Isolation and Characterization of Galectins in the Mammalian Retina

Abstract

purpose. Previous studies have suggested that galectins may be involved in retinal adhesion and photoreceptor cell survival. To elucidate the underlying mechanisms, the authors isolated retinal galectins, determined their types and distributions, and investigated the validity of the hypothesis, using rat models.

methods. An antibody was prepared against a bovine retinal lectin that was isolated by use of a lactose-agarose column. cDNA of the lectin was isolated by screening of a bovine retinal cDNA library, using the antibody, and then was sequenced. The cDNAs of rat retinal galectins were also isolated by means of polymerase chain reaction and used to produce an antibody against recombinant galectin-3. Using the described antibodies, the authors examined the distributions of galectins in bovine and rat retinas, morphologic changes of rat retinas induced by the antibodies, and distributional changes of galectins in constant-light–exposed rat retinas.

results. The cDNAs of bovine galectin-1, rat galectin-1, and rat galectin-3 were isolated. Galectin-1 was found in various regions, including the retinal pigment epithelium, outer limiting membrane, and outer plexiform layer in bovine and rat retinas. Galectin-3 was increasingly detected in the cytoplasm of Müller cells after constant light exposure after an increase in its transcript. Retinal detachment and vacuolation of the outer plexiform layer were induced in rat eyes by intravitreous injection of the anti-galectin-1 antibody.

conclusions. Galectin-1 may be involved in adhesion of the photoreceptor and outer plexiform layers by interacting with glycoconjugates withβ -galactoside residues in the interphotoreceptor matrix and synaptic cleft matrix. Galectin-3 may increase in Müller cells of a degenerative rat retina, probably through endogenous anti-apoptosis.

Various kinds of glycoconjugates with β-galactose residues have been detected in the retina, including the interphotoreceptor matrix (IPM), using galactose-binding plant lectins. ^{1 2 3 4 5} Glycoconjugates in the IPM play a number of important roles in the interactions between photoreceptors, retinal pigment epithelium (RPE) cells, and Müller cells, including mediation of intercellular adhesion ^{6 7 8} and transfer of substances between these cells. ^{8 9 10} We have observed that the IPM around photoreceptors is lost when the retinal tissue is incubated with an exogenous lectin in the presence of d-galactose. ¹¹ This finding suggests that galactose residues may be involved in retinal adhesion mediated by the IPM. Galectin is an endogenous animal lectin that exhibits an affinity for β-galactoside sugars and promotes cell–matrix adhesion by cross-linking cell surface and substrate glycoconjugates. ¹² These properties indicate that galectin may be a mediator of retinal adhesion. 

In addition to cell–matrix adhesion, another function of galectin has been reported. Galectin-3 inhibits apoptosis through the bcl-2 ¹³ or cysteine protease pathways. ¹⁴ We recently observed that intravitreous injection of an anti-galectin-3 antibody accelerates the rat photoreceptor degeneration due to constant light, whereas that of galectin-3 inhibits the light damage. ¹⁵ These findings suggest that increased expression of endogenous galectin-3 may represent an anti-apoptotic action in the rat retina after constant light (CL) exposure. 

An endogenous 16-kDa galactose-binding lectin has been isolated from the chicken retina ¹⁶ and shown to be distributed in the apical villi of Müller cells and some other regions of the retina. ¹⁷ Although its molecular size is similar to that of galectin-1, it has not yet been confirmed to be galectin-1 or another type of galectin. In the present study, therefore, we isolated, and determined the types and distributions of galectins in mammalian retinas. Then, we examined whether intravitreous injection of an anti-galectin antibody induces retinal detachment to confirm that galectin is involved in retinal adhesion. We also investigated whether the expression of galectin-3 increases in CL-exposed rat retinas, to elucidate the mechanism underlying the acceleration of photoreceptor degeneration by an intravitreous injection of an anti-galectin-3 antibody. 

