|Home | About | Journals | Submit | Contact Us | Français|
It is generally accepted that the glycans on the cell surface and extracellular matrix proteins play a pivotal role in the events that mediate re-epithelialization of wounds. Yet, the global alteration in the structure and composition of glycans, specifically occuring during corneal wound closure remains unknown. In this study, GLYCOv2 glycogene microarray technology was used for the first time to identify the differentially expressed glycosylation-related genes in healing mouse corneas. Of ~2000 glycogenes on the array, the expression of 11 glycosytransferase and glycosidase enzymes was upregulated and that of 19 was downregulated more than 1.5-fold in healing corneas compared with the normal, uninjured corneas. Among them, notably, glycosyltransferases, β3GalT5, T-synthase, and GnTIVb, were all found to be induced in the corneas in response to injury, whereas, GnTIII and many sialyltransferases were downregulated. Interestingly, it appears that the glycan structures on glycoproteins and glycolipids, expressed in healing corneas as a result of differential regulation of these glycosyltransferases, may serve as specific counter-receptors for galectin-3, a carbohydrate-binding protein, known to play a key role in re-epithelialization of corneal wounds. Additionally, many glycogenes including a proteoglycan, glypican-3, cell adhesion proteins dectin-1 and -2, and mincle, and mucin 1 were identified for the first time to be differentially regulated during corneal wound healing. Results of glycogene microarray data were confirmed by qRT-PCR and lectin blot analyses. The differentially expressed glycogenes identified in the present study have not previously been investigated in the context of wound healing and represent novel factors for investigating the role of carbohydrate-mediated recognition in corneal wound healing.
Wound re-epithelialization in many organs including cornea, skin, and mucous membranes occurs by the coordinated migration of adjacent epithelial cells over the wound surface (Singer and Clark 1999; Lu et al. 2001). Defects in the re-epithelialization process, including impaired migration or failure of migrated epithelium to remain attached to the substratum, cause serious clinical conditions such as persistent epithelial defects and ulceration (Singer and Clark 1999; Ma and Dohlman 2002). It is generally accepted that glycans on the plasma membrane of the corneal epithelium play a pivotal role in the events that modulate re-epithelialization (Gipson and Anderson 1980; Gipson et al. 1984; Panjwani et al. 1990; Panjwani, Zhao, et al. 1995; Yang et al. 1996; Cao et al. 2001). It has long been demonstrated that increased amounts of plant lectins, such as concanavalin A (ConA) and wheat germ agglutinin (WGA), bind to migrating epithelium as compared to normal nonmigrating epithelium (Gipson et al. 1983), and a number of glycoproteins and glycolipids are differentially expressed in migrating as compared to normal corneal epithelium (Panjwani et al. 1990; Panjwani, Ahmad, et al. 1995; Panjwani, Zhao, et al. 1995; Yang et al. 1996; Saika et al. 2000). While the studies conducted thus far using plant lectins and glycan-specific monoclonal antibodies have recognized that there are substantial changes in specific glycan structures during re-epithelialization of corneal wounds, the global alteration in the structure and composition of glycans during corneal wound closure remains unknown.
