Investigative Ophthalmology & Visual Science, Vol. 33, No. 10, September 1992Copyright © Association for Research in Vision and Ophthalmology

Quantitative and Qualitative Analyses of Transferrin inAqueous Humor From Patients With Primary

and Secondary Glaucomas

Ramesh C. Tripathi,* Navaneet S. C. Dorisuth,* Drenda J. Tripathi,* and Stelios 5. Gotsisf

By using a highly sensitive and specific radioimmunoassay and the slot-blot technique, transferrin wasquantified in fresh samples of aqueous humor from patients with primary open-angle glaucoma(POAG, n = 36) or secondary glaucoma (SG, n = 18). The levels were compared with those in aqueoushumor obtained from age-matched patients without glaucoma (n = 33) and in primary and secondaryaqueous humor from normal dogs (n = 10) in which breakdown of the blood-aqueous barrier wasinduced experimentally. The concentration of transferrin in the aqueous humor of human controlsubjects ranged from 0.3-3.4 mg/dl (mean ± standard deviation, 1.36 ± 0.66 mg/dl); in POAG sam-ples, from 0.34 to > 10 mg/dl (2.07 ± 1.90 mg/dl); and in SG samples, from 0.29 to > 10 mg/dl (2.78± 2.24 mg/dl). The level of transferrin in secondary aqueous humor samples obtained from dogs was asmuch as ninefold greater than that in primary aqueous humor. The protein concentration in the humanaqueous humor samples was 11.87 ± 4.47 mg/dl for control subjects, 62.11 ± 56.74 mg/dl for patientswith POAG, and 124.53 ± 152.67 mg/dl for those with SG. In dogs, the protein levels were 7.97 ± 3.12mg/dl for primary aqueous humor and 191.9 ± 149.8 mg/dl for secondary aqueous humor. A significantcorrelation (r = 0.744, P < 0.01) was found between total protein and transferrin contents in thesamples of aqueous humor from patients with glaucoma but not in the samples from age-matchedcontrol subjects. The authors propose that the breakdown of the blood-aqueous barrier in POAG, SG,and other ocular diseases is responsible for an influx of transferrin into the anterior segment of the eyeand that this may be implicated in the pathogenesis of various intraocular disorders. Invest OphthalmolVir "-! ---«"»" -°73,1992

The physical and chemical composition of aqueoushumor is determined primarily by the blood-aqueousbarrier (BAB),1 the stability of which defines the con-centrations of different growth-modulating factors inthe aqueous humor.2-3 Although previous investiga-tions found changes in the protein content of aqueoushumor after breakdown of the BAB,4"6 specific knowl-edge about the types and amounts of growth-regula-

From the *Department of Ophthalmology and Visual Science,The University of Chicago, Chicago, Illinois, and the tDepartmentof Ophthalmology, University of Athens, Athens, Greece.

Supported by Public Health Service award EY08707 from theNational Eye Institute, Bethesda, Maryland, and, in part, by theVision Research Foundation, Alcon R & D, and the GlaucomaFoundation, New York, New York. NSCB is the recipient of afellowship award from Merck Sharp and Dohme, Rahway, NewJersey.

Submitted for publication: February 2, 1992; accepted April 15,1992.

Reprint requests: Professor Ramesh C. Tripathi, MD, PhD, Vi-sual Sciences Center, The University of Chicago, 939 East 57thStreet, Chicago, IL 60637.

tory substances in this condition and in diseases suchas primary and secondary glaucomas has been limitedby the availability of fresh samples and the sensitivityof the techniques used.

Transferrin is a member of a large family of iron-binding proteins that transports ferric iron and possi-bly zinc between the sites of absorption, storage, andutilization.4 It is an ideal candidate molecule for studyfor several reasons. Because transferrin is present inblood at a concentration that is approximately 200-fold higher than in aqueous humor,7 it is a particu-larly good marker for studying the integrity of theBAB. Unlike other growth-promoting substances,transferrin is an essential requirement for cell matura-tion and growth,89 and it also is a growth factor inde-pendent of its iron-transporting properties.10>l1 Incombination with other growth-modulating sub-stances, transferrin regulates the growth and mainte-nance of many cells of the anterior segment of the eyein vivo and in vitro,1213 and it also has been impli-cated in the pathophysiologic changes of glaucoma.3

Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933164/ on 04/02/2018

No. 10 TRANSFERRIN IN AQUEOUS HUMOR / Triparhi er ol 2867

In this study, we used highly sensitive and specificradioimmunoassay (RIA) and slot-blot techniques todetermine the concentrations of transferrin in individ-ual samples of aqueous humor from patients with pri-mary open-angle glaucoma (POAG) and secondaryglaucomas (SG). We compared these levels to those innormal aqueous humor from age-matched patientswithout glaucoma. In addition, we investigatedchanges in the content of transferrin in a caninemodel in which breakdown of the BAB was inducedexperimentally. We present evidence that variationsin the concentration of transferrin in plasmoidaqueous humor correlate with the concentration ofprotein in this fluid and can be attributed to break-down of the BAB.

