Knockdown of Zebrafish Lumican Gene (zlum) Causes Scleral Thinning and Increased Size of Scleral Coats*

The lumican gene (lum), which encodes one of the major keratan sulfate proteoglycans (KSPGs) in the vertebrate cornea and sclera, has been linked to axial myopia in humans. In this study, we chose zebrafish (Danio rerio) as an animal model to elucidate the role of lumican in the development of axial myopia. The zebrafish lumican gene (zlum) spans ∼4.6 kb of the zebrafish genome. Like human (hLUM) and mouse (mlum), zlum consists of three exons, two introns, and a TATA box-less promoter at the 5′-flanking region of the transcription initiation site. Sequence analysis of the cDNA predicts that zLum encodes 344 amino acids. zLum shares 51% amino acid sequence identity with human lumican. Similar to hLUM and mlum, zlum mRNA is expressed in the eye and many other tissues, such as brain, muscle, and liver as well. Transgenic zebrafish harboring an enhanced GFP reporter gene construct downstream of a 1.7-kb zlum 5′-flanking region displayed enhanced GFP expression in the cornea and sclera, as well as throughout the body. Down-regulation of zlum expression by antisense zlum morpholinos manifested ocular enlargement resembling axial myopia due to disruption of the collagen fibril arrangement in the sclera and resulted in scleral thinning. Administration of muscarinic receptor antagonists, e.g. atropine and pirenzepine, effectively subdued the ocular enlargement caused by morpholinos in in vivo zebrafish larvae assays. The observation suggests that zebrafish can be used as an in vivo model for screening compounds in treating myopia.

Myopia is a very common ocular disorder, which is characterized by excessive elongation of the eyeball. In Taiwan, the prevalence of myopia is about 84% of schoolchildren aged 16 -18, and the prevalence of high myopia (less than Ϫ6.0 D) at 18 years of age is 24% in girls and 18% in boys (1). In contrast, the prevalence of high myopia is much lower in Western countries, about 1% of the general population (2). These studies imply that genetic susceptibility of ethnic differences may account for the high prevalence of myopia in Taiwanese.
Lumican, a member of the SLRP family, is one of the major extracellular components in interstitial collagenous matrices of corneal stroma, sclera, aorta, skin, skeletal muscle, lung, kidney, bone, cartilage, and intervertebral discs (13)(14)(15)(16)(17)(18)(19)(20). In corneal stroma, lumican is a KSPG, whereas lumican presents as an under-or unglycanated glycoprotein in other tissues (9,13,14,21). It has been proposed that the horseshoe-shaped lumican core protein binds collagen molecules to modulate collagen fibril diameter, whereas the N-linked glycosaminoglycan chains regulate fibril spacing and stromal hydration for the formation and maintenance of transparent corneas (22)(23)(24). The wide distribution of lumican implies that lumican may have multiple functions in tissue morphogenesis and maintenance of tissue homeostasis, besides serving as a regulatory molecule of collagen fibrillogenesis (24). Indeed, lumican plays essential roles in wound healing by modulating epithelial cell migration (25) and in epithelium-mesenchyme transition of the injured lens (26), in addition to regulating collagen fibrillogenesis (20,25). Lumican-null (Lum Ϫ/Ϫ ) mice and lumican-and fibromodullin-null (Lum Ϫ/Ϫ Fmod Ϫ/Ϫ ) mice showed alterations of collagen fibril arrangement in interstitial connective tissues. Lum Ϫ/Ϫ Fmod Ϫ/Ϫ mice exhibited elongated axial lengths and thin sclera, which are characteristics of high myopia, and lumicannull mice (Lum Ϫ/Ϫ ) also had a slight elongation of the eyeball (27,28). These findings indicated that these proteoglycans may be directly or indirectly involved in scleral development, resulting in the pathoetiology of high myopia.
The zebrafish is an excellent model to study vertebrate genetics and development (29 -31). As a disease model, transgenic zebrafish provide several advantages, such as shorter duration of embryonic development and optically transparent embryos; its development also provides easy access for observation and treatment schemes during embryogenesis in comparison with transgenic mouse models. In this study, we examined the structure, expression pattern, promoter activity, and function of zlum. Our data indicated that lumican is highly conserved between zebrafish and mammals (e.g. human and mouse) in respect to gene structure, expression patterns, and protein function. In particular, ocular enlargement and scleral thinning with changes in the ultrastructure of the sclera were noted when zLum expression was down-regulated by antisense zlum morpholinos (MO). Furthermore, in vivo zebrafish larvae assays were performed to elucidate the potential application of several muscarinic receptor antagonists to attenuate increased scleral coats caused by zlum knockdown with MO. Our results suggest that zebrafish can be used as an in vivo model for screening compounds of treating myopia.

EXPERIMENTAL PROCEDURES
Zebrafish Husbandry-Zebrafish were raised and maintained according to protocols established previously (32). Briefly, adult zebrafish and embryos were maintained at 28.5°C on a 14-h light and 10-h dark cycle. Embryos were sorted at the different stages required for each experiment and staged according to Kimmel et al. (33). Chorions were removed manually with Dumont Watchmaker's Forceps No. 5. All procedures were approved by the Institutional Animal Care and Use Committee of National Taiwan University and performed in compliance with the Association for Research in Vision and Ophthalmology Statement for Use of Animals in Ophthalmic and Vision Research.
