Exp. Eyu Res.(1990) 50. 45-50
45
Sialic AB KUIZENGA*,
Acid
in Human
Tear Fluid
ERIC J. van AGTMAAL, NICOLAAS AND AIZE KIJLSTRA”
J. van HAERINGEN
The Netherlands Ophthalmic Research Institute, Biochemical Laboratory, P.O. Box 72741, 7 700 AC Amsterdam, The Netherlands and =Department of Ophthalmology, University of Amsterdam, the Netherlands (Received
13 April
1989 and accepted in revised form 77 August
7989)
A simple assay for the determination of sialic acid (N-acetylneuraminic acid) in human tear fluid was evaluated. Sialic acid, terminally bound on carbohydrate side-chains of glycoproteins, was releasedafter treatment with neuraminidase and measured by an enzymatic calorimetric test. Tear fluid sampleswere collected from ten healthy adults, using glass capillaries and cellulose sponges.Sialic acid levels in tears collected with sponges (08-1.8 mmol 1-l) did not differ significantly from those found in capillary tears (09-1.8 mmol 1-l). Sialic acid, expressed as mmol g-’ protein, was significantly lower in tears collected with sponges (018-0.32 mmol g-‘) than with capillaries (0.19-0.42 mmol g-l). Recovery of sialic acid and protein after incubation of cellulose spongeswith tears was more than 99 %. Sialic acid levels in human tears, which had been centrifuged to remove insoluble material, remained unchanged. Furthermore, tear sialic acid activity did not pass a filter with a molecular weight cut-off point of 10000. Our data indicate that with the assay used in this report, sialic acid in tears is not due to secretory immunoglobulin A (sIgA), lactoferrin and lysozyme. The fact that the major tear proteins do not contribute to the sialic acid levels detectedin tears suggests that other as yet unknown soluble glycoproteins are involved. Key words: sialic acid: tear fluid; glycoproteins
1. Introduction
The preocular tear film covers the cornea1 and conjunctival epithelia. Its functions are to maintain structural integrity of the cornea and conjunctiva and to protect the eye against external influences. The composition of human tear fluid is very complex (van Haeringen, 1981) and contains various substances, including lipids, electrolytes, proteins and muco-glycoproteins. The understanding of the mucoglycoproteins on the ocular surface is gradually increasing. The mucous layer, covering the epithelial surface of the cornea and conjunctiva and containing mucous glycoproteins (mucins) (Moore and Tiiany, 1979) is considered to maintain the stability of and contribute to the viscosity of the tear film (Kaura and Tiffany, 1986). Mucous glycoproteins are macromolecules of varying molecular weight and consist of a large linear protein core, to which many oligosaccharide side chains attach, containing hexoses and sialic acid. They can exist in water-soluble form and in an insoluble high molecular weight gel-form. The serumtype glycoproteins have a lower molecular weight and contain fewer carbohydrate side chains, attached to a
more globular protein core. Although the conjunctival goblet cells are the main source for mucous glycoproteins (Kessing, 1968), glycoproteins have also been found in lacrimal gland acini (Allen, Wright and Reid, 1972) and glyco-
2. Materials and Methods Collection of Tear Fluid
Tear samples were collected from healthy adults (n = 10) using 25+1 glass capillaries (capillary tears). Production of tear fluid was stimulated by a jet of pressurized air onto the cornea. Within a period of 15 min tear fluid production
* For correspondence. 00144835/90/010045+06
proteins in tears have been demonstrated by their hexosamine and sialic acid content (Cabezaset al., 1964; Dohlman et al., 1976), suggesting lacrimal secretion of these proteins onto the ocular surface (Kreuger, Sokoloff and Botelho, 1976). Water-soluble glycoproteins were detected in tears by electrophoresis (Berta and Tiirak. 19 8 6) and may be altered in various pathological states of the eye (Dohlman et al., 1976 ; Halken, Bog-Hansen and Prause, 1986a). Glycoproteins of lacrimal, together with those of conjunctival origin, can be important in maintaining the physicochemical properties of the precorneal tear film. Until now no rapid and simple method was available to quantitate glycoprotein levels in human tears. Therefore, the determination of sialic acid in human tear fluid, using a simple enzymatic colorimetric method, was evaluated as a parameter for tear glycoproteins. The major tear proteins do not contribute in the sialic acid assay used and it is concluded that sialic acid in human tears is mainly due to soluble glycoproteins of still unknown identity.