Materials and Methods

Collection of Human Aqueous Humor

After informed consent, we collected samples ofaqueous humor in the operating room from patients(age range, 24-89 yr) who underwent filtration sur-gery for advanced POAG (36 patients) or surgery andanterior chamber tap for SG (18 patients). The lattergroup consisted of nine patients with inflammatoryfibrinous exudates after cataract surgery, three withexfoliative glaucoma, three with neovascular glau-coma, one with traumatic glaucoma, one with phaco-morphic glaucoma, and one with steroid-inducedglaucoma. For control purposes, we obtained samplesof aqueous humor from 33 age-matched human sub-jects without glaucoma who underwent elective cata-ract extraction. We used three criteria to obtain ashomogeneous a population as possible for the controlerouD: (1) the samples were chosen at random; (2) thepolypeptide profile and the concentrations of proteins(mean ± standard deviation, 12.4 ± 2.0 mg/dl) werenormal as defined;2 and (3.) the method of collectionof the samples conformed to that described.

Before cataract surgery, the pupil was dilated withtopical 2.5% epinephrine and 1% tropicamide. All pa-tients with POAG, exfoliative giaucoma, and steroid-induced glaucoma were receiving a therapeutic regi-men that consisted of topical pilocarpine (4% fourtimes a day), timolol (0.5% twice a day), and dipive-frin (0.1% twice a day). The patients with inflamma-tory fibrinous exudates received dexamethasonedrops (every 2 hr); those with neovascular glaucomareceived prednisolone (1% four times a day) and sco-polamine (0.25% twice a day). The patient with pha-comorphic glaucoma was treated with 1% predniso-lone (every 2 hr). The method used to collect aqueoushumor from control subjects and patients with POAGor SG was identical. All samples of aqueous humor

were obtained by limbal paracentesis. We used a 30-gauge needle attached to a tuberculin microsyringe(Becton Dickinson, Rutherford, NJ) and aspirated50-150 jtl of aqueous humor within 2-5 sec from thecentral pupillary area without touching the iris, lens,or corneal endothelium.2 The samples were placed im-mediately in liquid nitrogen and subsequently storedat -90°C until analyzed.

Collection of Primary and Secondary AqueousHumor From Dogs

We obtained fresh samples of aqueous humor fromten eyes of normal adult dogs anesthetized with pen-tobarbital. The. collection method was identical tothat used for obtaining aqueous samples in our hu-man subjects. For primary aqueous humor, approxi-mately 300 A*l of fluid was aspirated from the anteriorchambers of dogs by limbal paracentesis with a 30-gauge needle. Breakdown of the BAB was induced byan abrupt fall in the intraocular pressure caused bylimbal paracentesis,14"17 and the resulting secondaryaqueous humor was aspirated after time intervals thatranged from 7 min to 1 hr. All studies on animalsconformed to the ARVO Resolution on the Use ofAnimals in Research.

Total Protein Assay

The total protein content of all samples of aqueoushumor was measured with the Bio^Rad protein mi-croassay (Richmond, CA) as described previously.218

Briefly, 20 /xl of individual samples was diluted to 0.8ml with distilled water and combined with 0.2 ml ofdye reagent concentrate (Coomassie blue G-250).Spectrophotometric readings were taken at an opticaldensity of 595 nm with a Spectronic 20 UVD spectro-photometer (Milton Roy, Rochester, NY).

RIA for Transferrin

The RIA of individual samples of human aqueoushumor was done with a highly specific and sensitivehomologous system (Amersham, Arlington Heights,IL) according to methods described in the manufac-turer's protocol.19 The standards, antiserum, andtracer were reconstituted in an assay buffer (50 mmol/1 sodium phosphate, pH 7.4, with 0.2% bovine serumalbumin). We incubated 100 fA of standard trans-ferrin (iron-saturated or -nonsaturated) solution or ofthe diluted sample of aqueous humor with 100 fi\ ofsheep anti-human transferrin antiserum and 100 /il of125I-transferrin (approximately, 13,000 counts/min)for 24 hr at 4°C. Next, 100 n\ of polyclonal donkeyanti-sheep immunoglobulin G was added, and themixture was allowed to incubate for 2 hr at room tern-

Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933164/ on 04/02/2018

2868 INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / September 1992 Vol. 33

perature. After centrifugation at 1500 X g at 4°C for20 min, the supernatant was decanted, and the radio-activity of the precipitate was counted in a gammacounter (Packard, Meriden, CT).

