Abstract
purpose. To determine the diurnal intraocular pressure (IOP) response of Brown
Norway rat eyes after sclerosis of the aqueous humor outflow pathways
and its relationship to optic nerve damage.
methods. Hypertonic saline was injected into a single episcleral vein in 17
animals and awake IOP measured in both the light and dark phases of the
circadian cycle for 34 days. Mean IOP for light and dark phases during
the experimental period were compared with the respective pressures of
the uninjected fellow eyes. Optic nerve cross sections from each nerve
were graded for injury by five independent masked observers.
results. For fellow eyes, mean light- and dark-phase IOP was 21 ± 1 and
31 ± 1 mm Hg, respectively. For four experimental eyes, mean IOPs
for both phases were not altered. Six eyes demonstrated significant
mean IOP elevations only during the dark phase. Of these, five showed
persistent, large circadian oscillations, and four had partial optic
nerve lesions. The remaining seven eyes experienced significant IOP
elevations during both phases, and all had extensive optic nerve
damage.
conclusions. Episcleral vein injection of hypertonic saline is more likely to
increase IOP during the dark phase than the light. This is consistent
with aqueous outflow obstruction superimposed on a circadian rhythm of
aqueous humor production. Because these periodic IOP elevations
produced optic nerve lesions, both light- and dark-phase IOP
determinations are necessary for accurate correlation of IOP history to
optic nerve damage in animals housed in a light–dark
environment.
Although long considered a major risk factor for glaucomatous
optic neuropathy, the cellular mechanisms by which elevated intraocular
pressure (IOP) damages the optic nerve remain unknown. Acquiring this
knowledge relies heavily on the use of animal models of chronically
elevated IOP.
Although nonhuman primates are anatomically most appropriate for such
studies,
1 2 their use is limited by cost and decreased
availability. This has led to increased interest in developing a
well-characterized model of aqueous humor outflow obstruction in
rodents with identifiable, predictable optic nerve damage, along with
reliable methods for documenting IOP.
3 4 5 6 7 8 This model
should be economical enough to allow experimentation with sufficient
numbers of animals to provide a complete picture of both the
ultrastructural consequences and the cell biology of pressure-induced
optic nerve damage. Understanding these mechanisms will provide
important insights for developing new methods of counteracting the
effects of pressure on the optic nerve.
Such a model will also provide a system for assessing the efficacy of
treatments designed to protect the optic nerve, either during elevated
IOP or after IOP is controlled by medications or surgery.
9 However, successful evaluation of such neuroprotective strategies
requires a thorough understanding of the relationship between pressure
and optic nerve damage, beginning with careful documentation of the IOP
to which the optic nerve is subjected.
Our demonstration that IOP can be monitored in awake animals represents
a major advance in this regard,
10 because it avoids
general anesthetics that may in themselves alter the pressure–damage
relationship by artifactually lowering IOP.
11 12 We found
that Brown Norway rats normally experience a large circadian
fluctuation in IOP, ranging from approximately 20 mm Hg during the
light phase to 30 mm Hg in the dark.
10
We have devised a method of producing chronically elevated IOP in Brown
Norway rats by injecting hypertonic saline into episcleral
veins.
3 Because this treatment is likely to increase
resistance to aqueous outflow, we might expect these animals to
experience a larger than usual fluctuation in IOP. Such a fluctuation
could be missed if IOP were monitored only during the day or light
portion of the cycle.
In these experiments, we determined the effect of episcleral vein
injection of hypertonic saline on rat IOP during the light and dark
phases of the circadian cycle. The results provide important insight
into the pressure–damage relationship in this model and strongly
support our supposition that obstruction of aqueous humor outflow is
the primary mechanism of pressure elevation.
All experiments complied with the ARVO Statement for the Use of
Animals in Ophthalmic and Vision Research. Seventeen male Brown Norway
rats (Rattus norvegicus), weighing 300 to 400 g, were housed
in a standard animal room with food and water provided ad libitum and
the temperature maintained at 21°C. The room was lighted by
fluorescent lights (330 lux) that were turned on and off automatically
every 12 hours, with the lights on from 6:00 AM to 6:00 PM (light
phase) and off from 6:00 PM to 6:00 AM (dark phase). Animals were
weighed weekly to monitor their general health.
A TonoPen XL (Mentor, Norwell, MA) tonometer was used to
measure awake intraocular pressure, with one drop of 0.5% proparacaine
hydrochloride applied to each eye, as described
previously.
3 10 Ten IOP readings were obtained from each
experimental and fellow (control) eye daily by using firm contact with
the cornea and omitting readings obtained as the instrument was removed
from the eye (off readings). The mean of these readings for each eye
was recorded as the IOP for that date and used to calculate the mean
IOP for each eye during the dark and light phases over the course of
the experiment. Dark-phase measurements were made using a
long-wavelength 16-W bulb (Bright Laboratory Jr.; CPM Inc., Dallas,
TX), which emits light in the far-red range of the visible spectrum and
does not affect circadian rhythms.
