Expression and Localization of Angiogenic Inhibitory Factor, Chondromodulin-I, in Adult Rat Eye

Abstract

purpose. To determine the role in the eye of chondromodulin (ChM)-I, which has been identified in cartilage as an angiogenic inhibitor, the expression and localization and a possible function of ChM-I were investigated.

methods. Expression and localization of ChM-I in rat eyes were examined by RNase protection assay and in situ hybridization and by immunostaining, using an antibody against a synthetic peptide. The effect of recombinant ChM-I on tube morphogenesis of retinal endothelial cells was examined in culture.

results. The rat ChM-I gene was determined to encode the open reading frame of 334 amino acid residues, and ChM-I mRNA was exclusively expressed in cartilage, eye, and cerebellum in rats. ChM-I mRNA expression was evident in the iris–ciliary body, retina, and scleral compartments, but not in other compartments of the eye. In situ hybridization revealed mRNA expression in the ganglion cells, inner nuclear layer cells, and pigment epithelium in the retina and in the nonpigment epithelium of the ciliary body. Immunoreactive ChM-I was present in these cells and also in the vitreous body. Western blot analysis detected an ∼25-kDa band of ChM-I presumed as a secretory form in the aqueous humor and vitreous body and an ∼37-kDa band as a precursor form in the retina. Recombinant human ChM-I inhibited tube morphogenesis of human retinal endothelial cells in vitro.

conclusions. These observations indicate a potential role for ChM-I in inhibition of angiogenesis in the rat eye.

Vasculature is highly restricted in the eye, and intraocular neovascularization is a major concern in recent ophthalmic disease, because it leads to severe impediment of visual function. ^{1 2 3} Neovascularization in the eye is associated with several systemic or ocular disorders, including diabetes, prematurity, occlusion of the retinal vein, choroidal neovascularization (age-related macular degeneration), chronic retinal detachment secondary to tumor growth, and neovascular glaucoma. The control of neovascularization would aid in the treatment of these disorders. In spite of the presence in the eye of potent angiogenic molecules, such as basic fibroblast growth factor (bFGF), ^{4 5} vascular endothelial growth factor (VEGF), ^{6 7 8} and transforming growth factor (TGF)-β, ^{9 10} ocular tissues are maintained in the physiological condition without the occurrence of neovascularization, suggesting the presence of angiogenesis inhibitors in the tissue. Thus, a balance of inducers and inhibitors for endothelial cell proliferation and angiogenesis may regulate to promote or suppress vascularization in the eye. ¹¹ Little is known, however, about the angiogenic inhibitory factors. The existence of the angiogenic inhibitors has been predicted in retinal pigment epithelium, ^{12 13} the vitreous body, ^{14 15 16 17 18} and the lens, ¹⁹ although most inhibitors have not been isolated yet. Recently, pigment epithelium-derived factor (PEDF), a neurotrophic factor and a member of the serine protease inhibitor (serpin) supergene family, has been identified as an angiogenic inhibitor in the eye. ¹³  

Chondromodulin (ChM)-I, a 25-kDa glycoprotein was purified from bovine epiphyseal cartilage and identified as a chondrocyte growth factor trapped in the extracellular matrix. ^{20 21} Subsequently, ChM-I was demonstrated to inhibit proliferation and tube morphogenesis of vascular endothelial cells in vitro and angiogenesis in chick chorioallantoic membrane in vivo. ^{22 23 24 25}  

Because structural analogy has been noted between cartilage and eye in terms of their avascularity and components of extracellular matrix, ²⁶ it may be speculated that ChM-I also plays a role in maintenance of the avascular condition in the eye. To test this, we first cloned rat ChM-I cDNA and raised an antibody against synthetic rat ChM-I peptide to investigate the expression of ChM-I in the eye by RNase protection assay and in situ hybridization and its location by immunostaining and Western blot analysis. In addition, the effect of recombinant ChM-I on tube morphogenesis of retinal endothelial cells was examined in culture. 

