Origin and Biosynthesis of Human Tear Fluid Proteins

Peter T Jonssen and O. Paul van Dijsrerveld

In tear fluid, a large number of proteins can be detected with electrophoretic or immunologic techniques. The composition of serum proteins in tears resembles that of whole serum. By comparison of the protein patterns resulting from different sampling methods, it was shown that serum albumin is not present in the secretion of the lacrimal gland, but is mixed with the tear fluid in the conjunctival sac. The in vitro synthesis and excretion of more than 20 protein components by the lacrimal gland could be demonstrated. Among these were lactoferrin, tear-specific prealbumin, lysozyme, and secretory IgA. The complexity of the electrophoretic protein pattern of tear fluid can be explained from the combination of the secretory activity of the lacrimal gland and the leakage of serum proteins from the circulation into the tear fluid. Invest Ophthalmol Vis Sci 24:623-630, 1983

In electrophoretic analysis of tear fluid, at least 60 as their decrease after administration of calcium do- protein components were detected by Gachon et al.' besilate, a drug that reduces vasopermeability, we With some simplification, the tear fluid proteins can concluded that serum proteins are mixed with the be divided in three groups: proteins present in serum tear fluid by leakage from the conjunctival capil- as well as in tears, proteins present in tears but not laries.10 in serum, and proteins present in epithelial cells as The present study determines which components well as in tears. were the product of local synthesis in the lacrimal This last group of proteins can be detected in tear gland and which components originated from the samples, collected with filter paper or by other meth- serum within the complex protein pattern observed ods that cause slight epithelial damage.2 In general, after immunoelectrophoresis and SDS polyacryl- however, their concentration will be below the de- amide gel electrophoresis. tection limit of electrophoretic techniques, currently used for protein analysis. Materials and Methods Lactoferrin, lysozyme, and tearspecific prealbumin (TSPA) are the most important examples of proteins Reagents 13 present in tears, but not in serum to any extent. Unless stated otherwise all reagents were analytical The results of immunofluorescence studies of the lac- 4 5 grade, obtained from British Drug Houses Ltd., rimal gland ' and the simultaneous disappearance of United Kingdom. these proteins in the tear fluid of keratoconjunctivitis 36 sicca patients indicate a common origin of these Tear Samples proteins. Lysozyme was shown to originate from the lacrimal gland.7 Tear samples were collected from the conjunctival There is no agreement in the literature on the way sac on Whatmann 3 MM filter paper discs (6 mm in the serum proteins enter the tear fluid. Chao et al8 diameter) or in glass capillaries. Sometimes tear fluid could not exclude synthesis or modification of some was also collected in a glass capillary near the lacrimal of the serum proteins within the lacrimal gland. Grab- gland slightly above the lateral palpebral ligament or ner et al9 suggested that serum proteins are present from the meniscus above the lower lid. Volumes of in basic tear flow from the lacrimal gland. From the tear fluid collected were determined by weighing. strong coincidence of high serum albumin levels in A part of the tear samples was analyzed separately, tear fluid and hyperemia of the conjunctiva, as well and another part of the samples was pooled, dialysed by ultrafiltration on a Diaflo YM 5 membrane (Amicon Corp., USA), and lyophilized. From the Laboratory for Medical Anatomy and Embryology, State University Utrecht, Utrecht, and the Royal Dutch Eye Clinic, Utrecht, The Netherlands. , Antisera Submitted for publication June 1, 1982. Reprint requests to O. P. van Bijsterveld, MD, PhD, Royal Pooled normal serum was obtained from a clinical Dutch Eye Clinic, F. C. Dondersstraat 65, 3572 JE Utrecht, The laboratory, and normal tears were obtained from a Netherlands. group of volunteers. Human serum albumin was ob-

0146-0404/83/0500/623/$ 1.20 © Association for Research in Vision and Ophthalmology

