February 1991 Vol. 32/2 Investigative Ophthalmology & Visual Science Articles Immunolocalization of Tissue Plasminogen Activator in the Diabetic and Nondiabetic Retina and Choroid Gerard A. Lurry,* Kiyomi lkeda,f Carol Chandler,* and D. Scorr McLeod^: Retinal capillary closure is a common finding in many patients with diabetic retinopathy. The cause of this capillary occlusion is unknown. Since occlusions in microthromboembolic disease can occur because of deficiencies in tissue plasminogen activator (tPA) and since systemic tPA decreases with an increasing duration of diabetes mellitus, the immunohistochemical localization of tPA in the retinas and choroids of diabetic and nondiabetic patients was investigated. The localization of tPA was confined to arteries and arterioles in peripheral retinas from nondiabetics. Both veins and arteries were positive in these choroids. Two of three noninsulin-dependent diabetics had normal levels of immunoreactivity in their retinas, and all had normal levels of immunolocalization in their choroids. All but 2 of the 12 insulin-dependent diabetic eyes (IDDM), however, had reduced levels of retinal tPA immunoreactivity which was most pronounced in their peripheral retinas. Seven eyes from patients with IDDM had no reaction product in their peripheral retinas. Two such eyes also had reduced tPA immunoreactivity in their choroidal vessels. Some tPA-positive vessels were observed in the central retinas of these eyes, but the number of positive vessels and amount of reaction product was greatly reduced compared with eyes from nondiabetic patients. These observations suggest that IDDM patients have reduced fibrinolytic activity in their retinas, which might predispose them to thromboembolic disease. Invest Ophthalmol Vis Sci 32:237-245, 1991 to arteries and veins in the uterus, not to capillaries. The production of systemic tPA is increased by venous occlusion,5 physical exercise,6 or the administration of epinephrine7 or the vasoactive agents bradykinin and vasopressin.1 In vitro production of tPA by the capillary endothelium is increased in response to angiogenic factors and 12-0-tetradecanoyl phorbol 13-acetate.8 Decreased production of tPA occurs in response to corticosteroids9 and in systemic and cutaneous vasculitis.10 Changes in circulating levels of tPA and in its local production can have profound effects because of its key role in thrombolysis.' An increase in systemic tPA without a concomitant increase in the natural inhibitors of tPA can result in poor clotting rates. A deficiency in systemic tPA can be associated with thrombotic events that occur in such diseases as cutaneous vasculitis10 and thrombocytopenic purpura." Lower levels of circulating tPA and an increase in circulating tPA inhibitor activity have also been observed in diabetes mellitus.12 These studies and those by other authors13 suggest that reduced fibrinolytic activity occurs in long-standing diabetes which could promote the associated microthromboembolic disease. If this causal relationship is correct, then re- Tissue plasminogen activator (tPA) is a serine protease that cleaves plasminogen to plasmin; it thus plays a key role in the fibrinolytic system. Normal production of tPA is necessary for prevention of thromboembolic disease.' This enzyme has received much attention recently for its effect in restoring coronary artery patency after myocardial infarction. The vascular endothelium is the major source of circulating tPA.2 Using immunohistochemical techniques, tPA has been localized to the endothelium of capillaries, arteries, and veins in ovary, prostate, uterus, lung, liver, placenta, kidney, and heart by Balaton et al.3 Larsson et al4 observed tPA localized only From the *Wilmer Ophthalmologic Institute, Johns Hopkins University School of Medicine, Baltimore, Maryland, fDepartment of Ophthalmology, Keio University School of Medicine, Tokyo, Japan, and |Johns Hopkins University Applied Physics Laboratory, Laurel, Maryland. Supported by the Maryland Chapter of the American Diabetes Association. Submitted for publication: September 12, 1988; accepted August 20, 1990. Reprint requests: Gerard A. Lutty, 170 Woods Research Building, Johns Hopkins Hospital, 600 North Wolfe Street, Baltimore, MD 21205. 