Among the many factors that regulate the glycosylation pattern of cells are glycosyltransferases (Ohtsubo and Marth 2006). Alteration in the expression pattern of genes encoding glycosyltransferases has been shown to have enormous impact on cell behavior, morphology, and functions (Ohtsubo and Marth 2006). For example, extensive studies aimed at characterization of the role of glycosyltransferases in cancer cell migration have demonstrated that changes in expression of glycosyltransferases alter the glycan profiles and hence the functions of many glycoproteins (reviewed in Gorelik et al. 2001; Lau and Dennis 2008). Thus far, only a few studies have focused on the role of glycosyltransferases in the repair of wounds. For example, β1,4-galactosyltransferase-1, which synthesizes type 2 chain (Galβ1,4GlcNAc) on N-glycans and the core 2 branch in O-glycans, has been shown to participate in skin wound healing by recruiting leukocytes, and by promoting angiogenesis and collagen deposition at the sites of wound (Mori et al. 2004; Shen et al. 2008). To date, the expression profile of glycosyltransferases during re-epithelialization of corneal wounds has not been determined. In the present study, using a glycogene microarray approach that detects the transcript levels of enzymes regulating glycosylation, we report that many glycosyltransferases are differentially regulated during re-epithelialization of corneal wounds after transepithelial excimer laser keratectomy. Of particular interest is our finding that among the differentially expressed glycosyltransferases are enzymes such as β1,3-galactosyltransferase 5 (β3GalT5), mannoside acetylglucosaminyltransferase IVb (GnTIVb), T-synthase, and sialyltransferases, whose differential expression is likely to result in the expression of glycans on glycoproteins and glycolipids that serve as high-affinity ligands for galectin-3, a β-galactoside-binding protein, that is known to promote re-epithelialization of corneal wounds (Cao, Said, et al. 2002).
The total RNA extracted from the normal and healing corneas were run on an Agilent 2100 Bioanalyzer for assessment of quality and purity. The average yield of total RNA was 12.5 ± 1.0 μg and 8.8 ± 0.8 μg per 10 normal and healing corneas, respectively. The ribosomal RNA 28S/18S ratio ranged between 1.3 and 1.7 (supplementary Figure 1A). Supplementary Figure 1B shows representative electropherograms and gel-like images, wherein the gel bands and graph peaks correspond to 18S and 28S dominant components with insignificant amount of low-molecular-weight RNA. This suggests that RNA degradation in the samples used was minimal and the quality of RNA preparations was satisfactory for the microarray analysis.
Glycov2 array was hybridized with biotinylated cRNA probes prepared from total RNA from normal and healing corneas, and the RMA algorithm was used to obtain the expression signal values. Similarity in overall gene expression profiles between individual samples was assessed by an unsupervised hierarchial clustering. As shown in Figure Figure1,1, all normal corneal samples clustered together on one side of the dendrogram, whereas all three samples of healing corneas clustered on the other side of the dendrogram. This implies that the glycogene expression pattern in healing corneas was distinct from that of the normal corneas.
The expression of 36 mouse-specific genes was upregulated and that of 39 genes was downregulated more than 1.5-fold in healing compared with normal, uninjured corneas with parametric P-value of <0.01. The differentially expressed genes were visualized as a heat map using Cluster and Tree View software (Figure (Figure2A).2A). The color-based view also demonstrates that the healing corneas show a distinct gene expression profile from that of the normal, uninjured corneas. For clarity, these genes are grouped according to their involvement in specific cellular processes or functions as determined using David software, Entrez gene, and Gene Cards (Figure (Figure2B2B and supplementary Table 1). Of the 75 differentially expressed genes, 30 genes (40%) are glycosyltransferases and glycosidases, 36% fall under the growth factors and their receptors category, and 9% are cell adhesion proteins. Of the remaining 15% of differentially expressed genes, 4% were core proteins of proteoglycans and the rest (11%) are known to participate in various biological functions and were categorized as miscellaneous molecules (Figure 3B and supplementary Table 1). IL-1β, transforming growth factor-β1 (TGF-β1), amphiregulin, and intercellular adhesion molecule 1 (ICAM1), which are known to be upregulated in healing corneas (Planck et al. 1997; Sotozono et al. 1997; Chen et al. 2000; Zieske et al. 2000; Cao, Wu, et al. 2002), were all found to be induced in response to injury in the current study, thereby attesting to the validity of the analysis. The expression of many glycogenes including core proteins of proteoglycans, serglycin, glypican-3, cell adhesion proteins mincle, dectin-1 and -2, and MUC1 were identified for the first time to be differentially expressed in healing corneas (supplementary Table 1).
