Intraocular pressure and aqueous flow are
decreased by cholera toxin
Douglas Gregory, Marvin Sears, Larry Bausher, Hiromu Mishima, and Alden Mead
Delivery of 2.1 fxg of cholera toxin, a specific, irreversible activator of adenylate cyclase, via the
blood lowers IOPfrom 17.4 to 11.2 mm Hg in 8V2 hr, decreases net aqueous flow by about 50%
in 8 hr, and doubles blood flow to the anterior uvea at 8 to 13 hr. Intravitreal injection of 0.26
Hg of cholera toxin lowered IOP from 15.0 to 9.6 mm Hg, but heat-inactivated toxin had no
effect on IOP. The toxin activates adenylate cyclase from ciliary processes 2.2-fold and stimulates cyclic AMP production by ciliary processes 7.4 times. Absence of aqueous flare, normal
protein concentrations in the aqueous, and histologic examination all confirmed the functional
and structural integrity of the blood-aqueous barrier after cholera toxin infusion. The data
point to an important role for ciliary process adenylate cyclase in regulation of aqueous flow
and maintenance of IOP.
Key words: cholera toxin, adenylate cyclase, intraocular pressure, aqueous flow,
ocular blood flow, cyclic AMP, adrenergics
X he purpose of these experiments was to
learn how stimulation of adenylate cyclase in
vivo would affect intraocular pressure (IOP)
and aqueous flow. It is known that cholera
toxin induces its devastating diarrhea by
stimulating an intestinal adenylate cyclase,
with consequent efflux of electrolytes and water, '* 2 and does so with no appreciable
morphologic changes. 3 " 5 Cholera toxin has
proved useful to study the action of adenylate
cyclase in diverse systems. 6 " 12 The "irreversible" or prolonged action of the toxin makes it
From the Department of Ophthalmology and Visual Science, Yale University School of Medicine, New Haven, Conn.
Supported in part by U.S.P.H.S. grants EY-00237,
EY-00785, EY-01630, the Connecticut Lions Eye Research Foundation, Inc., and Research to Prevent
Blindness, Inc.
Submitted for publication May 21, 1980.
Reprint requests: Douglas Gregory, Ph.D., Department
of Ophthalmology and Visual Science, Yale University
School of Medicine, 310 Cedar St., P.O. Box 3333,
New Haven, Conn. 06510.
very useful for studies of regulation of aqueous flow and IOP.
Materials
ATP and cyclic AMP were obtained from P-L
Biochemicals, Inc., Milwaukee, Wise; creatine
phosphate, creatine kinase, and bovine serum albumin (BSA) from Sigma Chemical Co., St. Louis,
Mo.; [a-32P]ATP and [3H]cyclic AMP from New
England Nuclear, Boston, Mass.; AG50W-X4 cation exchange resin from Biorad Laboratories,
Richmond, Calif.; and Woelm neutral alumina
from Alupharm Chemicals, New Orleans, La. Purified cholera enterotoxin was obtained from
Schwarz-Mann, Orangeburg, N.Y.; the potency of
the toxin was approximately 23 Lb//u,g of protein, a
drug house standard related to the potency of the
toxin to induce vascular permeability. Each batch
was tested for its ability to stimulate cyclic AMP
production in intact ciliary processes in vitro.
Radiolabeled (141Ce,51Cr) microspheres (15 ± 1
/urn) were obtained from the 3M Co., St. Paul,
Minn.
Methods
Animals. Male New Zealand white rabbits,
weighing 4 to 5 pounds, were used.
0146-0404/81/030371+ ll$01.10/0 © 1981 Assoc. for Res. in Vis. and Ophthal., Inc.
Downloaded From: http://iovs.arvojournals.org/ on 06/17/2017
371
372 Gregory et al.
Statistics. All data are expressed as mean ±
S.E.M. (No. of eyes).
IOP. IOP was recorded with a calibrated applanation pneumotonometer.
lntravitreal injection. Cholera toxin was injected in a volume of 10 fi\ into the vitreous with a
25 ix\ Hamilton syringe with a Luer fitting and a
30-gauge needle. With this volume the technique
did not cause any noticeable reflux under the conjunctiva. At the low doses of toxin used, no conjunctival reaction was noticed. Needles were discarded after a single use.
Heat inactivation of cholera toxin. Cholera
toxin was inactivated by heating to 60° for 30 min.
Activity of the toxin was evaluated by determining
its ability to activate ciliary process adenylate cyclase in vitro.
Intra-arterial infusion. Animals were anesthetized with 4 ml of 25% urethane (ethyl carbamate) and 1 ml of sodium pentobarbital (50 mg/ml)
delivered via the marginal ear vein. After careful
dissection of the right external and internal carotid
and maxillary arteries and the lingual artery, a
cannula of PE 50 tubing was placed in the lingual
artery, a branch of the external maxillary artery.
The cannula was directed a short distance within
the arterial tree to the proximal end of the internal
maxillary artery. Ligatures were placed around
the lingual artery distal and proximal to the point
of insertion of the cannula to prevent bleeding and
to hold the tubing in place. In this way the carotid
and maxillary arteries were not compromised
when the infusion was complete. The infusion
usually consisted of 2 (xg of cholera toxin delivered
in a volume of 10 ml over a period of 15 min. The
lingual artery was then ligated distal and proximal
to the point of insertion of the cannula, the cannula was removed, and the initial incision was
sutured.