Materials and Methods

Purification of Bovine Retinal Lectins and Their Antibodies

Twelve bovine retinas were isolated and homogenized in 50 ml of 1.25% Triton X-100 in 0.01 M Tris-HCl buffer (pH 7.6), containing 0.15 M NaCl, 2 mM 2-mercaptoethanol, 2 mM EDTA, and 1 mM phenylmethylsulfonyl fluoride (PMSF) (extraction buffer), followed by centrifugation at 140,000g for 1 hour. The supernatant (Triton X-100–extracted sample) was decanted and applied to a column of lactose-agarose (column volume, 1 ml; Seikagaku Corp., Tokyo, Japan) equilibrated with 0.1% Triton X-100 in 0.01 M-Tris-HCl buffer (pH 7.6), containing 0.15 M NaCl, 2 mM 2-mercaptoethanol, 2 mM EDTA, and 1 mM PMSF (equilibration buffer) at 4°C. After the column had been washed with 20 ml equilibration buffer, the lactose binding lectins bound to the lactose column were eluted with 4 ml equilibration buffer containing 0.2 M lactose. Approximately 130 μg lectins was purified from the 12 bovine retinas (10.8 mg/retina), as determined with a protein assay (D_C Protein Assay System; Bio-Rad, Hercules, CA). Therefore, to obtain enough of the lectins for the immunization of two rabbits and affinity purification of antibodies, we purified 4.5 mg lectins from 440 bovine retinas by 10 repetitive cycles of the procedure, the column volume being 4 ml each time. The lectins were analyzed by Western blot analysis as described in the section, Protein Electrophoresis and Western Blot Analysis, after incubation with a biotinylated d-galactose probe (Seikagaku Corp.). Rabbit antisera were raised against the lectins by immunizing two New Zealand White rabbits, and then antibodies were affinity purified on columns of the lectins coupled to an NH₂-crosslinked agarose (Affi-Gel 10; Bio-Rad), according to a method previously described. ¹⁸  

Screening of a Bovine Retinal cDNA Expression Library

A cDNA library of bovine retinas (Uni-ZAP XR; Stratagene, La Jolla, CA) was screened with an affinity-purified antibody against bovine retinal lectins. Screening was performed according to the instructions provided by the manufacturer. Seven positive plaques were detected on the screening of approximately 1.0 × 10⁶ phages in the library. The DNAs of the positive clones were sequenced on both strands, using T3 and T7 universal primers (Stratagene), according to a terminator cycle sequencing protocol (Prism Big Dye; PE-Applied Biosystems, Foster City, CA). 

Polymerase Chain Reaction

To determine the kinds of galectins that are expressed in the rat retina, cDNA clones of rat retinal galectins were amplified by means of polymerase chain reaction (PCR) from the rat retinal cDNAs in the pAP3neo vector (mRNA source: pooled eyes from 100 Sprague-Dawley outbred males, aged 10–11 weeks; Takara Shuzo, Shiga, Japan), by using three pairs of oligonucleotides (5′-CGGGATCCTTCGCTTCAATCATGGCCTG-3′ and 5′-GGGCTGGGGCTGGCTGGCTTCACTC-3′ for galectin-1, ¹⁹ 5′-CGGGATCCAGGAAAATGGCAGACGGCTTC-3′ and 5′-GGGGTACCTCATAACACACAGGGCAGTTC-3′ for galectin-3, ²⁰ and 5′-CGAATTCCGACTCTCAAGATGGCCTATTG-3′ and 5′-GATTAGATGGAACTTGGGATCTCTCTGC-3′ for galectin-4 ²¹ ) as PCR primers. PCR was performed with a kit (Advantage cDNA PCR; Clontech Laboratories, Palo Alto, CA) according to the manufacturer’s instructions. After thirty cycles (1 minute at 94°C and 3 minutes at 68°C of amplification), the products were isolated from agarose gels for subcloning into a vector (Bluescript II SK[+]; Stratagene) using PCR DNA (GFX; Amersham Pharmacia Biotech, Piscataway, NJ) and a gel band purification kit (Amersham Pharmacia Biotech). After ligation, the DNA was sequenced on both strands, using T3 and T7 universal primers, according to the manufacturer’s protocol (PE Applied Biosystems). 

