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.