Aqueous flow determinations. At —100 min the
animal was anesthetized with 4 ml of 25% urethane and 1 ml (50 mg/ml) of sodium pentobarbital
delivered via the marginal ear vein. A 600 mg aspirin suppository was inserted rectally, and IOP
was recorded in both eyes. IOP was recorded
again at -85, -70, -55, and -40 min. At -40
min mixing needles were inserted into each eye
just anterior to the limbus and adjacent to the lateral canthus. These needles were constructed according to the description of Cevario.13 IOP was
monitored to ensure that it had stabilized. At zero
time 25-gauge needles were fired consecutively
into the right and left eyes with the needle gun
designed for that purpose.14 These needles were
attached to 25 /xl Hamilton syringes by a short
Downloaded From: http://iovs.arvojournals.org/ on 06/17/2017
Invest. Ophthalmol. Vis. Sci.
March 1981
length of PE 100 tubing. The syringes, tubing, and
needles were previously filled with a solution of
fluorescein (0.2 fxg//x\) in saline, and the needles
were inserted just anterior to the nasal limbus. A
10 /xl volume of the fluorescein solution was injected simultaneously into the anterior chambers
of both eyes. The total interval from firing of the
needles to injection of the 10 /xl offluoresceinsolution was less than 2 min. IOP was measured at
+5 min, and the stirring motors turned on. The
stirring motors were turned off at +10 min, and
IOP was recorded every 5 min. At +60 min a total
paracentesis of the anterior chamber of each eye
was executed with labeled, dry, preweighed
syringes fitted with 27-gauge needles. The syringes and needles were weighed after paracentesis,
and the anterior chamber volume calculated assuming a density of 1.0 gm/ml for aqueous humor.
Aqueous protein concentrations were determined
by the technique of Lowry et al.15 The concentration of fluorescein in the aqueous samples was
measured with an Aminco-Bowman spectrophotofluorometer (excitation wavelength 490 nm,
emission wavelength 520 nm). The total amount of
fluorescein in the anterior chamber was calculated
from the sample concentration and the anterior
chamber volume. It was assumed (1) that mixing of
the 10 /xl of fluorescein solution with anterior
chamber aqueous was complete, (2) that after the
initial mixing, newly formed aqueous did not mix
with anterior chamber aqueous containing fluorescein, (3) that paracentesis removed all anterior
chamber aqueous, and (4) that IOP was constant
between fluorescein injection and paracentesis.
The flow rates were calculated with zero-order kinetics and the equation
F _= a,, - at _V
a., " t
where a0 is the amount of fluorescein injected at
time zero (2.0 /xg), at is the amount of fluorescein
in the anterior chamber at t = 60 min (at =
Ct " V), V is the volume of the anterior chamber,
and t is 60 min.
Blood flow determinations. The regional blood
flow in the eye was measured with radioactively
labeled microspheres by the reference flow method of Aim and Bill.16 141Ce- or 51Cr-labeled plastic
microspheres were injected into the left heart ventricle. The radioactivity of the ocular tissues was
compared to the radioactivity of a reference blood
sample obtained from the femoral artery. Radioisotopes were determined in a Searle Model 1190
automatic gamma counter.
The animals were anesthetized with urethane,
Volume 20
Number 3
1.5 to 2 gm/kg intravenously, and were artificially
ventilated. The pH of the arterial blood was
checked, and values from 7.4 to 7.6 were considered normal. The mean arterial blood pressure
measured in a femoral artery during the blood flow
determinations was 72 ± 6 mm Hg for the entire
group of 14 animals.
Dissection of ciliary processes. Eyes were enucleated, placed in cold 0.85% saline, and then
opened from the posterior pole. Lens, lens capsule, zonules, and vitreous were removed, and
individual processes were cut off near the base of
the processes with McPherson-Vannas scissors
under a dissecting microscope. Ciliary processes
were placed in a small amount of cold Hanks' balanced salts until the dissection was completed.
Adenylate cyclase. A modification of the technique of Salomon et al. 17 was used to assay adenylate cyclase activity in crude particulate preparations from homogenates of ciliary processes.
Ciliary tips were homogenized in 0.5 ml of 0.10M
Tris HC1, 2.0 mM dithiothreitol (DTT), pH
7.5,—homogenizing medium—and centrifuged at
3000 x g (5000 rpm in a Sorvall SS34 rotor) for 5
min. The pellet was washed three times with
homogenizing medium. Prior to the final centrifugation a sample of the resuspended pellet was set
aside for protein determination. 15 The final pellet
was resuspended in 900 fx\ of a solution consisting
of 300 /Ltl of 0.10M Tris HCl, 8 mM DTT, 2 mg/ml
BSA (pH 7.5); 150 fA of 0.10M Tris HCl containing
5 mg/ml creatine kinase (100 U/mg) (pH 7.5); 30
ju.1 of 0.40M creatine phosphate in 0.10M Tris HCl
(pH 7.5); and 420 fA of 0.10M Tris HCl (pH 7.5).
Aliquots (30 /x,l) of this suspension were distributed to 12 by 75 mm glass test tubes, the tubes
were preincubated for 8 min at 30°, and then 10 fi\
of a substrate solution containing 4 mM [a32
P]ATP (27.5 mCi/mmol), 4 mM [3H]cAMP (7
mCi/mmol), and 20 mM MgSO4 were added to
initiate the reaction. The reaction was stopped
after 8 min by adding 100 fA of 2% sodium dodecyl
sulfate and 40 mM ATP. Cyclic AMP was separated from ATP by the technique of Salomon et
al. 17 [32P]cyclic AMP was determined in a Nuclear
Chicago Mark II liquid scintillation counter and
corrected for quenching and for recovery of
[3H]cyclic AMP.