Materials and Methods

Tissues and RNA

All experiments were conducted in accordance with the Animal Care and Use Committee guidelines and the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Wistar–Kyoto rats were anesthetized with diethyl ether and killed by cutting the axillary artery. Systemic organs (cerebrum, cerebellum, eye, submandibular salivary gland, heart, lung, liver, pancreas, small intestine, colon, spleen, testis, ovary, kidney, and smooth muscle) were removed from adult (3 months old), rib cartilage from newborn, and thymus from 1-month-old rats. Eyes were dissected into several compartments under a microscope: extraocular muscle, conjunctiva, cornea, iris–ciliary body, anterior chamber angle, lens, retina without retinal pigment epithelium, and sclera with choroid and retinal pigment epithelium. The tight association of the retinal pigment epithelium with scleral compartments and the negligible attachment of the epithelium to retinal compartments were certified by light microscopy. Total cellular RNA was purified from these organs and eye compartments by a modified acid guanidium thiocyanate phenol-chloroform extraction method (TRIzol; Gibco–Life Technologies, Rockville, MD). 

For immunohistochemistry and in situ hybridization, after enucleation the eyes and the cartilage were immersed in 4% paraformaldehyde in phosphate-buffered saline (PBS) for 4 hours and followed by immersion in 30% sucrose in PBS overnight. The tissues were embedded in optimal temperature cutting compound (Tissue-Tek; Miles, Elkhart, IN), frozen in liquid nitrogen, and cut in a cryostat at −20°C. Specimens were sectioned at 4 μm for immunohistochemistry and at 10 μm for in situ hybridization, and mounted on slides coated with poly-l-lysine. For immunohistochemistry of the vitreous body, whole eyes were fixed in methyl-Carnoy fixative for 16 hours, dehydrated in graded ethanol, embedded in paraffin, and sectioned at 4μ m on poly-l-lysine–coated slides. 

Isolation of Rat ChM-I cDNA

Rat ChM-I cDNA was obtained by the reverse transcription polymerase chain reaction (RT-PCR) and 5′/3′ rapid amplification of cDNA ends (RACE) methods using oligonucleotide ChM-I primers designed on the basis of the nucleotide sequence conserved in bovine ChM-I and human ChM-I ^{20 26} : sense primer, 5′-GGGAATTCGGAAGGCAAGATCATGCCAGT-3′, and antisense primer, 5′-GGGGATCCACACCATGCCCAGGATGCGG-3′. Rat cartilage RNA (1 μg) was reverse transcribed by using oligo (dT) primer and reverse transcriptase (Superscript II; Gibco–Life Technologies). The cDNA was amplified by PCR (35 cycles: 94°C, 30 seconds; 58°C, 1 minute; and 72°C, 1 minute) using 100 picomoles of the ChM-I primers. The PCR product of approximately 500 bp were subcloned into a vector (pGEM 4Z; Promega, Madison, WI) at the EcoRI and BamHI sites. Then, 3′ and 5′ RACE was performed with a kit (5′/3′ RACE Kit; Boehringer–Mannheim, Mannheim, Germany), according to the manufacturer’s instructions. The 5′ end of the rat ChM-I cDNA was amplified using an anchor primer and the nested primer GSP-1 (5′-AATAGGCAGGTCGCCACAGA-3′), an upstream primer of GSP-2 (5′-CGGGATCCAGCTGTTGTCCTTTACAGGC-3′), and GSP-3 (5′-CGGGATCCAGAAGTAGAAGGC-3′). The 3′ end of the cDNA was amplified using an anchor primer, and the nested primer GSP-5 (5′-GGGAATTCACCACAAGGAGACCACACAG-3′). The amplified polymerase chain reaction (PCR) products were subcloned into another vector (pGEM-T; Promega), and more than five clones of each part were sequenced by a DNA sequencer (Perkin Elmer, Urayasu, Japan). 