237 238 INVESTIGATIVE OPHTHALMOLOGY b VISUAL SCIENCE / February 1991 duced tPA could cause microangiopathy and result in the characteristic retinal ischemia that is prominent in diabetic retinopathy, a concept proposed by Little14 in 1981. Since local production of tPA is the first line of defense against fibrin deposition,1 the current imraunohistochemical study was undertaken to determine if there were changes in tPA immunoreactivity in retinas and choroids from diabetic versus nondiabetic patients and, if so, were the changes related to diabetic retinopathy. Materials and Methods Twenty-two human eyes were provided by the Medical Eye Bank of Maryland and the Wisconsin Lions Eye Bank. The identities of the donors remained anonymous. The cause of death, age, sex, postmortem time, and death-to-enucleation time for each subject are summarized in Table 1. Before freezing the eyes, a 1-cm incision was made 3 mm posterior to the limbus for intact globes; the iris and lens were removed when the posterior poles were received. The globes or posterior poles were placed in O.C.T. (Miles, Naperville, IL) and frozen in liquid nitrogen. All frozen tissue blocks were stored at -70°C. Streptavidin-peroxidase immunohistochemical localization of tPA was done on 12-^m frozen sections that were fixed in 2% paraformaldehyde in phosphate-buffered saline (PBS, 140 mM NaCl, 10 raM Vol. 32 NaPO4, pH 7.4) for 5 min at 4°C. After fixation, the sections were washed in PBS, permeabilized in absolute methanol for 1 min at -20°C, and then air dried. Endogenous peroxidases were inhibited by a 5-min incubation in 3% hydrogen peroxide. After washing in PBS, the tissue was blocked with 2% rabbit serum and then washed again in PBS. Nonspecific binding of streptavidin and biotin was prevented by using the ABC Blocking Kit (Vector, Burlingame, CA) as recommended by the manufacturer. Polyclonal goat antibody against human melanoma tPA (American Diagnostica, New York, NY) was diluted with 1% bovine serum albumin (BSA, Pentax Fraction V; Miles) in PBS and used at titers of 1:750, 1:500, and 1:100 for 30 min at room temperature. After incubation with primary antibody, the sections were washed in PBS, and then incubated with biotinylated rabbit anti-goat immunoglobulin (Ig) G (1:500; Kirkegaard and Perry, Gaithersburg, MD) for 30 min. This secondary antibody was preincubated for 30 min at 37°C with human serum (one part antibody to ten parts serum) before final dilution with PBS containing 1% BSA. After washing in PBS, the sections were incubated with streptavidin labeled with peroxidase (1:500; Kirkegaard and Perry) for 45 min. After washing with PBS, 3-amino-9-ethylcarbazole (AEC; Sigma, St. Louis, MO) was used as the peroxidase indicator: 12 ml of stock 8 mM AEC in absolute dimethyl sulfoxide was added to 100 ml of 0.1 M sodium acetate (pH 5.1) and 0.8 ml of 3% H2O2. After 15 min of developing, the slides were washed twice in Table 1. Summary of clinical histories and peripheral tPA localization Cause of death 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Cardiac arrest Malignant arrhythmia Cardiac arrest Cardiopulmonary edema Pulmonary edema Myocardial infarct Cardiac arrest Asystole Cardiac arrest Renal failure Respiratory arrest Cardiac arrest Respiratory arrest Cardiac arrest Peritonitis Renal failure Renal failure Pending autopsy Myocardial infarct Cardiac arrest Diabetes No No No No No No No NIDDM NIDDM NIDDM IDDM IDDM IDDM IDDM IDDM IDDM IDDM IDDM IDDM IDDM Evaluation of tPA localization in inferior and superior retina. Age/sex 29/M 47/M 60/F 74/M 77/F 78/F 81/M 58/F 64/M 70/M 7/M 31/F 39/F 45/M 47/M 67/M 72/F 72/F 78/F 83/F PMT/DET (hr) Retinal tPA* 8/2 + 22/12 11/2 20/1 + + + + + ± + 3/2 16/2 16/6 29/3 22/2 13/1 30/1.5 20/2 20/18 10/1.5 4/2 12/1.5 + + + — + - 5/2 OSOD± 8/2 + - 10/6 8/3 ODOS- Comments Coronary artery disease Alzheimer's Hypotension, pancreatitis Hypotension, lung cancer Undiagnosed diabetes Vitrectomy, photocoagulation Photocoagulation Photocoagulation