Alteration of Sialyl Lewis Epitope Expression in Pterygium

Abstract

purpose. Mucin-related antigens are abundantly expressed by the cells of the normal human conjunctiva. The pattern of these antigens in pterygium, and especially the role of Galβ1-3GlcNAcα 2,3-sialyltransferase (ST3Gal III), sialyltransferase necessary to build the sialyl-Le^a (Lewis^a) antigen, were studied.

methods. Immunoperoxidase staining was performed on 28 pterygia using different monoclonal antibodies: anti-M1 (against the peptidic core of gastric mucins encoded by MUC 5AC gene), anti-Le^a (7LE), anti–sialyl Le^a (NS 19-9), and anti-Le^b (2-25LE). A serologic Lewis determination was done in 18 patients. ST3Gal III sialyltransferase expression was also studied in 10 healthy conjunctiva and 10 pterygia by reverse transcriptase–polymerase chain reaction (RT–PCR). Glyceraldehyde-3-phosphate-dehydrogenase was used as an endogenous internal control.

results. First, Le^a, sialyl Le^a, and Le^b immunoreactivities either decreased or were no longer detectable in pterygium goblet cells as opposed to normal conjunctiva. Second, unlike in pterygium, the Lewis immunoreactivity, which is mainly located in the surface epithelial cells in the normal conjunctiva, was occasionally restricted to the epithelial cells of the deep layers. However, M1 mucins did show an identical pattern expression in a normal conjunctiva and pterygium. ST3Gal III expression was significantly lower in pterygium (0.20 ± 0.02 AU [arbitrary units]) than in normal conjunctiva (0.95 ± 0.12 AU).

conclusions. ST3Gal III gene is less expressed in pterygium than in normal conjunctiva. This observation could explain the decrease of sialyl Le^a expression observed in pterygium by immunohistology.

The preocular tear film is essential to the health, transparency, and optical quality of the ocular surface. The inner layer of the preocular tear film corresponds to a mucin-containing gel layer approximately 1-μg thick. ¹ Recently, the“ long-accepted” model of tear film composition has been challenged, suggesting that the true thickness of the film is between 35 and 40μ g and made up by this mucin-containing gel. ^{2 3} Mucin is the major constituent of this layer coating the ocular surface. ⁴ Conjunctival goblet cells are the primary source of ocular mucin, ⁵ but the entire ocular surface epithelium produces mucins for the tear film. ⁶  

Pterygium (from the diminutive of πτερυζ, small wing), is considered to be an invasion of the cornea, with dissolution of the Bowman’s membrane by a triangular segment of bulbar conjunctiva. ⁷ It shows basophilic degeneration with actinic or senile elastosis of the subepithelial substantia propria. The fully developed pterygium is covered by flat or cylindrical epithelial cells that are comparable to the conjunctival epithelium. The epithelium of pterygium also contains goblet cells whose morphology is similar to those found in the conjunctiva, but their densities are increased. ⁸  

Human conjunctival mucins consist of a peptidic core bearing a large number of oligosaccharide side chains ⁹ showing alternating poorly (naked) and highly glycosylated (T domain) regions. ¹⁰ Mucins may be considered as a mosaic of epitopes, ¹¹ some of which are associated with the peptidic core of the naked regions (M1 epitopes encoded by MUC 5AC gene) ¹² and others, with the peripheral parts of the oligosaccharide chains located in T domains displaying blood group–related epitopes (ABO, H and Lewis). ¹³ These histo-blood group antigens are built by the addition of monosaccharides onto precursor glycoprotein by specific glycosyltransferases such as galactosyltransferases, sialyltransferases, and fucosyltransferases. ¹⁴ These glycosyltransferases are encoded by the genes of ABH, Secretor (Se), and Lewis (Le) systems. ¹³ Secretor individuals show an active α 1-2 fucosyltransferase encoded by the secretor gene. ¹⁵  