Detection of Rat ChM-I mRNA

An RNase protection assay was conducted using ³²P-labeled antisense cRNA probes for ChM-I mRNA and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA, as reported previously. ²⁷ The templates were ChM-I cDNA (329 bp corresponding to nucleotides 348-676 of the full-length cDNA, as will be described later) or GAPDH cDNA (114 bp corresponding to nucleotides 673-787) as a housekeeping gene. Ten micrograms of total RNA isolated from the systemic organs or eye compartments were hybridized with these probes, and unhybridized probes were digested with ribonucleases A and T1. The probes protected from the RNase digestion were electrophoresed on 6% polyacrylamide gels, and the dried gels were exposed to x-ray films (Fuji, Kanagawa, Japan) for 7 days at −80°C. 

For in situ hybridization, digoxigenin-labeled antisense and sense cRNA probes were synthesized using ChM-I cDNA templates encoding nucleotides 631-1129 according to the manufacturer’s protocol (Boehringer–Mannheim). Cryostat sections (10 μm thick) of newborn rat rib cartilage and adult rat eyes were fixed in 4% paraformaldehyde in PBS, treated with 1.5 μg/ml proteinase K (Promega) for 15 minutes at 37°C, and then hybridized with the digoxigenin-labeled probes (10 ng/ml) overnight at 58°C. ²⁸ After washing in 2× SSC at 65°C, the sections were treated with 20 μg/ml RNase A for 30 minutes at 37°C. Thereafter, the sections were incubated with alkaline phosphatase-labeled anti-digoxigenin antibody (1:500 dilution; Boehringer–Mannheim) for 2 hours at room temperature and rinsed in PBS. The antibody-bound alkaline phosphatase was visualized by the reaction with nitroblue tetrazolium (0.34 mg/ml) and 5-bromo-4-chloro-3-indolyl phosphate (0.18 mg/ml). The sections were counterstained with stain solution (Kernechtrot; Muto Pure Chemicals, Tokyo, Japan). 

Preparation of Antibody

An oligopeptide corresponding to 26 amino acids of mature rat ChM-I (NH2-PSTTRRPHSEPRGNAGPGRLSNRTRP-COOH, double-underlined in Fig. 1 ) with an added cysteine at C terminus was synthesized and conjugated with keyhole limpet hemocyanin as a carrier protein. The sequence was selected because it is unique in rats among other species and is assumed to be hydrophilic. The conjugate of 1 mg was emulsified with complete Freund’s adjuvant (CFA) and injected subcutaneously into New Zealand White rabbits every 2 weeks three times. One week after the last injection, the blood was collected. Anti-ChM-I antibody was affinity purified by using the ChM-I synthetic peptide–conjugated column (Cellulofine; Seikagaku, Tokyo, Japan). The specificity of the antibody was tested by blocking of the immunostaining after absorption of the affinity-purified antibody (0.5 mg IgG) with the synthetic peptide (50 μg of the synthetic peptide, approximately 10 times excess to IgG at molar ratio) for 16 hours at 4°C. 

Western Blot Analysis

Reactivity of the ChM-I antibody and localization of ChM-I in eye compartments were examined by Western blot analysis. ²⁷ In brief, the samples were extracted from newborn rat rib cartilage, adult rat whole eye, cornea, retina without retinal pigment epithelium, and sclera–choroid with retinal pigment epithelium in 8 M urea buffer (8 M urea, 50 mM Tris-HCl [pH 8.0], 1 mM dithiothreitol, 1 mM EDTA). Aqueous humor was collected from the anterior chamber of adult rat eyes, and the protein was precipitated with methanol. Aqueous humor and vitreous body samples were dissolved in Laemmli buffer to apply on 4% to 20% polyacrylamide gels. To visualize the ChM-I in each sample, the loading amount of each sample was varied. The bands were transferred to polyvinylidene difluoride membrane and incubated with preimmune serum (diluted 1:2000, ∼5 μg IgG/ml) or the affinity-purified antibody (∼5 μg IgG/ml) at room temperature overnight. The membrane was then reacted with goat anti-rabbit immunoglobulins conjugated to peroxidase-labeled dextran polymer (1:5 dilution; Dako, Kyoto, Japan) and incubated with diaminobenzidine and hydrogen peroxide. 