Advanced diabet. retinopathy No. 2 239 rPA IN DIABETIC AND NONDIADETIC RETINAS / Lurry er ol distilled H2O and one of the two sections on each slide was counterstained with Harris' hematoxylin; the cover slips were applied with Kaiser's glycerogel. Rabbit antibody against human factor VIII antigen (1:1000; Dako, Santa Barbara, CA) was used on serial sections to label vascular endothelium in the retina and choroid. Sections incubated with this antibody were blocked with 2% goat serum and biotinylated goat anti-rabbit IgG was used as the secondary antibody (1:500; Kirkegaard and Perry). The hematoxylin counterstaining of one of the sections on each slide allowed arteries with their multilayered muscular wall to be distinguished from veins. Controls included sections incubated with no primary antibody and sections incubated with goat or rabbit serum (1:500) instead of primary antibody. Also binding of anti-tPA was blocked in some experiments by preincubating the antibody in a 1:1 molar ratio with single-chain melanoma tPA (American Diagnostica) overnight at 4°C. After incubation the antibody-antigen solution was centrifuged and the supernatant used as a primary antibody. The relative amount of tPA immunoreactivity in each section of inferior and superior retina and choroid was evaluated immediately after completion of the experiment, and only slides treated with antibody at 1:750 dilution were used for evaluation. This titration yielded a minimum of background with an ample signal in positive structures. The 1:100 and 1:500 titrations were used to detect structures with lower levels of immunoreactivity, but they yielded a background reaction product. Sixteen to 80 sections from each eye were evaluated. A three-point grading scale was used in which + indicated a normal level of immunoreactivity as observed in sections from the two youngest eyes from nondiabetics (eyes 1 and 2), ± indicated less immunoreactivity (either less overall reaction product or less vessels were found positive) than seen in model normal eyes, and - indicating no immunoreactivity. The sections were evaluated by three independent blinded observers, and the average scores are presented in Table 1. Morphometric analysis of tPA immunolocalization was conducted on five eyes: two eyes from nondiabetics (eyes 3 and 6) and three from patients with insulin-dependent diabetes mellitus (IDDM) (eyes 17-OD, 19, and 20-OD). This was used to determine if localization was the same in all quadrants of the retina. The eyes were sectioned until the central axis was reached, ie, the optic nerve head and macular region were included in the sections. Sixteen sections from each eye were included in the analysis. Alternating sections were reacted for factor VIII localization and tPA immunoreactivity (1:750 dilution). The total number of vessels or arteries were counted in these factor VIII sections, and tPA-positive vessels were counted in the next sections. No attempt was made to grade the tPA reaction product in this study, therefore even vessels with trace amounts of tPA immunoreactivity were considered positive. The central and peripheral retina were analyzed separately. The central retina included the optic nerve head and retina within 2 disc diameters on either side of the nerve head. The peripheral retina was considered to be from 2 disc diameters from the nerve head to the ora serrata. The results of this analysis are summarized in Table 2. Results The results of the study of the superior and inferior retina are summarized in Table 1. The youngest eyes from nondiabetics (eyes 1 and 2) were used as the standard (graded as +) against which relative levels of tPA immunolocalization were determined. The localization of tPA was limited in the normal peripheral retina to arteries and arterioles (Fig. 1A). Even at titrations of 1:100 which yielded a background reaction product, localization above background was limited to the arterial circulation and never occurred in the capillaries or veins (Fig. 1C). Immunoreactivity was never observed in any vessels in the secondary retinal vascular bed in the peripheral retina, the vasculature