Reverse Transcription-Polymerase Chain Reaction

To compare the mRNA-expression of galectin-3 after CL exposure (described in the Immunohistochemistry section), RT-PCR was performed with β-actin as an internal control. The cDNA for rat galectin-3 was amplified by RT-PCR using the same oligonucleotides as described under PCR, whereas that for β-actin was amplified with a pair of oligonucleotides, 5′-TTGTAACCAACTGGGACGATATGG-3′ and 5′-GATCTTGATCTTCATGGTGCTAGG-3′ (Clontech Laboratories). Because this pair of β-actin primers spans an intron, it seems to be a control for genomic contamination as well. Total RNA (1 μg/sample for each set of CL exposure conditions) was reverse transcribed using oligo dT primers (Boehringer–Mannheim, Mannheim, Germany) and reverse transcriptase (Superscript II; Life Technologies, Rockville, MD) according to the manufacturer’s instructions. An aliquot of the same RT-product from each sample (1/20 of the total volume) was used for PCR using the cDNA PCR kit (Advantage; Clontech Laboratories) following the method used for PCR. Five microliters of each PCR-product was electrophoresed on a 1% agarose gel. Photographs were taken under ultraviolet light, and the intensities of the PCR bands were determined by measurement with image analysis software (NIH Image; provided in the public domain by the National Institutes of Health [NIH], Bethesda, MD, and available at http://www.nb.nih.ncbi.gov). 

Purification of an Anti-Rat Galectin-3 Antibody

The cDNA encoding rat retinal galectin-3 was cloned into plasmid vectors (pMALc2; New England Biolabs, Beverly, MA, and pGEX-4T1; Amersham Pharmacia Biotech) to express rat galectin-3 as fusion proteins with maltose-binding protein (MBP) and glutathione S-transferase (GST) in Escherichia coli BL21 cells, respectively. The fusion proteins were induced and affinity-purified by methods previously described. ^{15 22} Rabbit antiserum was raised against the MBP/galectin-3 fusion protein in a New Zealand White rabbit, and the antibody against galectin-3 was affinity purified on a column of the GST/galectin-3 fusion protein coupled to Affi-Gel 10 (Bio-Rad), by using a method previously described. ^{15 18 22}  

Protein Electrophoresis and Western Blot Analysis

Protein electrophoresis and Western blot analysis were performed essentially as previously described, ^{18 22} using Laemmli’s sample buffer. ²³ The blotted lectins eluted from a lactose-agarose column were incubated with a biotinylated β-galactose probe (20 μg/ml PBS; Seikagaku Corp.) for 1 hour at room temperature. The blotted proteins (10 μg/10 μl each) of bovine and rat retinas (with or without CL exposure) were incubated with the anti-bovine retinal lectin or anti-rat galectin-3 antibody (all the antibodies were 1:50 diluted with PBS) for 1 hour at room temperature. The bands of galectins were then detected with biotinylated anti-rabbit IgG (Vector Laboratories, Burlingame, CA; 1:100 diluted with PBS) and an ABC kit (Vectastain Elite; Vector Laboratories), by a method previously described. ^{15 22} Color slides of the blotted proteins of rat retinas were obtained and transferred to a computer by a scanner (Coolscan; Nikon, Tokyo, Japan). The intensities of the protein bands were determined by measurement with NIH Image on the computer screen and compared. 

Immunohistochemistry

According to the procedures described in previous papers, ^{18 22} tissue sections of bovine and rat eyes for immunohistochemistry were prepared and reacted with anti-bovine retinal lectin (galectin-1; described in the Results section) or anti-rat galectin-3 antibodies (1:50 diluted with PBS) for 1 hour at room temperature. In rats, we also examined the distribution of galectins in the retina after CL exposure at an illuminance of 130 to 150 foot-candles for 1 day, 3 days, 1 week, or 2 weeks (at least four rats for each experiment). The antibody-binding sites were visualized with biotinylated anti-rabbit IgG (Vector Laboratories) and an ABC kit (Vectastain Elite; Vector Laboratories), according to a previously described method. ²²  

Intravitreous Injection of the Antibodies

Albino Wistar rats, 2 to 3 months of age, were anesthetized and 2 μl anti-bovine retinal lectin antibody (galectin-1; 1:50 diluted with PBS; five rats; described in the Results section), anti-rat galectin-3 antibody (1:50 diluted with PBS; five rats), or PBS (five rats) was injected into the vitreous by inserting a 32-gauge beveled needle for 2 mm through the inferior temporal equator of the left eye at an angle of approximately 45°. After the rats had been maintained under cyclic lighting (12-hours on–12-hours off) conditions for 3 days after the injection, they were killed with an overdose of carbon dioxide, after which their eyes were enucleated and immersed in a fixative comprising 2.5% glutaraldehyde and 2% paraformaldehyde. The eyes were bisected along the vertical meridian and embedded in paraffin, by using a procedure previously described. ^{15 18 22} The nasal halves were sectioned at 5μ m thickness and stained with hematoxylin-eosin. Sections along the vertical meridian of the eyes at approximately 20 μm from the bisected edge containing the entire retina were examined (one section per eye). Photographs were made of each section and printed. The horizontal lengths of the RPE, with or without retinal detachment, were determined by measurement with a flexible rule. The degree of retinal detachment was then determined by calculating the ratio of RPE length to retinal detachment/total RPE-length. 

