Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Investigative Ophthalmology & Visual Science, Vol. 30, No. 8, August 1989 Copyright © Association for Research in Vision and Ophthalmology Electrophoresis Combined with Immunologic Identification of Human Tear Proteins P. K. Coyle,* Patrick A. 5ibony,*f and Cynthia Johnson* The protein content of normal human tears from five subjects was examined by molecular weight separation using SDS-polyacrylamide gel electrophoresis (PAGE) and by charge separation using agarose isoelectric focusing (IEF) gels. After separation, specific proteins were identified by immunoblot and immunofixation. Tear proteins examined included albumin, IgA, IgG, prealbumin, lactoferrin, lysozyme, secretory component and transferrin. These techniques required 1 to 14 /il of unconcentrated tears. We found SDS-PAGE superior to agarose IEF to examine total tear protein pattern, and silver stain almost ten-fold more sensitive than Coomassie blue stain. Immunologic staining markedly enhanced protein detection in all tear samples and appeared to offer the definitive method to probe for a specific protein in tears. In this study prealbumin and a portion of the IgG were present in normal tears at higher than expected molecular weight, suggesting they were present in complexed form. Prealbumin and secretory component staining showed marked variability between subjects. These techniques should be applicable to examine tear proteins in a variety of ocular disease states. Invest Ophthalmol Vis Sci 30:1872-1878,1989 The role of the tear film in human health and disease has come under increasing scrutiny.1 Accurate and sensitive characterization of tear components is crucial to interpreting the changes found in ocular surface disorders. Yet the techniques applied in tear studies must be reliable on very small volumes. Although a variety of electrophoretic techniques have been applied to standardize tear proteins, newer and more sensitive immunologic probes such as immunoblot and immunofixation have not. In this study we examined normal human tears for eight protein components using immunologic probes and electrophoresis based on molecular weight and isoelectric point separation. Informed consent was obtained from all individuals after the nature of the procedures was fully explained. There were four females and one male aged 23 to 43 years. Sample Collection Reflex tears were stimulated with aromatic ammonia and collected using glass capillary tubes gently inserted into the tear pool of the lower lid margin as previously reported.2 Care was taken not to touch or irritate the eye directly. Tears were stimulated until at least 100 /A had been collected; total collection time was 2 to 3 min. Final volumes from the five subjects ranged from 100 to 150 /A. Tears were stored at -20° or -70°C until use. Materials and Methods Subjects Tear Proteins Five normal subjects without clinical signs of eye disease were studied. None were on any medications. Total protein concentration was measured on 3 nl of tears using the Bradford method, as previously reported.3 Bovine albumin was used as the protein standard. The following proteins were examined in tears using immunologic identification as outlined below: albumin, IgA, IgG, lactoferrin, lysozyme, prealbumin, secretory component and transferrin. From the Departments of *Neurology and fOphthalmology, Health Sciences Center, SUNY at Stony Brook, Stony Brook, New York. Presented in part at the Annual Meeting of the Association for Research in Vision and Ophthalmology, Sarasota, Florida, May 1-6,1988. Supported by the National Eye Institute (grant EY-05736). Submitted for publication: June 6, 1988; accepted February 14, 1989. Reprint requests: Dr. P. K. Coyle, Department of Neurology, HSC T-12, SUNY at Stony Brook, Stony Brook, NY 11794. Tear Electrophoresis Two different techniques were used to study total tear proteins. To separate proteins by molecular weight, fixed protein concentrations of tears were 1872 Downloaded From: http://iovs.arvojournals.org/ on 08/12/2017 No. 8 1873 ELECTROPHORESIS OF HUMAN TEARS / Coyle er ol electrophoresed on 12% separating SDS-polyacrylamide gels (PAGE) using 5% stacking gels as