Mucin alteration can be observed in several ocular diseases: aqueous tear deficiency, ocular rosacea, ¹⁶ atopic conjunctivitis, ¹⁷ and others. To the best of our knowledge, mucin expression in pterygium has only been studied using lectins. Nongoblet cells of pterygium were labeled with ulex europaeus agglutinin-1 (UEA-1), doliches biflorus agglutinin (DBA), and peanut agglutinin (PNA), but those of normal conjunctiva were not. Abnormal distribution of the UEA-1 staining was noted in the superficial layers of epithelial cells in pterygium. ⁸  

The aim of this study was to characterize the immunohistologic mucin modifications in the pterygium. First, we used monoclonal antibodies (Mabs) directed against the peptidic core and Mabs against saccharide moieties of mucin, which are known to be more specific than lectins. Second, in addition, we evaluated a sialyltransferase (ST3Gal III) ¹⁸ expression in pterygium and healthy conjunctiva to try to explain the mucin immunohistologic modification in the pterygium. We observed different patterns in sialyl Le^a immunoreactivity, which is confirmed by a decrease of the ST3Gal III gene expression. 

Materials and Methods

Immunohistology

Mabs against Peptidic Core of Mucin.

Anti-M1 Mabs directed against the peptidic core of gastric M1 mucins were used. M1 immunoreactivity was recently described as encoded by the MUC 5AC gene. ¹²  

Mabs against Saccharide Moieties of Mucin.

The following Mabs were used: anti-Le^a (7LE), ^{19 20} anti-Le^b (2-25LE), ^{21 22} anti–H type 2 (19-0LE) from our laboratory, ^{19 20} and anti–sialyl Le^a (NS 19-9). ²³ The biochemical structures recognized by these Mabs are shown in Table 1 . They all strongly reacted using immunoperoxidase methods with mucus cells of gastroduodenal mucosae according to the phenotype of the tissue donor: Anti-Le^a Mab reacted on nonsecretor individuals and anti–sialyl Le^a Mab on mucosae of Lewis-positive individuals. 

Tissue Samples.

A total of 28 fully developed pterygium tissues affecting the nasal cornea were obtained from 28 patients who ranged in age from 24 to 83 years. 

Preparation of Tissues.

Pterygium was fixed in 95% ethanol overnight, routinely processed, and then embedded in paraffin wax. Serial sections (3-μg thick) were cut with an Autocut (Reichert–Jung, Heidelberg, Germany). 

Immunoperoxidase.

The sections were deparaffinized with three successive baths of xylene and ethanol of 10 minutes each. The sections were rinsed with water for 10 minutes, preincubated for 3 minutes in phosphate-buffered saline containing 0.1% Tween-20 (PBS–Tween), and then incubated for 30 minutes with the Mabs (undiluted hybridoma supernatants). After 3 rinses in PBS–Tween, the sections were incubated for 30 minutes with mouse anti-Ig antibodies (diluted 1/200) linked to peroxidase (Amersham, Aylesbury, UK). The sections were washed three times with PBS–Tween and incubated for 4 minutes with amino-ethylcarbazole (Sigma, St. Louis, MO) containing H₂O₂. Before microscopic examination, the cell nuclei were stained with 1% hematein for 2 minutes. The specificity of the immunoreactivity was controlled by the inhibition of the staining after incubation of the hybridoma supernatant with the gastric mucin preparation (100 μg/ml supernatant), red blood cells, or tissue extracts containing M1 or Lewis antigens. 

ABO and Lewis Phenotype.

Blood samples were obtained from the 28 patients to determine ABO and Lewis erythrocyte phenotype by hemagglutination. 

RT–PCR

Tissue Samples.

The guidelines of the Declaration of Helsinki were followed. Conjunctival biopsy specimens were collected with fully informed written consent and authorization of the local ethical committee in 10 patients undergoing cataract surgery (age range, 35–65 years). Pterygia were obtained from 10 patients who ranged in age from 26 to 68 years. After removal, these specimens were immediately frozen in liquid nitrogen and stored at −80°C before further analysis. 

Extraction of RNA and Analysis of ST3Gal III Transcripts by RT–PCR.