of the inner nuclear layer. In the normal retinal arterial circulation, localization varied from exclusively vascular endothelial to reaction product throughout the arterial adventitia. Localization in the choroid was observed in both the arteries and veins, but the veins were less immunoreactive (Fig. 2). Immunoreactivity was observed in choroidal interstitial Table 2. Morphometric analysis of tPA localization Total vessels Peripheral Central Patient # n % tPA + n % tPA + 3 6 17-OD 19 20-OD 592 467 814 136 604 11% 12% 2% 2% 0.20% 1771 539 952 284 579 14% 12% 7% 8% 2% Total arteries 3 6 17-OD 19 20-OD 88 53 85 13 58 61% 87% 20% 8% 2% n = number of Factor VIII+ vessels or arteries;' blood vessels that display tPA immunoreactivity. 179 43 77 28 42 76% 90% 45% 17% 19% tPA+ = percentage of 240 INVESTIGATIVE OPHTHALMOLOGY b VISUAL SCIENCE / February 1991 Vol. 02 Fig. 1. Immunolocalization of tPA in 12 ^m frozen sections from the inferior retina in eye #3, a 60-year-old nondiabetic. (A) Localization (+) is primarily to the vascular endothelium of a large artery although there is some reaction product throughout the wall of this artery. (Antibody was diluted 1:75O and held overnight at 4°C before use so that antibody was comparable to that used in plate B.) (B) Localization in artery (*) was blocked by preincubating the antibody with single chain melanoma tPA (1:1 molar ratio) overnight at 4°C after final dilution of antibody 1:750. (C) Large retinal vein (V) in eye #3 with no reaction product and therefore graded (-), (Goat anti-human tPA, 1:750; lightly counterstained with hematoxylin; original magnifications X450.) tissue in scattered areas, but some of this appeared to be background as determined by experiments using tPA antibody that was preadsorbed with tPA antigen. In these sporadic areas, the choriocapillaris was positive for tPA localization, but overall it was negative. Reaction product in the scleral vasculature was observed in the arteries, veins, and capillaries (data not shown). Immunolocalization in all three areas of the globe was confined only to the vasculature. Factor VIII immunolocalization was used to identify all vascular endothelium (Fig. 3A). Factor VIII immunoreactivity was observed throughout the vessel walls, perhaps because factor VIII binds to noncollagenous extracellular matrix material of many cell types.15 The relative amount of immunoreactivity did not change with age from 29-78 yr of age (eyes 1-6) or Fig. 2. Immunolocalization of tPA in choroid of eye #6, a 78-year-old nondiabetic. Reaction product was observed in arteries (A), less prominently in veins (V), and was not observed in choriocapillaris (Ch). The dark material over the choriocapillaris is pigment in the pigment epithelium. (Goat anti-human tPA, 1:750; lightly counterstained with hematoxylin; X450.) A B Fig. 3. Immunoiocaiization in 12 ^m sections of inferior retina from eye #20, OD, an 83-year-old insulin-dependent diabetic. (A) Factor VIII (1:1000) localization in the primary and secondary retinal vasculature (arrowhead) and in choroidal vessels. (B) Localization of tPA (1:750) was negative in all retinal vessels and greatly reduced (±) in choroidal arteries and veins. (Lightly counterstained with hematoxylin, X295.) 242 INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / February 1991 sex. Neither postmortem times (PMT) (range, 3-22 hr) nor death-to-enucleation times (DET) of 1-12 hr seemed to affect normal localization of tPA (Table 1). The only nondiabetic with reduced levels of peripheral retinal tPA immunoreactivity was an 81-yr-old hypotensive patient (eye 7) treated with methylprednisolone for pancreatitis. No reaction product was observed in control sections incubated without primary antibody or in sections incubated with nonimmune goat serum instead of primary antibody. Binding of the goat anti-human tPA antibody was blocked by preincubating the antibody with single-chain melanoma tPA in a 1:1 molar ratio overnight at 4°C; therefore binding of the antibody appeared to be specific for tPA (Fig. IB). Two of the three noninsulin-dependent diabetics (NIDDM) studied had normal