All animal procedures conformed to the Guidelines of the Kagoshima University Faculty of Medicine for Animal Experiments and the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Results

Western Blot Analysis of Retinal Lectins

The lactose-binding lectins of bovine retinas, bound to a lactose-agarose column, were detected with a biotinylatedβ -galactose probe as two protein bands comprising an intense band corresponding to M_r 15 kDa and a faint band corresponding to M_r 31 kDa (Fig. 1 , lane A). The polyclonal antibody produced against bovine retinal lectins selectively bound to the M_r 15-kDa band for a bovine retina sample (Fig. 1 , lane B). This antibody cross-reacted with the M_r 15-kDa band for a rat retina lysate as well (Fig. 1 , lane C). This finding implies that this antibody was mainly produced against the major retinal lectin of M_r 15 kDa and thus was considered to be effective for the isolation of its cDNA through screening of a retinal cDNA library. The polyclonal antibody against rat galectin-3 reacted with the M_r 31-kDa band for a rat retina sample (Fig. 2) . The intensity of this band increased with 3 days of CL exposure (Fig. 2 , lanes A–D; band-intensity, light–dark cycle: 1.00; 1 day of CL exposure: 0.98; 3 days: 6.56; and 1 week: 10.84). In contrast, the intensity of the band that reacted with the anti-bovine lectin antibody did not increase after CL exposure (Fig. 2 , lanes E–H; band-intensity, light–dark cycle: 1.00; 1 day of CL exposure: 1.07; 3 days: 1.03; and 1 week: 0.62). 

cDNA and Predicted Amino Acid Sequence of Bovine Retinal Lectin

The cDNA encoding the bovine retinal lectin of M_r 15 kDa was cloned from the cDNA library of bovine retinas (Uni-ZAP XR; Stratagene). Each phagemid isolated was found to contain a cDNA insert of approximately 0.7 kbp, by agarose gel electrophoresis. The sequencing data for the cDNA indicated that the isolated clone contained a common open reading frame coding for 135 amino acids, which form a protein of 14.5 kDa. A search of GenBank (provided in the public domain by NIH, Bethesda, MD, and available at http://www.ncbi.nlm.nih.gov) for similar nucleotide sequences revealed that this sequence is homologous to that of bovine galectin-1 (accession number, X14330) isolated from other tissues (e.g., trachea fibroblasts). ²⁴ The predicted amino acid sequence encoded by the cDNA isolated in the present study was completely identical with that of bovine galectin-1. ²⁴  

Polymerase Chain Reaction

A DNA band corresponding to 450 bp was observed for the PCR product, which was amplified using primers for rat galectin-1, electrophoresed on an agarose gel, and stained with ethidium bromide (Fig. 3 , lane A). After the DNA was ligated into a vector (Bluescript II SK [+]; Stratagene) and sequenced, a search of GenBank revealed that the predicted amino acid sequence encoded by this DNA is identical with that of rat galectin-1 (accession number, M19036). ¹⁹ A DNA band corresponding to 900 bp was observed for the electrophoresed PCR product amplified by using the primers for rat galectin-3 (Fig. 3 , lane B). After ligation and sequencing, a search of GenBank revealed that the nucleotide sequence encoded by this cDNA is identical with that of galectin-3 from rat basophilic leukemia cells ²⁰ (accession number, J02962) except for one nucleotide (nt 59: G→A). No PCR-product was detected by agarose gel electrophoresis of the sample amplified by using the PCR primers for rat galectin-4 ²¹ (Fig. 3 , lane C). These results confirmed that the galectin-1 and galectin-3 transcripts are both expressed in the rat retina but that a galectin-4 transcript is not. 