Total cytoplasmic RNA was extracted from homogenized conjunctiva or pterygium tissue by guanidinium–phenol–chloroform technique ²⁴ with Trizol kit (Gibco–BRL, Grand Island, NY). Three micrograms of total RNA was converted into first-strand cDNA by a random priming technique ²⁵ in 20 μl of RT mixture: 8 μl RNA, 2 μl random hexamers (150 ng/μl), 10 μl RT buffer (40 mM/l Tris–HCl, 100 μM/l each dNTP, 6 U Rnasin, 50 nM/l dithiothreitol, 5 ng/ml bovine serum albumin, 8 mM/l MgCl₂). After RNA denaturation for 10 minutes at 70°C and then cooling on ice, 1 μl of Avian Myeloblastosis Virus reverse transcriptase was added, and the samples were incubated for 60 minutes at 37°C. All reagents were purchased from Promega (Madison, WI). Changes in the amount of ST3Gal III mRNAs were monitored by coamplification in the same tube of the target fragment and of a glyceraldehyde-3-phosphate dehydrogenase (GAPDH) housekeeping gene sequence as an internal standard. For reliable quantification, the concentration of each set of primers and the stringency of PCR conditions were adjusted so that the amplicons would have similar amplification kinetics and their exponential phases of amplification would overlap as previously described. ²⁶ Amplification was done with specific primer target sequences as follows ²⁷ : for ST3Gal III, 5′-CGGATGGCTTCTGGAAATCTGT-3′ and 3′-AGTTTCTCAGGACCTGCGTGTT-5′ and for GAPDH, 5′-ACCACAGTCCATGCCATCAC-3′ and 3′-TCCACCACCCTGTTGCTGTA-5′. The PCR reaction was set up in a25-μl mixture consisting of 5 μl cDNA solution diluted in 2.5 μl PCR buffer (1 mM/l Tris–HCl, pH 8.3, 1 mM/l KCl, 1 mM/l each dNTP, and 15 mM/l MgCl₂), 2 μl ST3Gal III primers, 0.2 μl Taq polymerase (Biotaq; Bioprobe, Montrevil, France), and an adjusted volume of distilled water. After 10 minutes of denaturation at 95°C, 36 cycles of amplification were carried out on a PHC3 thermocycler (Techne, Cambridge, UK) followed by a final 10-minute extension at 72°C. The cycling parameters were the following: denaturation at 94°C for 1 minute, annealing at 63°C for 1 minute, and extension at 72°C for 2 minutes. 

After amplification, 10 μl of PCR products was electrophoresed in a 2% ethidium bromide–stained agarose gel for quantitation of fluorescence by image analysis under UV light (UVP Transiluminator, Cambridge, UK). 

The PCR amplification yield of target sequences was expressed in arbitrary units (AU) as a ratio of ST3Gal III/GAPDH electrophoretic band optical density. 

Statistics

All values given were mean ± SEM. An unpaired t-test was performed to compare the different mean values. Values were considered significant if P ≤ 0.05. 

Results

Immunohistology

Mucin Peptide Epitopes.

M1 mucins showed identical pattern expression in normal conjunctiva and pterygium. The cytoplasm of goblet cells alone was strongly stained with anti-M1 gastric mucin (Fig. 1) . 

Lewis-Related Epitopes.

Goblet Cells. Lewis-related epitopes were detected in normal conjunctiva of Lewis-positive donors, exclusively. Yet, the intensity of the staining with anti-Le^a, anti–sialyl Le^a, and anti-Le^b Mabs decreased in goblet cells in pterygium (Table 2) . One of 6 Lewis^a-positive patients and 3 of 20 Lewis^b-positive patients did not express Le^a epitope. Sialyl Le^a was no longer detectable in the goblet cells of 7 Lewis-positive patients (7/26; Fig. 2 ). 

Epithelial Cells. As for the goblet cells, the immunoreactivities of anti-Lewis Mabs decreased on the epithelial cells of some pterygia. However, these cells were stained with anti-Le^b (2-25 LE) and anti–H-Type 2 (19-0LE) Mabs. Moreover, the sialyl Le^a immunoreactivity displayed peculiar patterns, which are not observed in the normal conjunctiva: Some areas showed staining in the deeper epithelial cells but not on the surface (Fig. 3) . We could not determine whether this aspect was related to their limbal localization. 

Comparison of the Lewis Phenotype using Hemagglutination and Immunohistology.