In addition, ten samples of aqueous humor fromfive glaucomatous patients and five without glau-coma and transferrin standards at three concentra-tions (1, 5, and 10 ng) were deglycated by incubationfor 5 hr at 37°C in a buffer (50 mmol/1 sodium ace-tate, 154 mmol/1 sodium chloride, 9 mmol/1 calciumchloride, 0.05% [w/v] sodium azide, 25 mg/1 bovineserum albumin, pH 5.5) containing 0.01 U ml"1 ofneuraminidase from Vibrio cholera (Boehringer-Mannheim, Indianapolis, IN). The treated samplesthen were assayed according to the RIA proceduredescribed. The specificity of the antiserum was deter-mined against various growth-modulating factors andhormones, such as epidermal growth factor (Diagnos-tic Systems Laboratories, Webster, TX), transforminggrowth factor-j81 (Oncomembrane, Seattle, WA), and-a (Oncogene Science, Manhasset, NY), and basic fi-broblast growth factor (R & D Systems, Minneapolis,MN). To assess the accuracy of the technique further,we did a recovery study. Selected samples of normaland glaucomatous aqueous humor that contained dif-ferent levels of endogenous transferrin were spikedwith various amounts of standard transferrin (0.5, 10,and 100 ng) and assayed.

We also conducted control studies to determinewhether the drugs that were used topically by the hu-man subjects could interfere with the binding be-tween transferrin and the antiserum in the RIA. Forthese experiments, transferrin standards were spikedwith each drug at a concentration in excess of thatexpected in the aqueous humor.2021

Sodium Dodecyl Sulfate-Polyacrylamide GelElectrophoresis and Western Blotting

Selected individual samples of human aqueous hu-mor (10 fi\) or pure transferrin standard (Boehringer-Mannheim) were solubilized in one volume of Trisbuffer (70 mmol/1, pH 6.8), 2% sodium dodecyl sul-fate, 3% 2-mercaptoethanol, 10% glycerol, and 0.01%Pyronin-Y (Sigma, St. Louis, MO). Before electropho-resis, the solubilized samples were heated at 100°C for5 min. The polypeptide f ions were separated ongradient 10-15% T minigeis (43 X 50 X 0.45 mm).Each lane was loaded with 4 /A of aqueous humorsamples, and these underwent electrophoresis withPhastgel buffer strips (Pharmacia, Piscataway, NJ)composed of 0.20 M tricine, 0.20 M Tris, and 0.55%(w/v) sodium dodecyl sulfate for 60 V-hr at 10.0 mA/

3.0 W/15°C. The polypeptides separated on the un-stained gels then were blotted by diffusion onto nitro-cellulose at 70°C overnight and immunostained asdescribed.

Slot-Blot Technique

Because the antiserum used in the RIA does notcrossreact with transferrin in canine samples ofaqueous humor, we used a rabbit antitransferrinserum (Dakopatts, Copenhagen, Denmark) and aslot-blot technique to quantify transterrin in both hu-man and dog samples.2223 Individual samples (5 n\) ofaqueous humor or pure transferrin (range, 1.95-1000ng) were suspended in 50 /ul of 0.1% (w/v) nonfat drymilk in Tris-buffered saline (TBS, 0. i5 mol/1 NaCland 0.01 mol/1 Tris Cl, pH 7.6). The samples and puretransferrin were pipetted onto nitrocellulose (0.2 nm)in a "slot" configuration by means of individual wellsin a filtration manifold (Hoefer, San Francisco, CA).We applied additional samples of 0.1% nonfat drymilk in TBS to the nitrocellulose as a blank test spotto exclude the possibility of nonspecific backgroundbinding. A low vacuum (approximately, 5 mmHg)was applied for 5-10 min for pulling the samplesthrough the nitrocellulose. The individual wells thenwere washed (three times for 10 min each) with 100 /xlof TBS. The nitrocellulose filter, which containedrows of slots, was removed from the manifold and cutinto thin strips for all subsequent immunostaining.