The differentially expressed glycogenes that regulate glycosylation were further categorized according to their specific function and are listed in Table TableI.I. Among the glycosyltransferase and glycosidase enzymes, the expression of 11 glycogenes was upregulated, whereas that of 19 was downregulated in healing compared to normal corneas. In an earlier study, we have established that a carbohydrate-binding protein, galectin-3, but not galectin-1, plays a role in re-epithelialization of corneal wounds (Cao, Said, et al. 2002). It was, therefore, of interest to assess whether injury-specific glycosyltransferases identified in the current study have the potential to synthesize glycans that serve as ligands of galectin-3. In this respect, most notably, the expression of β3GalT5, GnTIVb, and T-synthase was upregulated. Interestingly, these glycosyltransferases have the potential to synthesize high-affinity ligands for galectin-3 (Yoshida et al. 1998; Glinsky et al. 2000; Zhou et al. 2000; Salvini et al. 2001; Khaldoyanidi et al. 2003; Holgersson and Lofling 2006). In contrast, the expression of GnTIII and ST6GalI sialyltransferases, which synthesize glycans that block galectin-3 binding to its ligands (Patnaik et al. 2006; Zhuo et al. 2008), was downregulated in healing corneas. Based on these data, it appears that the differential expression of the glycosyltransferases in healing corneas leads to the upregulation of the high-affinity glycans for galectin-3 on glycoproteins and glycolipids.
Gene-specific qRT-PCR was performed to confirm the differential expression of 11 selected genes. Of these 11 genes, we chose 7 glycosyltransferases because they are likely to have an impact on the expression of glycan ligands of galectin-3, a lectin known to play a role in corneal wound healing (Cao, Said, et al. 2002). The expression level of a housekeeping gene, RPL19, that was used as a reference gene in the current study, was similar in both normal and healing corneas (fold change: 1.06). For the most part, the qRT-PCR data were in agreement with those obtained by array hybridization. For genes measured by qRT-PCR, the mRNA levels were shown to change in the same direction as the changes measured by the microarray detection for 10 of the 11 genes tested (~91%) (Table (TableII).II). Consistent with the microarry data, we measured significant upregulation of GnTIVb, β3GalT5, T-synthase, ST3GalI (representative gene from the differentially expressed ST3 sialyltransferases), MUC1, and IL-1β, and significant downregulation of GnTIII, ST6GalI (representative gene from the differentially expressed ST6 sialyltransferases), ST8SiaIV (representative gene from the differentially expressed ST8 sialyltransferases), and glypican-3, in healing corneas by qRT-PCR. While modest downregulation of the lumican gene was detected in healing corneas by microarray analyses (1.6-fold), qRT-PCR analysis contradicted this result (2.9-fold induction), thereby confirming data in an earlier published report (Saika et al. 2000).
We used lectin blot analysis to determine whether differential expression in transcript levels lead to corresponding alterations in the glycan products in healing corneas. Among various differentially expressed glycosyltranferases identified in the current study (Table (TableI),I), we randomly chose three glycosyltransferases, two genes (GnTIVb and ST3GalI) representing the upregulated group, and one gene (GnTIII) representing the downregulated group, as examples to assess the potential of differentially expressed transcripts to modulate the expression of corresponding glycans. Plant lectins, E-PHA, DSL, and MAA, have been widely used to analyze the glycan products of GnTIII (bisecting GlcNAc-branched N-glycans), GnTIV (β1,4-GlcNAc branched N-glycans), and ST3 sialyltransferases (α2,3-sialic acids), respectively (Wang and Cummings 1988; Ohtsubo et al. 2005; Patnaik et al. 2006; Abbott et al. 2008). Lectin blot analyses using these plant lectins showed that the expression level of DSL- and MAA-reactive glycoproteins is upregulated, whereas that of E-PHA-reactive glycoproteins is downregulated in healing corneas compared to the normal controls (Figure (Figure3).3). These data are consistent with the qRT-PCR data (Table (TableII)II) showing that the mRNA levels of the corresponding glycosyltransferases, GnTIVb, and ST3GalI are increased, but that of GnTIII is decreased in healing corneas.