Immunohistochemistry and Immunofluorescence Microscopy

Frozen tissues or paraffin-embedded tissues of newborn rat rib cartilage and adult rat eyes were sectioned at a thickness of 4 μm. The paraffin-embedded tissue sections were deparaffinized with xylene and then rehydrated through ethanol and distilled water. These sections were treated with 0.3% H₂O₂ in methanol for 30 minutes, to reduce endogenous peroxidase activity, and washed in PBS. Subsequently, the sections were treated with testicular hyaluronidase (520 U/ml, type IV; Sigma–Aldrich, Tokyo, Japan) in PBS for 20 minutes at 37°C. ²³ After they were rinsed in PBS, the sections were incubated with (1) 5% normal goat serum for 60 minutes, (2) affinity-purified anti-ChM-I antibody (∼10 μg/ml) or preimmune serum (1:1000 dilution) or affinity-purified anti-ChM-I peptide antibody absorbed with synthetic ChM-I for 16 hours, and (3) goat anti-rabbit immunoglobulins conjugated to peroxidase-labeled dextran polymer (1:5 dilution; Dako) for 60 minutes or fluorescein isothiocyanate (FITC)-conjugated goat anti-rabbit IgG (Seikagaku) for 30 minutes. The peroxidase reaction products were visualized with diaminobenzidine and hydrogen peroxide. 

Determination of Effect of ChM-I on Tube Morphogenesis of Retinal Endothelial Cells in Culture

Human retinal endothelial cells were obtained from Applied Cell Biology Research Institute (Kirkland, WA) and cultured in minimum essential medium (MEM) containing endothelial cell growth supplement and 0.5% fetal bovine serum (FBS) on basement membrane matrix (Matrigel; Becton Dickinson Labware, Bedford, MA) at 37°C under 5% CO₂ in air. The cells of the subconfluent monolayer were harvested by adding tripsin-EDTA solution (Sigma–Aldrich) and were suspended at 4 × 10⁴ cells/ml in MEM containing 0.5% FBS. They were then incubated with recombinant human ChM-I prepared as reported previously ²⁵ at a final concentration of 0, 1, and 5μ g/ml for 20 minutes at room temperature. Matrigel of 200 μl was placed in 48-well culture plate (Costar, Corning, NY) and maintained at 37°C for 60 minutes for gelling. Then, 500 μl of the cell suspension was applied on top of the gel in triplicate and incubated at 37°C in 5% CO₂ for 6 hours. The tube morphogenesis of human retinal endothelial cells was observed under a light microscope. ²⁹  

To quantify the effects of recombinant human ChM-I on tube morphogenesis of retinal endothelial cells, the resultant cultures were photographed, and the total length of capillary-like structures per field (230 × 340 μm) was measured on computer, by using an image processing and analysis program (NIH Image version 1.61, provided in the public domain by the National Institutes of Health, Bethesda, MD, and available at http://www.nih.gov/od/oba). Each experiment was performed four times, and the data were analyzed statistically using one-way analysis of variance (ANOVA), followed by Scheffé’s multicomparison test. 

Results

Cloning of Rat ChM-I cDNA

Rat ChM-I cDNA consisted of a 127-bp 5′ untranslated sequence preceding an initiation site (Kozac consensus), a 1002-bp open reading frame deducing 334 amino acid residues, and a 293-bp 3′ untranslated sequence containing a polyadenylation consensus (Fig. 1) . Identity of the overall sequence with the bovine counterpart was 87%. The precursor form was presumably processed by proteolytic enzyme(s) at the Arg-Glu-Arg-Arg sequence, which had been demonstrated as a processing-signal motif for precursor cleavage catalyzed by furin in the constitutive secretory pathway, ³⁰ locating in the C-terminal part to release mature ChM-I (120 amino acid residues). The Arg-Glu-Arg-Arg sequence is well preserved in the human, bovine, and mouse counterparts, suggesting that they are processed similarly. ²⁰ The amino acid sequence identity of mature ChM-I was 82% with the bovine counterpart. This gene has been registered in GenBank (accession number AF051425). 