$03.00/O
ranged from 25 ~1 to
0 1990 Academic Press Limited
A. KUIZENGA
46
ET AL
100 ,uI. Samples were stored in 400-,~rl polyethylene tubes at -20°C. After at least 2 hr, tear samples were collected from the same subjects by means of surgical cellulose sponges (Sugi, Kettenbach, GDR) as described earlier (van Agtmaal et al., 1988) and the samples (sponge tears) were stored at - 20 “C. A pool of capillary tears was used for determination of the precision of the sialic acid assay.
Molecular Weight Limit) low-binding PLGC membrane (Millipore Corporation, USA) and the filtrates were assayed for sialic acid.
Assay of Sialic Acid in Human Tear Fluid
Recovery Experiments
Sialic acid was assayed, using a commercially available test kit (Boehringer, Mannheim, FRG). Briefly, the principle of this method is as follows : sialic acid, as a covalently bound terminal sugar residue of oligosaccharide side chains, is enzymatically hydrolyzed by neuraminidase (source : Arthrobacter ureafaciens ; specificity : hydrolyses glycosidically bound N-acetyl- and N-glycolylneuraminic acid) to free sialic acid and then cleaved to N-acetylmannosamine and pyruvate by N-acetylneuraminic acid aldolase. Formed pyruvate is then oxidized by pyruvate oxidase to acetylphosphate, CO, and H,O,. The amount of H,O, formed, which is equivalent to the free sialic acid, is determined by a peroxidase reaction, where a chromogen is converted to a red dye, of which the absorbance is measured at 550 nm. In our experiments assays were performed in l-ml microcuvettes (Sarstedt, FRG) and the quantities of reagents and samples were only a quarter of those stated in the manufacturer’s protocol. In each assay 5 ~1 of tear or control protein sample was used (incubation volume: 255 ~1; assay volume after addition of stabilizer: 755 ~1). A serum sample provided with the kit, containing a known amount of bound sialic acid, was used as a standard. The blank consisted of 5 ,~l of water. Proteins analyzed in the sialic acid assay included 5 g 1-l samples in phosphate buffered saline (PBS) of purified milk lactoferrin, secretory immunoglobulin A from human colostrum, chicken egg white lysozyme (all from Sigma Chemical Company, St Louis, MO) and human serum albumin. In a separate experiment the tear samples were assayed for free pyruvate, using an available test kit (Boehringer, Mannheim, FRG) and the sialic acid values were corrected by subtraction of the pyruvate values. In order to investigate if any insoluble material in human tears could contribute to sialic acid levels in this fluid, tear samples from four healthy volunteers, obtained with the capillary method, were centrifuged (15 min, 12 000 g, 20°C) and the supernatant was assayed for sialic acid together with the samples prior to centrifugation. Parallel to this experiment. the original four capillary tear samples were treated by centrifugal ultrafiltration (15 min, 12 000 g, 20°C) through an ultraFree-MC filter unit, 10 000 NMWI, (Nominal
To investigate whether adsorption of sialic acid containing tear proteins takes place to the cellulose material of the sponges, used for tear collection, the following experiment was performed. Reflex tear fluid was collected with glass capillaries, and sialic acid and protein content were measured afterwards. Five cellulose sponges (mean weight &-S.E.M. : 10.4 k 0.2 mg) were placed in 400~,~l polyethylene tubes and 50 ,~l of tear fluid was added and incubated at room temperature for 30 min. Subsequently, tear fluid was regained from the sponges by centrifugation as described earlier (van Agtmaal et al., 1987) and assayed for sialic acid and total protein. As a control, sponges were soaked with 50 ,~l of water.
Assay of Total Protein in Human Tear Fluid Total protein was assayed with Bradford’s method (Bradford, 19 76) modified according to Jongeneel (1982), using human serum albumin as a standard.
Statistics Significance of differences between groups of data was determined by using the Man-Whitney U-test.