The overall goal of the present study was to identify differentially expressed glycosyltransferases in healing corneas that may enable us to predict wound healing-specific glycan structures and identify novel genes that are differentially expressed in response to injury. A comparison of gene expression profiles of normal and healing corneas revealed that healing corneas have a unique glycogene expression pattern that defines them as a group distinct from the normal, unwounded corneas (Figures 2 and and3).3). In general, the structural variations in glycans on glycoproteins could alter their interaction with the endogenous carbohydrate-binding proteins and this, in turn, could modulate the role of the lectins in cell adhesion and signal transduction events required for cell motility (Ohtsubo and Marth 2006), a key event in wound healing. In an early study, we have established that a carbohydrate-binding protein, galectin-3, plays a role in re-epithelialization of corneal wounds (Cao, Said, et al. 2002). Specifically, we have demonstrated that (i) among the members of the galectin family, galectin-3 is the most abundantly expressed lectin in corneal epithelium (experiment ID: MAEXP_248_072104; http://www.functionalglycomics.org/ glycomics/publicdata/microarray.jsp), (ii) migrating epithelium of healing mouse corneas express elevated levels of galectin-3 protein compared to nonmigrating epithelium of normal corneas, (iii) exogenous galectin-3 stimulates re-epithelialization of corneal wounds, and (iv) the rate of re-epithelialization of corneal wounds is significantly slower in galectin-3-deficient mice compared to the wild-type mice. We have also demonstrated that while galectin-1 is expressed in healing corneal stroma, it is not expressed in corneal epithelium, and there is no difference in corneal epithelial wound closure rates between galectin-1-deficient and wild-type mice (Cao, Said, et al. 2002). Also, unlike exogenous galectin-3, exogenous galectin-1 does not stimulate re-epithelialization of corneal wounds. Thus, our specific goal was to identify injury-specific glycosyltransferases, which are likely to synthesize glycans that serve as ligands of galectin-3.
A major finding of the present study is that the expression of Mgat3 that encodes N-acetylglucosaminyltransferase-III (GnTIII) was downregulated. GnTIII introduces a bisecting β1,4GlcNAc to the β4-linked core mannose residues of N-glycans (Figure (Figure4A)4A) (Ohtsubo and Marth 2006), and a number of studies have reported that the presence of bisecting GlcNAc sugars on N-glycans is accompanied with a reciprocal change in the expression of β1,6GlcNAc-branched N-glycans, which are synthesized by N-acetylglucosaminyltransferase-V (GnTV) (Isaji et al. 2004; Zhao et al. 2006; Kariya et al. 2008; Pinho et al. 2009). It is thought that GnTIII dominantly competes with GnTV for the modification of the same protein, and, therefore, leads to the inhibition of the function of GnTV (Isaji et al. 2004; Zhao et al. 2006). These findings in conjunction with the studies reporting that β1,6GlcNAc-branched N-glycans serve as substrates for the synthesis of poly N-lactosamines, the high-affinity counter-receptors for galectin-3 (Demetriou et al. 2001; Partridge et al. 2004; Lagana et al. 2006), lead us to propose that downregulation of Mgat3 gene expression in healing corneas observed in the current study may result in the increased addition of poly N-lactosamine residues on N-glycans to synthesize high-affinity counter-receptors for galectin-3 that, as described above, plays a role in re-epithelialization of corneal wounds. Indeed, Patnaik et al. (2006) have demonstrated that the binding of galectin-3 to Lec10 CHO cells, which overexpresses GnTIII, was reduced compared to that of wild-type CHO cells (Patnaik et al. 2006).