levels of tPA immunoreactivity in their superior and inferior retinas (eyes 8-10, Table 1). All three had comparable levels of immunoreactivity in their choroidal and scleral vessels. The eyes from IDDM patients studied were comparable to the nondiabetic eyes in age, PMTs, and DETs. The exception was the youngest example (a 7-yr-old boy, eye 11) with a PMT of 30 hr who deserves further comment. This patient was admitted to the hospital with ketoacidosis and was found to have previously undiagnosed diabetes mellitus. Insulin was administered in an attempt to stabilize the patient who subsequently died from respiratory arrest. This was the patient's first exposure to therapeutic insulin. The rest of the subjects in the IDDM group were maintained on therapeutic insulin. The group as a whole had dramatically diminished levels of tPA immunoreactivity in their superior and inferior retinas compared with the nondiabetics. Eighty-three percent of the eyes (10 of 12) had little or no retinal tPA immunoreactivity (Table 1): 25% (3 of 12) of the eyes from IDDM patients had reduced localization of tPA in their retinas, and 58% of the eyes (7 of 12) had no reaction product in their peripheral retinal vasculatures (Fig. 3B). Only two eyes from IDDM patients (eyes 11 and 18) had normal levels of tPA retinal immunolocalization. Three of the 12 eyes had been treated with photocoagulation, two had reduced levels of tPA immunoreactivity (eyes 12 and 15), and the third had no retinal localization of tPA in inferior retina (eye 13). Two of the 12 eyes from IDDM patients had reduced levels of choroidal tPA immunoreactivity (eyes 19 and 20-OD). Immunolocalization of tPA was normal in the scleral vasculature of all eyes from IDDM patients. In summary, one of seven nondiabetic patients had reduced (±) tPA localization (14%), and eight often IDDM cases had reduced ( or ±) tPA localization (80%). Vol. 32 The results of the morphometric analysis of five eyes are presented in Table 2. The percentage of all vessels that were positive for tPA was 14% or less. The three eyes from IDDM patients (eyes 17-OD, 19, and 20-OD) had less tPA-positive vessels overall than those from nondiabetics. Less total vessels were positive in the peripheral than in the central retina in all five cases. In the central retina some nonarterial vessels were positive for tPA, but in the peripheral retina, only the arteries and arterioles were positive. More arteries were positive in the central than in the peripheral retina (Table 2). The one eye from an IDDM patient graded (±) in the initial analysis (eye 17-OD) had a substantially greater number of tPA-positive arteries than the two other such eyes graded (-) in the initial analysis (eyes 19 and 20-OD). The reason that the percent tPA-positive vessels was so large in the eyes from IDDM patients was that all vessels, even with trace amounts of reaction product, were considered positive for this analysis. Overall, the amount of reaction product in all three eyes from diabetic patients was greatly reduced compared with the two nondiabetic eyes (eyes 3 and 6). Figure 4 illustrates the reduction in tPA localization observed in the optic nerve head from eye 19 compared with eye 3 (nondiabetic patient). The morphometric analysis summarized in Table 2 illustrates that eyes with no tPA immunoreactivity in the superior and inferior retina have some reactivity in the optic nerve head area, albeit considerably less. Discussion The normal peripheral retinal vasculature appears to be unique in terms of tPA production since localization was confined to arteries and arterioles in the primary retinal vasculature and not capillaries or veins. In addition, the secondary retinal vasculature of the inner nuclear layer was negative in tPA immunoreactivity. Balaton et al,3 using immunohistochemical methods, observed tPA localized in arteries, veins, and capillaries in the ovary, prostate, uterus, lung, placenta, kidney, and heart. This universal vascular localization was observed in the sclera of the eye in the current study. Todd,16 in his classic study using tissue sections with a "fibrin