3. Results Sialic Acid Assay in Human Tear Fluid Within the assay, the precision, expressed as coefficient of variation, was 2.1 Y0 (n = 20); day-today precision (duplicate assay, n = 14) was 3.5%. Detection limit in the sialic acid assay for human tears was approximately 0.1 mm01 1-l. To evaluate the effect of tear collection method on sialic acid levels, tears were collected from ten healthy individuals with capillaries after stimulation with air and with cellulose sponges. Sialic acid values (corrected for free pyruvate) in capillary tears (range: 0.9-1.8 mmol 1-l) did not differ significantly from those found in sponge tears (range : 0.8-1.8 mmol 1-l) [Fig. l(a)]. Also no significant differences were found for free tear pyruvate (range: O-O.131 mm01 1-l : not shown in Fig. l), when comparing the two types of tears. Protein levels [Fig. l(b)] in sponge tears (range: 3.6-10.2 g l-l), however, were significantly higher as compared with capillary tears (range: 2.7-5.4 g 1-l) and consequently the amount of sialic acid per gram of protein was significantly lower [Fig. l(c)]. To investigate whether known tear proteins con-
SIALIC
ACID
IN HUMAN
TEAR
47
FLUID
(c 1
(b)
a)
0.4c
6
0.3t
4
0.25
om
3
0.20
0.6(
FIG. 1. Comparison of sialic acid (a), total protein (b) and sialic acid levels per g of protein (c) in tear fluid collected by the capillary (0) and the sponge method (0). Significance of difference of capillary vs. sponge: (a) not significant. (b) P < 0.005, (c) P < 0~005.
tribute to the sialic acid assay used, sIgA, lactoferrin, lysozyme and serum albumin were assayed for sialic acid. As can be seen in Table I these proteins were negative in the assay.
Centrifugation and Ultrafiltration
Sialic acid in capillary tear samples of four volunteers was measured before and after centrifugal TABLE I
Recovery Experiments In order to investigate if adherence of sialic acidcontaining material in tears takes place to the cellulose material, used for tear collection, sponges were soaked with capillary tears and sialic acid was determined in the regained fluid after centrifugation of the sponges. Table II shows both sialic acid and protein values in the regained tear fluid as a percentage of the original values, found in the tears prior to sponge incubation, No significant change in sialic acid and protein values in tears was observed after incubation with sponges,
of Human Tears
Sialic acid analysis in various protein samples Protein sample Tears Lactoferrin Secretory-IgA Lysozyme Serum albumin
.~
Sialic acid (mm011-l) -___ .-~~~~-.--~~_ 1.2
< < <
45
Sialic AB KUIZENGA*,
Acid
in Human
Tear Fluid
ERIC J. van AGTMAAL, NICOLAAS AND AIZE KIJLSTRA”
J. van HAERINGEN
The Netherlands Ophthalmic Research Institute, Biochemical Laboratory, P.O. Box 72741, 7 700 AC Amsterdam, The Netherlands and =Department of Ophthalmology, University of Amsterdam, the Netherlands (Received
13 April
1989 and accepted in revised form 77 August
7989)
A simple assay for the determination of sialic acid (N-acetylneuraminic acid) in human tear fluid was evaluated. Sialic acid, terminally bound on carbohydrate side-chains of glycoproteins, was releasedafter treatment with neuraminidase and measured by an enzymatic calorimetric test. Tear fluid sampleswere collected from ten healthy adults, using glass capillaries and cellulose sponges.Sialic acid levels in tears collected with sponges (08-1.8 mmol 1-l) did not differ significantly from those found in capillary tears (09-1.8 mmol 1-l). Sialic acid, expressed as mmol g-’ protein, was significantly lower in tears collected with sponges (018-0.32 mmol g-‘) than with capillaries (0.19-0.42 mmol g-l). Recovery of sialic acid and protein after incubation of cellulose spongeswith tears was more than 99 %. Sialic acid levels in human tears, which had been centrifuged to remove insoluble material, remained unchanged. Furthermore, tear sialic acid activity did not pass a filter with a molecular weight cut-off point of 10000. Our data indicate that with the assay used in this report, sialic acid in tears is not due to secretory immunoglobulin A (sIgA), lactoferrin and lysozyme. The fact that the major tear proteins do not contribute to the sialic acid levels detectedin tears suggests that other as yet unknown soluble glycoproteins are involved. Key words: sialic acid: tear fluid; glycoproteins
1. Introduction