623

Downloaded from jov.arvojournals.org on 09/25/2021 624 INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / May 1983 Vol. 24

tained from Sigma Chemical Corp., USA, and pu- moved from an antiserum against tear proteins by rified further by gel filtration on Sephadex G-100 passing the antiserum at a flow rate of 2 ml/hr (Pharmacia, Sweden) in a buffer containing 0.05 mol/ through the immobilized serum proteins. 1 Tris-HCl pH 7.5 and 0.15 mol/1 NaCl. Lysozyme Tear fluid was fractionated on the Sepharose-cou- was isolated from leukemic urine by the method of pled, 7-globulin fraction of an antiserum against hu- Alderton et al." As a final purification step, the ly- man serum proteins. Protein from pooled normal sozyme was chromatographed on Ultrogel Ac A 54 tears was dissolved in immunoadsorbent buffer "and (LKB, Sweden) in a buffer containing 0.05 mol/liter passed through the immunoadsorbent at a flow rate Tris-HCl pH 7.0 and 0.3 mol/1 NaCl. of 2 ml/hr, followed by buffer. The unbound protein Antisera against human serum proteins, serum al- fraction was dialysed by ultrafiltration on Diaflo YM bumin, tear fluid proteins, and lysozyme were raised 5 (Amicon) and lyophilized. Bound protein was in New Zealand White rabbits as described previ- eluted in the reverse direction with 0.1 mol/1 glycine- ously.3 HC1 pH 2.5 at a flow rate of 10 ml/h. The eluate was Antisera against human lactoferrin, secretory com- neutralized immediately with 1 mol/1 Tris with phe- ponent, and the a-chain of IgA were purchased from nol red as indicator, dialysed on Diaflo YM 5 and Behringwerke, German Federal Republic. lyophilized.

Immunodiffusion Tissue Culture A surgical specimen of tear gland tissue, obtained Immunodiffusion was performed on 75 X 25 mm from an epiphora patient was divided in pieces of glass slides covered with 4 ml of 1.5% purified Bacto 3 approximately 3 mm and washed three times at 4 Agar (Difco Laboratories, USA) in phosphate buff- C with Hanks' balanced salt solution (Flow Labora- ered saline pH 7.3. Sample wells were made with a tories, UK). Twelve pieces of tissue were transferred Gelman gelpuncher. Diffusion lasted for 48 hrs, after to a culture dish containing a medium of 0.6 ml (1.1 which the slides were washed for 24 hrs in several 14 MBq) [U- C] protein hydrolysate, specific activity portions of saline. The slides were dried under filter 2.18 GBq/milliatom C (Radiochemical Center, UK), paper at 40 C, stained with a solution of 5 g/1 Coom- and 2.4 ml Hanks' balanced salt solution with vita- assie Blue (Serva, GFR) in 9:9:2 water-ethanol-acetic mins for minimal essential medium (Flow) and 100 acid, and destained in the same solvent mixture. IU/ml penicillin and streptomycin (Difco Laborato- Serum albumin concentrations in tear samples were ries, USA). The tissue was incubated at 37 C for 24 determined by radial immunodiffusion.10 hrs. After this, the medium was dialysed by ultrafiltration on Diaflo YM 5 and lyophilized (pulse me- Electrophoresis dium). Six pieces of tissue were homogenized in 200 Immunoelectrophoresis and SDS-polyacrylamide fA H2O with an ultrasonic cell disruptor (Kontes, gelelectrophoresis was performed as described pre- USA) and centrifuged, after which the precipitate was 3 viously. extracted once more with 200 ^1 H2O. Pooled su- pernatants were lyophilized (tear gland extract [pulse]). Immunoadsorbents The remaining six pieces of tear gland tissue were transferred to a culture dish with 1 ml medium 199 Pooled normal serum was dialysed against water (Difco), containing 100 IU/ml penicillin and strep- and lyophilized. tomycin and solidified with three drops of 5% agar. The 7-globulin fraction of an antiserum against After incubation for 24 hrs at 37 C, the tissue was human serum proteins was isolated by precipitation extracted as described above (tear gland extract with 34% satured ammonium sulphate of pH 6.5. [chase]). The medium was homogenized with a Tef- After three precipitation cycles the precipitate was lon potter and centrifuged. The precipitate was ex- dissolved in water, dialysed, and lyophilized. tracted once more with 1 ml H2O, and pooled su- Twenty milligrams of protein was coupled to 1 g pernatants were dialysed by ultrafiltration on YM 5 CNBr activated Sepharose 4B (Pharmacia, Sweden) and lyophilized (chase medium). by the basic procedure, recommended by the manufacturer. The immobilized protein was equilibrated Autoradiography with immunoadsorbent buffer containing 0.04 mol/ Radioactive protein samples were analysed by im- 1 Tris-HCl pH 7.3, 0.15 mol/1 NaCl, 1 g/1 Triton X- munodiffusion and electrophoresis. Immunodiffu- 100, and 0.2 g/1 sodium azide and transferred to a sion and immunoelectrophoresis slides were dried 5-ml disposable syringe, plugged with cotton wool. routinely in air; SDS-polyacrylamide gels were dried Antibodies against serum components were re- under vacuum on Whatmann 3 MM filterpaper. Au-