Immunostaining Procedure and QuantitativeAnalysis of Transferrin

The nitrocellulose strips were blocked for 1 hr atroom temperature in TBS containing 0.5% nonfat drymilk. After one washing with a saline solution (NaCl0.15 mol/1), the strips were incubated, with continu-ous rocking, in rabbit anti-human transferrin anti-body (1:50 dilution in 0.1% nonfat dry milk in TBSfor 3 hr at 25°C; Dakopatts). Next, the strips werewashed in saline (five times for 5 min each) and ex-posed to peroxidase-conjugated swine anti-rabbit im-munoglobulins (1:300 in 0.1% nonfat dry milk in TBSfor 1.5 hr at 25°C). After repeated washing in TBS,the nitrocellulose was incubated with diaminobenzi-dine chromogenic solution (40 mg in 100 ml of TBSwith 0.03% H2O2). A positive reaction was indicatedby the precipitation of a brownish-red chromogen.The specificity of this antibody against pure trans-ferrin, serum and aqueous humor transferrin, and thetau fractions of aqueous humor and cerebrospinalfluid was characterized previously.27

Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933164/ on 04/02/2018

No. 10 TRANSFERS IN AQUEOUS HUMOR / Triporhi er ol 2869

Table 1. Comparison of protein and transferrinlevels in primary and secondary aqueous humorobtained surgically from dogs

Dog no.

123456789

10

Interval(min)

60201515306030257

30

Total protein ofaqueous humor

(mg/dl)

Primary

9.65.05.95.07.75.0

12.613.96.57.5

Secondary

410.9219.2415.0330.0123.044.23

171.323.924.5

157.0

Transjernnchange(%)

210250880340270ND200110180180

Interval refers to the time between paracenteses of primary and secondaryaqueous humor. ND = no detectable difference in transferrin levels. A statis-tically significant correlation (r = 0.758, P < 0.05) exists in the increase in theconcentration of protein and transferrin between the primary and secondarysamples.

To induce transparency, the slot-blot nitrocellu-lose strips were immersed in xylene for 1 min andthen analyzed with a scanning laser densitometer(Pharmacia LKB Biotechnology, Piscataway, NJ) at612 nm and 0-4 optical density. The absorbancepeaks and peak areas were stored as data files, ana-lyzed on Gel-scan XL software (Pharmacia), and re-plotted with the Slidewrite Plus software package (Ad-vanced Graphics Software, Sunnyvale, CA).

Results

Total Protein Assay

The protein content of human aqueous humorranged from 6.0-19.0 mg/dl (mean, 11.87 ± 4.47 mg/dl) in control samples, 6.9-226.45 mg/dl (mean,62.11 ± 56.74 mg/dl) in POAG samples, and 10.2-675.91 mg/dl (mean, 124.53 ± 152.67 mg/dl) in SGsamples. For dog eyes, the protein content of primaryaqueous humor ranged from 5-13.9 mg/dl (mean,7.97 ±3.12 mg/dl) and that of secondary aqueoushumor ranged from 23.92-415.04 mg/dl (mean,191.9 ± 149.8 mg/dl). The increase in total proteincontent from primary to secondary samples ofaqueous humor from the same animals ranged from2-to 70-fold (Table 1).

RIA

We verified the accuracy, precision, and reproduc-ibility of the homologous RIA by calculating the in-tra- and interassay coefficients of variation, which

were determined for aqueous humor to be 8.27%(mean of seven replications each) and 14.62% (meanof five separate runs), respectively. The sensitivity ofthe assay is 0.2 Mg/dl. In the samples of aqueous hu-mor spiked with standard transferrin or various levelsof the drugs that had been used topically, we obtainedcomplete recovery of the known amounts of trans-ferrin. The antiserum used in the RIA detected iron-free and iron-saturated transferrin equally, and it didnot react with the various growth factors tested orwith transferrin from canine or other animal species.In the samples of human aqueous humor that weredeglycated selectively for removal of terminal sialicacid moieties by neuraminidase digestion, we de-tected differences of up to 10% between the levels oftransferrin in treated versus nontreated samples. To-tal binding was approximately 50%, and nonspecificbinding was 4%.