The preocular tear film covers the cornea1 and conjunctival epithelia. Its functions are to maintain structural integrity of the cornea and conjunctiva and to protect the eye against external influences. The composition of human tear fluid is very complex (van Haeringen, 1981) and contains various substances, including lipids, electrolytes, proteins and muco-glycoproteins. The understanding of the mucoglycoproteins on the ocular surface is gradually increasing. The mucous layer, covering the epithelial surface of the cornea and conjunctiva and containing mucous glycoproteins (mucins) (Moore and Tiiany, 1979) is considered to maintain the stability of and contribute to the viscosity of the tear film (Kaura and Tiffany, 1986). Mucous glycoproteins are macromolecules of varying molecular weight and consist of a large linear protein core, to which many oligosaccharide side chains attach, containing hexoses and sialic acid. They can exist in water-soluble form and in an insoluble high molecular weight gel-form. The serumtype glycoproteins have a lower molecular weight and contain fewer carbohydrate side chains, attached to a
more globular protein core. Although the conjunctival goblet cells are the main source for mucous glycoproteins (Kessing, 1968), glycoproteins have also been found in lacrimal gland acini (Allen, Wright and Reid, 1972) and glyco-
2. Materials and Methods Collection of Tear Fluid
Tear samples were collected from healthy adults (n = 10) using 25+1 glass capillaries (capillary tears). Production of tear fluid was stimulated by a jet of pressurized air onto the cornea. Within a period of 15 min tear fluid production
* For correspondence. 00144835/90/010045+06
proteins in tears have been demonstrated by their hexosamine and sialic acid content (Cabezaset al., 1964; Dohlman et al., 1976), suggesting lacrimal secretion of these proteins onto the ocular surface (Kreuger, Sokoloff and Botelho, 1976). Water-soluble glycoproteins were detected in tears by electrophoresis (Berta and Tiirak. 19 8 6) and may be altered in various pathological states of the eye (Dohlman et al., 1976 ; Halken, Bog-Hansen and Prause, 1986a). Glycoproteins of lacrimal, together with those of conjunctival origin, can be important in maintaining the physicochemical properties of the precorneal tear film. Until now no rapid and simple method was available to quantitate glycoprotein levels in human tears. Therefore, the determination of sialic acid in human tear fluid, using a simple enzymatic colorimetric method, was evaluated as a parameter for tear glycoproteins. The major tear proteins do not contribute in the sialic acid assay used and it is concluded that sialic acid in human tears is mainly due to soluble glycoproteins of still unknown identity.
$03.00/O
ranged from 25 ~1 to
0 1990 Academic Press Limited
A. KUIZENGA
46
ET AL
100 ,uI. Samples were stored in 400-,~rl polyethylene tubes at -20°C. After at least 2 hr, tear samples were collected from the same subjects by means of surgical cellulose sponges (Sugi, Kettenbach, GDR) as described earlier (van Agtmaal et al., 1988) and the samples (sponge tears) were stored at - 20 “C. A pool of capillary tears was used for determination of the precision of the sialic acid assay.
Molecular Weight Limit) low-binding PLGC membrane (Millipore Corporation, USA) and the filtrates were assayed for sialic acid.
Assay of Sialic Acid in Human Tear Fluid
Recovery Experiments
Sialic acid was assayed, using a commercially available test kit (Boehringer, Mannheim, FRG). Briefly, the principle of this method is as follows : sialic acid, as a covalently bound terminal sugar residue of oligosaccharide side chains, is enzymatically hydrolyzed by neuraminidase (source : Arthrobacter ureafaciens ; specificity : hydrolyses glycosidically bound N-acetyl- and N-glycolylneuraminic acid) to free sialic acid and then cleaved to N-acetylmannosamine and pyruvate by N-acetylneuraminic acid aldolase. Formed pyruvate is then oxidized by pyruvate oxidase to acetylphosphate, CO, and H,O,. The amount of H,O, formed, which is equivalent to the free sialic acid, is determined by a peroxidase reaction, where a chromogen is converted to a red dye, of which the absorbance is measured at 550 nm. In our experiments assays were performed in l-ml microcuvettes (Sarstedt, FRG) and the quantities of reagents and samples were only a quarter of those stated in the manufacturer’s protocol. In each assay 5 ~1 of tear or control protein sample was used (incubation volume: 255 ~1; assay volume after addition of stabilizer: 755 ~1). A serum sample provided with the kit, containing a known amount of bound sialic acid, was used as a standard. The blank consisted of 5 ,~l of water. Proteins analyzed in the sialic acid assay