plate" technique, observedfibrinolysisonly in veins from various human tissues, but this assay did not consider the as yet undiscovered fast-acting inhibitor of tPA. Heterogeneity in the distribution of tPA in the retinal arterial wall was observed in the current study. There are conflicting reports in the literature on this point. Casslen et al17 and Larsson et al4 observed tPA immunoreactivity throughout the muscular layers of arteries and only in the endothelium of veins in the uterus and No. 2 rPA IN DIABETIC AND NONDIABETIC RETINAS / Lurry er o\ 243 '•, '#, • » V O * 4 H D N Fig. 4. Immunolocalization of Factor VIII (A, B) and tPA (C, D) at the optic nerve head in nondiabetic eye #3 (A, C) and in insulin-dependent eye #19 (B, D). There is one large artery positive for tPA and trace amounts of reaction product in some other vessels in the 1DDM eye (D), but overall the tPA immunoreactivity is greatly reduced when compared to tPA immunoreactivity in the nondiabetic (C). These meridional sections of the optic disc include the optic nerve (N), anterior to the lamina cribosa, and the vitreous-retina interface (V). (Rabbit anti-human Factor Vlll, 1:1000; goat anti-human tPA, 1:750; no counterstain; original magnifications XI50.) umbilical cord, respectively; Balaton et al3 found it restricted only to the endothelium in the uterus and several other tissues. The choroidal localization reported in our study differs from these reports in that tPA localization was found throughout the walls of both arteries and veins. In all of these studies, however, tPA localization was confined to the vasculature. To our knowledge there has been only one other immunohistochemical study of tPA in the eye.18 In this study, conducted with a monoclonal antibody on paraffin sections of human and monkey eyes, tPA immunolocalization was observed in the retinal and choroidal vasculature, but the authors did not distinguish among arteries, veins, and capillaries. They also observed tPA immunolocalization in many extravascular structures. This apparent contradiction with our data may be caused by technical differences; the Tripathi18 study used paraffin sections, a monoclonal antibody, and the peroxidase anti-peroxidase staining technique. Although this is the first report on relative levels of tPA immunolocalization in the diabetic retina to our knowledge, the observation of decreased levels of tPA in diabetic tissue is not unexpected. Aimer et al,12 in a series of studies published in 1975, found lower levels of circulating tPA in diabetic blood than in that from normal patients. Using the technique of Todd,16 they found decreased levels of tPA activity in arteries and veins from the hands of diabetics, with the lowest levels being detected in vessels from obese diabetics.19 When diabetics with proliferative retinopathy were compared with those that were preproliferative, Aimer et al20 found elevated plasma fibrinolytic activity (both spontaneous and in response to venous cuff) in the preproliferative group (almost normal levels) and diminished activity in the proliferative group. Levels of fibrinolytic activity diminished as the duration of proliferative retinopathy increased. This agrees with our observation of normal levels of tPA immunoreactivity in two of the three retinas 244 INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / February 1991 from NIDDM patients and in eye 11 from a 7-yr-old recently diagnosed diabetic. Conclusions concerning tPA levels in eyes from NIDDM versus IDDM patients are not possible because of the limited sample size of eyes from NIDDM patients. The eyes from the only other subjects whose duration of diabetes was known were 12 (19 yr and graded ±), 14 (22 yr and graded - ) , and 15 (35 yr and graded ±). Although our immunohistochemical data on the diabetic retina appear to parallel the findings of Aimer using the fibrinolytic assay1219'20, immunohistochemical localization of tPA does not give a complete picture of fibrinolytic activity in the retina. Looking only at the activator precludes awareness of the levels of the fast-acting tPA inhibitor produced by smooth muscle21 and endothelial