Downloaded from jov.arvojournals.org on 09/25/2021 No. 5 HUMAN TEAR FLUID PROTEINS / Jonssen ond von Oijsrerveld 625

Fig. 1. Immunoelectrophoretic analysis of serum proteins in tear fluid. A, a: tear fluid; b: antihuman serum proteins antiserum; c: human serum. B, a: antitear proteins antiserum; b: human serum; c: antihuman serum proteins antiserum. The arrows indicate the position of 2: serum albumin; 5: IgG. B

toradiography was performed on dried slides or gels, Proteins Originating from the Lacrimal Gland using X-Omat x-ray film (Kodak, USA). The filmwa s Immunoadsorbents: The protein pattern of tear processed with DX-80 developer and FX-40 fixer fluid collected near the lacrimal gland is dominated (both from Kodak) according to the manufacturer's by three proteins: lactoferrin, tearspecific prealbu- instructions. mins, and lysozyme (Fig. 2). When tear fluid is passed through an immunoadsorbent of immobilized im- Results munoglobulin from an antiserum against human Serum Proteins in Tears With an antiserum against human serum proteins, Table 1. Serum albumin concentrations measured the immunoelectrophoretic separation patterns of in tear fluid collected with a glass pipette near the tear fluid and serum are very similar (Fig. 1A). Only excretory duct of the lacrimal gland resp. on a filter a few lines are lacking or are less prominent in the paper disc from the conjunctival sac. tear sample, indicating that it is not a highly specific Serum albumin process by which some proteins from the serum are concentration (mg/mf) in: passed to the tear fluid. The presence of many serum proteins in tear fluid Lacrimal Conjunctival can also be demonstrated by the presence of anti- Patient no. gland fluid sacjluid bodies against these proteins in an antitear fluid an- 1 OD* <0.1 0.7 tiserum. The immunoelectrophoretic separation pat- 1 OSj <0.1 0,8 tern of human serum with antitear fluid antiserum 2OD <0.1 0.5 is very similar to that developed with antihuman 2 OS <0.1 0.6 serum antiserum as shown in Figure IB. 3 OD 0.2 1.3 The serum proteins do not enter the tear fluid to- 3 OS 0.6 1.1 gether with the lacrimal secretion. Serum albumin is 4OD <0.1 0.6 not present in tear fluid collected with a glass pipette 4 OS <0.1 1.0 near the exit of the lacrimal gland, but it is readily 5 OD <0.1 1.5 detected by radial immunodiffusion in tear fluid col- 5 OS <0.1 3-0 lected on filter paper discs from the conjunctival sac. 6OD 0.4 1.3 Albumin concentrations in tear samples collected 6 0S <0.1 1.4 near the lacrimal gland and from the conjunctival sac 7 OD <0.1 0.4 of both eyes of 11 volunteers are compared in Table 7OS <0.1 0.1 1. The remarkable difference is not explained by the 8OD <0.1 <0.1 slight epithelial damage known to result from an in- 8 0S <0.1 0.6 sertion of a filter paper disc, as tear fluid collected 9OD <0.1 1,3 with a glass pipette from the conjunctival sac also 9OS <0.1 0.1 contains serum albumin, as shown in Figures 2A and 10 OD <0.1 0.5 B. The protein pattern of tear fluid collected by glass 10 OS <0.1 1.4 pipette from the tear meniscus above the lower lid, 11 OD <0.1 0.5 however, resembles much more that of lacrimal gland 11 OS <0.1 2.3

fluid than that of conjunctival sac fluid, especially • OD = right eye. after stimulation of the tearflow (Fig. 2C). t OS = left eye.