The content of transferrin in the aqueous humorfrom control subjects ranged from 0.3-3.4 mg/dl(mean, 1.36 ± 0.66 mg/dl); in patients with POAG,0.34 to > 10 mg/dl (2.07 ± 1.90 mg/dl); and in pa-tients with SG patients, 0.29 to > 10 mg/dl (2.78± 2.24 mg/dl, Fig. 1). The scatter distribution of theconcentrations of transferrin and total protein for indi-vidual samples of aqueous humor is depicted in Fig-ure 2. The difference between the mean values oftransferrin in POAG and control samples using thestudent two-sample t-test was significant statistically(P < 0.01); that between the mean transferrin valuesin patients with SG and control subjects also was sig-nificant (P < 0.001). For the samples of aqueous hu-mor from patients with POAG and SG, there was ahigh degree of correlation (r = 0.620 and 0.891, respec-tively) between the protein levels and transferrin con-centrations (one-tailed test, P < 0.05). When the find-

CONTROLS POAQ

Fig. 1. Dot plots showing distribution of transferrin in aqueoushumor samples from controls and patients with primary open-an-gle and secondary glaucomas. Solid line and hatched areas repre-sent mean concentrations and standard deviations, respectively.

Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933164/ on 04/02/2018

2870 INVESTIGATIVE OPHTHALMOLOGY 6 VISUAL SCIENCE / September 1992 Vol. 33

50 100 150

Total Prolein (mg/dl)

200 250

100 150

Total Protein (mg/dl)

200 250

Fig. 2. Scatter plots of concentrations of transferrin versus thecorresponding protein concentrations of aqueous humor samples,(a) Nonglaucomatous control patients. Solid line represents linearregression equation for the control samples (y = 0.017x + 1.155).Coefficient of correlation: r = 0.119. There is no significant correla-tion (P> 0.05) between protein concentration and transferrin levelsin aqueous samples from control patients with cataract, (b) Patientswith primary open-angle glaucoma (closed circles) and secondaryglaucoma (triangles). Regression equations are plotted for POAG(solid line) and SG (dashed line) patients. For POAG: y = 0.015x+ 0.903 (r = 0.620, P < 0.01). For SG: y = 0.0131x + 1.151 (r= 0.891, P < 0.001). The correlation between the concentrations ofprotein and transferrin are highly significant.

Western Blotting

By using the rabbit antiserum to transferrin, we de-tected transferrin as an immunoreactive fraction atapproximately 80 kD in aqueous humor samplesfrom control subjects and glaucomatous patients. Pu-rified transferrin co-migrated with these polypeptideson the same blots (Fig. 3). No binding of the antibodywas seen in the control blots when nonimmune serumwas substituted for the primary antibody.

Slot-Blot Method and Quantitative Analysis

All absorbance units (AU) generated for the trans-ferrin standards were 0.11-0.49 AU. The minimumsensitivity of the slot-blotting procedure was 0.95 ng,as graded quantitatively in absorbance area units (AUX mm) by a laser scanning densitometer. For the hu-man aqueous humor, the individual samples gave apositive reaction that was comparable to that gener-ated by the RIA. No reaction product was visualizedwhen 0.1% nonfat dry milk in TBS was applied to thenitrocellulose in place of aqueous humor. As evalu-ated by the scanning laser densitometer in absorbancearea units, the secondary canine aqueous samplesgave a positive reaction product that was as much asninefold greater than the primary counterparts (Table1). There was a statistically significant correlation (r= 0.758, P < 0.05) between the increase in the con-centrations of protein and transferrin in the primaryand secondary canine aqueous humor.

ings in the patients with POAG and SG were groupedtogether, there was a significant correlation (r= 0.744, P < 0.01) between the concentration of pro-tein and that of transferrin. No significant correlationwas detected between the protein content and thelevel of transferrin in the control aqueous humor sam-ples (r = 0.119, P > 0.05) or between the age of thepatients and the concentration of transferrin in con-trol, POAG, or SG samples.

Fig. 3. Western blots showing the immunologic detection oftransferrin in aqueous humor samples from patients with POAG(lanes 1 and 2) and nonglaucomatous controls (lanes 3 and 4). Theantisera are the same as we used previously against human andfeline aqueous humor, and are specific for purified transferrin(lane 5).

Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933164/ on 04/02/2018

No. 10 TRANSFERRIN IN AQUEOUS HUMOR / Triporhi er ol 2871

DiscussionThe concentration of transferrin that we obtained

for the nonglaucomatous control aqueous humorsamples (1.36 ± 0.66 mg/dl) corresponded closelywith those reported previously.3'5'24"26 For representa-tive samples of aqueous humor, quantitative analysisby slot blotting of transferrin confirmed the valuesobtained by RIA. Our study also found that a signifi-cant amount of a transferrin-like immunoreactivepolypeptide was present in the aqueous humor of pa-tients with POAG and SG. Because the RIA detectsboth native transferrin and polypeptide fragmentsthat still retain their antigenic determinants (epi-topes),1927 we believe that the levels of transferrin wemeasured represent the total amount of transferrin invivo in the aqueous humor.