Reverse Transcription–Polymerase Chain Reaction

DNA bands corresponding to galectin-3 (Fig. 4 , lanes A–D) and β-actin (Fig. 4 , lanes E–H) were observed for the RT-PCR products. The ratio of the intensity of the DNA bands (galectin-3 and β-actin), which was calculated by measurement with NIH Image, was 0.75 after the light–dark cycle, 1.03 after 1 day of CL exposure, 1.68 after 3 days, and 1.28 after 7 days. These findings showed that the RT-PCR product for galectin-3 increased between 0 and 3 days of CL exposure (Fig. 4 , lanes A–C), whereas it decreased between 3 and 7 days (Fig. 4 , lanes C, D). 

Immunohistochemical Distribution of Galectins in Bovine and Rat Retinas

The sequencing data for the cDNA coding the bovine retinal lectin of M_r 15 kDa, which was detected as a single band with the anti-bovine retinal lectin antibody in Western blot analysis, showed that this antibody can be used for the detection of galectin-1 in the retina, RPE cells, and choroid (Figs. 5A 5B 5C 5D) , although we cannot absolutely rule out the presence of antibodies against other minor lectins, which are not detectable on Western blot analysis. The antibody prominently reacted with the apical halves of photoreceptor inner segments (IS), the outer limiting membrane (OL), the outer plexiform layer (OP), the inner plexiform layer (IP), the nerve fiber layer (NF), and vessels in the bovine retina (Fig. 5A) . The antibody also bound to RPE cells, choroidal arteries, and veins (Fig. 5C) . This labeling pattern was significantly inhibited by preincubation of the antibody with the bovine retinal lectin (Figs. 5B 5D) . Pigmented choroidal fibroblasts were observed in both sections with and without this preincubation (Figs. 5C 5D ; arrows). 

In the rat retina, almost the same antibody-binding pattern was observed as in the bovine retina, although the binding to regions of the inner retina was relatively weak (Fig. 6A) . This labeling pattern did not change after CL exposure (Fig. 6B : 1 day of CL exposure, and Fig 6C : 1 week), and was significantly inhibited by preincubation of the antibody with the bovine retinal lectin (Fig. 6D : 1 week of CL exposure). 

When rats were maintained with a light–dark cycle, galectin-3 was faintly detected in the OL, OP, IP, ganglion cell layer (GC), NF, and inner limiting membrane (IL; Fig. 7A ). After 1 day of CL exposure, the cytoplasm of certain cells with projections in the inner nuclear layer (IN) and strands in the outer nuclear layer (ON) reacted with the antibody, in addition to the increased labeling in the OL, OP, IP, GC, NF, and IL (Fig. 7B) . After 3 days of CL exposure, the immunoreactivity detected was increased, especially in the ON and IN (Fig. 7C) . The reaction products appeared as strands extending from the IL to the OL across the neural retina with side branches through the OP. This staining pattern closely matched the distribution of Müller cells and their processes. ²⁵ After 1 and 2 weeks of CL exposure, this labeling pattern remained unchanged although the thickness of the ON had gradually decreased (Fig. 7D : 1 week of CL exposure, and Fig 7E : 2 weeks). This immunoreactivity to galectin-3 was significantly inhibited by preincubation of the antibody with the GST/galectin-3 fusion protein, suggesting that this reaction was specifically induced by the antibody (Fig. 7F : 1 week of CL exposure). 

Anti-galectin Antibody–Induced Changes in the Rat Retina

Retinal detachment and vacuolation of the OP were observed in all five rat eyes at 3 days after vitreous injection of the anti-galectin-1 antibody, whereas the morphology of the detached photoreceptor outer (OS) and inner segments (IS) was well organized (Figs. 8A 9A) . In the case of anti-galectin-3 antibody injection, little morphologic change was observed at 3 days after injection in most regions of the eye (Figs. 8B 9B) , although retinal regions with torn photoreceptor outer segments were occasionally detected (Fig. 9D) . In these regions, because residual fragments of the OS were attached to the RPE (Fig. 9D) , we did not include them with those with retinal detachment. In the case of PBS injection, little morphologic change was observed as well (Figs. 8C 9C) . After PBS injection in one eye, vitreous and subretinal hemorrhages were observed as a severe complication. In this retina, disorganized photoreceptor OS and IS were detected, whereas no morphologic change was observed in the OP (Figs. 9E 9F) . Although significant differences were obvious between anti-galectin-1–injected eyes and either anti-galectin-3 antibody– or PBS-injected eyes, we compared the ratio of RPE length with retinal detachment/total RPE length (RD ratio: mean ± SD) in those eyes to ascertain the retinal detachment–inducing effect of the anti-galectin-1 antibody. The ratio was 0.82 ± 0.15 after anti-galectin-1 antibody injection (n = 5), 0.16 ± 0.10 after anti-galectin-3 antibody injection (n = 5), and 0.18 ± 0.10 after PBS injection (n = 5). These results showed that the RD ratio after anti-galectin-1 antibody injection was greater than that after the injection of either the anti-galectin-3 antibody or PBS (P < 0.05). 