Another key finding of the current study is that the genes encoding the glycosyltransferases, GnTIVb, β1,3GalT5, and T-synthase were all upregulated in healing corneas. The GnTIVb enzyme generates distinct β1,4GlcNAc branches on α3-linked core mannose of N-glycans (Figure (Figure4A)4A) that also serve as a substrate for polylactosamine glycans (Yoshida et al. 1998). The glycosyltransferase, β1,3GalT5 generates type I lactosamine chain (Galβ1,3GlcNAc) elongation of core 2 and core 3 O-glycans, N-glycans (Salvini et al. 2001; Holgersson and Lofling 2006), lactoceramides (Amado et al. 1999), and globosides (Zhou et al. 2000). Since galectin-3 binds to glycans containing Galβ1,3GlcNAc disaccharides (Sato and Hughes 1992; Hirabayashi et al. 2002; Brewer 2004), the upregulated expression of β1,3GalT5 emphasizes that increased levels of galectin-3 ligands may be synthesized during healing of corneal wounds. T-Synthase generates the Thomsen–Friedenreich (TF) antigen (Galβ1,3GalNAc, i.e., core 1 dissacharide) and initiates the synthesis of core 1-derived O-glycans (Figure (Figure4B)4B) (Van den Steen et al. 1998). The core 1 dissacharide is a substrate for a number of glycosyltransferases including ST6GalNAcI that synthesize the sialylated TF antigen (Takashima 2008). However, the expression level of ST6GalNAcI was downregulated in healing corneas (see below). Clearly, the enhanced expression of T-synthase coupled with downregulated ST6GalNAcI sialyltransferase (discussed below) may produce more unsubstituted TF antigen in healing corneas. The unsubsituted TF antigen is also a known ligand for galectin-3, and it has been shown that the interaction between galectin-3 and unsubstituted TF antigen on the cell surface mediates the homotypic aggregation of breast carcinoma cells and adhesion of breast carcinoma cells to endothelium for metastasis (Glinsky et al. 2003; Khaldoyanidi et al. 2003). Thus, the increased unsubstituted TF antigen in healing corneas may facilitate galectin-3-mediated cell–cell and cell–matrix interactions. In this respect, it is noteworthy that in the current study, the expression of MUC1, a transmembrane O-glycosylated protein and a major carrier for the TF antigen (Yu et al. 2007), was upregulated in healing corneas. In epithelial cancer cells, elevated expression of the MUC1 protein is associated with an increased expression of the unsubstituted TF antigen, and it has been shown that galectin-3 interactions with MUC1 through the unsubstituted TF antigen facilitates the adhesion of carcinoma cells to endothelium (McGuckin et al. 1995; Yu et al. 2007).
We found that sialyltransferases, which add sialic acids to terminal sugars either by α2,6- and α2,8-linkages, including ST6GalI (N-acetyllactosaminide α2,6 sialyltransferase), ST6GalNAcI (sialyl-Tn, and mono- and di-sialyl T-synthase), ST8SiaIV (polysialyltransferase), and ST8SiaVI (sialylates O-glycans) (Takashima 2008), are largely downregulated in healing corneas. Recent studies have demonstrated that sialylation distinctively modulates the recognition of cell surface glycans and biological signaling by different galectins (Amano et al. 2003; Toscano et al. 2007; Stowell et al. 2008a; Zhuo et al. 2008). Of particular relevance to the current study are the reports demonstrating that the presence of α2,6-linked sialic acids in the glycans substantially reduce their ability to interact with galectin-3 (Hirabayashi et al. 2002; Brewer 2004) and that ST6GalI-mediated α2,6-sialylation of β1 integrins prevents galectin-3-induced apoptosis of colon tumor cells (Zhuo et al. 2008). In a different study, it was shown that the removal of α2,8-sialic acids by neuraminidase primes the neutrophils for stimulation by galectin-3 (Almkvist et al. 2004). Collectively, these observations lead us to propose that downregulation of ST6GalI, ST6GalNAcI, ST8SiaIV, and ST8SiaVI sialyltransferases that synthesize α2,6- and α2,8-sialic acids may assist in the enhanced interaction of galectin-3 with its counter-receptors during re-epithelialization of corneal wounds. It is noteworthy that the expression of ST3 sialyltransferases (ST3GalI and ST3GalIV) that synthesize α2,3-linked sialic acids was upregulated in healing corneas (Takashima 2008). Coincidentally, it has been demonstrated that while galectin-3 does not prefer α2,6-sialylated glycans, it binds well to α2,3-sialylated glycans (Hirabayashi et al. 2002; Brewer 2004).