Characterization of zlum-To identify the zebrafish expressed sequence tag clone encoding a putative protein sharing high sequence similarity with the human and mouse SLRP family proteins, we applied the Basic Local Alignment Search Tool (BLAST) analysis of the GenBank TM database using the fulllength human lumican cDNA sequence. An ϳ4.6-kb NotI/ MluI zebrafish genomic DNA fragment containing the 5Ј portion of the zebrafish lumican gene was amplified by PCR and subcloned into the pBluescript SK vector (Stratagene, La Jolla, CA). The insert was sequenced, using T 3 , T 7 , and walk-in primers, by the DNA core of the Department of Molecular Genetics, National Taiwan University.
The 5Ј-and 3Ј-ends of the zlum mRNA were amplified using the 5Ј-rapid amplification of cDNA end (5Ј-RACE) and 3Ј-RACE systems, respectively (Invitrogen). For the 5Ј-RACE experiment, 1 g of total RNA from zebrafish eyes was reversetranscribed with a lumican-specific primer (5Ј-AAGTAGAG-GTATTTGATTCCGGTC-3Ј) corresponding to a sequence in exon 2 of the zlum gene. The RNA templates were degraded by treatment with an RNase mix. A poly(dCTP) tail was added to the 3Ј-end of the cDNAs with terminal deoxynucleotidyltransferase. The cDNA was amplified with a second gene-specific primer (5Ј-GCACAAGAAGGTGATGAAACG-3Ј) corresponding to a sequence from the junction between exon 1 and 2 in conjunction with the abridged anchor primer (5Ј-GGCCA-CGCGTCGACTAGTACGGGIIGGGIIGGGIIG-3Ј). The resulting PCR products were diluted 100-fold and used as templates to be reamplified with a third gene-specific primer (5Ј-CAGACTTAGAAGTCCAGCCAAC-3Ј) in conjunction with the universal amplification primer (5Ј-CUACUACUAC-UAGGCCACGCGTCGACTAGTAC-3Ј). For 3Ј-RACE, PCRs were performed using a gene-specific primer (5Ј-GCCTCAG-AGATCATCTTTGAATAG-3Ј) corresponding to a sequence in exon 3 of the zlum gene. The cycling conditions were as follows: 34 cycles of 94°C for 1 min, 55°C for 1 min, and 72°C for 3 min followed by a 10-min extension at 72°C at the end of the cycles. Finally, the 5Ј-RACE and 3Ј-RACE PCR products were gel-purified, and the sequences were determined with a dideoxy sequencing protocol. The transcription initiation and termination sites of the zlum gene were determined by a sequence comparison between genomic DNA, the 5Ј-RACE product, and the 3Ј-RACE product, respectively.
Sequence Alignment and Phylogenetic Analysis of zLum-The amino acid sequences of the open reading frames (ORFs) were initially aligned using the ClustalW program, as described previously (34). The aligned amino acid sequences used were subsequently analyzed with the Neighbor-Joining distance analysis method to construct a phylogenetic tree by using Geneious Pro software (version 4.7). Bootstrap values were calculated from 100 replicates, and values Ͼ50% are indicated at each divergence point.
RT-PCR-Unless otherwise specified, RT-PCR reagents used in this procedure were purchased from Promega (Madison, WI). RevertAid TM H Minus First Strand cDNA synthesis kit was purchased from Fermentas (St-Leon-Rot, Germany). Zebrafish cDNA was synthesized using 40 l of 5ϫ reverse transcription buffer, 20 l of 0.1 M dithiothreitol, 8 l of 25 mM dNTPs, 10 l of RNasin (40 units/ml), 10 l of 50 mM random hexamers (GE Healthcare), 10 l of avian myeloblastosis virus reverse transcriptase (9.5 units/l), and 1 g of heat-denatured corneal poly(A) ϩ RNAs. Diethylpyrocarbonate-treated water was added to bring the final reaction volume to 200 l, and the reaction was incubated at room temperature for 10 min, 42°C for 90 min, 100°C for 2 min, and 0°C for 5 min. Twenty micro-liters of each of the above RT reactions was added to 80 l of a PCR mixture containing the following: 8 l of 10ϫ PCR buffer without MgCl 2 , 8 l of 25 mM MgCl 2 , 10 l of 20 ng/l primers, 0.5 l of Taq polymerase (5 units/l), and 45.5 l of H 2 O. The cycling conditions were as follows: 35 cycles of 94°C for 1 min, 57°C for 1 min, and 72°C for 1 min followed by a 15-min extension at 72°C at the end of these cycles. Primers used were as follows: CCGCTCGAGCGGATGTTTGCTCTGGGATCC-ATTC (forward 5Ј-XhoI cut site) and TCCCCGCGGGGACT-ATTCAAAGATGATCTCTGAGG (reverse 3Ј-SacII cut site). The PCR product was confirmed by an appropriate restriction enzyme digestion and analyzed by electrophoresis on a 1.5% agarose gel.