previously reported.4 Because 12% gels will not give reliable resolution of proteins above 160 kilodaltons (kD), 7% separating gels were also used to analyze higher molecular weight tear proteins. Samples were run both reduced and unreduced, and a Tris-HCl with glycine buffer system was used. Commercial molecular weight standards (BRL Inc, Gaithersberg, MD; and Sigma Chemical Co, St. Louis, MO) were run in tandem and used to calibrate the molecular weights of tear bands. Gels were stained with Coomassie blue (CB) or silver as outlined below. To separate proteins by isoelectric point, fixed volumes of tears were run rather than fixed protein concentrations. In preliminary studies volume standardization produced optimal gel results, whereas protein standardization gave inconsistent results. Tears were diluted to 20 /A in deionized water and run on agarose isoelectric focusing (IEF) gels purchased from IsoLab, Inc. (Akron, OH). For the anode solution 0.5 M acetic acid was used, and for the cathode solution 0.5 M NaOH. Gels with a pH range of 3-10 were used to examine total proteins, IgG and prealbumin. Gels with a pH range of 7-10 were used to examine lysozyme. In all other cases gels from pH 3-7 were used. Gels were run on the commercial Resolve electrophoresis unit according to the manufacturer's instructions and stained as outlined below. Gels were calibrated by measuring the pH at 1 cm intervals from the anode using a micro-combination electrode. Gel Staining SDS-PAGE gels were stained with CB or silver. For CB staining 8 ng of total tear protein was run. This was reduced to 1 ng for silver staining. CB gels were fixed and stained in 0.1% CB, 50% ethanol, 10% acetic acid for 1 hr and destained in two to three changes of 30% isopropanol, 10% acetic acid. For silver staining gels were first stained with CB, then washed overnight in four changes of 30% isopropanol, 10% acetic acid. The staining method of Merril was then used.5 Agarose IEF gels were washed in 5% glycerol, dried at 70°C for 70 min, and then silver-stained using the Resolve-CSF kit reagents and method (IsoLab, Inc., Akron, OH). Protein Identification Specific proteins were identified in tears by immunologic probes. For these assays commercial human protein standards were obtained as follows: albumin (Sigma Chemical Co.), IgA and IgG (Tago, Inc., Burlingame, CA), lactoferrin, lysozyme and prealbumin (Calbiochem, La Jolla, CA), secretory component Downloaded From: http://iovs.arvojournals.org/ on 08/12/2017 (Cappel, Malvern, PA), and transferrin (Jackson Immunoresearch, West Grove, PA). Commercial primary antisera (affinity-purified antibodies were used when available) were obtained as follows: goat antialbumin, goat anti-IgA and goat anti-IgG (Tago, Inc.), rabbit anti-lactoferrin (Jackson Immunoresearch), rabbit anti-lysozyme, rabbit anti-prealbumin and rabbit anti-secretory component (Dako Corp., Santa Barbara, CA), and goat anti-transferrin (Cappel). In preliminary experiments cross-reactivity between the standards and antisera was examined by gel electrophoresis followed by immunoblot. No crossreactivity was noted except for the prealbumin antisera, which reacted with albumin, lactoferrin, lysozyme and transferrin. Prior to use the prealbumin antisera was preabsorbed with these standards until all cross-reactivity (by immunoblot) was removed. Immunoblot was used to identify tear proteins separated by SDS-PAGE.6'7 Five micrograms of total tear protein were run for all immunoprobes except for IgG, in which case 9 /*g were used. Proteins were electrophoretically transferred to a 0.1 ^m nitrocellulose membrane (Schleicher & Schuell, Keene, NH) at 100 vc for 1 hr. The membrane was blocked by exposure to 3% bovine serum albumin-phosphate-buffered saline for 90 min at 37°C and incubated with an optimal dilution of the primary antisera for 30 min at 37°C and 30 min at room temperature (RT). After several washes in PBS-Tween, the membrane was then incubated with horseradish peroxidase-conjugated anti-goat or anti-rabbit IgG antibodies (depending on the species-specific primary antisera) for 1 hr at RT. After washing the membrane was developed using 0.03% diaminobenzidine and 0.015% H2O2 in deionized water