Downloaded from jov.arvojournals.org on 09/25/2021 626 INVESTIGATIVE OPHTHALMOLOGY 6 VISUAL 5CIENCE / Moy 1983 Vol. 24

A B II II

Fig. 2. Influence of sample collection method on gelelectrophoretic separation pattern of tear proteins. Samples were obtained from three healthy volunteers (A, B, and C) by means of I: a capillary pipette placed slightly above the lateral palpebral ligament; II: a capillary pipette from the conjunctiva] sac; III: a capillary pipette from the tear meniscus above the lower eye lid; and IV: a filterpape r disc from the conjunctival sac. The arrows indicate the position of 1: lactoferrin, 2: serum albumin, 3: tearspecific prealbumin (TSPA), and 4: lysozyme.

serum proteins, a protein pattern results that is iden- I I Iff tical to that of lacrimal gland fluid. In the fraction that elutes at low pH from the immunoadsorbent serum albumin can be detected among other proteins (Fig. 3). Antibodies reacting with serum components were removed from an antiserum against tear proteins by passing this antiserum through an immunoadsorbent of serum proteins coupled to CNBr-ac- Fig. 3. Immunoadsorbent tivated Sepharose. Immunoelectrophoresis of tear I fractionation of tear pro- fluid with this absorbed antiserum resulted in the for- teins. I: total tear proteins; mation of only four precipitation lines (Fig. 4). Apart II: tear proteins passing an from these four, more or less anodal lines, Gachon immunoadsorbent of im- et al1 could detect two other lines in the cathodal mobilized 7-gIobulins of an antiserum against serum region of the separation pattern, that were identified proteins; III: proteins bound as lysozyme and a cathodal line of PMFA (proteins i to the immunoadsorbent, moving faster than albumin in nondenaturing acryl- eluting at low pH. amide gelelectrophoresis). Our antitear antiserum fails to result in a visible precipitation line of lysozyme after immunoelectrophoresis, probably due to a low titer of the corresponding antibody; a cathodal PMFA line could not be identified.3 Tissue culture—SDS gelelectrophoretic analysis: Tear gland tissue was incubated for 24 hrs in Hanks' balanced salt solution, containing antibiotics, vita- mins, and 14C-labeled aminoacids (pulse), followed

Downloaded from jov.arvojournals.org on 09/25/2021 No. 5 HUMAN TEAR FLUID PROTEINS / Janssen and von Dijsrerveld 627

Fig. 4. Immunoelectropho- a retic analysis of tear proteins, not present in serum. A, a: tear A b fluid; b: antitear protein antiserum; c: human serum. B, a: c tear fluid; b: antitear protein antiserum, absorbed with serum proteins; c: human serum. The arrows indicate the positions of a 1: lactoferrin; 2: serum albumin; 3: TSPA; 5: IgG. B b c

by a 24-hr culture in medium 199 (chase). Tear fluid In pulse medium (Fig. 5-IV), apart from these pro- collected before surgery, as well as media and tissue teins, considerable amounts of serum albumin and extracts obtained ater 24- or 48-hr cultures, were an- of two protein bands migrating faster than lysozyme alyzed by SDS polyacrylamide gel electrophoresis, (molecular masses: 15,500 and 14,800) are found. followed by autoradiography (Fig. 5). These proteins are also present in tear gland extracts Incorporation of radioactivity is observed for nearly (Fig. 5-III, 5-V), but are not synthesized there as can all the protein bands of the tear gland extract, that be concluded from the corresponding autoradio- are visible after staining (Figs. 5-III, 5-V). When chase grams. Their low molecular mass and the equal stain- medium (Fig. 5-VI) is added to the preoperative tear ing intensity of the two bands suggested that these sample a blackening of the x-ray film is observed from bands were the a- and /3-chains of hemoglobin from lactoferrin, TSPA, lysozyme, and a group of protein erythrocytes, apparently not completely removed bands below albumin (Fig. 5-II). during the washing procedure of the lacrimal gland

I I IE BE 3L IT S S A SASASASA Fig. 5. Tissue culture of lacrimal gland. Lacrimal gland tissue was incubated 10 for 24-hr in a medium con- 8 taining l4C-labeled amino acids (pulse) followed by a 24-hr culture on a cold medium (chase). 1: preoperative tear sample; II: mixture of preoperative tear sample and chase medium (VI); III: lacrimal gland extract (pulse); IV: pulse culture medium; V: lacrimal gland extract (chase); VI: chase culture medium. S indicates the separation pattern after staining of the gel with Coomassie Blue; A indicates the autoradiogram obtained after exposing an x-ray film to the dried gel.