ChM-I Gene Expression in Systemic Organs and Ocular Compartments

The distribution of ChM-I mRNA in rat systemic organs was examined by RNase protection assay and ChM-I mRNA expression was demonstrated in the eye, cerebellum, and rib cartilage (Fig. 2A ) as detected in the mouse by Northern blot analysis. ³¹ No ChM-I mRNA expression was detected in other organs. An RNase protection assay using RNA samples purified from several eye compartments showed the expression of ChM-I mRNA in the iris–ciliary body, the retina without retinal pigment epithelium, and the sclera with choroid and retinal pigment epithelium. However, no expression was detected in RNA samples from extraocular muscle, conjunctiva, cornea, lens, anterior chamber angle, or lacrimal gland (Fig. 2B) . The expression was most intense in the scleral compartment including choroid and retinal pigment epithelium, among other eye compartments examined. 

In situ hybridization elucidated ChM-I mRNA expression in the resting, proliferating, and early hypertrophic zones of epiphyseal cartilage in newborn rat ribs (Fig. 3A ) as demonstrated in bovine, human, and mouse cartilage in the previous studies. ^{22 25 31} No hybridization signal was detected in the late-hypertrophic and calcified zone of the cartilage. In the eye, hybridization signal was found intensely in ganglion cells and inner nuclear layer of retina (Fig. 3C) , in the nonpigment epithelium of ciliary body (Fig. 3E) , and, weakly but significantly, in the retinal pigment epithelium (Fig. 3C) . No apparent hybridization was found in these sites when hybridized with the sense probe (Figs. 3B 3D 3F) . 

Western Blot Analysis and Immunohistochemistry

To explore the localization sites of ChM-I protein, we raised an antibody against the synthetic oligopeptide and examine by immunohistochemistry or immunofluorescence microscopy. The specificity of the anti-ChM-I antibody was examined by Western blot analysis. As seen in Figure 4a broad band of ∼25 kDa was specifically stained in the cartilage sample, and the predicted molecular mass was comparable with that of the bovine ChM-I mature form found in the fetal epiphyseal cartilage. ^{20 22}  

The distinct sharp band of ∼25 kDa was also immunostained in the aqueous humor and vitreous body and was presumed to be a mature glycosylated form of ChM-I. In addition to the ∼25-kDa band, another weak band of ∼30 kDa was immunoblotted in the vitreous body and was presumed to be a form of mature ChM-I of ∼25 kDa that was differently glycosylated. In the retina and whole eye samples, an ∼37-kDa band was apparently stained and was comparable to the molecular mass of ChM-I precursor form consisting of 334 amino acid residues. A minor∼ 33-kDa band was found in the retina compartment. These observations suggest that the ∼37-kDa form of ChM-I is unglycosylated, is an unprocessed transmembrane-type form, ²⁰ and is localized in cytoplasmic membrane of cells. 

An immunoperoxidase method using the anti-rat ChM-I antibody showed intense staining in interterritorial matrix in the proliferating cartilage zone, the resting cartilage zone, and the early hypertrophic cartilage zone of the newborn rat rib cartilage, but no staining in the late hypertrophic and calcified cartilage zone (Fig. 5A ). The staining in rib cartilage was eliminated when the antibody preincubated with the synthetic ChM-I peptide was used (Fig. 5B) . The observation suggested the presence of the mature ChM-I protein in the extracellular matrix of the cartilage zones where ChM-I mRNA was intensely expressed (Fig. 3A) . 

In the eye, faint but distinct immunostaining for ChM-I was observed in the nonpigment epithelium of the ciliary body (Fig 5C , large arrowheads), the ganglion cells (Fig 5C , small arrowheads), inner nuclear layer cells (Fig 5C , small arrows), and retinal pigment epithelium (Fig 5C , large arrows) by immunohistochemistry. No staining was observed in the sclera (Fig. 5C) or Bruch’s membrane (Fig. 5E , arrows). Immunofluorescence microscopy showed ChM-I staining in retinal pigment epithelium (Fig. 5F , large arrows), inner nuclear layer cells (Fig. 5F , small arrows), the ganglion cells (Fig. 5H , arrowhead), and vitreous body (Fig. 5H , asterisk). No staining was observed in these tissues when the antibody preincubated with the synthetic ChM-I peptide was used (Figs. 5B 5D 5G 5I) . 