Generation of an Epitope-specific Anti-zebrafish Lumican Antibody-To develop an anti-zLum antibody, an oligopeptide deduced from zLum cDNA was synthesized (N-terminal peptide, CNERNLKFIPIVPTGIKY). The peptides were conjugated to keyhole limpet hemocyanin for antibody production in rabbits. The antibodies were purified through an immune absorbent column of the above zebrafish lumican oligopeptide conjugated to SulfoLink gel (Pierce) according to the manufacturer's instructions. Fractions containing purified anti-zebrafish lumican antibody were pooled and concentrated, and the protein concentration was measured by spectrophotometry at 280 nm.
In Situ Hybridization and Immunohistochemistry-Embryos were fixed in 4% paraformaldehyde in 1ϫ PBS overnight at 4°C, rinsed with PBS three times, transferred into 100% methanol, and stored at Ϫ20°C until use. To prevent melanization, embryos raised to time points beyond the 24-h post-fertilization (hpf) stage were treated with 0.003% phenylthiourea. Whole mount RNA in situ hybridization was carried out as described previously (34). Sense and antisense digoxigenin-labeled oligonucleotide probes were obtained from Bio Basic Inc. (Ontario, Canada). The oligonucleotide sequence (5Ј-3Ј) was GTTTCCATCCAAGCGCAGGGTCCTCAGTCTAGAGTA-GTTGACCGGTGAGCTAAATCTGCA. The hybridization signals were visualized with anti-digoxigenin antibody-alkaline phosphatase conjugates using procedures recommended by Roche Applied Science. The sections were counterstained with 0.5% neutral red and mounted. Images were obtained using an AxioCam digital camera on a dissection microscope (Zeiss, Germany).
Adult zebrafish corneal tissues were fixed with 4% paraformaldehyde in PBS and then embedded in paraffin. Paraffinized sections (5 m) of adult zebrafish corneal tissue were placed on slides and processed for deparaffinization and immunohistochemistry. Before immunostaining, tissue sections were incubated with 0.5 unit/ml keratanase at 37°C overnight. After blocking with 3% hydrogen peroxide for 30 min, the samples were incubated with either the primary affinity-purified antizebrafish lumican antibody (0.1 g/ml) or monoclonal antikeratan sulfate antibody, washed in PBS, and then incubated with a biotinylated secondary antibody (goat anti-rabbit IgG). After washing in PBS, sections were incubated with streptavidin-HRP (DAKO, Carpinteria, CA), then washed in PBS, and incubated with the 3Ј,3-diaminobenzidine chromogen for 5-10 min. Negative control samples were obtained using preimmune rabbit IgG.
Analysis of zlum Promoter by Transgenic Zebrafish-Genomic DNA, both 1.7 and 0.48 kb from the 5Ј-untranslated region of the zlum gene, were amplified with specific PCR primers and inserted into the multiple cloning sites of pBluescript II SK vectors (Stratagene, La Jolla, CA) containing an EGFP sequence. PCR primers are as follows: 5Ј-ATAAGAATGC-GGCCGCTCCATTAATTCGACAGACCAG-3Ј and 5Ј-ATAAGAATGCGGCCGCAGGTAGACAACACGGTTA-TGT-3Ј (forward primer) and 5Ј-CGACGCGTGGCTGCAC-AACTTAAATTAAACCT-3Ј (reverse primer). The recombinant plasmids were propagated in Escherichia coli DH5␣ and purified with a Qiagen plasmid purification maxi kit. Purified plasmid DNA was adjusted to 50 ng/l in distilled water and microinjected into one-cell stage zebrafish embryos under a dissecting microscope. Embryos with GFP expression were observed and imaged under a fluorescence microscope.
zlum Knockdown by Morpholino Injection-A morpholinoantisense oligonucleotide (Gene Tools, Philomath, OR) was designed to target the 5Ј-untranslated and/or -flanking regions, including the translation start codon of the respective genes. The MO sequence was designed as follows: zlum-MO, 5Ј-GATCCCAGAGCAAACATGGCTGCAC-3Ј. This oligonucleotide complemented the sequence from Ϫ8 through ϩ17 with respect to the translation initiation codon. A search of the database did not identify any sequence similarity of known zebrafish genes to zlum-MO. A random sequence MO (RS-MO) serves as a control for zlum-MO, 5Ј-CCTCTTACCTCA-GTTACAATTTATA-3Ј. This RS-MO was obtained from Gene Tools as a standard control oligonucleotide with no target specificity. Solutions were prepared and injected at the 1-4-cell stage as described by Nasevicius and Ekker (35).