cells.22'23 Local mean fibrinolytic levels are related to production of both the activator and its inhibitor. Hegt21 noted, however, that in disorders like endotoxin shock, hyaline membrane disease, and Waterhouse-Friderichsen syndrome, the decrease in tPA production is concomitant with an increase in circulating plasmin inhibitor. The end result is even less systemic fibrinolytic capacity. In the current study this point seems inconsequential since 58% of the diabetic eyes had no demonstrable tPA localization in inferior and superior retina. Therefore fibrinolytic activity in those retinas was probably reduced regardless of the local levels of tPA inhibitor or circulating levels of plasmin inhibitor. The relative levels of tPA localization we observed may be an indicator of the thrombolytic capacity of several ocular vasculatures. The choroidal vasculature, with the exception of the choriocapillaris, was rich in tPA immunoreactivity in all eyes except those from two IDDM patients, and thrombotic occlusion occurs less frequently in the choroid than in the retina. The peripheral retina, however, had tPA limited to the arterial circulation and not in the veins or capillaries where most thrombotic obstruction occurs. Thrombotic occlusion in the retina is a major cause of decreased vision in adults in the United States. In the normal retina, fibrin deposition in the capillaries and veins may be minimized by tPA produced and secreted by the arterial endothelium upstream. This implies that eyes from IDDM patients may be at greater risk for fibrin deposition since even arterial localization was reduced. The cause of this reduction in tPA localization is not clear, but it does not appear to be viability of the vascular endothelium. Factor VIII localization in the retinal endothelium was comparable in all eyes studied, and it is a marker for viable vascular endothelium. One possible explanation for the decrease in immunoreactive tPA in IDDM patients is blood flow. Vol. 32 Diamond et al24 recently demonstrated that increases in wall shear stress generated by an elevated flow causes greater production of tPA by the vascular endothelium. If retinal blood flow is decreased in IDDM, as suggested by several groups,25'26 this could explain the decrease in tPA production we observed in the eyes from these patients. Furthermore, venous levels of shear stress to monolayers of endothelial cells24 resulted in basal-level (no flow) production of tPA; arterial levels of shear stress resulted in a threefold increase in tPA production. These differences between arterial and venous shear stress in vivo could explain our observations that immunoreactive tPA is predominantly found in the arteries in the human retina. Considering the greater adhesiveness of diabetic platelets13 and the apparent reduced fibrinolytic capacity of the diabetic retinal vasculature, the IDDM patient would seem to be at high risk for retinal thrombus formation and subsequent capillary occlusion. The resultant thrombi may contribute to the retinal ischemia that is frequently observed in diabetic retinopathy. Key words: tissue plasminogen activator, retina, choroid, immunolocalization, diabetic retinopathy Acknowledgments The authors thank the Medical Eye Bank of Maryland and the Wisconsin Lions Eye Bank for providing the tissue for this study, Dr. Ann Hanneken for her assistance in acquiring tissue and medical histories of some patients and her critical review of this manuscript, Sylvia Crone for her technical assistance, Dr. Michael Berman for his insightful discussions about tPA and critical review of this manuscript, and Arnall Patz and Daniel Finkelstein for their critical review of this manuscript. References 1. Bachmann F and Kruithof EK: Tissue plasminogen activator: Chemical and physical aspects. Semin Thromb Hemost 10:6, 1984. 2. Rijken DC, Wijngaards G, and Welbergen J: Relationship between tissue plasminogen activator and the activators in blood and vascular wall. Thromb Res 18:815, 1980. 