for 10 min at RT, blotted dry, and photographed on Panatonic-X black and white film using a Nikon FE 35 mm camera equipped with a Mikro-Nikkor 55 mm f/2.8 lens and green filter. Immunofixation was used to identify tear proteins separated by agarose-IEF as previously reported.8 Gels were overlayed with a cellulose acetate membrane saturated with undiluted primary antisera for 1 hr at RT. Proteins not bound by antibody were then washed out of the gel over a 24 to 48 hr period with 0.15 M NaCl, 5% glycerol. After a final wash with 5% glycerol, the gels were dried and silver-stained as reported above. Results Tear Protein Content Total protein concentration of the five samples fell within a fairly narrow range: 4.6, 6.2 (two subjects), 6.3 and 6.9 mg/ml. 1874 Vol. 30 INVESTIGATIVE OPHTHALMOLOGY 6 VISUAL SCIENCE / Augusr 1989 CB Stain Ag Stain 290 230 210- 1175-* 74 68 55' 39 31- 17.5 16.5' 14.5 14 13.5' 4C i V* Fig. 1. (A) Eight micrograms of normal tears were run on nonreducing 12% SDS-PAGE and stained with CB; three representative samples are shown. Seven bands were routinely visualized with occasional additional bands at 31 and 16 kD. (B) One microgram of normal tear was run as in A, then stained with silver (Ag). Three representative samples are shown. Five additional bands were noted routinely, with an occasional sample showing a faint 175 kD band. 0 A Tear Analysis Using SDS-PAGE Eight micrograms of total tear proteins were run on SDS-PAGE and stained with CB (Fig. 1A). Seven bands were routinely visualized: a very dense 14.5 kD band, moderately dense bands at 16.5, 17.5 and 68 kD, and faint bands at 55, 74 and 230 to 290 kD. Additional bands were noted in two samples at 31 kD and in one sample at 16 kD. Because of the increased sensitivity of silver staining, tears were run at a total protein concentration of 1 /ig (Fig. 1B). Silver staining accentuated the seven bands noted with CB as well as the 31 kD band, which now appeared in four samples. In addition all samples showed faint but discrete bands in the region of 13.5, 14, 39, 44 and 210 kD. Two samples showed an additional faint band at 175 kD. When a 7% separating gel was used (Fig. 2), there was improved resolution at the top of the gel with three discrete bands noted from 280-380 kD. However, bands below 35 kD were lost. Reduced tear samples run on both 7% and 12% separating gels showed less intense staining of bands at the top of the gel, with the appearance of a new dense staining at approximately 80-88 kD (Fig. 2). This would be in the region where lactoferrin would run. 7% 12% Protein Identification by Immunoblot 20097.468- 200- 43- 97.4 - 25.7- 68- I 18.414.3- 43- N N Fig. 2. Nonreduced (N) and reduced (R) tears were electrophoresed on 12% and 7% SDS-PAGE, then silver-stained. Known molecular weight standards are indicated on the left of each gel pair. Reduced tears show fainter staining at the top of the gel, and a new band at the approximate molecular weight of lactoferrin (arrow). The 7% gel gives better resolution of proteins at the top of the gel. Downloaded From: http://iovs.arvojournals.org/ on 08/12/2017 Tears were electrophoresed on SDS-PAGE, transferred to nitrocellulose and immunostained for specific proteins. The overall patterns noted are outlined in Table 1. Representative samples are shown in Figure 3 for all proteins separated on 12% gels, and in Figure 4 for higher molecular weight proteins (IgA, IgG, secretory component) separated on 7% gels. There were several unexpected findings. IgG stained at higher than expected weights (310 and 350 kD) in addition to the 150 kD for monomeric immunoglobulin (Fig. 4). This was not an artifact of storage, since fresh tear samples gave identical results. To exclude the possibility that IgG reagents were reacting with IgA, specific immunoblot studies were carried out at several concentrations and documented no cross-reactivity with up to 1 Mg of purified IgA standard. Prealbumin showed a marked variability in staining between tear samples. All samples stained faintly at 70 kD, higher than the 1.3 kD reported for prealbu- 1875 ELECTROPHORESIS OF HUMAN TEARS / Coyle er ol No. 8 Table 1. Tear proteins identified by SDS-PAGE and immunoblot* Protein