It is known that different members of the galectin family have overlapping (Hirabayashi et al. 2002; Maeda et al. 2003; Brewer 2004; Friedrichs et al. 2007; Stowell et al. 2007, Stowell et al. 2008a; Paclik et al. 2008) as well as unique (Fukumori et al. 2003; Sturm et al. 2004; Hernandez et al. 2006; Stillman et al. 2006; Lu et al. 2007; Stowell et al. 2008a, 2008b; Diskin et al. 2009) saccharide binding specificity. It is therefore, reasonable to expect that injury-specific glycosyltransferases discussed above are also likely to influence, at least to a degree, the expression of counter-receptors of other members of galectin family besides galectin-3. Another galectin that has been implicated in the process of corneal wound healing is galectin-7 (Cao, Said, et al. 2002; Cao et al. 2003). Precisely to what extent galectins-3 and -7 share counter-receptors remains to be determined. Our early study demonstrating that galectin-7, but not galectin-3, accelerated re-epithelialization of wounds in Gal3−/− corneas (Cao, Said, et al. 2002) in conjunction with the findings that unlike galectin-3, galectin-7 does not show affinity to the unsubstituted TF antigen (http://www.functionalglycomics.org/ glycomics/HFileServlet?operation=downloadRawFilefileType =DATsideMenu=noobjId=1000348) may suggest that the two lectins may bind to distinct cell surface and/or ECM counter-receptors. However, based on the glycan array, thermodynamic binding (Brewer 2004), and frontal affinity chromatographic (Hirabayashi et al. 2002) studies showing that like galectin-3, galectin-7 also binds to Galβ1, 3GlcNAc, internal lactosamine, α2,3-sialylated glycans but not α2,6-sialylated glycans, it is reasonable to expect that the two lectins share at least some counter-receptors.
Although, the focus of the current study was glycosyltransferases, we report here for the first time that a number of other genes encoding core proteins of proteoglycans and cell adhesion proteins were also differentially expressed in healing corneas. These genes include mincle, dectin-1, dectin-2, glypican-3, and serglycin (supplementary Table 1). These differentially expressed genes have not previously been investigated in the context of wound healing and represent the novel factors for the further study of the mechanism of wound healing.
The current functional genomics approach is the first study aimed at identifying alterations in the expression of enzymes regulating glycosylation during corneal wound healing. In summary, using glycogene microarray technology, we demonstrated that corneal wound healing response is characterized by the differential expression of a number of glycosyltransferases. The glycans produced by these differentially regulated enzymes in healing corneas are likely to render the glycoproteins and glycolipids as high-affinity counter-receptors for galectin-3, a carbohydrate binding protein, known to play a key role in re-epithelialization of corneal wounds.
Three groups (10 animals/group) of 6- to 8-week old mice (C57BL/6 and 129 mixed genetic backgrounds) were used. All procedures were approved by the Institutional Animal Care and Use Committee (IACUC) of Tufts University and were performed in accordance with the Association for Research in Vision and Ophthalmology Resolution on the Use of Animals in Vision Research and the recommendations of the National Institutes of Health Guide for the Care and Use of Laboratory Animals. The mice were anesthetized by an intraperitoneal injection of 1.25% avertin (0.2 mL/10 g body weight) (Aldrich Chemical Co., Milwaukee, WI) and proparacaine eye drops (Alcain, Alcon Labs, Inc., Fort Worth, TX) were applied to the cornea as a topical anesthetic. Transepithelial excimer laser ablations (2 mm optical zone; 42 to 44 micron ablation depth, phototherapeutic keratectomy mode) were performed on the right eye of each animal using Summit Apex Plus Excimer Laser (Waltham, MA). The left eye of each animal was used as normal, uninjured cornea. After surgery, all animals received buprenorphine (intramuscular, 0.2 mL of 0.3 mg/mL Buprenex, Reckitt and Colman Pharmaceuticals, Inc., Richmond, VA) as a painkiller. Antibiotic ointment (Vetropolycin, Pharmaderm, Melville, NY) was applied, and the corneas were allowed to partially heal in vivo for 18–22 h. At the end of the healing period, animals were euthanized, and the corneas of both eyes were excised and immediately placed in liquid nitrogen until use.