Adenylate cyclase activity was also estimated
indirectly by measuring the production of cyclic
AMP by intact ciliary processes in the presence of
an inhibitor of cyclic AMP phosphodiesterase. Intact processes were incubated at 37° in Hanks' balanced salts containing 10 mM theophylline for 10
min, and the tissue was homogenized in 6%
Downloaded From: http://iovs.arvojournals.org/ on 06/17/2017
Cholera toxin reduces aqueous flow 373
Table I. Comparison of experimental and
model flow rates
Experimental
flows
(fd/min)
Normal eyes
Cholera toxininfused eyes
Difference
Model 3
(equal
Model 1
(no mixing) sampling)
(fd/min)
(id/min)
2.38
1.14
2.05
0.50
2.55
0.62
1.24
1.55
1.93
trichloroacetic acid (TCA) containing approximately 20,000 dpm of [3H]cyclic AMP. The
homogenates were centrifuged at 3000 X g for 5
min, and the supernatants were extracted five
times with 2 ml of water-saturated diethyl ether,
heated to 55° for 2 hr, then stored frozen until
assayed. The pellets were dissolved in IN NaOH,
and total protein was determined. 15 Aliquots of the
ether-extracted supernatants were diluted with
0.05M acetate buffer (pH 6.2) and acetylated with
acetic anhydride, and cyclic AMP was determined
by the radioimmunoassay technique of Steiner et
al. 18 with kits purchased from New England Nuclear. Cyclic AMP values were corrected for recovery of the tracer [3H]cyclic AMP and expressed
as picomoles of cyclic AMP per milligram of
TCA-precipitated protein.
Histology. Six groups of albino adult rabbits that
had been infused with 2 /xg of cholera toxin via the
internal maxillary artery were studied. Rabbits
were sacrificed at 3, 6, 12, 24, 72, and 168 hr after
cholera toxin infusion. Immediately after enucleation the globes were divided equatorially, and
the anterior halves immersed in 2.5% phosphatebuffered glutaraldehyde solution for 45 min. The
zonular fibers were then cut with fine scissors near
the equator of the lens, and the lens was gently
removed. Small meridional sectors of the entire
ciliary body, about 2 to 3 mm wide, were cut with
a razor blade and refixed in the same buffered
2.5% glutaraldehyde solution for 2 hr. For evaluation on light microscopic analysis, the specimens
were dehydrated with graded alcohols and embedded in Epon 812 and Araldite 506. At every
time interval a total of at least 50 processes were
taken from several animals. The processes were
selected from each of the four quadrants and included samples of iridial and ciliary processes. The
detailed morphology of the toxin effects will be a
subject of a future report.
Discussion of error of the method for aqueous
flow determinations. One of the assumptions implicit in the equation used to calculate flow rates
374
Invest. Ophthalmol. Vis. Sci.
March 1981
Gregory et al.
INTRAOCULAR PRESSURE AFTER INTRAVITREAL CHOLERA TOXIN
T
£
°" O °A O
0
2
0
D
-2
•••
A*
•
•
O
-4
"
c °
•
•
•
A
I
-6
A
D
0
6
•
•A
"
•
qCT
0
-8
-
A
0
Q
D
•
0
A
•
D
Q
-10
.015
.02
.04
.15
.2
0
2.0
A
-12
14
1
1
1
I
i
12
16
1
1
, , 1
1
1
1
1
7
9
II
I
i
13 15
DAYS
HOURS
TIME
Fig. 1. A 10 fx\ volume of cholera toxin (CT) was injected into the right vitreous at time zero; 10
/u.1 of saline were then injected into the left vitreous. IOP in right eyes became significantly
lower than that in left eyes in about 6 to 8 hr. Animals were treated with 20 fi\ of 0.5 mg/dl
indomethacin in 10% DMSO topically at —60, —45, —30, -15, and 0 min, then every hour (•,
• , A) or with intraperitoneal injections of indomethacin as follows: —24 and —12 hr, 25 mg; 6
hr, 125 mg; 12 hr, 25 mg; 24 hr, 50 mg (D, O, A).
from the concentration of anterior chamber fluorescein is that 10P remains constant during the
time between fluorescein injection and paracentesis. Of course, IOP increased a few millimeters
of mercury just after the fluorescein solution was
injected into the anterior chamber but returned to
normal within 5 min and remained constant until
paracentesis 1 hr later. To evaluate the effect of an
"elevated" 10P during the first 5 min and the effect of stirring during the second 5 min on estimates of aqueous flow, fluorescein concentrations
were measured at 10 min after its injection into 12
normal anterior chambers, and flows calculated
with the zero-order equation
~ &60
V
where t = 50 min. The mean flow rate through
normal eyes calculated in this way was 2.35 /u-1/
min. This is virtually identical to the flow rate in
normal eyes cited in Table II.
Thefluoresceinconcentration was measured 90
min after injection of fluorescein into two normal
eyes, and the aqueousflowrate through these two
eyes calculated as in Table II. The flow, 2.16 ±
0.02 /il/min, is close to that in the 12 normal eyes
(Table II) where measurements were done at 60
Downloaded From: http://iovs.arvojournals.org/ on 06/17/2017
min. Kottler et al.19 related aqueous flow to anterior chamber fluorescein concentration, assuming that newly formed aqueous mixes completely
and continuously with fluorescein-containing anterior chamber aqueous: F = In (ao/at) • V/t.