Transmission Electron Microscopy-Corneas of zlum-MOinjected, RS-MO-injected, and wild type zebrafish at 7 and 12 days post-fertilization (dpf) stage were fixed in 50 mM phosphate buffer (pH 7.2) containing 2.5% glutaraldehyde and 2% paraformaldehyde for 24 h at room temperature. After re-fixation in 1% osmium tetraoxide for 4 h at room temperature, the samples were washed in phosphate buffer, dehydrated, and embedded in Epon 812 epoxy resin. Semi-thick sections (100 nm) were stained with toluidine blue. Ultrathin 50-nm sections were collected on 75 mesh copper grids and stained with uranyl acetate and lead citrate, and images were photographed with a Hitachi 7100 transmission electron microscope (TEM) (Hitachi, Tokyo, Japan) equipped with an AMT digital camera.
Image Analysis-Collagen fibril diameters and scleral thickness were measured with Image-Pro Plus version 4.5. Corneal stroma, anterior sclera, and posterior sclera from six zlum-MO and six wild type zebrafish at each 7-and 12-dpf stage were analyzed. Six to 12 areas of collagen fibrils were analyzed for each region in each group, generating 772-1678 measurements for each condition, and 27-84 measurements were generated by analysis of scleral thickness for anterior and posterior sclera in each group at the 12-dpf stage. Collagen fibers and scleral thickness (in micrometers) are represented as the mean Ϯ S.D., and all measurements were analyzed using Student's t tests assuming unequal variances.
In Vivo Zebrafish Larvae Assays for the Eye Size Regulation by Muscarinic Receptor Antagonists-To apply our zebrafish model for myopia drug screening, different muscarinic receptor antagonists, atropine, pirenzepine (M1), and methoctramine (M2) purchased from Sigma were used according to protocols previously described with some modifications (36 -38). Briefly, before drug screening test, the maximal sublethal concentration of each drug was determined, and the drug concentration was selected by the survival rate of embryos greater than 60% at 6 dpf stage. Zebrafish larvae were placed in a 96-well plate at 1 fish per well. In this assay, the zlum-MO-injected larvae were exposed to 0.5% atropine, 0.25% pirenzepine, and 0.01% methoctramine starting at 3 dpf for 48 h, respectively, after which the larvae were anesthetized with Tricaine, immobilized in 3% methylcellulose, and observed under dissecting microscopy. The axial length of eyeball, diameter of scleral coat, and the ratio of RPE diameter/scleral coat diameter were measured. Student's t test was used to compare these parameters between control group and treated groups, respectively. The difference of significance was defined as p Ͻ 0.05.

RESULTS
Primary Structure of zlum and Alignment Analysis of the Zebrafish Lumican Amino Acid Sequence-We identified zlum by performing a BLAST search of the publicly available zebrafish databases with human LUM. The zlum gene is 11 kb upstream of zkera (keratocan) (Fig.  1). We have isolated clones representing the full open reading frames (ORF) of zebrafish lumican. A cDNA clone encoding the lumican ORF was subcloned into the pBluescript SK vector (Stratagene, La Jolla, CA). The ORF of the zlum gene was 1032 bp long and encoded 344 amino acid residues. The proved nucleotide sequence of the zebrafish lumican gene (zlum) gene was submitted to the GenBank TM database under GenBank TM accession number GQ376197. This sequence has been scanned against the database and is significantly related to the sequence NM_ 001002059.1, which is derived from BC071347.1 and has not yet been subject to final NCBI review (39).
The entire genomic DNA sequence of zlum is shown in supplemental Fig. 1. The zlum gene spans ϳ4.6 kb (4610 bp) and contains three exons and two introns (Fig. 1). Sequence analysis revealed that exon 1 contains 26 untranslated nucleotides, exon 2 contains 24 noncoding and 880 coding nucleotides, and exon 3 contains 152 bases of coding sequence and 1106 bases of 3Ј-untranslated sequence. The transcription initiation site marked ϩ1 was determined by 5Ј-RACE, as described under "Experimental Procedures." The first translation initiation ATG codon is located at the 844th base downstream of the beginning of exon 1. There was no TATA consensus sequence found in the ϳ2.58 kb of proximal 5Ј-flanking region. The full-length zlum cDNA clone (ϳ1.9 kb) contains a 1032-bp ORF of 344 amino acids of zLum core protein. Like other SLRP core proteins such as bovine lumican (40), bovine mimecan (41), and mouse keratocan (42), zebrafish lumican consists of three distinct domains, a highly conserved central leucine-rich repeat region flanked by hypervariable N-and C-terminal regions. As shown in supplemental Fig. 1, after the signal peptide (Met to Tyr, the first 20 amino acids), the negatively charged N-terminal domain contains a possibly sulfated tyrosine and four possibly conserved cysteine residues that form intramolecular disulfide bonds, followed by nine tandem leucine-rich repeats (LXXLXLXXNX(L/I)) that are similar to lumican of other species and might mediate binding to other extracellular matrix components and C-terminal globular domains containing two conserved cysteine residues.
Sequence Alignment and Phylogenetic Analysis-For comparison, multiple alignment analysis of the predicted amino acid sequences of zLum with those of other species is shown in supplemental Fig. 2. Using the sequence of zLum, a search for homologous sequences was made with BLAST. The predicted amino acid sequences showed a high homology among lumican core proteins of different species. Zebrafish lumican shared 51% amino acid identity to human, 50% amino acid identity to mouse, and 53% amino acid identity to chicken lumican, respectively. Human lumican shared 86% amino acid identity to mouse and 68% amino acid identity to chicken lumican.