Although the normal aqueous humor is known tocontain at least three isoforms of transferrin (based onvariations in the amount of iron and/or on glycationpolymorphisms),57'28 with our techniques, we de-tected only the total concentration of transferrin. Weconsidered the possibility that the antiserum used inthe RIA and in the slot-blotting experiments mayhave had various affinities for different glycated iso-forms or iron-saturated forms of transferrin. How-ever, our deglycation experiments (neuraminidasemethod) indicated that the antiserum used in the RIAwas able to detect both the sialated and desialatedforms of transferrin with approximately the same af-finity. Furthermore, the antiserum that we used forthe Western blot and slot-blotting experiments hasbeen characterized extensively and recognizes the taufraction and the sialated forms of transferrin.27 Thesimilarity in values obtained with RIA and slot-blot-ting methods provides additional confirmation of thereliability of our assays.

The results obtained using sodium dodecyl sulfate-polyacrylamide gel electrophoresis, followed by im-munoblotting, showed that aqueous humor fromcontrol subjects and patients with glaucoma con-tained isoforms of transferrin that were indistinguish-able from purified human transferrin. It is likely that,in most aqueous samples from glaucomatous pa-tients, the observed increase in transferrin levels iscaused by an influx of serum transferrin. This hypoth-esis is strengthened by several lines of evidence. Thesignificant correlation (P < 0.05) between mean totalprotein levels and mean transferrin concentrations inaqueous humor from patients with POAG and SG,respectively, indicated that a breakdown of the BABwas responsible for the observed increase in trans-ferrin and total protein concentration. In POAG,there appeared to be wide qualitative and quantitative

variations in the protein composition of aqueous hu-mor, which were attributable to many factors, includ-ing the chronicity of the disease process and the insta-bility of the BAB secondary to topical antiglaucoma-tous medications.229"33 The content of transferrin inserum is approximately 200-fold greater than that innormal aqueous humor, and serum transferrin mightenter the anterior chamber of the eye in excessiveamounts in conditions of breakdown of the BAB.4

Previous studies found that, in neovascular glaucomaand inflammatory disorders such as iridocyclitis anduveitis, the breakdown of the BAB results in an in-crease in certain growth-regulatory polypeptides ofthe aqueous, including epidermal growth factor,3'634

insulin-like growth factor-1,35 and serum trans-ferrin.4-5-25

Because of the known stability of the BAB in thecanine eye, we used the dog to establish an animalmodel for the breakdown of the BAB in humans. Thismodel permitted us to study the changes in transferrinconcentrations in samples from the same animal, anapproach not feasible in humans. Our slot-blot exper-iments indicated that the correlation between thechange in the concentration of protein and transferrinin both primary and secondary samples of canineaqueous humor was significant statistically. The sec-ondary aqueous sample with a 70-fold increase in to-tal protein concentration concurrently showed thegreatest change in total transferrin content. The shorttime (7 min to 1 hr) between collection of primaryand secondary aqueous humors further substantiatedthat an influx of serum transferrin, and not a massiveincrease in intraocular synthesis of transferrin, wasresponsible.

Elucidating the quantities of transferrin in theaqueous humor in health and disease and its interac-tions with the target cells in the eye is necessary for usto understand the role of this growth-modulating sub-stance in the progression of anterior segment dis-orders. The increased mean level of transferrin in theaqueous humor of patients with SG may be responsi-ble in part for the abnormal hyperplastic response ofthe lens epithelium that frequently follows inflamma-tory conditions and/or traumatic insults to theeye.343637 The increased concentration of transferrinin the aqueous humor of patients with POAG wasimplicated in the proliferative activity of Tenon's cap-sule fibroblasts that often results in the scarring andfailure of the filtering bleb after glaucoma filtrationsurgery.338 Flow-cytometric analysis of Tenon's cap-sule fibroblasts showed the presence of transferrin re-ceptors; these decrease in number after these cellsreach confluence.38

Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933164/ on 04/02/2018

2872 INVESTIGATIVE OPHTHALMOLOGY 6 VISUAL SCIENCE / September 1992 Vol. 33

Investigations in our laboratory found 24,000transferrin receptors on cultured trabecular cells;these decrease to 5000 as the cells progress from con-fluence to postconfluence.39 However, despite thepresence of transferrin and its receptor-binding sitesin intraocular tissues, the lack of a significant prolifera-tive activity in the tissues that border the anteriorchamber of the eye probably can be attributed to thebioavailability of transferrin to its target cells. Thisprocess might depend on circulating transferrin re-ceptors shed by the cells that maintain the capacity tobind transferrin39"41 and on the interactive effects ofother growth-modulating substances that, not onlyregulate the expression of transferrin receptors on thecell surface,4243 but also are increased quantitativelyin plasmoid aqueous humor.3-635'44 Because the bal-ance among these regulatory factors varies in healthand disease, an understanding of the pathogenesis ofseveral ocular disorders of the anterior and posteriorchambers of the eye will require elucidation of theresulting combination and action of growth factors atthe target tissue.