Discussion

Exogenous lectins, which recognize specific sugar sequences, have been used for the detection and isolation of glycoconjugates in the retina. ^{1 2 3 4 5 11} Several kinds of endogenous animal lectins have been isolated and characterized. ^{12 19 20 21 26} Most animal lectins can be classified into four distinct families: C-type lectins including selectins, P-type lectins, pentraxins, and galectins. ¹² Some of these are believed to be involved in cell–cell or cell–matrix interactions, including cell adhesion. ^{12 26} Galectins, previously known as S-type lectins, specifically bind to β-galactoside residues and share a carbohydrate-binding domain. ^{12 26} They are found in many animal species, and nine galectins have been isolated so far. ²⁷ Expression analyses have revealed that galectins-2, -5, and -7 exhibit a restricted distribution, whereas galectins-1, -3, -8, and -9 exhibit a broad tissue distribution. ^{26 27} Galectin-1 has been reported to be abundant in the cardiac, smooth, and skeletal muscles; motor and sensory neurons; thymus; kidneys; and placenta. In the nervous system, galectin-1 expression has been observed in the brain, cochlea, and olfactory system. ²⁸ The present study revealed that the cDNA of galectin-1 can be isolated either by screening of a retinal cDNA library using an antibody or by PCR using primers for galectin-1. The present immunohistochemical study also confirmed that galectin-1 is distributed in bovine and rat retinas, and this finding indicates that the retina should be added to the list of sites of galectin-1 expression in neural tissues. 

The present Western blot analysis and cDNA sequencing showed that the lactose-binding lectins eluted from a lactose column consisted of two proteins: galectin-1 and a protein of M_r 31 kDa. We did not try to isolate the cDNA encoding the lectin of M_r 31 kDa by increasing the screened number of phages in the cDNA library, because the anti-bovine retinal lectin antibody did not bind to the M_r 31-kDa band in Western blot analysis, and the amount of this lectin in the retina appeared to be much smaller than that of galectin-1 in both Western blot analysis and immunohistochemistry. Therefore, the cDNA for the 31-kDa lectin could not be isolated by the cDNA library screening method. Previous studies showed that the molecular size of galectin-3 is between 26.2 kDa and 30.3 kDa, ²⁰ whereas that of galectin-4 is 36.3 kDa. ²¹ The molecular weights of both galectins are thus close to 31 kDa. The present study involving PCR revealed that the cDNA for galectin-3 was amplified, whereas that for galectin-4 was not. The present immunohistochemical study also showed that galectin-3 is distributed in the retina. These findings suggest that the lactose-binding lectin other than galectin-1 in the retina is galectin-3. 

The chicken 16-kDa lectin that was previously isolated ¹⁶ is considered to be galectin-1, based on the molecular weight similarity. However, the distribution of the 16-kDa lectin in the chicken retina ¹⁷ is partially different from that of galectin-1 in the mammalian retinas observed in the present study. The chicken galectin has been reported to be distributed in Müller cells, including their apical villi, and in the OL membrane and cone IS. ¹⁷ The distribution of galactin-1 in the OL membrane and cone IS observed in the present study was the same as that of the chicken galectin. However, the other labeling patterns were different between the chicken galectin and galectin-1. The chicken galectin has been detected in the cytoplasmic and nuclear compartments of Müller cells throughout the different retinal layers. ¹⁷ In the present study, galectin-3 was more markedly detected in Müller cells throughout the retina than was galectin-1 in the CL-exposed retinas. In this respect, the chicken galectin is very similar to galectin-3. It is possible that this inconsistency is partially due to the species difference between mammals and birds. 