molecular weight (kD) Tear volume Protein examined used (nl) Protein staining pattern Albumin IgA IgG Lactoferrin Lysozyme Prealbumin Secretory component Transferrin 0.8 0.8 1.5 0.8 0.8 0.8 0.8 0.8 one band one band three bands two dense bands one dense band; two faint bands two faint bands two-three bands one band Intersubject variability minimal moderate moderate minimal minimal marked moderate minimal 55 260-360 150,310,350 74,81 14.5, 23, 24 31,70 59, 78, 260-360 75 * Tears from five normal subjects were electrophoresed on nonreducng 12% or 7% SDS-PAGE gels, transferred to nitrocellulose, and immunoblotted for eight specific proteins. Molecular weight was calculated by co-elec- trophoresing known molecular weight standards. Intersubject variability was judged to be minimal, moderate or marked depending on the degree of conformity between thefivesamples. min subunits9 or the 21 to 24 kD reported for tearspecific prealbumin. 10 " Three samples showed an additional band at 31 kD and in one sample this band was quite intense. Secretory component in the 7% gel stained diffusely at 260-360 kD (bound to IgA), as well as at 78 kD and very faintly at 59 kD (free or precurser forms).12 Based on comparison of molecular weight and band intensity, it appeared that in the 12% gels the 14.5 kD band in CB- or silver-stained gels represented lysozyme; the 31 kD band represented prealbumin; the 55 kD band represented albumin; the 68 kD band represented lactoferrin and possibly prealbumin; the 74 kD band represented lactoferrin and possibly secretory component; and the 230 to 280 kD band represented IgA and IgG. The other CB- and silver-stained bands were not identified as any of the eight proteins probed for. Because of the enhanced gel staining noted on reduced tears, an immunoblot for lactoferrin was carried out on the same tear sample nonreduced and reduced. There was a marked increase in lactoferrin staining in the reduced sample (data not shown). (Fig. 5). Multiple bands were noted in all samples from pH 3.6 to 6.4. Staining was almost homogeneous from 5.0 to 6.0. There were major bands at pH 8.0 and 9.7, and occasional faint bands in the region ofpH7.3and8.4. Tear Analysis Using Agarose-IEF Ten microliters of tears were run on agarose-IEF gels (pH 3-10) and silver-stained for total protein Fig. 3. Tears were electrophoresed on nonreducing 12% SDS-PAGE, transferred to nitrocellulose and immunostained. Specific proteins examined were albumin (Alb), IgA, IgG, lactoferrin (Lact), lysozyme (Lys), prealbumin (PAlb), secretory component (SC) and transferrin (Trans). Representative samples for each protein are shown, along with the calibrated molecular weights. Protein Identification by Immunofixation Optimal volumes of unconcentrated tears were electrophoresed on agarose-IEF gels, then immunofixed for specific proteins. The volume for each protein had been determined in preliminary studies as giving optimal gel results. Tear patterns are noted in Table 2 and representative samples shown in Figure 6. Single bands were noted for albumin, lysozyme and transferrin. Lysozyme stained at pH 9.7 in this gel system, which only measured up to pH 10. This slightly underestimated its true pH range, which has been reported as 10 to 11." IgA showed multiple bands of varying intensity ranging from pH 4.4 to 6.1, whereas IgG by this technique showed a limited number of very faint bands in the pH range 4.7 to 8.0. Lactoferrin stained diffusely and very intensely over the pH 3.9 to 5.3 region. In this less sensitive technique tear prealbumin was highly variable: one subject showed no prealbumin, three had a single band at 270* 200** 240 >•190*- 310* 97-* 79-* 8174" 55 70- 75-** 2 ^ 23-* 14.5* Alb Downloaded From: http://iovs.arvojournals.org/ on 08/12/2017 IgA IgG Lact Lys PAlb SC Trans 1876 INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / August 1989 360260- pH 4.4 (consistent with the standard) and one subject had three faint bands at pH 4.4, 4.6 and 4.9. Secretory component showed marked variability when separated by isoelectric point with obvious variations in amount of staining. Multiple bands were noted from pH 4.15 to pH 6.7, with the densest staining in the pH 5.3 to 5.9 region. 