Effect of ChM-I on Tube Morphogenesis of Retinal Endothelial Cells in Culture

The effect of recombinant human ChM-I on morphologic differentiation of retinal endothelial cells into capillary-like structure (tube morphogenesis) was examined in culture. The cells cast on Matrigel changed their morphology to form capillary-like tube structure after incubation for 6 hours as reported previously (Fig. 6A ). ²⁹ The capillary-like structure was apparent but less prominent in the cultures of the cells that were preincubated with recombinant human ChM-I than in the untreated control cells (Figs. 6B 6C) . The total length of the capillary-like structures per field was significantly shorter in a dose-dependent manner (Fig. 6D) . 

Discussion

RNase protection assay demonstrated ChM-I mRNA expression not only in the rib cartilage of newborn rats but also in the eye and cerebellum of adult rats in concordance with the expression in mice. ³¹ Therefore, we examined the site of expression and localization of ChM-I in the eye by in situ hybridization, Western blot analysis, and immunohistochemistry. 

ChM-I mRNA expression was detected in the iris–ciliary body, retina compartment without retinal pigment epithelium, and the scleral compartment with choroid and retinal pigment epithelium by RNase protection assay. In situ hybridization localized these ChM-I expression sites to nonpigment epithelium in the ciliary body, ganglion cells and inner nuclear layer cells in the retinal compartment, and pigment epithelium in the scleral compartment. Subsequently, immunohistochemistry confirmed the localization of ChM-I protein in the nonpigment epithelium in the ciliary body and ganglion cells, inner nuclear layer cells, and pigment epithelium in the retina. It is of interest that ChM-I protein was also demonstrated in the vitreous body and aqueous humor by immunostaining and Western blot analysis. These observations suggest that ChM-I is synthesized in nonpigment epithelium in the ciliary body and ganglion cells and inner nuclear layer cells in the retina, and that a significant amount of the synthesized ChM-I is released from these cells to the vitreous body or aqueous humor. Molecular mass of the predominant form of ChM-I existed in aqueous humor and vitreous body was ∼25 kDa, which was comparable to the mature glycosylated secretory form found in the cartilage. ²²  

The minor bands of ∼30 kDa detected in the vitreous body and ∼33 kDa in the retina compartment may be other forms of ChM-I differently glycosylated. We have also observed such microheterogeneity of the glycosylated ChM-I form isolated from cartilage of bovine fetus. The significance of the difference in glycosylation remains to be elucidated. On the other hand, a larger form of ChM-I (∼37 kDa) was immunoblotted in the retina compartment by Western blot analysis. The molecular mass was comparable to that of the precursor form of ChM-I deduced from the amino acid sequence (334 amino acid residues). This ChM-I is presumably unglycosylated and of the unprocessed transmembrane type in the ganglion cells and inner nuclear layer cells, as shown by immunohistochemistry. 

The immunodetection of ChM-I by immunohistochemistry and Western blot analysis in the nonpigment epithelial cells of the ciliary body was faint, whereas ChM-I mRNA expression was intense in the ciliary body by RNase protection assay. The reason for the discrepancy may be that most of ChM-I synthesized in the nonpigment epithelial cells of ciliary body is secreted to aqueous humor as the soluble ∼25-kDa form of ChM-I. Because aqueous humor is generated at the epithelium of the ciliary body, ChM-I synthesized at that location may migrate without difficulty to aqueous humor. Similar migration of ChM-I has been demonstrated from chondrocytes which synthesized and secreted ChM-I to the surrounding extracellular matrix. ²²  

The soluble ChM-I delivered by aqueous humor may prevent angiogenesis into the vitreous body or at the iris. Because ChM-I protein of the same molecular mass (∼25 kDa) was also detected in the vitreous body by Western bot analysis and the aqueous humor diffuses into the vitreous body, the ChM-I detected in the vitreous body is assumed to be the soluble form in the aqueous humor. However, the amounts of ChM-I were much higher in the vitreous body than in the aqueous humor (unpublished data). This observation may indicate that ChM-I is concentrated in the vitreous body by fixation to the extracellular matrix as in the cartilage. ²² The ∼30-kDa ChM-I found in the vitreous body may be concentrated by higher affinity of this form than the ∼25-kDa form to extracellular matrix, or it may be released from the ganglion cells and inner nuclear layer cells in the retina. 