Total RNA from corneas was extracted using RNeasy mini kit according to the manufacturer's recommendations (Qiagen, Chatsworth, CA). Briefly, the frozen corneas were crushed, suspended in the guanidine thoiocyanate buffer (Buffer RLT), and further homogenized using a QIAshredder column. The eluent was loaded onto a silica gel base column and the bound RNA was eluted with RNase-free water. The yield and quality of RNA were analyzed using the Bioanalyzer 2100 with RNA Pico Lab Chips (Agilent, Waldbronn, Germany).
The glycogene microarray, GLYCOv2, is an oligonucleotide microarray, custom designed by Affymetrix (Santa Clara, CA) for the Consortium for Functional Glycomics at the Scripps Institute, La Jolla, CA. The array contains approximately 2000 mouse and human glycogenes including glycosyltransferases, glycosidases, enzymes involved in nucleotide-sugar synthesis and transport, proteoglycans, and glycan binding proteins. A complete list of probe sets and annotation for the GLYCOv2 oligonucleotide array is available at (http:// www.functionalglycomics.org/static/consortium/resources/ resourcecoree.shtml). In this array, transcripts are detected by three identical probe sets for each glycogene. Each probe set consisted of 11 probe pairs, with each probe pair made of one 25-bp perfect match oligonucleotide that matches the sequence of the targeted transcript, and one 25-bp oligonucleotide designed with a mismatch at the center position.
To prepare hybridization probes, total RNA (100 ηg) from each sample was amplified, and cDNA was synthesized according to a modified Baugh/Harvard protocol (http://www.scripps.edu/ researchservices/dna_array/pdf%20files/Affymetrix%20Small% 20Sample%20Protocol.pdf). The resultant cDNAs were transcribed in vitro in the presence of biotin-labeled ribonucleotides, and the labeled cRNA was hybrized to GLYCOv2 microarrays, and scanned using Affymetrix Scanner 3000 (www.affymetrix.com). Quantitation of expression signal values, quantile normalization, and background subtraction were performed as described earlier (Diskin et al. 2006). The gene expression patterns in the three replicates of each group (healing and normal corneas) were then analyzed by hierarchical clustering using BRB ArrayTools 3.2.2 (http://linus.nci.nih.gov/BRB- ArrayTools.html).
The differential expression of genes in healing compared to normal, uninjured corneas was analyzed as described by Diskin et al. (2006). Briefly, the transformed expression values for the replicated probe sets were averaged to get a single expression value for each probe set on each array. Then, statistically significant changes in gene expression were identified using BRB ArrayTools 3.2.2 software. The class comparison test was conducted using a univariate alpha level cutoff of 0.001 and a multivariate permutation-based false discovery rate calculation. The predicted proportion of false discoveries was preset at 10% and a false discovery rate calculation was set at a confidence level of 80%. The gene microarray data from this study has been deposited in the Consortium of Functional Glycomics database (http://www.functionalglycomics.org/glycomics/publicdata/ microarray.jsp).
The differentially expressed genes were visualized using Cluster and Tree View software for heat map creation (Eisen Lab, UC Berkeley; http://rana.lbl.gov/EisenSoftware.htm). For this, each gene was compared to median unlogged RMA signal intensity values using normal and healing arrays in the study. Differences from the median were represented in varying intensities of green (decreased fold-change compared to median) and red (increased fold-change compared to median).
The Database for Annotation, Visualization and Integrated Discovery (DAVID) software (Diskin et al. 2006), GeneCards (http://www.genecards.org), and Entrez gene (http://www.ncbi.nlm.nih.gov/entrez) were used to gain insight into the biological functions of differentially expressed genes.