Flow rates in normal eyes calculated with this
equation and data from paracentesis done at 60
and 90 min are 3.80 ± 0.22 and 5.92 ± 0.16
/u,l/min, respectively. These high flow rates
and the discrepancy between the rates calculated
with the 60 and the 90 min anterior chamber
fluorescein concentrations indicate that the zeroorder equation used to calculate flows in Table II
describes washout offluoresceinfrom the anterior
chamber in these experiments better than the
first-order equation above. These control experiments support the validity of the assumptions that
initial mixing is complete and that there is no or
little further mixing of newly formed aqueous and
fluorescein-containing aqueous.
Another assumption implicit in the equation for
flow is that paracentesis always removed all anterior chamber aqueous and therefore accurately
measured the anterior chamber volume. Partial withdrawal of the anterior chamber aqueous
would lead to underestimates of V and at and
would affect estimates of flow rates. It was as-
Volume 20
Number 3
Cholera toxin reduces aqueous flow 375
INTRAOCULAR PRESSURE AFTER CLOSE ARTERIAL INFUSION
OF 2.1/ig CHOLERA TOXIN
20
18
_
l6
x
I '2
RE O
LE
10
•
8
5
hours
10
5
7
days
TIME
Fig. 2. Cholera toxin (2.1 /xg) was infused via the right internal maxillary artery of six rabbits,
and IOP was recorded (see Methods).
sumed (and the data support the assumption) that
in this model for fluorescein displacement from
the anterior chamber, newly formed aqueous entering from the posterior chamber does not mix
with anterior chamber aqueous already containing
fluorescein. Therefore incomplete paracentesis
could also introduce an error into determinations
of ct. There are two limiting models for fluorescein
sampling if paracentesis is not complete. (1)
Paracentesis removes all the newly formed, fluorescein-free anterior chamber aqueous near the
pupil, and the rest of the sample is made up with
fluorescein-containing aqueous, and (2) paracentesis removes allfluorescein-containingaqueous,
and the rest of the sample is made up from newly
formed, fluorescein-free anterior chamber aqueous. The latter is unlikely because the open end of
the needle faces the pupil and the posterior
chamber during paracentesis, and the sampling is
more likely to favor aspiration of newly formed
aqueous emerging through the pupil (model 1,
Table I). Fluorescein-containing aqueous is most
probably located toward the chamber angle. A
third model (model 3), which simplifies estimation
of the error introduced by incomplete paracentesis, is one in which fluorescein-free and fluorescein-containing aqueous are equally sampled
by paracentesis and ct in the paracentesis sample is
identical to the average concentration of fluorescein over the entire anterior chamber. "True" flow
rates for normal and ipsilateral cholera toxininfused eyes were calculated, assuming that estimates of the anterior chamber volume are 80% of
Downloaded From: http://iovs.arvojournals.org/ on 06/17/2017
IOP AFTER ARTERIAL INFUSION OF SALINE
-*i,*i**H* *
S 12
10
8 10
TIME (hours)
Fig. 3. A 10 ml volume of saline was infused via
the right internal maxillary artery of four rabbits,
and IOP was recorded (see Methods), o, Right
eye; •, left eye.
the true volume and that fluorescein sampling is
described by models 1 and 3. Another assumption
in these calculations is that the anterior chamber
volume is the same in eyes of control and toxininfused animals. The anterior chamber volumes
determined by complete paracentesis show no
significant difference between these two groups
(control: 231 ± 3.4 /xl (12); toxin infused: 222 ±
3.4 ix\ (6)). Incomplete paracentesis can lead to
errors in estimates of aqueous flow, especially at
Invest. Ophthalmol. Vis. Sci.
March 1981
376 Gregory et al.
BLOOD FLOW IN ANTERIOR UVEA AND IOP AFTER INFUSION OF
CHOLERA
3.0
TOXIN INTO THE
RIGHT
2.1ug/l0ml
INTERNAL MAXILLARY ARTERY
-n
1.0
•
2.6
BLOOD FLOW
O IOP
0.9
2.2
5
1.8
0.8
o
O
o
o
oo
1-4
0.7
1.0
0.6
0.6
1-5
TIME
AFTER
8-13
17-20
INFUSION, HOURS
Fig. 4. Blood flow at each time interval is the mean of measurements from four rabbits. IOP
data (from Fig. 2) are included for comparison of the time course of the two parameters.
Table II. Aqueous flow in untreated and
acetazolamide-treated eyes and in eyes
ipsilateral to arterial cholera toxin infusion
Aqueousflow(fd/min) ± S.E.M.(n)
Normal
Acetazolamide
Cholera toxininfiised
2.38 ± 0.09 (12)
1.58 ± 0.12(12)
1.14 ± 0.15(6)
Aqueous flow was estimated as outlined in Methods. Acetazolamide-treated animals received 250 mg of acetazolamide (50
mg/ml) intravenously (marginal ear vein) 100 min before intracameral injection of fluorescein (see Methods). Cholera toxininfused animals received 2.1 /xg of cholera toxin in 10 ml of
0.85% NaCl (2.5 X 10"9M) via the right internal maxillary
artery, and aqueous flow was determined 8 hr later. Flow values
in the treated groups are significantly different from the controls, p < 0.01.
low flow rates. However, it is clear that these errors lead to underestimates of the differences between normal and decreased flow rates. Therefore
errors in flow estimates associated with incomplete paracentesis do not contribute to the decreased flows recorded in eyes from cholera
toxin-infused and acetazolamide-treated animals.
The observed differences are probably underestimates.