Key words: radioimmunoassay, slot-blot assay, blood-aqueous barrier, canine model, primary open-angle glau-coma, wound healing

Acknowledgments

The authors thank Professor Robert Ritch, MD, of theNew York Eye and Ear Infirmary (New York, NY) whoprovided five samples of aqueous humor from glaucoma-tous patients.

References

1. Davson H: The intraocular pressure. In The Eye, Davson H,editor. London, Academic Press, 1984, pp. 391-532.

2. Tripathi RC, Millard CB, and Tripathi BJ: Protein composi-tion of human aqueous humor: SDS-PAGE analysis of surgicaland postmortem samples. Exp Eye Res 48:117, 1989.

3. Tripathi RC, Borisuth NSC, and Tripathi BJ: Growth factors inthe aqueous humor and their therapeutic implications in glau-coma and anterior segment disorders of the human eye. DrugDevelopment Research 22:1, 1991.

4. Dernouchamps JP: The proteins of the aqueous humor. DocOphthalmol 53:193, 1982.

5. Yu T-C and Okamura R: Quantitative study of characteristicaqueous humor transferrin, serum transferrin, and desializedserum transferrin in aqueous humor. Jpn J Ophthalmol32:268, 1988.

6. Tripathi RC, Borisuth NSC, Tripathi BJ, and Fang VS: Analy-sis of human aqueous humor for epidermal growth factor. ExpEye Res 53:407, 1991.

7. Tripathi RC, Millard CB, Tripathi BJ, and Noronha A: Taufraction of transferrin is present in human aqueous humor andis not unique to cerebrospinal fluid. Exp Eye Res 50:541, 1990.

8. Welch S: A comparison of the structure and properties ofserum transferrin from 17 animal species. Comp BiochemPhysiol [A] 97:417, 1990.

9. Harris WR: Thermodynamic binding constants of the zinc-hu-man transferrin complex. Biochemistry 22:3920, 1983.

10. Neckers LM: Transferrin receptors regulate proliferation ofnormal and malignant B cells. Curr Top Microbiol Immunol113:62, 1984.

11. May WS Jr and Cuatrecasas P: Transferrin receptor: Its biologi-cal significance. J Membr Biol 88:205, 1985.

12. Tripathi BJ, Tripathi RC, Livingston A, and Borisuth NSC:The role of growth factors in the embryogenesis and differen-tiation of the eye. American Journal of Anatomy 192:442,1991.

13. Hyldahl L: Factor VIII expression in the human embryoniceye: Differences between endothelial cells of different origin.Ophthalmologica 191:184, 1985.

14. Neupert JR and Lawrence C: Protein release during aqueouswithdrawal in rabbits. Invest Ophthalmol 9:865, 1970.

15. Stjernschantz J, Uusitalo R, and Palkama A: The aqueous pro-teins of the rat in normal eye and after aqueous withdrawal.Exp Eye Res 20:255, 1973.

16. Raviola G: Effects of paracentesis on the blood-aqueousbarrier: An electron microscopic study on Macaca mulatta us-ing horseradish peroxidase as a tracer. Invest Ophthalmol13:828, 1974.

17. Unger WG, Cole DF, and Hammond B: Disruption of theblood-aqueous barrier following paracentesis in the rabbit.Exp Eye Res 20:255, 1975.

18. Bradford M: A rapid and sensitive method for the quantitationof microgram quantities of protein utilizing the principle ofprotein-dye binding. Anal Biochem 72:248, 1976.

19. Feldkamp CS and Smith SW: Practical guide to immunoassayevaluation. In Immunoassay: A Practical Guide, Chan DW,editor. Orlando, FL, Academic Press, 1987, pp. 49-95.

20. Tripathi BJ and Tripathi RC: Effect of epinephrine in vitro onthe morphology, phagocytosis, and mitotic activity of humantrabecular endothelium. Exp Eye Res 39:731, 1984.