To analyze the evolutionary relationships, a bootstrapped neighbor-joining tree using Jukes-Cantor model was used to illustrate the relationship between different 18 species (Fig. 2). The numbers at the nodes represent the statistical confidence estimated by the bootstrap procedure. Bootstrap values were calculated from 100 replicates, and values of Ͼ50% are indicated at each divergence point. The bootstrap values allow inspection of the relationships among clades with low or no ambiguity. The mouse lumican appears to be more closely related to the lumican of human and primates. The zebrafish and chicken lumican are closer to those of Salmo salar and Taeniopygia guttata as compared with human and other mammalian lumican, although they displayed similar structure and Ͼ50% homology with those of other species.
Spatial Distribution of Zebrafish Lumican mRNA-To further confirm the expression patterns of zlum, RT-PCR analysis was carried out using template cDNAs that were reverse-transcribed from total RNAs from various tissues. The results of RT-PCR revealed that an ϳ1-kb RT-PCR product was abundant in most tissues examined, e.g. eye, brain, liver, muscle, and fin (Fig. 3). Whole mount in situ hybridization was used to analyze the expression of zLum mRNA during development. The results revealed that zLum mRNA was widely expressed in many tissues throughout the fish body, and it is highly expressed in the embryonic fore-, mid-, and hindbrain, anterior spinal cord, and eyes at 1 and 3 dpf (Fig. 4, A-C). The corresponding control sense riboprobes showed negative staining in samples (Fig. 4D). Lumican could be found in the corneal stromal layer of adult human and mouse corneas (43). In the zebrafish eye, zLum mRNA was also expressed mainly in the corneal stromal layer (Fig. 4, E and F) and sclera (Fig. 4, G and  H), which was similar to other species (17,44). The corresponding control sense riboprobes showed negligible hybridization signals in samples (Fig. 4, I-L).
Characterization and Spatial Distribution of Zebrafish Lumican Protein-An affinity-purified anti-zLum antibody against a synthetic peptide N-terminal peptide (CNERNLKFIP-IVPTGIKY), corresponding to the 18 N-terminal amino acid residues deduced from the zlum cDNA, was generated to detect zebrafish lumican. Immunohistochemistry with anti-zLum antibodies showed that similar to human and mouse lumicans, which were both found in the corneal stromal layer (25), zLum was also present mainly in corneal stromal layers (Fig. 5, A and B) and scleral tissue (Fig. 5, C and D), and weak immunostaining signal was also detected in the corneal epithelium and surrounding tissues, such as the iris and ciliary body. No immunoreactivity could be detected in samples incubated with preimmune rabbit IgG (Fig. 5, E-H). Tissue sections probed with anti-keratan sulfate (KS) antibody demonstrated that keratan sulfate-glycosaminoglycan was primarily present in corneal stroma, and little or no immune reactivity could be seen in the corneal epithelium (Fig. 5, I and J). Keratanase treatment abolished the immune reactivity seen in the stroma by the anti-KS antibody (Fig. 5, M and N). It was of interest to note that no immune reactivity was seen in the scleral tissue by the anti-KS antibody (Fig. 5, I, K, and L). This observation indicates that keratan sulfate-glycosaminoglycan is not present in the scleral tissue, consistent with the notion that zLum exists as KSPG in corneal stroma and under-glycanated glycoprotein in other tissues, such as the corneal epithelium and sclera.
To further demonstrate that the lumican in zebrafish cornea is a proteoglycan, total lysates from zebrafish eyes were treated with or without keratanase or endo-␤-galactosidase digestion and were then subjected to Western blotting analysis. Fig. 6   shows that a major band of 50 kDa and multiple high molecular mass bands ranging from 60 to 170 kDa were labeled by the antibodies (lane 1) in specimens prepared from eyes without enzyme digestion. By contrast, in specimens treated with keratanase (Fig. 6, lane 2) or endo-␤-galactosidase digestion (lane 3), the high molecular mass bands diminished, but the major band at 50 kDa remained. These data indicated that the zLum protein isolated from eyes exists in two forms, as a KSPG and an under-glycanated core protein.
zLum Promoter Activity in Transgenic Zebrafish-To confirm that the genomic DNA clone of the isolated zlum was indeed derived from a functional allele, we examined promoter activity of the 5Ј-flanking sequence of the zlum genomic DNA, using transgenic zebrafish microinjected with zlumpr1.7-EGFP-bpA and zlumpr0.5-EGFP-bpA, respectively (Fig. 7, B and C). Live transgenic EGFP-positive embryos were screened and selected under a fluorescent microscope. In transgenic zebrafish, 86% (212:246) of the 1.7-kb zlum promoter fragment-injected embryos expressed the EGFP transgene throughout the fish body, including the eyes, at 3 and 7 dpf (Fig.  7, D-F). However, the 0.5-kb promoter fragment of zlum failed to drive EGFP expression in transgenic zebrafish (0:113) (Fig.  7G). These results indicated that the 1.7-kb zlum promoter is able to drive EGFP expression in the transgenic fish, whereas deletion of the 5Ј 1.2 kb of the sequence completely eliminates EGFP expression, suggesting that the 1.2-kb region between the 1.7-and 0.5-kb 5Ј-flanking sequence consists of the necessary regulatory elements controlling lumican promoter activity for zlum expression in vivo.