Downloaded from jov.arvojournals.org on 09/25/2021 628 INVESTIGATIVE OPHTHALMOLOGY 6 VISUAL SCIENCE / May 1983 Vol. 24

Fig. 6. Immunoelectrophoretic analysis of tissue culture medium. I: a: pulse and chase culture medium, mixed with pooled normal tear fluid; b: anti-tear protein antiserum; II: a: same as I-a; b: antitear protein antiserum, absorbed with serum proteins; III: a: same as I-a; b: antiserum proteins antiserum. S indicates the stained separation pattern; A indicates the corresponding autoradiogram. The numbering of the precipitation lines is as in Figure 4. Arrow 6 indicates the position of IgA, Arrow might indicate tysozyme. Note that the electrophoresis time for III was a longer than that for I and II in order M O to obtain an optimal separation. b

tissue. This conclusion was supported by SDS gel zyme), another 18 components can be distinguished electrophoresis of human erythrocyte extract and tear in the autoradiogram of chase medium. Their pres- fluid, in which the positions of the a- and /3-globin ence cannot be caused by necrosis of tissue as certain chains were found to be the same as those in Figure intensely stained and labeled tissue components are 5-1V. The incomplete removal of blood from the tis- not detected in the medium. Therefore, a selective sue can also explain the presence of serum albumin excretion of these components by the lacrimal gland in pulse medium. tissue into the culture medium must be assumed. From the autoradiograms of pulse and chase media Tissue culture. Immunologic analysis: Equal (Figs. 5-IV, 5-VI), it can be concluded that lysozyme amounts of pulse and chase media were mixed with and TSPA are rapidly synthesized and excreted in pooled normal tear fluid and subjected to immuno- vitro by lacrimal gland tissue. The incorporation of electrophoresis with antisera against tear fluid pro- radioactive label in lactoferrin, TSPA, and lysozyme teins, human serum proteins, and an antiserum found in chase medium is similar in magnitude, while against tear fluid proteins absorbed with serum pro- in the pulse medium only a small amount of radio- teins (Fig. 6). Radioactive label can be detected in active lactoferrin can be detected in comparison to about eight precipitation lines of the immunoelectro- the other two proteins. This is not caused by an ac- phoresis pattern developed with antitear fluid pro- cumulation of lactoferrin within the tissue because teins antiserum. They include among others TSPA, the amount of lactoferrin there (Fig. 5-III) is also low lactoferrin, IgA, IgG, and, faintly, serum albumin. with respect to that of lysozyme and TSPA. There- The radioactivity detected near the antiserum through fore, it can be concluded, that the rate of in vitro in the 7-region probably corresponds to lysozyme. synthesis and excretion of lactoferrin is less than that After absorption of this antiserum with serum pro- of lysozyme and TSPA. teins, the four precipitation lines remaining in the The separation pattern of chase medium after separation pattern all contained radioactive label. staining as well as autoradiography (Fig. 5-VI) resem- After immunoelectrophoresis with an antiserum bles closely that of lacrimal gland fluid (Fig. 2: A-I, against serum proteins, a very faint blackening of the B-I and C-I) and that of tear fluid after removal of x-ray film was found from albumin, IgA, and a line serum components (Fig. 3B). Apart from the three in the /3-region. major components (lactoferrin, TSPA, and lyso- The same antigen mixture was used in an immu-

Downloaded from jov.arvojournals.org on 09/25/2021 No. 5 HUMAN TEAR FLUID PROTEINS / Jonssen and van Dijsrerveld 629

Fig. 7. Immunodiffusion analysis of tissue culture medium. Sam- ple wells a-f contained antisera, sample well g contained the same antigen as Figure 61-a. a: antiserum proteins; b: anti-lgA-a-chain;c: antisecretory component; d: anti-lactoferrin; e: anti-lysozyme; f: antiserum albumin. S indicates the stained immunodiffusion pattern; A indicates the corresponding autoradiogram.