Previous studies have demonstrated structural and component similarity of the extracellular matrix between the vitreous body and the cartilage. ^{26 32 33 34 35 36} A three-dimensional network of collagen fibers, stabilized by the spongy hyaluronic acid molecule between the fibers is thought to maintain the vitreous gel structure of the cartilage matrix. ³³ In addition, both the vitreous gel and the cartilage matrix consist of similar components—type II and IX collagen, ^{34 35} the cartilage-specific proteins, ²⁶ —and are rich in hyaluronic acid. ³⁶ Therefore, it is highly probable that most of the ChM-I protein present in the vitreous gel is anchored to the matrix components to exert an anti-angiogenic function, as was shown in the cartilage matrix. 

Iris rubeosis and neovascular glaucoma are the most common postoperative complications after vitrectomy for treatment of diabetic retinopathy, ³⁷ suggesting the presence of anti-angiogenic agents in the vitreous body. Recombinant human ChM-I actually inhibited the tube morphogenesis of cultured human retinal endothelial cells in culture at a dose similar to that of other angiogenesis inhibitors in endothelial cells. ^{38 39} This may indicate that ChM-I is one of the angiogenesis-inhibiting factors in the eye. 

Recently, another antiangiogenic factor, PEDF was discovered in the retinal pigment epithelium, where ChM-I was also identified. ¹³ These antiangiogenic factors may prevent vascular invasion from choroid to retina in age-related macular degeneration at the pigment epithelium layer. ^{40 41} Because ChM-I was absent in Bruch’s membrane, this membrane may serve as a structural barrier to vascular invasion, but the pigment epithelial layer may serve as a functional barrier by releasing ChM-I and PEDF. Identification of these factors provides a clue to analyzing the molecular basis of the angiogenesis-inhibiting mechanism in the eye. 

Further studies are needed to clarify the roles of ChM-I in the development of vasculature in the eye in pathologic conditions and to test the possibility of therapeutic application of ChM-I for angiogenesis prevention in some ocular disorders. ChM-I was also expressed in the cerebellum and may have an angiogenesis-inhibiting function there; however, its localization and role remain to be investigated. 

 Supported in part by grants-in-aid for Science Research from the Ministry of Education, Science, Sports and Culture and the Japan Society for the Promotion of Science (Research for the Future Program).

 Submitted for publication January 18, 2000; revised December 11, 2000; accepted January 8, 2001.

 Commercial relationships policy: N.

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked“ advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Nucleotide and deduced amino acid sequence of rat ChM-I cDNA. The putative one-transmembrane domain is denoted by a dashed line. Mature ChM-I (underlined) is at the 3′ end of the coding sequence. Potential processing signal is boxed. Asterisk: putative N-glycosylation site. A polyadenylation consensus sequence is denoted by a wavy line. The sequence of peptide used for antibody preparation is double underlined.

Figure 2.

 

Rat ChM-I mRNA expression in rat systemic organs. An RNase protection assay demonstrated remarkably intense expression of ChM-I mRNA in the newborn rat rib cartilage and less intense expression in the adult rat eye and cerebellum (A). Lane P, probe; lane 1, cerebrum; lane 2, cerebellum; lane 3, eye; lane 4, submandibular salivary gland; lane 5, heart; lane 6, lung; lane 7, liver; lane 8, pancreas; lane 9, small intestine; lane 10, colon; lane 11, spleen; lane 12, testis; lane 13, ovarium; lane 14, kidney; lane 15, thymus; lane 16, smooth muscle; and lane 17, newborn rib cartilage. In the eye, ChM-I mRNA was expressed in iris–ciliary body, retina without pigment epithelium, and sclera with pigment epithelium and choroid (B). Lane P, probe; lane 1, whole eye; lane 2, extraocular muscle; lane 3, conjunctiva; lane 4, cornea; lane 5, lens; lane 6, iris–ciliary body; lane 7, anterior chamber angle; lane 8, retina without pigment epithelium; lane 9, sclera with pigment epithelium and choroid; lane 10, lacrimal gland; and lane 11, newborn rib cartilage.