360260- 350- 1507859- IgA IgG SC Fig. 4. Tears were electrophoresed on nonreducing 7% SDSPAGE, transferred to nitrocellulose and immunostained. Specific proteins examined on this gel were IgA, IgG and secretory component (SC). Representative samples and calibrated molecular weights are shown. Ag Stain PH 3.6 6.4 7.3 8.0 8.4 9.7 Fig. 5. Ten microliters of normal tears were run on agarose IEF gels, pH 3-10, and silver-stained. Three representative samples are shown. The pH gradient measured is noted at the left. Downloaded From: http://iovs.arvojournals.org/ on 08/12/2017 Vol. 30 Discussion Tears are made up of an aqueous layer covered by a thin lipid film. Their chemical composition is quite complex and includes over sixty proteins and polypeptides as well as electrolytes, enzymes, lipids, metabolites and mucin.13 Tears are produced by the innervated lacrimal gland along with contributions from minor accessory glands and the conjunctiva. With a baseline production rate of 1 to 2 jul per minute, and only 5 to 10 ^1 available in the eye at any given time,14 the absolute volume of tears is quite small. Therefore reflex tears produced by neural stimulation are generally used to provide a workable volume. Even so, the absolute amount obtainable from subjects varies, and this can become a limiting factor in tear studies. A number of investigators have analyzed tear proteins using a variety of methods. Electrophoretic studies have included disc electrophoresis,15 electroimmunoassay,16 dissociating and nondissociating PAGE,11131517"19 agar and agarose gel immunoelectrophoresis,11-15'19'20-23 IEF,1518 and two-dimensional electrophoresis.11'317 When immunologic means have been used to identify tear proteins they have not been used in conjunction with techniques that distinguished molecular weight or isoelectric point. To our knowledge, the current study is the first systematic analysis of tear proteins using the newer technology of silver stain, immunoblot and immunofixation. These studies were carried out on unconcentrated tears, avoiding the potential loss of protein. And they required only 1 to 14 /il volumes, compared to prior studies which had to pool samples or use up to 100 /ul oftears.13'15-17-19-21 We found separation of tear proteins by weight rather than charge to be more useful and interpretable as a screening examination, because many tear proteins had similar charges. This required two different separating gels (7% and 12%) since proteins over 160 kD were not sufficiently resolved on the 12% gel. As expected, silver staining was much more sensitive than CB. Silver staining permitted almost a 10-fold reduction in sample protein, yet resulted in enhanced detection. However, it was clear that immunologic staining was superior even to silver, and offered the definitive way to probe for specific tear proteins. It is simply not reliable to depend on co- 1877 ELECTKOPHORE5IS OF HUMAN TEARS / Coyle er QI No. 8 Table 2. Tear proteins identified by agarose IEF and immunofixation* Protein examined Tear volume used (pi) Albumin one band multiple bands faint bands homogeneous dense stain one band one-three faint bands multiple bands one band 5 5 14 0.5 3 10 6 10 IgA IgG Lactoferrin Lysozyme Prealbumin Secretory component Transferrin Protein pH range Protein staining pattern Intersubject variability moderate minimal minimal minimal minimal marked marked moderate 4.5 4.4-6.1 4.7-8.0 3.8-5.3 9.7 4.4, 4.6, 4.9 4.15-6.7 5.1 * Optima) unconccntrated tear volumes from five normal subjects were electrophorescd on agarose IEF gels and immunofixed for eight specific proteins. The pH of stained bands was measured using a microelectrode probe. Intersubject variability between thefivesamples was judged as noted in Table 1. electrophoresis to identify proteins, as prior studies have done. The immunologic staining was able to enhance detection of all proteins looked at, even those in low concentration which could not be identified on gels stained for total tear proteins. Our protein weights and charges were in fairly close agreement with those reported in the literature although there were some differences. We noted tear IgA to stain as a diffuse band from 260 to 360 kD