Effects of zlum Knockdown on Embryonic Development-To investigate zlum function during zebrafish development, MO was microinjected into fertilized zebrafish eggs. Western blot showed that zLum protein decreased after zlum-MO was injected. (Fig. 8A). Retarded development was seen in the zlum-MO embryos as compared with wild type embryos at the 22-hpf stages (Fig. 8, C and D). Significant morphological variations could be readily identified between zlum-MO embryos and control embryos beginning at 6 dpf stage, although there were no identifiable differences in phenotypes between the control RS-MO and wild type embryos. About 58.6% of zlum-MO-injected 7-dpf larvae (277:473) exhibited severe major defects, and less than 1% of the RS-MO negative control group (1:105) and 2.1% of the wild type group with mock injection of PBS alone (9:438) had minor morphological changes. A higher dose of zlum-MO (20 ng/2 l) produced more severe abnormalities in embryos and led to a higher mortality rate (187 dead embryos out of 211 total injected). The most common abnormalities of the zlum MO-injected group were enlarged eyes, enlarged pericardium, and deformed body shape (Fig. 8, E and  F). In particular, significantly enlarged ocular eyeballs were  (Fig. 8H). The control group injected with RS-MO had normal eye morphologies like the mock-injected control group.
Decreased Lumican Synthesis Results in Abnormal Collagen Architecture of the Corneal Stroma and Scleral Tissue-To elucidate the function of lumican in regulating collagen fibrillogenesis in the fish eye, eyes of zlum-MO fish were subjected to morphological analysis with TEM at 7 and 12 dpf (n ϭ 6, in each group). Fig. 9 shows the ultrastructural changes of collagenous matrix at corneal stroma (CS), anterior sclera (AS), and posterior sclera (PS) area (as shown in Fig. 9A) between the zlum-MO and wild type group at 12 dpf stage. The fibril architecture of corneal stroma and anterior sclera collagen of the MO-treated group differed from that of the control group and appeared irregularly, with large variations in fibril diameter and fibril spacing in comparison with WT control (compare Fig. 9, C with D and E with F). However, there was no significant difference in the diameter of collagen fibrils at the posterior sclera area between both groups (compare Fig. 9, G with H). For summary of the data of fibril diameter measurement, Fig. 9B shows that there was significant difference in the collagen fibril diameter of corneal stroma and anterior sclera matrix between the wild type and zlum-MO group at 12 dpf stage (CS, 13  Interestingly, a larger size of eyeball accompanied by thinner sclera was noted in the zlum MO-injected larva (Figs. 8F, 9, and 10). Lower magnification of electron microscopy showed that posterior scleral tissue from the wild type group consisted of ϳ3-4 layers of fibroblastic cells with newly formed collagen fibrils in relatively regular rearrangements at 7 dpf stage in the posterior sclera (Fig. 10A). However, there was only about one layer of fibroblastic cells in the posterior sclera of MO-treated larva at 7 dpf stage (Fig. 10B), which consisted of very few and irregular collagenous matrix with large variegations in fibril diameters and fibril spacing (Fig. 9H).  (Fig. 10C).
In conclusion, our results demonstrated that the ultrastructural changes of scleral tissue caused by decreased lumican synthesis with morpholinos were similar to the features found in sclera of mammalian model of high myopia. Taking the advantage that zebrafish eye development can be easily manipulated, e.g. administration of morpholinos and analyzed, we have attempted to develop a protocol of using zebrafish as an experimental model to test the efficacy of  It contains a 1.7-kb 5Ј-regulatory region of the zlum gene, the untranslated region of exon 1 (844 bp), an SV40 polyadenylation signal, and the pEGFP vector sequence. C, structure of zlumpr0.5-EGFPSV40 (2.0 kb). It contains a 0.5-kb 5Ј-regulatory region of the zlum gene, the untranslated region of exon 1 (844 bp), an SV40 polyadenylation signal, and the pEGFP vector sequence. D, EGFP expression was observed in the 3 dpf zebrafish after injecting linearized zlumpr1.7-EGFP SV40 DNA fragment. E, under different observation view of the same fish (3 dpf), strong EGFP expression was detected in corneal tissue. F, EGFP was expressed at the 7 dpf stage. G, as a control, no EGFP was detected after injecting the linearized zlumpr0.5-EGFP SV40 DNA fragment.
compounds, e.g. antagonist of muscarinic receptors, which can potentially modulate the pathoetiology of myopia as described below.