Aqueousflowwas measured in eyes which were
treated with acetazolamide to test the general validity of the technique used for estimating aqueous
flow. Flow in these eyes decreased by one third
Downloaded From: http://iovs.arvojournals.org/ on 06/17/2017
(Table II). In this experiment aqueous flow was
measured early, at a time when the effect of acetazolamide was not maximal. Furthermore, as discussed above, the technique used for estimating
flow undoubtedly underestimated the difference
between flow in eyes from control animals and
acetazolamide-treated animals.
Results
Intravitreal toxin. Intravitreal injection of
cholera toxin caused a remarkable decrease
in ipsilateral IOP, usually after 6 hr and always by 24 hr, compared with an unchanged
pressure in controls injected with intravitreal
saline (Fig. 1). The rate of the decrease in
IOP varied in general with the dose of cholera toxin; higher doses resulted generally in
more rapid decreases in IOP (Fig. 1). After
intravitreal injection of cholera toxin, recovery of IOP did not occur until after several
days. Heat-inactivated cholera toxin was not
effective. Mean IOP in six animals was, right,
15.0 ± 0 . 9 mm Hg and, left, 15.2 ± 0.9,
immediately prior to right intravitreal injection of 0.26 (xg of native cholera toxin and the
same dose of heat-inactivated toxin into left
eyes. By 22 hr after injection IOP was
9.6 ± 0.2 and 15.0 ± 0.1 mm Hg, right and
left, respectively.
Volume 20
Number 3
Cholera toxin reduces aqueous flow 377
* 4
Fig. 5. Representative light micrograph of ciliary processes treated as in Methods, taken from
left contralateral control and right eye of ipsilateral arterial infusion of toxin done 20 to 24 hr
earlier. Note intact epithelial cell layers with stromal edema without hemorrhage or
inflammatory reaction. Left, Control. Right, Toxin-treated. (X160.)
The blood-aqueous barrier was intact in
these eyes. The aqueous protein concentration in eyes treated intravitreally with 0.20 to
0.26 /Ltg of toxin was 156 ± 45 (17) mg/dl
compared to 160 ± 59 (17) for saline controls
and 104 ± 25 (9) for heated toxin controls.
Biornicroscopic examination of these eyes revealed a negative aqueous flare. Eyes that
were injected with 2 fig of toxin showed an
aqueous protein concentration of 570 ± 25
(6) mg/dl.
Intra-arterial toxin
Pressure and aqueous flow. After 4 to 6 hr
IOP decreased dramatically in eyes ipsilateral to arterial infusion of cholera toxin (Fig.
2). Recovery of IOP began after 24 hr and
returned to normal after several days. IOP in
the contralateral eyes was unchanged. Arterial infusion of 10 ml of saline alone had no
Downloaded From: http://iovs.arvojournals.org/ on 06/17/2017
effect on IOP of either eye (Fig. 3). Arterial
infusion of cholera toxin caused a 50% decrease in the net rate of aqueous formation
(Table II). The effect of systemic (intravenous) acetazolamide was to reduce aqueous
flow by 30% (Table II).
Blood flow. Bloodflowto the anterior uvea
and choroid was estimated after arterial infusion of cholera toxin (Fig. 4). Ocular blood
flow increased as the IOP fell and became
twice the contralateral control blood flow at
the protocol time when aqueous flow measurements were done (Table II). Return of
blood flow values toward normal levels occurred by 17 to 20 hr after infusion of toxin.
At this time, IOP was still in a recovery
phase. The arterial infusion technique allowed delivery of cholera toxin by the primary and major blood supply to the rabbit
Invest. Ophthalmol. Vis. Sci.
March 1981
378 Gregory et al.
Table III. Activation ofadenylate cyclase
in intact ciliary processes by cholera
toxin in vitro
200
• Extracellular
160
o Intracellular
Adenylate cyclase activity
(pmol/minlmg of protein ± S.E.M. (n))
120
t(hr)
Control
3 X 10~7M cholera toxin
22.2 ±2.5 (4)
27.9 ± 3.2 (4)
27.7 ± 4.0(4)
26.0 ±3.9 (4)
45.5 ± 5 . 1 (4)
58.1 ± 3.7(4)
80
40
10"
CHOLERA TOXIN CONCENTRATION (M)
Fig. 6. Dose-response relationship between cholera toxin concentration and cyclic AMP production by intact ciliary processes, (n = 3). Ciliary
processes were incubated in Hanks' balanced salt
solution at 37° with the indicated concentrations of
cholera toxin for 5 hr. Then theophylline was
added to a final concentration of 10 mM. The processes were incubated an additional 10 min, an
aliquot of the medium was removed, 100% (w/v)
TCA was added to the tissue suspensions to a final
concentration of 6%, and the tissue was homogenized. Intracellular and extracellular cyclic AMP
were assayed as outlined in Methods.
eye, the internal maxillary artery, without
manipulation of the eye and without compromising its blood supply.
Blood-aqueous barrier. As in the intravitreal experiments, there was neither ocular
inflammation nor breakdown of the bloodaqueous barrier by clinical examination, nor
did concentrations of protein in the aqueous differ from normal controls. In the flow
study groups protein values were normals:
122 ± 12 mg/dl (12); acetazolamide-treated:
102 ± 8 (12); and ipsilateral right eyes of
cholera toxin-infused animals (170 ± 46 (6).
Light microscopic examination of the ciliary
processes taken at varying intervals after arterial infusion of 2 /xg of cholera toxin showed
that the ciliary epithelia, both layers, were
intact and normal in appearance. The stroma
of the processes showed progressive edema
(Fig. 5) without hemorrhage or inflammation.