The frequencies of different Lewis phenotypes were the same as those observed in the French population: 6 of 28 Lewis^a (21%), 20 of 28 Lewis^b (72%), and 2 of 28 Lewis (7%). We found a percentage of Lewis phenotype in the red cells and the tissues similar to that of normal conjunctiva. 

RT–PCR

RT–PCR indicated that ST3Gal III expression was present in all the samples of conjunctiva or pterygium. ST3Gal III expression was 0.20 ± 0.02 AU in pterygium, whereas it was 0.95 ± 0.12 A.U. in normal conjunctiva (Fig. 4) . There was a statistically significant difference between pterygium and normal conjunctiva (P < 0.0001). 

Discussion

Mucus produced by goblet cells in the pterygium does not differ from mucus secreted by goblet cells from normal conjunctiva when assessed by conventional histochemical methods. On the other hand, pterygium epithelial cells react with UEA-1 (specific for fucose), DBA (specific for N-acetylgalactosamine), and PNA (specific for Galβ1-3GalNAc disaccharidic sequence) lectins, whereas normal conjunctiva epithelial cells react with WGA (specific for N-acetylglucosamine and N-acetylneuraminic acid) lectin exclusively. ⁸ Additionally, lectin reactivities are present in the apical portion of the pterygium epithelium. Some authors suggested that nongoblet cells of pterygium are involved in the production of abnormal mucus glycoproteins containing fucose and N-acetylgalactosamine residues on the oligosaccharide side chains. 

The present study also showed modifications in the carbohydrate structures because pterygium mucins secreted by the goblet cells present variations on the oligosaccharide chains but not on the peptidic core (M1 antigen) encoded by MUC-5AC gene. ¹² It also indicated a lower expression of the blood group–related antigens, suggesting an abnormal expression of the fucosyl- and sialyltransferases involved in the biosynthesis of these antigens. The anti–blood group–related antigen Mabs we used characterize the alteration of glycosylation associated with glycoconjugates expressed by the conjunctiva including the mucins. Indeed, Mab NS 19-9 recognizes the sialyl Le^a tetrasaccharide structure (Table 1) . The absence of the neuraminic acid or fucose on this sialyl Le^a structure because of a decrease of sialyl- or fucosyltransferase involves a loss of the sialyl Le^a immunoreactivity. Concerning the Le^b structure, the absence of one of the two fucoses on the tetrasaccharide Le^b molecule (Table 1) , due to a decrease of fucosyltransferases, involves also a loss of Le^b immunoreactivity (2-25LE). The decrease of staining suggested that sialyltransferases, fucosyltransferases, or both are less efficient in the goblet cells of pterygium. 

In the epithelial cells, sialyl Le^a immunostaining is occasionally restricted to the deep layer. Such an abnormal pattern was not observed in the normal conjunctiva. Restriction of abnormal patterns to surface epithelium observed with UEA-1 lectin is not in contradiction with our results. Indeed, on the surface epithelium, the decrease of the sialyltransferase we observed could favor the abnormal fucosyltransferase activity of the Se gene. 

Thus, we studied the Galβ1-3GlcNAc α2,3-sialyltransferase (ST3Gal III) expression in pterygium. Sialyltransferases form a subclass of glycosyltransferases that catalyze the transfer of sialic acid (NeuAc) from CMP–NeuAc to galactose, N-acetylgalactosamine, or another sialic acid residue. Indeed, to date, 13 different sialyltransferase cDNAs have been cloned, ²⁸ and at least 8 different enzymes are involved in the biosynthesis of the sialylated oligosaccharidic chains that define the mucin sialyl epitopes. ²⁹ Each sialyltransferase can be distinguished enzymatically by its specificity for the sequence of the acceptor oligosaccharide and the anomeric linkage formed between the sialic acid and the sugar to which it is attached. ³⁰ The fine substrate specificity of four α2,3-sialyltransferases (ST3Gals I, II, III, and IV) was recently defined by the use of soluble recombinant enzymes. ³¹ This study confirmed that ST3Gal III, which preferentially transfers sialic acid in α2,3 linkage to the Galβ1-3GlcNAc disaccharidic sequence, was the best candidate for the biosynthesis of the sialyl Le^a epitope. The recent progress obtained in the molecular cloning of human sialyltransferase cDNA allowed the development of an RT–PCR–based method that will assay the expression of these enzymes in biological samples. ²⁷ We showed that ST3Gal III expression is significantly lower in pterygium than in normal conjunctiva. These findings could explain the decrease of sialyl Le^a expression observed in pterygium by immunohistology. 