nodiffusion test with several different antisera (Fig. radioactive label in some serum proteins was also 7). The radioactive label incorporated in the precip- observed: slightly in albumin and IgG and more pro- itation lines formed with antisera against lysozyme nouncedly in IgA. This confirmed the results of Chao and lactoferrin confirms local synthesis within the et al,8 who also reported an incorporation of labeled lacrimal gland. The radioactivity in the precipitation glucosamine in immunodiffusion lines of the serum line formed with anti-IgA-or-chain and antisecretory proteins transferrin and ceruloplasmin. From this component antisera indicates the local production of they concluded that these serum-type glycoproteins secretory IgA. The spur in the latter precipitation line, were synthesized or modified by the lacrimal gland. pointing in the direction of the antisecretory com- After SDS gel electrophoresis, however, we could no ponent antiserum well, results from the presence of longer detect radioactive label in the serum albumin IgA in the antigen mixture, originating from the cir- band. This indicates that the radioactivity associated culation and, therefore, lacking secretory component with albumin after immunodiffusion or immuno- and of free IgA, produced locally, as indicated by the electrophoresis results from absorbed and not from radioactivity incorporated in the spur. Free secretory incorporated radioactive amino acids, thus ruling out component, resulting in the formation of a spur in local synthesis. This observation is in accordance with the precipitation line pointing towards the anti-IgA- the transport function of albumin for various com- a-chain antiserum well could not be detected. The pounds. horizontal precipitation line formed between the an- Whether the same phenomenon is responsible for tiserum wells containing antiserum albumin and an- the radioactivity in IgG immunoprecipitation lines ti lysozyme was caused by an interaction between cannot be excluded completely. The possibility of lo- these two antisera. A faint radioactive labeling was cal synthesis by plasma cells, however, is strongly sup- found once more for serum albumin. ported by the results of immunofluorescence localization of IgG and IgA in the lacrimal gland.45 In Discussion vitro synthesis of IgG and IgA in conjunctiva was observed by Lai A Fat et al.15 Although Balik12 reported the absence of proteins The local production of secretory IgA forms an in tear fluid collected from a ductus aberans, the idea important specific defense mechanism against micro- that the lacrimal gland is a major source for tear fluid bial attack of the mucous epithelium.l6 Therefore, the proteins is generally accepted. Indications for the se- IgA/IgG ratio in external secretions, such as tears, cretory activity of the lacrimal gland were obtained differs markedly from that in serum.17 Secretory IgA by immunofluorescence4'5 and by measurement of is not only synthesized in the conjunctiva,15 but also certain proteins—notably lysozyme—in tear fluid of in the lacrimal gland, as indicated by the strong in- patients suffering from the degeneration of the tear corporation of radioactivity, detected with antisera gland.13'14 against IgA-a-chain and secretory component (Figs. Proteins secreted by the lacrimal gland do not orig- 6, 7). No evidence was obtained for the presence of inate from the blood circulation, but are the product free secretory component in the culture medium. of local synthesis, as appears from the incorporation With electrophoretic techniques, more than 60 of radioactive label in these proteins. For lysozyme protein components can be detected in normal tear this confirms the study of Covey et al.7 In immu- fluid.1 We could show that more than 20 of these nochemical analysis of the proteins secreted by the components are secreted by the lacrimal gland. Apart lacrimal gland in vitro, however, an incorporation of from these, many serum proteins could be detected

Downloaded from jov.arvojournals.org on 09/25/2021 630 INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / Moy 1983 Vol. 24