included 5 g 1-l samples in phosphate buffered saline (PBS) of purified milk lactoferrin, secretory immunoglobulin A from human colostrum, chicken egg white lysozyme (all from Sigma Chemical Company, St Louis, MO) and human serum albumin. In a separate experiment the tear samples were assayed for free pyruvate, using an available test kit (Boehringer, Mannheim, FRG) and the sialic acid values were corrected by subtraction of the pyruvate values. In order to investigate if any insoluble material in human tears could contribute to sialic acid levels in this fluid, tear samples from four healthy volunteers, obtained with the capillary method, were centrifuged (15 min, 12 000 g, 20°C) and the supernatant was assayed for sialic acid together with the samples prior to centrifugation. Parallel to this experiment. the original four capillary tear samples were treated by centrifugal ultrafiltration (15 min, 12 000 g, 20°C) through an ultraFree-MC filter unit, 10 000 NMWI, (Nominal
To investigate whether adsorption of sialic acid containing tear proteins takes place to the cellulose material of the sponges, used for tear collection, the following experiment was performed. Reflex tear fluid was collected with glass capillaries, and sialic acid and protein content were measured afterwards. Five cellulose sponges (mean weight &-S.E.M. : 10.4 k 0.2 mg) were placed in 400~,~l polyethylene tubes and 50 ,~l of tear fluid was added and incubated at room temperature for 30 min. Subsequently, tear fluid was regained from the sponges by centrifugation as described earlier (van Agtmaal et al., 1987) and assayed for sialic acid and total protein. As a control, sponges were soaked with 50 ,~l of water.
Assay of Total Protein in Human Tear Fluid Total protein was assayed with Bradford’s method (Bradford, 19 76) modified according to Jongeneel (1982), using human serum albumin as a standard.
Statistics Significance of differences between groups of data was determined by using the Man-Whitney U-test.
3. Results Sialic Acid Assay in Human Tear Fluid Within the assay, the precision, expressed as coefficient of variation, was 2.1 Y0 (n = 20); day-today precision (duplicate assay, n = 14) was 3.5%. Detection limit in the sialic acid assay for human tears was approximately 0.1 mm01 1-l. To evaluate the effect of tear collection method on sialic acid levels, tears were collected from ten healthy individuals with capillaries after stimulation with air and with cellulose sponges. Sialic acid values (corrected for free pyruvate) in capillary tears (range: 0.9-1.8 mmol 1-l) did not differ significantly from those found in sponge tears (range : 0.8-1.8 mmol 1-l) [Fig. l(a)]. Also no significant differences were found for free tear pyruvate (range: O-O.131 mm01 1-l : not shown in Fig. l), when comparing the two types of tears. Protein levels [Fig. l(b)] in sponge tears (range: 3.6-10.2 g l-l), however, were significantly higher as compared with capillary tears (range: 2.7-5.4 g 1-l) and consequently the amount of sialic acid per gram of protein was significantly lower [Fig. l(c)]. To investigate whether known tear proteins con-
SIALIC
ACID
IN HUMAN
TEAR
47
FLUID
(c 1
(b)
a)
0.4c
6
0.3t
4
0.25
om
3
0.20
0.6(
FIG. 1. Comparison of sialic acid (a), total protein (b) and sialic acid levels per g of protein (c) in tear fluid collected by the capillary (0) and the sponge method (0). Significance of difference of capillary vs. sponge: (a) not significant. (b) P < 0.005, (c) P < 0~005.
tribute to the sialic acid assay used, sIgA, lactoferrin, lysozyme and serum albumin were assayed for sialic acid. As can be seen in Table I these proteins were negative in the assay.
Centrifugation and Ultrafiltration
Sialic acid in capillary tear samples of four volunteers was measured before and after centrifugal TABLE I
Recovery Experiments In order to investigate if adherence of sialic acidcontaining material in tears takes place to the cellulose material, used for tear collection, sponges were soaked with capillary tears and sialic acid was determined in the regained fluid after centrifugation of the sponges. Table II shows both sialic acid and protein values in the regained tear fluid as a percentage of the original values, found in the tears prior to sponge incubation, No significant change in sialic acid and protein values in tears was observed after incubation with sponges,
of Human Tears
Sialic acid analysis in various protein samples Protein sample Tears Lactoferrin Secretory-IgA Lysozyme Serum albumin
.~
Sialic acid (mm011-l) -___ .-~~~~-.--~~_ 1.2
< < <