3. Balaton A, Angeles-Cano E, and Sultan Y: Tissue plasminogen activator (t-PA) or vascular plasminogen activator (v-PA)? Its localization in normal human tissues as denned by a monoclonal antibody. In Progress in Fibrinolysis, Vol VII, Davidson J, Donati M, Coccheri S, editors. New York, Churchill Livingstone, 1985, pp. 182-184. 4. Larsson A and Astedt B: Immunohistochemical localization of tissue plasminogen activator and urokinase in the vessel wall. J ClinPathol 38:140, 1985. 5. Wojta J, Turcu L, Wagner OJ, Korninger C, and Binder BR: Evaluation offibrinolyticcapacity by a combined assay system for tissue-type plasminogen activator antigen and function using monoclonal anti-tissue-type plasminogen activator antibodies. J LabClin Med 109:665, 1987. No. 2 rPA IN DIABETIC AND NONDIABETIC RETINAS / Lurry er ol 6. Marsh N and Gaffney P: Some observations on the release of extrinsic and intrinsic plasminogen activators during exercise in man. Haemostasis 9:238, 1980. 7. Kjaeldgaard A and Kjaeldgaard M: In vitro stimulation of plasminogen activator release from vein walls by adrenalin. J ClinPathol 39:1241, 1986. 8. Gross JL, Moscatelli D, and Rifkin DB: Increased capillary endothelial cell protease activity in response to angiogenic stimuli in vitro. Proc Natl Acad Sci U S A 80:2623, 1983. 9. Roblin RO, Vetterlein DA, McLellan WL, Young PL, Bell TE, and Eichorn B: Regulation of plasminogen activator activity in embryonic and tumor-derived human cell cultures. In Progress in Fibrinolysis, Vol V, Davidson J, Nilsson I, and Astedt B, editors. New York, Churchill Livingstone, 1981, pp. 90-95. 10. Jordan JM, Allen NB, and Pizzo SV: Defective release of tissue plasminogen activator in systemic and cutaneous vascultis. Am J Med 82:397, 1987. 11. Glas-Greenwalt P, Hall JM, Panke TW, Kant KS, Allen CM, and Pollack VE: Fibrinolysis in health and disease: Abnormal levels of plasminogen activator, plasminogen activator inhibitor, and protein C in thrombotic thrombocytopenic purpura. J LabClinMed 108:415, 1986. 12. Aimer LO and Nilsson IM: Onfibrinolysisin diabetes mellitus. ActaMedScand 198:101, 1975. 13. Fuller JH, Keen H, Jarrett RJ, Omer T, Meade TW, Chakrabarti R, North WR, and Stirling Y: Haemostatic variables associated with diabetes and its complications. Br Med J 2:964, 1979. 14. Little HL: Alterations in blood elements in the pathogenesis of diabetic retinopathy. Ophthalmology 88:647, 1981. 15. Wagner DD, Urban-Pickering M, and Marder VJ: Von Willebrand protein binds to extracellular matrices independently of collagen. Proc Natl Acad Sci U S A 81:471, 1984. 245 16. Todd AS: The histochemical localization offibrinolysinactivator. J Pathol 78:281, 1959. 17. Casslen B, Larsson A, and Astedt B: Localization of tissue plasminogen activator (t-PA) in human vessels and uterus. Haemostasis 14:135, 1984. 18. Tripathi BJ, Geanon JD, and Tripathi RC: Distribution of tissue plasminogen activator in human and monkey eyes. Ophthalmology 96:1434, 1987. 19. Aimer LO, Pandolfi M, and Aberg M: The plasminogen activator activity of arteries and veins in diabetes mellitus. Thromb Res 6:177, 1975. 20. Aimer LO, Pandolfi M, and Nilsson IM: Diabetic retinopathy and thefibrinolyticsystem. Diabetes 24:529, 1975. 21. Hegt VN: Relations between activation and inhibition of fibrinolysis in the walls of human arteries and veins. Thromb Haemost 38:407, 1977. 22. Loskutoff DJ and Edgington TS: Synthesis of a fibrinolytic activator and inhibitor by endothelial cells. Proc Nat Acad Sci USA74:3903, 1977. 23. Hanss M and Collen D: Secretion of tissue-type plasminogen activator and plasminogen activator inhibitor by cultured human endothelial cells: Modulation by thrombin, endotoxin, and histamine. J Lab Clin Med 109:97, 1987. 24. Diamond SL, Eskin SG, and Mclntire LV: Fluid flow stimulates tissue plasminogen activator secretion by cultured human endothelial cells. Science 243:1483, 1989. 25. Rimmer T, Fallon TJ, and Kohner EM. Long-term follow-up of retinal blood flow in diabetes using the blue light entoptic phenomenon. Br J Ophthalmol 73:1, 1989. 26. Feke GT, Buzney SM, Goger DG, Spack NP, and Gabbay KH. Variation of retinal blood flow with duration of diabetes in Type I diabetes. ARVO Abstracts. Invest Ophthalmol Vis Sci 30(Suppl):5, 1989.