein, was found to be present in the corneal basal cell layers in the WT epithelium, but was absent in the Hspg2 ^−/−-Tg epithelium by immunohistochemistry (Figs. 6B1, 6B2). Real-time PCR for Cx43 in the Hspg2 ^−/−-Tg epithelium showed a 41% decrease in RNA levels compared with that in the WT epithelium (n = 5; P = 0.4698; Fig. 6E). The expression of Notch1 in the Hspg2 ^−/−-Tg corneal epithelium was significantly decreased compared with that of the WT epithelium by immunohistochemistry (Figs. 6C1, 6C2). Real-time PCR showed that the Hspg2 ^−/−-Tg mutation caused a significantly decrease in Notch1 RNA levels in the corneal epithelium, compared with that in the WT mice (n = 5; P = 0.0159; Fig. 6E). The expression of Pax6, a developmental regulator marker, was shown by immunohistochemistry to be significantly decreased in the corneal epithelium in the Hspg2 ^−/−-Tg mice compared with that of the WT mice (Figs. 6D1, 6D2). Real-time PCR for Pax6 in the Hspg2 ^−/−-Tg epithelium showed a significant decrease in RNA levels compared with that in the WT epithelium (n = 5; P = 0.0159; Fig. 6E). 

Figure 6.

 

Expression of differentiation and developmental regulator markers in the 8-week-old Hspg2 ^−/−-Tg corneal epithelium demonstrated that the expression of cytokeratin12 (K12) in the Hspg2 ^−/−-Tg epithelium was significantly decreased compared with that in the WT epithelium (A1, A2). Hspg2 ^−/−-Tg corneal epithelium showed no expression of Connexin43 (Cx43; B1, B2). The expression of Notch1 in Hspg2 ^−/−-Tg corneal epithelium was significantly decreased compared with that in the WT epithelium (C1, C2). Hspg2 ^−/−-Tg corneal epithelium showed decreased Pax6 expression compared with that in the WT epithelium (D1, D2). Quantification of RNA levels for differentiation and developmental regulator markers in corneal epithelium (±SEM, n = 5; E; Mann-Whitney U test: *P < 0.05, **P < 0.01, ***P < 0.0001). Scale bar: 50 μm. Low magnification: A1, B1, C1, D1; high magnification: A2, B2, C2, D2.

Discussion

In this study, perlecan was identified in corneal epithelial BM and the epithelium was shown to be thin and poorly differentiated in perlecan-deficient mice (Hspg2 ^−/−-Tg) and accompanied by the downregulation of Ki67, K12, Cx43, Notch1, and Pax6. However, the gross morphology of the corneal epithelium was not retarded in the Hspg2 ^−/−-Tg mice, suggesting that perlecan is not critically necessary in this process. Therefore, perlecan may be essential for the structure but not the development of corneal epithelium. In normal corneal epithelium, epithelial cells in the last phase of their differentiation undergo apoptosis as they reach the superficial cell layer. Since the cell death rate of the corneal epithelial cells in the Hspg2 ^−/−-Tg mice was similar to that in WT mice, the failure of those cells to form multilayered corneal epithelium must be due to the apparent decrease in the proliferation and differentiation rates in corneal epithelial cells. In this present study, we revealed that the expression of Ki67, K12, Cx43, Notch 1, and Pax6, which are markers of cell proliferation and differentiation, was reduced in the Hspg2 ^−/−-Tg mice, compared with that of the WT mice. Therefore, our findings revealed that perlecan in the BM of corneal epithelium may be critical for normal epithelial formation and terminal differentiation. 

It has been reported that K12 is essential for the differentiation and maintenance of corneal epithelium integrity., ^18,19 Targeted deletion of K12 in a mouse model showed fewer cellular layers in the corneal epithelium and corneal fragility. ¹⁹ The findings of this study showed that downregulation of the expression of K12 at protein and RNA levels may be one of the causes of aberrant differentiation in the Hspg2 ^−/−-Tg corneal epithelium. From another aspect, it has been reported that the gap junction marker Cx43 mediates the intercellular diffusion ions and other small molecules, ^{20 –22} thereby contributing to the regulation of tissue differentiation and homeostasis. ²³ Of particular interest, the expression of Cx43 was noted in the corneal epithelial basal cells in the WT corneal epithelium, but not in the Hspg2 ^−/−-Tg epithelium, thus suggesting that the basal cell environment is impaired by gap junction functional decline. Therefore, the downregulation of Cx43 in the Hspg2 ^−/−-Tg mice most likely impairs the differentiation and structure of the corneal epithelium. 