Administration of Muscarinic Receptor Antagonists Reverts the Enlargement of Scleral Coats Caused by zlum Knockdown
with Morpholinos-To evaluate if the zebrafish could be an experimental drug screening model for treatment of induced myopia, we tested several different muscarinic receptor antagonists that have been used in treating myopia. As shown in Fig.  11, two drugs out of the three muscarinic receptor antagonists tested, atropine (0.5%) and pirenzepine (0.25%), produced promising outcomes as there was significant reduction in excessive axial elongation and enlarged scleral coats observed in the treated zlum-MO group as compared with those of untreated zlum-MO group (Fig. 11, D, E versus  C). However, no obvious improvement was noted in the eye of fish treated with 0.01% methoctramine (M2 receptor antagonist) and injected by zlum-MO at 7 dpf stage (Fig. 11F). To better evaluate the effects of drugs, we measured the diameters of the retina and sclera of the experimental fish (as shown in Fig. 12, A and B). Fig. 12C (1st to 5th  lanes) showed a significant reduction in the excessive axial elongation found in the atropine-and pirenzepine-treated groups as compared with those of the untreated zlum-MO group at 7 dpf stage ((205.01 Ϯ 21.10 m (atropinetreated) and 208.27 Ϯ 23.51 m (pirenzepine-treated) versus 248.71 Ϯ 20.41 m (untreated), both have a p Ͻ 0.001, n ϭ 55, 64, and 70, respectively)), and there was no significant change in the axial length in the methoctraminetreated group (242.05 Ϯ 21.68 m, p Ͼ 0.05, n ϭ 56). Fig. 12C (6 to 10th  lanes) showed a significant decrease in scleral diameters in the atropineand pirenzepine-treated groups as compared with those of the untreated zlum-MO group at 7 dpf stage ((323.51 Ϯ 46.83 m (atropine-treated) and 326.18 Ϯ 32.88 m (pirenzepine-treated) versus 399.07 Ϯ 59.68 m (untreated), both have p values Ͻ0.001, n ϭ 55, 64, and 70, respectively)), whereas there was no significant reduction of scleral diameter in methoctramine-treated group at 7 dpf stage (381.47 Ϯ 43.74 m, p Ͼ 0.05, n ϭ 56). Another index of eye enlargement is the ratio of retina diameter to sclera diameter (the smaller the ratio, the larger the eye). Fig. 12D showed that there was a significant decrease in the ratio of the RPE/scleral coat during ocular enlargement caused by the knockdown of zlum ((97.22 Ϯ 3.86% (WT) versus 56.11 Ϯ 4.64% (MO), p Ͻ 0.001)). The administration of muscarinic receptor antagonists, 0.5% atropine and 0.25% pirenzepine, effectively attenuate the decrease in the ratio of the RPE/ scleral coat caused by zlum knockdown group ((71.29 Ϯ 8.52% (atropine-treated) and 71.33 Ϯ 9.37% (pirenzepine-treated) versus 56.11 Ϯ 4.64% (MO), both have p values Ͻ 0.001, respectively)), whereas there was no obvious reversion of the decreased ratio of the RPE/scleral coat in the 0.01% methoctra-

DISCUSSION
In this study, we have demonstrated that zebrafish lumican has all the structural features of SLRPs, i.e. a central domain of nine leucine-rich repeats flanked by N-and C-terminal domains with conserved cysteine residues (45). Interestingly, zlum does not have a TATA box in its promoter. Unlike the TATA-less promoters of housekeeping genes, the zebrafish lumican promoter does not have a GC-rich sequence. As shown in Fig. 7, our results implied that zLum exists as a KSPG in corneal stroma but as unglycanated glycoprotein in sclera.
The sclera, a fibrous extracellular matrix, contains irregularly arranged lamellae of collagen fibrils, proteoglycans, elastin, and matrix-secreting fibroblasts (46). The changes of collagen matrix in the sclera caused by increased or decreased proteoglycans may lead to changes in the biomechanical properties of the sclera and may ultimately lead to  A, top is adjacent to the retina. Two to three layers of scleral fibroblastic cells with collagen fibril formation between the layers was found at the posterior sclera of the WT fish at 7 dpf stage. B, top is adjacent to the retina. Only one to two layers of fibroblastic cells at the posterior sclera of the zlum-MO-injected fish at 7 dpf stage is shown. C, scleral thinning was observed obviously in the zlum-MOinjected fish at 7 dpf stage. The phenomenon was much more prominent in the zlum-MO-injected fish at 12 dpf stage. In particular, significant scleral thinning was observed in the posterior sclera of the zlum-MO-injected fish at 7 and 12 dpf stage as compared with wild type group. (Scale bar, A and B, 1.5 m.) the alteration of the eyeball shape seen in myopic patients (46 -48). Therefore, the sclera is a dynamic tissue rather than a static container of the eye. Moring et al. (49) found that biglycan and lumican mRNA levels were lowered in the sclera during experimentally induced myopia and increased during recovery in the tree shrew model. Lum Ϫ/Ϫ mice and Lum Ϫ/Ϫ Fmod Ϫ/Ϫ mice presented altered collagen fibril diameter in sclera (28,29), suggesting that these proteoglycans play an important role in the biomechanical properties of the sclera. Our zebrafish model indicated that the altered collagen fibril diameter in corneal stroma and anterior sclera by zlum knockdown with morpholinos is similar to some features of the Lum Ϫ/Ϫ mice (Fig. 10). Although the collagen fibril diameter was altered in lumican-null mice, there was no ocular enlargement found in Lum Ϫ/Ϫ mice. Young and co-workers (50) proposed that the phenotypes of Lum Ϫ/Ϫ mice and Lum Ϫ/Ϫ Fmod Ϫ/Ϫ mice may be the possibility of false positive results due to the "hitchhiker" effect (51). However, ocular enlargement and scleral thinning resembling characteristics of high myopia could be readily recognized in our zebrafish model of decreased synthesis of lumican caused by zlum knockdown with morpholinos. This differ-ence can be explained in part by the fact that in mice there are more layers of lamellar organization in anterior and posterior sclera tissue than that of zebrafish, and thus mouse sclera may be more resilient to the loss of lumican and maintain a stable sclera structure seen in Lum Ϫ/Ϫ mice.