on 7% gels. This is close to its expected weight of 380 kD, since tear IgA is known to be dimeric with an attached joining chain and secretory component.12 It is likely that some of the IgA entering our SDS-PAGE gels was partially dissociated. IgG is present at low levels in tears and is presumed to be monomeric. However in addition to the 150 kD diffuse band, there were two consistent bands at 310 and 350 kD. They did not appear to be due to aggregation, since the bands ran above an aggregated IgG standard. Tear IgG has been said to originate from serum.24 Our results suggest that part of the IgG in tears is routinely complexed. This would support it being tear-specific and emphasize a possible functional role. Fullard has pointed out that both lactoferrin and lysozyme may bind to immunoglobulins.25 We feel this is unlikely to account for our IgG results since lactoferrin or lysozyme did not stain in this region. It is apparent that a large proportion of tear lactoferrin is complexed, since it was of sufficient weight not to enter a gel until it was reduced. In the current study lactoferrin showed a more anodal charge range than previously reported.1115 We also found differences from previous reports with regard to tear prealbumin, which stained at 31 and 70 kD. Although our antiserum was directed against serum prealbumin, in preabsorption studies we removed all nonspecific staining. And the isoelectric point values we obtained for prealbumin were in close agreement with those reported in the literature. 10 " 15 Our 31 kD band suggests that "protein G" identified by Gachon may actually be prealbumin.17 We could not confirm a prealbumin 24 kD band. Our 70 kD band may represent joined subunits, since prealbumin, at least in serum, is known to exist frequently as four joined subunits.9 Alternatively, it could be that prealbumin is bound to albumin. This has been reported to occur for 3 to 7% of prealbumin isolated from human ocular mucus.26 To explain our prominent staining in tears, however, the percent bound would have to be much higher. 3.8 * 4.4 ... 4.' 4.5- 4.15* I* 4.9 Fig. 6. Tears were electrophoresed on agarose-IEF gels, then immunofixed for the eight specific proteins noted in Figure 2. Representative samples for each protein are shown along with the pH measured for stained bands. 5.1 5,3 *• 6.0* 67> — I 9-7 Alb Downloaded From: http://iovs.arvojournals.org/ on 08/12/2017 IgA IgG Loct Lys PAlb SC Trans INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / Augusr 1989 1880 incubation, 100 iA normal rabbit serum, 100 y\ antirabbit IgG goat serum and 200 /A 15% polyethylene glycol were added to each tube, followed by further incubation at room temperature for 1 hr. After centrifugation at 1600 g for 30 min, radioactivity in the precipitates was measured by a gamma counter. Intra-assay as well as inter-assay variations in RIA, estimated using samples of varied concentrations of standard hEGF, were within acceptable ranges. The minimum detectable dose was 0.1 ng/ml. No crossreactivity was observed with rat and mouse EGFs. Platelet-derived growth factor, human transforming growth factors-a and -0, human insulin-like growth factor-I, human insulin, porcine glucagon and ACTH also showed no cross-reactivities in this assay system. ng/ml 10- Vol. 30 OO 5- Radioreceptorassay (RRA) OO Tear non-reflex reflex IIIIIIIB Aqueous Humor Fig. 1. Concentrations of hEGF in tears and aqueous humor. Bars indicate mean values ± standard deviation. dilution ratio of urine (•—•) x80 x40 x20 x10 x5 95i dilution ratio of tears ( x4 x2 x1 xi/2 A431 cells derived from human epidermoid carcinoma were used as the EGF receptor source. Synthetic hEGF was used as standard and substrate for 125 I-hEGF. Standard diluent used for RRA was 50 raM sodium phosphate buffer (pH 7.4) containing 25 mM EDTA, 140 mM NaCl, 0.5% bovine serum albumin and 0.02% sodium azide. A mixture of 200 ^1 test sample or standard hEGF, 100 ix\ 125I-labeled hEGF and 200 /xl formalin-fixed A431 cell suspension (2 X 105 cells) in each assay tube was incubated at 25°C for 16 hr. After incubation, 1 ml ice-cold standard diluent was added to each tube. Following centrifugation at 1600 g for 30 min at 4°C, radioactivity