Regulation mechanisms of sialyltransferases are not well established. Variations in complex carbohydrate structures were observed during development, during differentiation, in disease processes, and between different normal tissues, leading to the conclusion that the expression of terminal sequences was strictly controlled both spatially and temporally. ³⁰ The modification of cellular glycosylation is a common phenotypic change in malignancy, but only a limited subset of biosynthetic pathways is frequently altered in cancer. Increasedβ 1,6 branching, increased sialyl Le^a and sialyl Le^a epitopes, accumulation of H and Le^y antigens, and a general increase in sialylation are commonly observed in N- and O-linked oligosaccharides of carcinoma cells. ^{32 33} These changes in glycosylation are correlated with grade, with invasion and metastasis, and with a poor prognosis. As an example, increased sialyl Le^a expression occurred in colon cancer cells and has been associated with the acquisition of a high metastatic capacity. ^{34 35 36}  

The increase in sialylation of the metastatic tumor cell surface can result in a decreased attachment to basement membrane protein collagen type IV and fibronectin, predisposing the tumor cells to an increased mobility and a decreased growth control by substratum contact. ³⁷ Sialyl Le^a epitope is also a ligand for E-selectin. ^{38 39 40} It is an attractive hypothesis that some carcinoma cells would use that selectin–carbohydrate interaction in the cascade of events involved in metastasis. By contrast, pterygium is a tumor without degenerative processes, ⁴¹ in which ST3Gal III is less expressed. This could be explained by the fact that this tumor is benign. 

These abnormalities also suggest an abnormal expression of the fucosyl- and sialyltransferases and suggest changes during cell differentiation or maturation. Such a change seems specific to pterygium and, as far as we know has not yet been observed in other ocular diseases. 

To the best of our knowledge, no study has already tested ST3Gal III expression on the ocular surface. We demonstrated that ST3Gal III expression is lower in pterygium than in normal conjunctiva, which partially explains the modifications of mucin carbohydrate epitopes in pterygium. In fact, glycosyltransferases other than ST3Gal III could be altered in pterygium and be involved in these modifications. Moreover, we showed an abnormal pattern of cell differentiation characterized by sialyl Le^a expression, which is specific to pterygium. 

 Reprint requests: Catherine Creuzot–Garcher, Department of Ophthalmology, General Hospital, 3 rue du Faubourg Raines, 21000 Dijon, France.

 Presented in part at the annual meeting of the Association for Research in Vision and Ophthalmology, Ft. Lauderdale, Florida, May 1998.

 Submitted for publication June 17, 1998; revised February 11, 1999; accepted March 23, 1999.

 Proprietary interest category: N.

 

Table 1.

 

Blood Group–Related Structures Recognized by the Mabs against Blood Group–Related Antigens

Epitopes	Structures	Monoclonal Antibodies
Lea	\batchmode \documentclass[fleqn,10pt,legalpaper]{article} \usepackage{amssymb} \usepackage{amsfonts} \usepackage{amsmath} \pagestyle{empty} \begin{document} \(\begin{array}{r}\\\mathrm{Gal{\beta}1-3GlcN}\\\mathrm{{\Vert}}\\\mathrm{Fuc{\alpha}1-4}\end{array}\) \end{document}	7LE
Leb	\batchmode \documentclass[fleqn,10pt,legalpaper]{article} \usepackage{amssymb} \usepackage{amsfonts} \usepackage{amsmath} \pagestyle{empty} \begin{document} \(\begin{array}{r}\\\mathrm{Fuc{\alpha}1-2Gal{\beta}1-3GlcN}\\\mathrm{{\Vert}}\\\mathrm{Fuc{\alpha}1-4}\end{array}\) \end{document}	2-25LE
Sialyl Lea	\batchmode \documentclass[fleqn,10pt,legalpaper]{article} \usepackage{amssymb} \usepackage{amsfonts} \usepackage{amsmath} \pagestyle{empty} \begin{document} \(\begin{array}{r}\\\mathrm{NeuAc{\alpha}2-3{\beta}Gal1-3GlcN}\\\mathrm{{\Vert}}\\\mathrm{Fuc{\alpha}1-4}\end{array}\) \end{document}	NS 19-9
H type 2	Fucα1-2Galβ1-3GlcNAcβ1-R	19-0LE