The changes in refractive status in developing myopia are greatest from the elongated axial length of the eyeball rather than from the changes of corneal and anterior lens curvature. In humans, thinning of the scleral tissue, particularly at the posterior pole following the elongation of eyeball, is a characteristic of high myopia, which is compatible with our observations in this zebrafish model in that the most immature collagen fibrils are located in the posterior sclera (46). Therefore, a decrease and/or disruptions of the arrangement of collagen fibrils in this region are highly probable to weaken sclera strength and integrity and result in ocular enlargement seen in our zebrafish model.
In this study, we also found that the number of scleral fibroblasts are significantly reduced at the posterior sclera of zlum-MO injected fish (Fig. 10B). Thus, the lumican protein may also affect the functions of scleral fibroblasts during early development. Kao et al. (24,52) has suggested that lumican is a FIGURE 11. Zebrafish larvae assay for drug screen. A, normal phenotype of WT fish at 7 dpf stage. B, normal phenotype of RS-MO-injected embryos at 7 dpf stage. C, significantly enlarged eyeball of zlum-MO-injected fish at 7 dpf stage. D, significant decrease in ocular enlargement was noted in the zlum-MO-injected larvae at 7 dpf stage after treating with 0.5% atropine for 2 days. E, decrease in ocular enlargement was also found in the zlum-MO-injected larvae at 7 dpf stage after treating with 0.25% pirenzepine (P) for 2 days. F, no obvious changes in the phenotypes of zlum-MO-injected fish at 7 dpf stage after treating with 0.01% methoctramine (M). matrikine and can modulate fibroblast activities in addition to serving as a regulator of collagen fibrillogenesis.
In this study, we have also established a brand new platform of using the zlum-MO zebrafish as a screen tool for identifying drugs that can affect myopic progression. Recently, Barathi et al. (53) confirmed the presence of all five (M1 to M5) muscarinic acetylcholine receptors in human and mouse scleral fibroblasts. Muscarinic acetylcholine receptors have also been found in the brain of the zebrafish by radioligand binding techniques (37). Liu et al. (54) demonstrated that muscarinic antagonists acted directly on the sclera to prevent myopia development induced by form deprivation in guinea pigs in which myopia progression was accompanied by overexpression of muscarinic acetylcholine receptors (M1 and M4). Furthermore, atropine, a nonselective muscarinic antagonist, and pirenzepine, an antag-onist specific for M1, have been found to be effective in clinical trials to prevent myopia progression (55)(56)(57)(58). In experimentally induced myopia, including chicks and tree shrews, several muscarinic antagonists have been applied to control ocular growth in different species, although the mechanism of drug effects against myopia remain unclear (59 -61). However, variable effects on experimental myopia were shown by applying these muscarinic antagonists when administered intravitreously or by subconjunctival injection (36,59). In this study, we chose these muscarinic antagonists, i.e. 0.5% atropine, 0.25% pirenzepine, and 0.01% methoctramine, according to a previous study with minor modifications (36). In our results, we showed that the effect of ocular enlargement due to the reduction of zLum protein was blocked by atropine and pirenzepine but not by the M2-selective receptor antagonist methoctramine. Therefore, the observations are similar to the effects of muscarinic receptor antagonists in other myopic animal models and clinical trials. With our model, one can quickly screen drugs that may have therapeutic effects on controlling myopia progression within a week. Another advantage of our model is it can be easily and conveniently carried out by adding the drugs to the embryos kept in 96-well tissue culture plates without any injection procedures to the experimental animals.
In conclusion, we have identified and characterized the zebrafish lumican gene. Lumican is highly conserved during evolution. zlum shares high homology with human and mouse lumican with respect to gene structure, expression pattern, and promoter function. Moreover, we present evidence of ultrastructural changes of ocular enlargement, which are similar to the features of high myopia in humans. We have also established a protocol of drug screening tests for the identification of compounds that are of potential in modulating scleral growth. Therefore, the zebrafish can serve as a potential vertebrate model for studying the early development of corneal and scleral diseases.