21. Havener WH: Ocular Pharmacology, 5th ed. St. Louis, CVMosby, 1983, p. 691.

22. Kranz RG and Gennis RB: A quantitative radioimmunologi-cal screening method for specific gene products. Anal Biochem127:247, 1982.

23. Pappas MG, Hajkowski R, and Hockmeyer W: Dot enzyme-linked immunosorbent assay (Dot-ELISA): A micro techniquefor the rapid diagnosis, of visceral leishmaniasis. J ImmunolMethods 64:205, 1983.

24. Inada K, Baba H, and Okamura R: Quantitative determinationof human aqueous proteins by crossed immunoelectrophore-sis. Jpn J Ophthalmol 28:1, 1984.

25. Inada K, Murata T, Baba H, Murata Y, and Ozaki M: Increaseof aqueous humor proteins with aging. Jpn J Ophthalmol32:126, 1988.

26. Yu T-C and Okamura R: Comparative study of native proteinsin aqueous humor and serum detection of characteristicaqueous humor proteins. Jpn J Ophthalmol 31:2235, 1987.

27. Roitt I: Essential Immunology, 6th ed. Oxford, Blackwell Scien-tific Publications, 1988, p. 286.

28. Parker W and Beam A: Studies on the transferrins of adultserum, cord serum, and cerebrospinal fluid. J Exp Med 115:83,1962.

29. Krause U and Raunio V: The protein of the pathologic humanaqueous humor. Ophthalmologica 16:280, 1970.

30. Beardsley TL and Shields MB: Effect of timolol on aqueous

Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933164/ on 04/02/2018

No. 10 TRANSFERRIN IN AQUEOUS HUMOR / Triporhi er ol 2873

humor protein concentration in humans. Am J Ophthalmol95:448, 1983.

31. Stur M, Grabner G, Huber-Spitzy V, Schreiner J, and HaddadR: Effect of timolol oh aqueous humor protein concentrationin the human eye. Arch Ophthalmol 104:899, 1986.

32. Brubaker RF: The Friedenwald Lecture: Flow of aqueous hu-mor in humans. Invest Ophthalmol Vis Sci 32: 3145, 1991.

33. Mori M, Araie M, Sakuri M, and Oshika T: Effects of pilocar-pine and tropicamide on blood-aqueous barrier permeabilityin man. Invest Ophthalmol Vis Sci 33:416, 1992.

34. Tripathi RC, Borisuth NSC, Tripathi BJ, and Fang VS: Radio-immunoassay of epidermal growth factor in human lenses atvarious stages of development of cataract. Exp Eye Res 53:759,1991.

35. Tripathi BJ, Borisuth NSC, and Tripathi RC: Concentrationsof growth factors in primary and secondary aqueous humors.Invest Ophthalmol Vis Sci 33(Suppl):l 180, 1992.

36. Spencer WH: Lens. In Ophthalmic Pathology, 3rd ed, Vol 1,Spencer WH, editor. Philadelphia, WB Saunders, 1985, pp.423-479.

37. Yanoff M and Fine BS: Lens. In Ocular Pathology: A Text andAtlas, 3rd ed, Yanoff M and Fine BS, editors. New York,Harper and Row, 1988, pp. 347-376.

38. Smyth R, KJtada S, and Lee DA: Expression of transferrin re-ceptors on fibroblasts from Tenon's capsule of the human eye.ARVO Abstracts. Invest Ophthalmol Vis Sci 31(Suppl):63,1990.

39. Tripathi RC, Koili SP, Borisuth NSC, and Tripathi BJ: Identi-fication and quantification of transferrin receptors on trabecu-lar cells. Invest Ophthalmol Vis Sci (in press) 1992.

40. Rowers CH, Skikne BS, Covell AM, and Cook JD: The clinicalmeasurement of serum transferrin receptor. J Lab Clin Med114:368, 1989

41. Tedder TF: Cell-surface receptor shedding: A means of regu-lating function. Am J Respir Cell Mol Biol 5:305, 1991.

42. Wiley HS and Kaplan J: Epidermal growth factor rapidly in-duces a redistribution of transferrin receptor pools in humanfibroblasts.Proc Natl Acad Sci U S A 81:7456, 1984.

43. Davis RJ and Czech MP: Regulation of transferrin receptorexpression.at the cell surface by insulin-like growth factors,epidermal growth factor, and platelet-derived growth factor.EMBOJ 5:653, 1986.

44. Borisuth NSC, Tripathi BJ, and Tripathi RC: Identificationand partial characterization of TGF-j81 receptors on trabecularcells. Invest Ophthalmol Vis Sci 33:596, 1992.

Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933164/ on 04/02/2018