in tear fluid, and, therefore, the complexity of the excellent technical help, Mr. T. Hulskes and Mr. A. M. van protein pattern might very well be explained by the Egeraat for the illustrations, and Mrs. J. C. J. G. Heidt-de combination of the secretory activity of the lacrimal Bruijn for the preparation of the manuscript. gland and the leakage of serum proteins from the circulation into the tear fluid.10 References Lysozyme concentration in tear fluid is determined 1. Gachon AM, Verrelle P, Betail G, and Dastugue B: Immu- easily by enzymatic assays, and, therefore, lysozyme nological and electrophoretic studies of human tear proteins. is used as a parameter for tear gland function.1314 As Exp Eye Res 29:539, 1979. 2. Van Haeringen NJ and Glasius E: The origin of some enzymes several other proteins are synthesized in the lacrimal in tear fluid, determined by comparative investigation with gland apart from lysozyme, it could be that among two collection methods. Exp Eye Res 22:267, 1976. them there are some that can perform this function 3. Janssen PT and van Bijsterveld OP: Comparison of electro- equally well or perhaps even better than lysozyme. phoretic techniques for the analysis of human tear fluid pro- Grabner et al9 and Mackie and Seal18 reported that teins. Clin Chim Acta 114:207, 1981. 4. Franklin RM, Kenyon KR, and Tomasi TB Jr: Immunohis- in normal tear samples collected from the conjunc- tologic studies of human lacrimal gland: Localization of im- tival sac on filter paper, there is no relation between munoglobulins, secretory component and lactoferrin. J Im- the volume of tear fluid absorbed by the disc and its munol 110:984, 1973. lysozyme concentration; ie, tear lysozyme concentra- 5. Gillette TE, Allansmith MR, Greiner JV, and Janusz M: His- tion is not dependent on the rate of tear production. tologic and immunohistologic comparison of main and ac- Such a relation, however, does exist for serum al- cesory lacrimal tissue. Am J Ophthalmol 89:724, 1980. 6. Janssen PT and van Bijsterveld OP: Pathophysiology of the bumin levels; high concentrations of that protein are tear film:protei n patterns in health and disease. In Chronische measured in "slow tear producers," low concentra- Conjunctivitis-trockenes Auge, Marquardt R, editor. Wien. tions in "fast tear producers." Grabner et al9 sug- New York, Springer Verlag, 1982, pp. 73-87. gested, that in fast tear producers "basic" tear fluid, 7. Covey W, Perillie P, and Finch SC: The origin of tear lysozyme. containing serum albumin, is diluted by a plasma Proc Soc Exp Biol Med 137:1362, 1971. 8. Chao CW, Vergnes JP, Freeman IL, and Brown SI: Biosyn- nitrate, containing only proteins, smaller than serum thesis and partial characterization of tear film glycoproteins. albumin. From the present results it can be con- Incorporation of radioactive precursors by human lacrimal cluded, that "basic" tears and reflex tears, as secreted gland explants in vitro. Exp Eye Res 30:411, 1980. by the lacrimal gland, are free of serum albumin. 9. Grabner G, Formanek I, Dorda W, and Luger T: Human tear Furthermore, dilution of "basic" tears by a plasma lysozyme. A comparison of electro-immunodiffusion, radial immunodiffusion and a spectrophotometric assay. Graefes ultrafiltrate, would lead to a decrease in the concen- Arch Clin Exp Ophthalmol 218:265, 1982. tration of lysozyme, which actually has not been 10. van Bijsterveld OP and Janssen PT: The effect of calcium do- 918 found. Therefore, we suggest that "basic" and re- besilate on albumin leakage of the conjunctival vessels. Curr flex tears, as secreted by the lacrimal gland, are es- Eye Res 1:425, 1981. sentially the same in chemical composition. The low 11. Alderton G, Ward WH, and Fevold HL: Isolation of lysozyme from egg white. J Biol Chem 157:43, 1945. serum albumin concentrations generally found in tear 12. Balik J: A note on the chemical analysis of the secretion of samples collected from the conjunctival sac of "fast the lacrimal gland. Cesk Oftalmol 8:167, 1952. tear producers" can be explained by a strong dilution 13. Regan E: The lysozyme content of tears. Am J Ophthalmol of serum proteins, leaking from the conjunctival ves- 33:600, 1950. sels, in a large volume of tear fluid, secreted by the 14. van Bijsterveld OP: Diagnostic tests in the sicca syndrome. Arch Ophthalmol 82:10, 1969. lacrimal gland. 15. Lai A, Fat RFM, Suurmond D, and van Furth R: In vitro synthesis of immunoglobulins, secretory component and com- Key words: tears, muramidase, lysozyme, lactoferrin, tear- plement in normal and pathological skin and the adjacent specific prealbumin, biosynthesis, tissue culture, gel elec- mucous membranes. Clin Exp Immunol 14:377, 1973. trophoresis, serum proteins, immunoglobulins, lacrimal 16. Knopf HLS, Blacklow NR, Glassman MI, Cline WL, and gland Wong VG: Antibody in tears following intranasal vaccination with inactivated virus. III. Role of tear and serum antibody in experimental vaccinia conjunctivitis. Invest Ophthalmol Acknowledgments 10:760, 1971. 17. McClellan BH, Whitney CR, Newman LP, and Allansmith The authors wish to thank Professor W. J. van Door- MR: Immunoglobulins in tears. Am J Ophthalmol 76:89, enmaalen for his interest in our work, Dr. H. G. Parren, 1973. from whom a surgical specimen was obtained, Mrs. J. B. 18. Mackie IA and Seal DV: Quantitative tear lysozyme assay in de Gruiter-Verschuure and Miss L. A. C. Weimar for their units of activity per microlitre. Br J Ophthalmol 60:70, 1976.

Downloaded from jov.arvojournals.org on 09/25/2021