It has been reported that the Notch signaling pathway, another corneal homeostasis marker, limits cell proliferation and promotes differentiation. ^{24 –27} In this study, the expression of Notch1 was decreased in the Hspg2 ^−/−-Tg mice, compared with that in the WT mice. Recently, Vauclair et al. ²⁷ demonstrated that Notch1-deficient corneal cells lose their ability to heal and repair wounded corneal epithelium. The findings of that study showed that instead of generating new corneal epithelium after injury, those cells repair the wound by forming a hyperproliferative epidermislike epithelium. This process involves the secretion of FGF-2 through Notch1 signaling in the epithelium. ²⁷ It is well known that FGF-2 is a growth factor of corneal epithelial cells. ^28,29 Loss of Notch 1 in the corneal epithelium resulted first in upregulation of FGF-2 by the corneal epithelium, suggesting that Notch1 signaling repressed its expression. ²⁷ Despite the decreased expression of Notch1, a hyperproliferative change of corneal epithelium was not observed in the Hspg2 ^−/−-Tg mice. Since FGF-2 is a ligand of perlecan, there may be a possibility that a high dose of FGF-2 could not be maintained in the BM of the corneal epithelium of Hspg2 ^−/−-Tg mice. ^{30 –35} Reportedly, FGF-7 is also a ligand of perlecan. ¹ In a recent study, Lovicu et al. ³⁶ showed hyperproliferation of embryonic corneal epithelial cells in transgenic mice engineered to overexpress human FGF-7 in the eye. Chikama et al. ³⁷ analyzed the effects of excess FGF-7 on both the proliferation and differentiation of corneal epithelium in an FGF-7 transgenic mouse model in which cornea-specific FGF-7 was overexpressed. In that study, the mice exhibited epithelial hyperplasia, accompanied by the downregulation of K12. According to these results, the mechanism of the poor differentiation of the epithelium in Hspg2 ^−/−-Tg mice is due to the lack of the FGF-2 or FGF-7 that links to perlecan in the BM. Therefore, the strong correlation between the presence of perlecan in the BM and the formation of normal corneal epithelium suggests that perlecan functions as a reservoir for soluble factors involved in the proliferation and differentiation of corneal epithelial cells. 

It should be noted that the Hspg2 ^−/−-Tg mice had microphthalmos. This condition has been reported in Pax6-deficient mice. ^{38 –41} These reports suggest that Pax6 is a key developmental regulator and that it is generally essential for morphogenesis in the eye. Pax6 has autonomous roles in all eye tissues, where it is expressed at several developmental stages. Recently, a report by Garcia-Villegas et al. ⁴² revealed that Pax6 is the earlier differentiation marker expressed by corneal epithelial cells and that it is the main driver of the differentiation of corneal epithelial cells, as the expression of Pax6 promotes the differentiation of corneal epithelial cells. On the other hand, transgenic mice overexpressing Pax6 in the corneal epithelium also showed abnormal epithelial cell morphology. These results indicate that a correct Pax6 dosage for the normal development of corneal epithelium may be important. In this present study, we demonstrated that the corneal epithelium of Hspg2 ^−/−-Tg mice was thinner and not well differentiated and that the phenotypes became more severe with age. The corneal epithelial phenotype was similar to that of Pax6-deficient mice. Thus, the downregulation of Pax6 in the corneal epithelium of Hspg2 ^−/−-Tg mice is likely to be a factor in the observed microphthalmos and thinner epithelium. We theorize that the downregulation of K12, Cx43, Notch1, and Pax6 probably occurs to prevent the proliferation and the differentiation from basal cells to wing cells, thus making the corneal epithelium of Hspg2 ^−/−-Tg mice thinner than that of WT mice with downregulation of the expression of Ki67. 

In summary, by using perlecan-deficient mice (Hspg2 ^−/−-Tg) we demonstrated for the first time that perlecan is essential for the structure of corneal epithelium, as it controls the expression of markers for the proliferation or differentiation of corneal epithelial cells. Our findings revealed that perlecan in the BM of corneal epithelium were critical for normal epithelial structure and terminal differentiation. 

Footnotes

 Disclosure: T. Inomata, None; N. Ebihara , None; T. Funaki , None; A. Matsuda, None; Y. Watanabe, None; L. Ning, None; Z. Xu, None; A. Murakami, None; E. Arikawa-Hirasawa, None

The authors thank Glenn Longenecker and Ashok B. Kulkarni for help in creating the mutant mice, Yoshihiko Yamada for critical reading of the manuscript, and Saori Ito for secretarial assistance.