The edema was striking by the absence of any
cellular or vascular response (see Discussion).
Activation ofadenylate cyclase. In vitro in-
Downloaded From: http://iovs.arvojournals.org/ on 06/17/2017
Ciliary processes were incubated with 25 /ug/ml cholera toxin
(3 X 10~7M) at 30° in Hanks' balanced salt solution (BSS). At
the times indicated, processes were washed three times with
BSS and homogenized in 0.5 ml of homogenizing medium, and
washed particulate fractions were prepared (see Methods).
Adenylate cyclase activity in each washed particulate fraction
was assayed by the technique outlined in Methods. Assay conditions were 1.0 mM ATP, 1.0 mM cAMP, 5.0 mM MgSO4, 2.0
mM DTT, 10 mM creatine phosphate, 60 U/ml creatine kinase,
and 0.5 mg/ml BSA, pH 7.5.
Table IV. Stimulation of cyclic AMP
accumulation in intact ciliary
processes by cholera toxin in vitro
Cyclic AMP produced by ciliary
processes (pmol/tng of protein ± S.E.M. (n))
t(hr)
0
1
2
3
4
5
Control
45.1
36.3
61.0
49.4
63.9
78.7
± 7.4 (3)
(1)
± 3.9 (3)
± 4.1(3)
± 11.2(2)
± 12.4 (2)
3 X 10 7M cholera toxin
42.8
40.0
41.0
161
148
305
±
±
±
±
±
7. 3(3)
(1)
17. 7(3)
54 (3)
3 (2)
101 (2)
Ciliary processes were incubated with 25 /Ag/ml cholera toxin
(3 X 10~7M) at 30° in Hank's balanced salts containing 10 mM
theophylline. At the times indicated the processes were homogenized in TCA, and total cyclic AMP was assayed as outlined in Methods.
cubation of intact ciliary processes with cholera toxin activated adenylate cyclase. Activation of the enzyme could be demonstrated
directly by measuring the rate of conversion
of [a-32P]ATP to [32P]cyclic AMP in the particulate fraction of a broken cell preparation
(Table III) or indirectly by measuring cyclic
AMP produced by intact processes in the
presence of an inhibitor of cyclic AMP phosphodiesterase (Table IV, Fig. 6). The dose
dependence of cyclic AMP production on
cholera toxin concentration showed an abrupt
increase above 3 X 10~10M cholera toxin
(Fig. 6). At all doses of cholera toxin, 90% of
the total cyclic AMP produced was intracellu-
Volume 20
Number 3
lar except at the highest dose (3 X 10~7M),
where 70% was intracellular.
Discussion
Continued interest in regulation of IOP by
adrenergic receptors has been assured by
recognition of the virtually unmatched reduction in IOP induced by the beta-adrenergic blocker timolol.20"22 Reduction of IOP
in glaucoma by beta blockade is a finding
difficult to reconcile with studies of betaadrenergic agonists that also reduce IOP.
Weekers23 first demonstrated that a beta-adrenergic agonist, isoproterenol, reduced eye
pressure in humans. Eakins24 could demonstrate that intravitreal administration of
otherwise lethal doses of isoproterenol lowered IOP by suppressing aqueous inflow. In a
report on the toxicity of topical isoproterenol,
Ross and Drance25 showed that isoproterenol
reduced IOP without affecting gross outflow
facility. This finding also implied that betaadrenergic agonists lower IOP and do so by
reducing aqueous humor formation. These
results were not completely compatible with
those obtained by Bill26 in monkeys, who perfused isoproterenol through the anterior
chamber and found a 30% increase in aqueous humor formation. More recently, however, Gaasterland et al.27 showed that betaadrenergic stimulation by topical isoproterenol caused a 60% reduction of aqueous inflow
in four young, normal human males.
The pioneering work of Orloff et al.28 and
Orloff and Handler29 implicated synthesis of
cyclic AMP as the "second messenger" for
vasopressin-stimulated water and Na+ transport in toad urinary bladder. Subsequently
cyclic AMP has been shown important in
regulation of transport processes in other systems.30 These findings provide a base for
studies of cyclic AMP and adenylate cyclase
in regulation of aqueous flow. Evidence that
adenylate cyclase plays a role in regulation of
aqueous flow is circumstantial. Waitzman
and Woods31 first demonstrated a catecholamine-stimulated adenylate cyclase in the
ciliary processes of rabbits, and Neufeld and
Sears32 demonstrated that in vitro accumulation of cyclic AMP in the rabbit iris-ciliary
Downloaded From: http://iovs.arvojournals.org/ on 06/17/2017
Cholera toxin reduces aqueous flow 379
body preparations was stimulated by catecholamines. Tsukahara and Maezawa33 presented cytochemical evidence for adenylate
cyclase activity in nonpigmented ciliary
epithelia of rabbits. Although suggestive,
these studies have not shown how stimulation of adenylate cyclase affects the steadystate level of IOP.
In the current work delivery of cholera
toxin to the ciliary processes from either side
of the blood-aqueous barrier, from the blood
supply, or from the vitreous caused a dramatic decrease in IOP (Figs. 1 and 2). Delivery of
cholera toxin via the blood supply is particularly interesting because the apical membranes of the nonpigmented epithelial cells
face the bloodstream. By analogy with other
systems, this is the membrane side likely to
be sensitive to the toxin. Furthermore, any
endogenous regulators of aqueous formation
would be delivered to the epithelial cells via
the bloodstream. The decrease in IOP induced by arterial cholera toxin was associated
with and undoubtedly caused by drastically
reduced aqueous flow (Table II). The data
reported here with the use of cholera toxin to
investigate the role of adenylate cyclase in
regulation of IOP indicate that stimulation of
ciliary process adenylate cyclase reduces net
aqueous flow with a consequent reduction in
IOP.