 Fuc, L-fucose; Gal, D-galactose; GlcNAc, D-N-acetylglucosamine; NeuAc, N-acetylneuraminic acid.

Figure 1.

 

Anti-M1 staining with immunoperoxidase methods and hematein counterstaining on samples fixed in ethanol and then embedded in paraffin. Anti-M1 (mixture of eight different Mabs) strongly stained conjunctival goblet cells and did not stain epithelial cells. (A) Normal conjunctiva (magnification, ×250). (B) Pterygium (magnification, ×100).

Table 2.

 

Summary of the Immunoperoxidase Reactions Obtained with Non-A and Non-B Anti-Carbohydrate Antigens on the 28 Pterygia as Compared to the 89 Conjunctival Biopsies

Phenotype of Tissue Donor											Epithelial Cells							Goblet Cells
Status	Le	Lewis	Secretor	N, o	%	Le, a	S-Le, a	Le, b	Ht2	Le, a	S-Le, a	Le, b	Ht2
N	a	Positive	No	17	19	+++	+++	+	−	+++	+++	+	—
P	a	Positive	No	6	21	+/++	+++	+/++	−	+/++	+/−	—	—
N	b	Positive	Yes	63	71	—	+++	+++	++	+++	+++	+	—
P	b	Positive	Yes	20	71	—	+++	+++	+/+++	++/−	++/−	—	—
N	−	Negative	Yes	9	10	—	—	—	+	+	—	—	—
P	−	Negative	Yes	2	8	—	—	—	—	—	—	—	—

 Status: P, pterygium; N, normal conjunctiva; Le

^a  , anti-Le

^a  (7LE); S-Le

^a  , anti–NeuAc-Le

^a  (NS 19-9); Le

^b  , anti-Le

^b  (2-25LE); Ht2, anti–H type-2 (19-0LE); Le, abbreviation combining Lewis and secretor phenotypes. Positive reactions from + to +++ represent an increasing number of stained cells and the intensity of staining. N

^o  , number of individuals tested; %, percentage of individuals among the studied population.

Figure 2.

 

Anti–sialyl-Le^a (Mab NS 19-9) staining of a Lewis-positive secretor individual. (A) Normal conjunctiva: Mab NS 19-9 stained goblet cells and epithelial cells. (B) Pterygium: Epithelial cells are stained in contrast with goblet cells. Magnification, (A) ×250; (B) ×400.

Figure 3.

 

Anti–sialyl-Le^a (Mab NS 19-9) staining. (A) Normal conjunctiva: Mab NS 19-9 staining is located in the superficial layers. (B) Pterygium: Abnormal immunoreactivity of some epithelial cells; (B1) Lewis reactivity is mainly located in the deeper layers of the epithelial cells. (B2) Transitional zone between the abnormal Lewis immunoreactivity of the deeper layers of the epithelial cells (*) and the normal reactivity of the other layers (**). Magnification, (A) ×250; (B1)× 400; (B2) ×100.

Figure 4.

 

Analysis of ST3Gal III expression in normal conjunctiva and pterygium. Ten microliters of the PCR mixture of the coamplification of ST3Gal III– and GAPDH-specific fragments was electrophoresed through a 2% ethidium bromide–stained agarose gel and analyzed by densitometry under UV light. Among the 10 normal conjunctiva and pterygium, 3 representative samples of each are presented (lanes A, B, C, normal conjunctiva; lanes D, E, F, pterygium). The ratio of ST3Gal III/GAPDH signals was 0.95 ± 0.12 and 0.20 ± 0.02 in normal conjunctiva and pterygium, respectively. MM, size of PCR products (in base pairs).