in the precipitates was measured by a gamma counter. The sensitivity of this RRA was 0.4 ng/ml. Gel Chromatography Procedure for Pooled Human Tears o A Sephadex G-50 Superfine (Pharmacia, Piscataway, NJ) column (1.6 X 37 cm) was equilibrated with 1 M acetic acid. Aliquots of tears and urine were collected from one individual. Two milliliter test sample or standard hEGF was applied to the column and eluated at a flow rate of 66 ml/hr. Fractions (0.9 ml each) were collected and lyophilized; hEGF in each fraction was measured by RIA as described above. m 5) 50" Results 5- 10 20 40 80 160 320 640 1280 hEGF p g / t u b e ( •—• ) Fig. 2. Competitive binding curve of hEGF in tears in radioimmunoassay. Dilution curves generated by tear and urine samples, from one normal volunteer, and standard hEGF. B/Bo: bound/ total bound ratio. Downloaded From: http://iovs.arvojournals.org/ on 08/12/2017 Immunoreactive (IR) hEGF was detected in all human tear samples tested (n = 26). The level of IR-hEGF in reflex tears (n = 15) varied from 0.7 ng/ml to 8.1 ng/ml, with a mean (±SD) value of 3.4 (±2.9) ng/ml (Fig. 1). IR-hEGF was detectable at similar concentrations (1.9 to 9.7 ng/ml, with a mean value of 5.3 (±2.9) ng/ml), in non-reflex tears (n = 11) as well. In contrast, IR-hEGF was undetectable in aqueous humor by the present RIA system. 1881 EGF IN HUMAN TEARS / Ohoshi er ol No. 8 Figure 2 shows the standard curve of the hEGF RIA and the dilution curves of urine and tear samples obtained from one individual. As demonstrated, the curves for such samples paralleled the standard curve, indicating that the activity in the samples measured by the present RIA system was indistinguishable from that of standard hEGF. The competitive binding curves of the urine and tear samples in the present RRA paralleled that for standard hEGF as well (Fig. 3), demonstrating that tear EGF was biologically active. The concentrations of tear EGF estimated by RIA and RRA were at similar levels: 0.73 ng/ml and 0.68 ng/ml, respectively. Gel filtration study using Sephadex G-50 Superfine revealed that the IR-hEGF in tears was eluated as a single peak at the same position as standard hEGF or urine EGF (Fig. 4). recombinant hEGF Vo Vt I * Urine 1- > o (0 <D O § 0.15-1 Tears E 0.10-1 Discussion The current study clearly demonstrates that an EGF which is biologically, immunologically and biochemically indistinguishable from urine EGF or standard hEGF is present in human tears. The level of EGF in tears is lower than in urine,10 saliva and milk," almost comparable to the level in plasma,12 dilution ratio of urine (•• X64 95-, X32 X16 X8 dilution ratio of tears (AX2 X1 X1/2 50ID 5J 80 160320 640 1280 2560 5120 hEGF pg/tube (•—•) Fig. 3. Competitive binding curve of hEGF in tears in radioreceptorassay. Dilution curves generated by tear and urine samples, from one normal volunteer, and standard hEGF. B/Bo: bound/ total bound ratio. Downloaded From: http://iovs.arvojournals.org/ on 08/12/2017 0.05OJ 30 40 50 60 70 80 fraction No. Fig. 4. Sephadex G-50 Superfine gel exclusion chromatography of tear and urine samples from one normal volunteer. Concentration of hEGF was determined by radioimmunoassay. Vo: void volume, recombinant hEGF: elution volume of l25I-labeled recombinant hEGF. V,: total volume. and higher than in cerebrospinal fluid13 and pancreatic juice.14 Human tear EGF was shown by RRA to be biologically active. In contrast to our results, Elliott previously failed to demonstrate EGF immunoreactivity in human tear film.17 This discrepancy may simply be due to the fact that the present RIA system is 10 times more sensitive than that used in Elliott's study. The above finding is of great importance in view of the fact that the ocular surface epithelium is always in contact with tear fluid. Similarly, vitamin A has been shown to be present in tears as a form of retinol.18 Thus, tear fluid may constitute an environment in which ocular surface epithelia come into contact with a variety of nutritional and/or growth factors. It has been shown that corneal epithelial cells have receptors for EGF19 and proliferate upon topical application of EGF.3'4'6-7 It may be that tear EGF is related to the maintenance or