Total RNA (300 ηg) was reversed transcribed using the High Capacity kit (Applied Biosystems (ABI), Foster City, CA) according to manufacturer's instructions. Real-time PCR was performed (Mx4000 real-time PCR machine, Stratagene, La Jolla, CA) in triplicates using 5 μL of cDNA (derived from 15 ηg total RNA), TaqMan MGB probes, primers specific for the selected genes, and TaqMan Universal PCR master mix (ABI). Reactions performed in the absence of template served as negative controls. The ABI primer sets used included: ribosomal protein L19 (RPL19) (Mm020601633_g1), mannoside acetylglucosaminyltransferase III (GnTIII) (Mm00447798_m1), GnTIVb (Mm00521482_m1), β3GalT5 (Mm00473621_s1), T-synthase (Mm00473986_m1), β-galactoside α2,6-sialyltransferase I (ST6GalI) (Mm00486119_m1), β-galactoside α2,3-sialyltransferase I (ST3GalI) (Mm00501493_m1), α-N-acetylneuraminide α-2,8-sialyltransferase (ST8SiaIV) (Mm00456300_m1), interleukin-1β (IL-1β) (Mm004324228_m1), lumican (Mm00500510_m1), mucin 1 (MUC1) (Mm00449604_m1), and glypican-3 (Mm00516722_m1). For amplification, Amplitaq Gold DNA polymerase was activated (95°C for 10 min) and the reactions were subjected to 50 cycles involving denaturation (95°C for 15 s) and annealing plus extension (60°C for 1 min). The fluorescent signals were recorded using a FAM detector and data analysis was performed using Mx4000 software version 2 (Stratagene). The FAM fluorescent signals were measured against the ROX (internal reference dye) signal to normalize the non-PCR-related fluctuations; amplification plots showing the increase in FAM fluorescence with each cycle of PCR (ΔRn) were generated for all samples, and the threshold cycle values (Ct) were calculated from the amplification plots. The Ct value represents the cycle number at which the fluorescence was detectable above an arbitrary threshold, based on the variability of the baseline data during the first 15 cycles. All Ct values were obtained in the exponential phase. Quantification data of each gene were normalized to the expression of a housekeeping gene, RPL19. A value of 1.0 was assigned to the expression level of each gene in the normal, uninjured corneas. The values for healing corneas were expressed as a change in expression levels with respect to normal corneas.
Corneal wounds (2 mm) were produced on the right eye of four mice by transepithelial excimer laser ablations. The wounds were allowed to partially heal in vivo for 18–22 h, and the corneas were excised as described above. The corneas from left eyes served as normal controls. Protein extracts (5 μg) of normal and healing corneas prepared in the radioimmunoprecipitation assay (RIPA) buffer (50 mM Tris-HCl (pH 8.0), 150 mM NaCl, 0.1% Nonidet P-40, and 0.5% deoxycholic acid) were electrophoresed on 10% SDS–polyacrylamide gels and transferred to nitrocellulose membranes. The protein blots of the gels were stained with Ponceau S (Sigma, MO) to ensure equal loading of samples and were then probed with various biotinylated plant lectins (1 μg/mL) including Phaseolus vulgaris erythroagglutinin (E-PHA), Datura Stromonium lectin (DSL), and Maakia Amurensis agglutinin (MAA) (Vector Labs, Burlingame, CA). The lectin-reactive components were then visualized using streptavidin-HRP (ABC kit, Vector Labs) and a chemiluminescence detection system (PerkinElmer Life Sciences, Waltham, MA). Films were scanned and densitometric analysis was preformed using ImageJ.
Supplementary data for this article is available online at http://glycob.oxfordjournals.org/.
National Eye Institute (EY007088 to N.P.), New England Corneal Transplant Fund, Mass Lions Eye Research fund, and a challenge grant from Research to Prevent Blindness.
The gene microarray analysis was conducted by the Gene Microarray (E) Core of the Consortium for Functional Glycomics funded by the National Institute of General Medical Sciences grant GM62116.