What is the site of action of cholera toxin?
Could cholera toxin decrease aqueousflowby
an affect on the ciliary vasculature? The results of the blood flow studies in Fig. 4 show
that cholera toxin increases the blood supply
to the anterior uvea. This means that the decrease in aqueous flow is not caused by decreased blood flow to the ciliary processes
and suggests further that cholera toxin may
act directly on the ciliary epithelial cells. The
decrease in IOP observed when the toxin is
delivered by intravitreal injection also supports this idea. At least part of the increased
blood flow at 8 to 13 hr after toxin infusion
could be a consequence of the decrease in
IOP, i.e., hypotony. Other studies done in
the intestine indicate that cholera toxin also
doubles local blood flow.34
Can the reduction in IOP be accounted for
Invest. Ophthalmol. Vis. Sci.
March 1981
380 Gregory et al.
by an increase in outflow facility? The decrease in IOP is probably too large to be accounted for by an increase in outflow facility.
Mean IOP prior to arterial infusion of cholera
toxin was 17.4 mm Hg (Fig. 1). If episcleral
venous pressure were 8 mm Hg and constant
and if facility did not change after infusion of
the toxin, IOP should have been 12.5 mm Hg
at the time the flow measurements were
done (about 8 hr). IOP was 11.2 ± 1.5 mm
Hg at 8V2 hr after cholera toxin infusion (Fig.
2). Furthermore, as discussed above, the
technique used to estimate flow rates probably overestimates flow at low flow rates.
Compensating for this overestimate would
decrease the calculated IOP even further for
the toxin-treated eyes and eliminate the need
to postulate an increase in outflow facility.
Direct measurements of outflow facility are
in progress, even though it would appear that
decreased aqueous flow can account for the
entire decrease in IOP recorded in cholera
toxin-treated eyes.
Absence of aqueous flare, normal protein
concentrations in the aqueous, and histologic
examination all confirmed the functional and
structural integrity of the blood-aqueous
barrier in eyes ipsilateral to the cholera toxin
arterial infusion or intravitreal injection. Furthermore, the lack of an effect on IOP after
intravitreal injection of heat-inactivated cholera toxin rules out any nonspecific inflammatory response to a foreign protein. Finally,
the ciliary epithelial cells were intact by light
microscopic examination (Fig. 5). Interestingly, after a time, many of the processes
were swollen. This stromal edema may reflect
inhibition of net aqueous transport across the
ciliary epithelial cells and may represent the
in vivo counterpart of the "shrinking ciliary
process" system devised in vitro for other
pharmacologic studies by Berggren.35
The evidence developed here indicates
that cholera toxin delivered by close arterial
infusion (or by intravitreal injection) causes a
profound reduction in IOP by decreasing net
aqueous flow through the eye.36 These results are consistent with an important role for
adenylate cyclase in maintenance of IOP by
regulation of aqueous flow because cholera
Downloaded From: http://iovs.arvojournals.org/ on 06/17/2017
toxin is a well-characterized irreversible activator of adenylate cyclase that can stimulate
adenylate cyclase in ciliary processes (Fig. 6,
Tables III and IV). This conclusion supports
the idea that catecholamines, especially betaadrenergic agents, can lower IOP by decreasing aqueous flow and integrates biochemical
information about the adenylate cyclase of
ciliary processes with physiologic data on
beta-adrenergic influences that lower IOP
and reduce aqueous flow.
We thank Marion Stoj who helped with the preliminary intravitreal injections of cholera toxin, Johan
Stjernschantz who performed the blood flow measurements, Dave Keller who helped with the cyclase
assays, and Susan Fleischmann-O'Hara who prepared
the manuscript.
REFERENCES
1. Pierce NF, Greenough WB, and Carpenter CCJ:
Vibrio cholerae enterotoxin and its mode of action.
Bacteriol Rev 35:1, 1971.
2. Finkelstein RA: Cholera. CRC Crit Rev Microbiol
2:553, 1973.
3. Gangarosa EJ, Beisel WR, Benyajati C, Sprinz H,
and Piyaratn P: The nature of the gastrointestinal
lesion in Asiatic cholera and its relation to pathogenesis: a biopsy study. Am J Trop Med Hyg 9:125,
1960.
4. Elliott HL, Carpenter CCJ, Sack RB, and Yardley
JH: Small bowel morphology in experimental canine
cholera. Lab Invest 22:112, 1970.
5. Parkinson DK, Ebel H, DiBona DR, and Sharp
GWG: Localization of the action of cholera toxin on
adenylate cyclase in mucosal epithelial cells of rabbit
intestine. J Clin Invest 51:2292, 1972.
6. Epstein MH, Feldman AM, and Brusilow SW: Cerebrospinal fluid production: stimulation by cholera
toxin. Science 196:1012, 1977.
7. Cramer H, Hammers R, Maier P, and Brusilow SW:
Cyclic 3',5'-adenosine monophosphate in the choroid plexus: stimulation by cholera toxin. Biochem
Biophys Res Commun 84:1031, 1978.
8. Mertens RB, Wheeler HO, and Mayer SE: Effects
of cholera toxin and phosphodiesterase inhibitors on
fluid transport and cyclic-AMP concentrations in
rabbit gall bladder. Gastroenterology 67:898, 1974.