differentiation of surface epithelia at rest, and to the acceleration of epithelial regeneration after surface wounding. The origin and biological significance of EGF in tear fluid awaits further investigation. IR-hEGF was undetectable in human aqueous humor. Corneal endothelial cells have been shown to possess EGF receptors20 and to undergo mitosis when 1882 INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / August 1989 cultured in the presence of EGF.8 It seems possible, therefore, that corneal endothelial cells might proliferate when EGF is introduced into the anterior chamber, although there might be risk of Schlemm's canal or trabecular cell responses as well. Key words: epidermal growth factor (EGF), tear, aqueous humor, human, radioimmunoassay, radioreceptorassay 9. 10. 11. Acknowledgment The authors wish to thank Mr. Robert Brady for his editorial assistance. References 1. Cohen S: Isolation of a mouse submaxillary gland protein accelerating incisor eruption and eyelid opening in the new-born animals. J Biol Chem 237:1555, 1962. 2. Cohen S and Elliott GA: The stimulation of epidermal keratinization by a protein isolated from the submaxillary gland of the mouse. J Invest Dermatol 40:1, 1963. 3. Ho PC, Davis WH, Elliott JH, and Cohen S: ICinetics of corneal epithelial regeneration and epidermal growth factor. Invest Ophthalmol 13:804, 1974. 4. Savage CR Jr and Cohen S: Proliferation of corneal epithelium induced by epidermal growth factor. Exp Eye Res 15:361, 1973. 5. Gospodarowicz D, Mescher AL, Brown KD, and Birdwell CR: The role offibroblastgrowth factor and epidermal growth factor in the proliferative response of the corneal and lens epithelium. Exp Eye Res 25:631, 1977. 6. Daniele S, Frati L, Fiore C, and Santoni G: The effect of the epidermal growth factor (EGF) on the corneal epithelium in humans. Graefes Arch Klin Exp Ophthalmol 210:159, 1979. 7. Singh G and Foster CS: Epidermal growth factor in alkaliburned corneal epithelial wound healing. Am J Ophthalmol 103:802, 1987. 8. Gospodarowicz D, Mescher AL, and Birdwell CR: Stimulation Downloaded From: http://iovs.arvojournals.org/ on 08/12/2017 12. 13. 14. 15. 16. 17. 18. 19. 20. Vol. 30 of corneal endothelial cell proliferation in vitro by fibroblast and epidermal growth factors. Exp Eye Res 25:75, 1977. Fabricant RN, Salisbury JD, Berkowitz RA, and Kaufman HE: Regenerative effects of epidermal growth factor after penetrating keratoplasty in primates. Arch Ophthalmol 100:994, 1982. Dailey GE, Kraus JW, and Orth DN: Homologous radioimmunoassay for human epidermal growth factor (urogastrone). J Clin Endocrinol Metab 46:929, 1978. Starkey RH and Orth DN: Radioimmunoassay of human epidermal growth factor (urogastrone). J Clin Endocrinol Metab 45:1144, 1977. Hirata Y, Moore GW, Bertagna C, and Orth DN: Plasma concentration of immunoreactive human epidermal growth factor (urogastrone) in man. J Clin Endocrinol Metab 50:440, 1980. Hirata Y, Uchibashi M, Nakajima H, Fujita T, and Matsukura S: Presence of human epidermal growth factor in human cerebrospinal fluid. J Clin Endocrinol Metab 55:1174, 1982. Hirata Y, Uchibashi M, Nakajima M, Fujita T, and Matsukura S: Immunoreactive human epidermal growth factor in human pancreatic juice. J Clin Endocrinol Metab 54:1242, 1982. Yanaihara N, Yanaihara C, Nokihara K, Iguchi K, Fukata S, Tanaka M, Yamamoto Y, and Mochizuki T: Immunochemical study on PHI/PHM with use of synthetic peptides. Peptides 5:247, 1984. Ohgai H, Kumakura T, Adachi S, Ohshiden K, Tanino M, Kawai S, Mihara S, Yanaihara C, and Yanaihara N: Genetic engineered human epidermal growth factor and its derivatives. Can J Physiol Pharmacol 64:suppl. p56, 1986. Elliott JH: Epidermal growth factors: In vivo ocular studies. Trans Am Ophthalmol Soc 78:629, 1980. Ubels JL and MacRae SM: Vitamin A is present as retinol in the tears of human and rabbits. Curr Eye Res 3:815, 1984. Frati L, Daniele S, Delogu A, and Covelli I: Selective binding of the epidermal growth factor and its specific effects on the epithelial cells of the cornea. Exp Eye Res 14:135, 1972. Fabricant RN, Alpar AJ, Centifanto YM, and Kaufman HE: Epidermal growth factor receptors on corneal endothelium. Arch Ophthalmol 99:305, 1981.