Figure 3.

 

In situ hybridization for ChM-I mRNA expression in newborn rat rib cartilage and adult rat eye. ChM-I mRNA expression was detected in the cartilage and the eye. In the cartilage (A, B), intense signals were found in the chondrocytes. No signals were found in the late hypertrophic zone ( ). In the eye (C, D), hybridization signals were found in ganglion cells (small arrows) and the inner nuclear layer (arrowheads), and faintly in the retinal pigment epithelium (large arrows) in the retina. The nonpigment epithelium of the ciliary body (E, F) also had significant signals (arrowheads). No signals were found at these sites with the sense probe. Magnification, (A, B) ×65; (C, D, E, and F)× 135.

Figure 4.

 

Western blot analysis of ChM-I. A broad band of ∼25 kDa appeared in the immunoblot of the cartilage and the aqueous humor samples. In the vitreous body, an apparent band of ∼25 kDa and a minor band of ∼30 kDa were observed. Bands of ∼37 kDa appeared in the immunoblot of the whole eye and in the retina without pigment epithelium, where a minor∼ 33-kDa band was visible. Lane 1, cartilage; lane 2, aqueous humor; lane 3, vitreous body; lane 4, whole eye; lane 5, retina without pigment epithelium; lane 6, iris–ciliary body–anterior chamber angle; lane 7, sclera with retinal pigment epithelium and choroid.

Figure 5.

 

Immunohistochemical localization of ChM-I in newborn rat rib cartilage and adult rat eye. In the cartilage (A, B), specific strong staining for ChM-I was observed in the interterritorial matrix by immunohistochemistry. ChM-I was faintly detected in the nonpigment epithelium (C, D) of the ciliary body (large arrowheads), retinal pigment epithelium (large arrows), and vitreous body ( ) by immunoperoxidase staining. Distinct immunostaining was observed in the ganglion (small arrowheads) and inner nuclear layer cells (small arrows). The sclera was not stained with the anti-ChM-I antibody. At higher power magnification (E), Bruch’s membrane (arrows) was not stained with anti-ChM-I antibody. (F, G) The inner nuclear cell layer (small arrows) and pigment epithelium of the retina (large arrows) was stained faintly, but the sclera (S) was not stained with the anti-ChM-I antibody. (H, I) By immunofluorescence microscopy, ChM-I was expressed intensely in ganglion cells (arrowheads) and less intensely in the vitreous body ( ). No immunostaining was observed when the antibody preabsorbed with synthetic ChM-I peptide was used. S, sclera; C, choroid; RPE,retinal pigment epithelium. Magnification, (A, B) ×105; (C, D) ×210; (E) 840; (F, G, H, and I) ×420.

Figure 6.

 

Suppression of morphogenesis of tubelike cellular networks of human retinal endothelial cells by recombinant ChM-I in vitro. A representative micrograph shows a tubelike cellular network formation of human retinal endothelial cells 6 hours after cultivation on cell culture plates in MEM containing 0.5% FBS at 37°C (A). The tube morphogenesis of these cells was suppressed by pretreatment with recombinant human ChM-I (B: 1.0 μg/ml; C: 5.0 μg/ml). (D) Quantitative evaluation of suppression by image computer analysis. Data are means ± SD of four experiments. No recombinant ChM-I versus 1.0 μg/ml: P < 0.001; no recombinant ChM-I versus 5.0 μg/ml: P < 0.001; 1.0μ g/ml versus 5.0 μg/ml: P = 0.035 (one-way ANOVA and Scheffé’s multicomparison test). Magnification, (A, B, and C) ×290.

The authors thank Itaru Kihara, Daisuke Kondo, Hidehiko Fujinaka (Faculty of Medicine, Niigata University) and Kaori Mitsui (Researcher Center, Mitsubishi Chemical Corporation, Kanagawa) for their comments, and Kan Yoshida and Masaaki Nameta for technical assistance.