Galectin-1 was isolated as a homodimer with a subunit molecular weight of approximately 14.5 kDa. ^{12 21} In its active form, galectin-1 exists as a dimer, which enables it to cross-link ligands on apposing cell surfaces or to facilitate the adherence of cells to the extracellular matrix. ¹² The principal interaction of the binding site of galectin-1 is with terminal N-acetyllactosamine residues. ¹² Laminin, lysosome-associated membrane glycoproteins, and antigen CD45, which contain polylactosamine chains, have been shown to be good ligands for galectin-1. ¹² We preliminarily found that the mucinlike glycoprotein associated with photoreceptor cells (MLGAPC), which has N-acetyllactosamine residues, ²² also binds to galectin-1 in the retina, on affinity chromatography on an immobilized galectin-1 column. The previous immunohistochemical study showed that MLGAPC is distributed in the photoreceptor layer and OP. ²² The present study showed that this region, positive for galectin-1, includes the OL, IS, RPE, and OP. These observations suggest that galectin-1 may interact with MLGAPC in the photoreceptor layer and OP. However, the origin of galectin-1 in these regions, unlike galectin-3, cannot be restricted to Müller cells, because galectin-1 was detected in various regions including the IS, RPE, and vessels in the present study. 

In connection with this surmised phenomenon, the histologic changes of the rat retina induced by an injection of an anti-galectin-1 antibody were notable. When a fixative containing glutaraldehyde is used for the fixation of rat eyes, retinal detachment does not usually occur during the preparation of tissue sections. The retinal detachment observed in the sections of all five rat eyes after the injection of the anti-galectin-1 antibody was thought to have been induced by the weak adhesion between the RPE and neural retina. In addition to retinal detachment, an intravitreous injection of the antibody against galectin-1 produced vacuolation of the OP. This type of change was not induced by an injection of either the anti-galectin-3 antibody or PBS, even in severely injured cases with vitreous and subretinal hemorrhage. Therefore, this morphologic change of the OP was also thought to have been induced by the anti-galectin-1 antibody. This finding is notable in relation to the common distribution of galectin-1 and MLGAPC described earlier. Not only in the photoreceptor layer but also in the OP, galectin-1 and MLGAPC may be involved in the mediation of interphotoreceptor and synaptic cleft adhesion. 

The reason that the anti-galectin-3 antibody did not induce retinal detachment is that only a small amount of galectin-3 was expressed in the retina under a normal light–dark cycle. Galectin-3 may play a physiological role other than retinal adhesion in the retina. The present study revealed that the expression of galectin-3 increased in the Müller cells of CL-exposed retinas. The present findings also show that an increase in the galectin-3 transcript preceded that in the galectin-3 protein, implying that galectin-3 expression is upregulated in Müller cells in response to CL exposure. These findings are consistent with the acceleration of rat photoreceptor degeneration due to CL exposure with the anti-galectin-3 antibody and the inhibition of light damage by galectin-3 observed in the previous study. ¹⁵ These findings suggest that the increased expression of endogenous galectin-3 observed in the present study may represent an anti-apoptotic action in CL-exposed rat retinas, as in other tissues including T-cells. ^{13 14}  

 Supported by Grant-in-Aid for Scientific Research (C) from the Japanese Ministry of Education, Science, Sports and Culture (13671844).

 Submitted for publication January 9, 2001; revised May 11, 2001; accepted May 31, 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.

 

Western blot analysis of retinal lectins in bovine retinas. Lane A: Lectin bands of 15 kDa and 31 kDa were detected with a biotinylated β-galactose probe for the bovine retinal eluate from a lactose-agarose column. Lane B: A 15-kDa band was selectively detected with a polyclonal antibody against the bovine retinal lectin for a bovine retina sample. Lane C: A 15-kDa band was detected with this antibody in a rat retina sample as well.

Figure 2.

 

Western blot analysis of galectin-3 and galectin-1 in rat retinas. Lanes A–D: A 31-kDa band reacted with the anti-galectin-3 antibody. Its intensity increased after CL exposure. Lane A: 0 days of CL exposure; lane B: 1 day; lane C: 3 days; and lane D: 1 week. Lanes E–H: A 15-kDa band was constantly detected with the anti-galectin-1 antibody independently of the period of CL exposure. Lane E: 0 days of CL exposure; lane F: 1 day; lane G: 3 days; and lane H: 1 week.