9. Field M, Fromm D, Al-Awqati Q, and Greenough
WB: Effect of cholera enterotoxin on ion transport
across isolated ileal mucosa. J Clin Invest 51:796,
1972.
10. Carpenter CCJ and Greenough WB: Response of
canine duodenum to intraluminal challenge with
cholera exotoxin. J Clin Invest 47:2600, 1968.
11. Friedler RM, Kurokowa K, Coburn S, and Massry
H: Renal action of cholera toxin. I. Effects of
Volume 20
Number 3
urinary-excretion of electrolytes and cyclic-AMP.
Kidney Int 7:77, 1975.
12. Feldman AM and Brusilow SW: Effects of cholera
toxin on cochlear endolymph production: model for
endolymphatic hydrops. Proc Natl Acad Sci USA
73:1761, 1976.
13. Cevario SJ: A leakproof needle for anterior chamber
mixing or for multiple injections in acute experiments. INVEST OPHTHALMOL 12:464, 1973.
14. Sears, ML: Miosis and intraocular pressure changes
during manometry. Arch Ophthalmol 63: 707, 1960.
15. Lowry OH, Rosebrough NJ, Farr AL, and Randall
RJ: Protein measurement with the Folin phenol
reagent. J Biol Chem 193:265, 1951.
16. Aim A and Bill A: The oxygen supply to the retina.
II. Effects of high intraocular pressure and of increased arterial carbon dioxide tension on uveal and
retinal blood flow in cats. Acta Physiol Scand 84:306,
1972.
17. Salomon Y, Londos C, and Rodbell M: A highly
sensitive adenylate cyclase assay. Anal Biochem
58:541, 1974.
18. Steiner AL, Parker CW, and Kipnis DM: Radioimmunoassay for cyclic nucleotides. I. Preparation of
antibodies and iodinated cyclic nucleotides. J Biol
Chem 247:1106, 1972.
19. Kottler M, Brubaker R, and Macri F: The decay
kinetics of substances which leave the anterior
chamber by bulk flow. INVEST OPHTHALMOL 9:758,
1970.
20. Vareilles P, Silverstone D, Plazonnet B, LeDouarec
J-C, Sears ML, and Stone CA: Comparison of the
effects of timolol and other adrenergic agents on intraocular pressure in the rabbit. INVEST OPHTHALMOL VIS SCI 16:987, 1977.
21. Katz IM, Hubbard WA, Getson AJ, and Gould AL:
Intraocular pressure decrease in normal volunteers
following timolol ophthalmic solution. INVEST OPHTHALMOL 15:489, 1976.
22. Zimmerman TJ and Kaufman HE: Timolol, a betaadrenergic blocking agent for the treatment of glaucoma. Arch Ophthalmol 95:601, 1977.
23. Weekers R, Delmarcelle Y, and Gustin J: Treatment
of ocular hypotension by adrenalin and diverse sympathomimetic amines. Am J Ophthalmol 40:666,
1955.
Downloaded From: http://iovs.arvojournals.org/ on 06/17/2017
Cholera toxin reduces aqueous flow 381
24. Eakins K: The effect of intravitreous injections of
norepinephrine, epinephrine and isoproterenol on
the intraocular pressure and aqueous humor dynamics of rabbit eyes. J Pharmacol Exp Ther 140:79,
1963.
25. Ross RA and Drance SM: Effects of topically applied
isoproterenol on aqueous dynamics in man. Arch
Ophthalmol 83:39, 1970.
26. Bill A: Effects of norepinephrine, isoproterenol and
sympathetic stimulation on aqueous humor dynamics in vervet monkeys. Exp Eye Res 10:31, 1970.
27. Gaasterland D, Kupfer C, Ross K, and Gabelnick
HL: Studies of aqueous humor dynamics in man.
III. Measurements in young normal subjects using
norepinephrine and isoproterenol. INVEST OPHTHALMOL 12:267, 1973.
28. Orloff J, Handler JS, and Preston AS: The similarity
of effects of vasopressin, adenosine-3',5'-phosphate
(cyclic-AMP) and theophylline on the toad bladder.
J Clin Invest 41:702, 1962.
29. Orloff J and Handler JS: The role of adenosine
3',5'-phosphate in the action of antidiuretic hormone. Am J Med 42:757, 1967.
30. Strewler GJ and Orloff J: Role of cyclic nucleotides
in transport of water and electrolytes. Adv Cyclic
Nucleotide Res 8:301, 1977.
31. Waitzman MB and Woods WD: Some characteristics of an adenyl cyclase preparation from rabbit
cilia/y process tissue. Exp Eye Res 12:99, 1971.
32. Neufeld AH and Sears ML: Cyclic-AMP in ocular
tissues of the rabbit, monkey, and human. INVEST
OPHTHALMOL 13:475, 1974.
33. Tsukahara S and Maezawa N: Cytochemical localization of adenyl cyclase in the rabbit ciliary body.
Exp Eye Res 26:99, 1978.
34. Cedgard S, Hallback DA, Jodal M, Lundgren O,
and Redfors S: The effects of cholera toxin on intramural blood flow distribution and capillary hydraulic conductivity in the cat small intestine. Acta
Physiol Scand 102:148, 1978.
35. Berggren L: Direct observation of secretory pumping in vitro of the rabbit eye ciliary processes. INVEST OPHTHALMOL 3:266, 1964.
36. Sears ML: The aqueous. In Adler's Physiology of the
Eye, ed. 7, Moses RA, editor. St. Louis, 1980, The
C. V. Mosby Co., p. 223.