Figure 3.

 

DNAs of rat galectins amplified by PCR. Lane A: A 450-bp DNA-band was detected for the electrophoresed PCR product, amplified using primers for rat galectin-1. Lane B: A 900-bp DNA band was observed for the electrophoresed PCR product, amplified using primers for rat galectin-3. Lane C: No PCR-product was detected in the sample amplified using PCR primers for rat galectin-4. DNA molecular weight marker is shown in the extreme left and right lanes.

Figure 4.

 

DNAs of rat galectin-3 and β-actin amplified by RT-PCR from RNAs of CL-exposed rat retinas. Lanes A–D: A 900-bp DNA band was observed in the electrophoresed PCR product, amplified using primers for rat galectin-3. Lane A: light–dark cycle; lane B: 1 day after CL exposure; lane C: 3 days after; and lane D: 1 week after. Lanes E–H: A 764-bp DNA band was detected in the electrophoresed PCR product, amplified using primers for rat β-actin. Lane E: light–dark cycle; lane F: 1 day after CL exposure; lane G: 3 days after; and lane H: 1 week after.

Figure 5.

 

Immunohistochemistry of a bovine retina, RPE and choroid using the anti-galectin-1 antibody. (A, B) Retina; (C, D) RPE and choroid. (A) The apical halves of photoreceptor IS, the OL membrane, the OP layer, the IP layer, the NF layer, and vessels were labeled. (C) The RPE, almost the whole thickness; choroidal arteries; and veins were labeled with the antibody. (B, D) Weak labeling was observed after preincubation of the antibody with the bovine retinal lectin. (C, D) Pigmented choroidal fibroblasts (arrows) were observed in both sections, with and without preincubation. AR, choroidal artery; VE, choroidal vein; PE, pigment epithelium. Scale bar, 30 μm.

Figure 6.

 

Immunohistochemistry of rat retinas using the anti-galectin-1 antibody. (A–C) The apical halves of photoreceptor IS, the OL membrane, the OP layer, and vessels intensely reacted with the antibody. The IP and NF layers were weakly labeled. This reaction pattern remained unchanged, although the thickness of the ON layer gradually decreased after CL exposure (A: 0 days of CL exposure; B: 1 day; C: 1 week). (D) A small reaction was observed after preincubation of the antibody with the bovine retinal lectin (1 week of CL exposure). Scale bar, 30 μm.

Figure 7.

 

Immunohistochemistry of rat retinas using the anti-galectin-3 antibody. (A–E) The labeling of Müller cells increased with CL exposure (A: 0 days of CL exposure; B: 1 day; C: 3 days; D: 1 week; E: 2 weeks). (F) This immunoreactivity to galectin-3 was inhibited by preincubation of the antibody with the GST-galectin-3 fusion protein (1 week of CL exposure). Arrows, OP layer; OL, outer limiting membrane; ON, outer nuclear layer; IN, inner nuclear layer; IP, inner plexiform layer; GC, ganglion cell layer; IL, inner limiting membrane. Scale bar, 30 μm.

Figure 8.

 

Hematoxylin-eosin staining of rat retinas after intravitreous injection of anti-galectin antibodies or PBS. (A) Retinal detachment was observed after vitreous injection of an anti-galectin-1 antibody. (B, C) Little morphologic change was observed after injection of an anti-galectin-3 antibody (B) or PBS (C). Scale bar, 200 μm.

Figure 9.

 

Hematoxylin-eosin staining of rat retinas after intravitreous injection of anti-galectin antibodies or PBS. (A) Retinal detachment and vacuolation of the OP layer were observed after vitreous injection of the anti-galectin-1 antibody. (B, C) Little morphologic change was observed after injection of the anti-galectin-3 antibody (B) or PBS (C). (D) Even in the anti-galectin-3 injected retinal region with torn photoreceptor OS, no morphologic change was observed in the OP. (E, F) Even in a PBS-injected eye with hemorrhage, and disorganized photoreceptor OS and IS segments, no morphologic change was observed in the OP. OS, outer segments; IS, inner segments; ON, outer nuclear layer; arrows, OP; IN, inner nuclear layer; PE, pigment epithelium. Scale bars, (A–D, F) 20 μm; (E) 80 μm.