10 13
The animals were acclimated to daily handling and awake IOP measurement
before episcleral injection. Animals were confirmed to have normal IOP,
21 ± 1 mm Hg and 31 ± 1 mm Hg (mean ± SD) for the
light and dark phases, respectively. One eye of each animal was then
injected with 50 μl 1.75 M hypertonic saline solution through an
episcleral vein, using a force sufficient to just blanch the limbal
artery.
3 IOPs were determined daily from the first day
after injection throughout the experimental period, alternating between
light-phase (9–11 AM) and dark-phase (7–9 PM) measurements. These
time windows were chosen because our previous experience demonstrated
that IOP in these animals reaches its peak or trough within 1 hour of
the light change and varies little during each phase.
After 34 days of IOP monitoring, animals were anesthetized with
halothane and perfused transcardially with 5% glutaraldehyde, as
previously described.
9 Optic nerve segments 1 mm from the
back of the globe were dissected, washed, postfixed, dehydrated, and
embedded. Sections (0.70 μm) were cut on a microtome (Ultra cut E;
Reichert, Vienna, Austria) and stained with 1% toluidine blue.
Optic nerve cross sections from injected eyes and fellow eyes were
masked and assessed by light microscopy for damage by five independent
observers. For each nerve, a cross section from approximately 2 mm
behind the globe was chosen at random, and each observer based the
assessment of injury on an evaluation of the entire cross section. A
graded scale of nerve injury ranging from 1 (normal) to 5 (total
degeneration) was used, based on prior observations of a stereotyped
pattern of injury in this model.
3
As illustrated and described in
Figure 1 , the extent of injury varies in both the intensity of axonal
degeneration, shown in the photomicrographs, and the extent of the
nerve cross section affected. In general, early mild lesions (grade 2)
begin in one, focal region of the nerve, with moderate concentration of
degenerating axons. Greater injury (grade 3) is characterized by an
increasing density of degenerating axons, with spread to other portions
of the optic nerve. As the entire nerve becomes affected, the intensity
of damage progresses, until degenerated and normal-appearing axons are
equal in number (grade 4). Finally, nearly all axons are degenerating
across the entire optic nerve (grade 5). From this analysis, each eye
was then assigned a grade of injury determined by calculating the mean
of the five independent grade scores.
For each experimental eye, the light- and dark-phase IOP
(mean ± SD) from injection to death was compared with the
corresponding IOP of the fellow eye by Student’s t-test.
Eyes were also categorized into three groups by extent of optic nerve
injury and the mean light- and dark-phase IOPs for each group compared
with the mean light- and dark-phase IOPs of the fellow eyes, by using
the same test.
Interrater agreement for the analysis of optic nerve cross-section
injury was evaluated by calculating the kappa statistic according to
Fleiss.
14 The kappa is a chance-corrected measure of
interrater agreement with values ranging from −1 to +1. The kappa
statistic has been divided into categories describing varying degrees
of agreement
15 : less than 0, poor agreement; 0.0 to 0.20,
slight agreement; 0.21 to 0.40, moderate agreement; 0.41 to 0.60,
moderate agreement; 0.61 to 0.80, substantial agreement; and 0.81 to
1.0, almost perfect agreement. All analyses were performed using a
statistical software package (Excel 97; Microsoft, Redmond, WA) except
the kappa analysis, which was performed with KAPPA (Shareware, Allison
Park, PA) by David Zubrow.
Mean IOPs for light and dark phases of the circadian cycle for
each experimental eye from the time of injection to death are shown in
Table 1 , along with the SD and grade of optic nerve injury. When mean
light- and dark-phase IOPs for each experimental eye were compared with
the respective mean IOPs of their control eyes, the animals segregated
into three groups.
Four animals (group 1) demonstrated at most a 1 mm Hg increase in IOP
over that of the uninjected, fellow eyes for the corresponding phase.
All these eyes had an injury grade of 1.0, indicating no evidence of
optic nerve damage.
Six eyes (group 2) showed significantly (
P < 0.05)
increased IOP only during the dark phase. The mean dark-phase IOPs were
at least 3 mm Hg higher than the corresponding IOP of the control
group. The difference between the mean dark and light IOP was equal to
or greater than 13 mm Hg, with a mean of 16 ± 3 mm Hg. Large
circadian IOP fluctuations persisted throughout the experimental
period, without sustained elevation during the light phase.
Figure 2 illustrates the IOP history and the histologic appearance of the optic
nerve cross section of one experimental eye that was typical of this
group. Of these eyes, four had unequivocal, partial optic nerve
lesions.
All seven eyes in the third group had mean IOPs significantly elevated
(
P < 0.001) relative to the fellow eye in both light
and dark phases. Their corresponding optic nerves all had extensive
lesions with injury grades of 4.9 or higher. The IOP history record and
experimental optic nerve section illustrated in
Figure 3 demonstrates a typical example from this group.
When all 17 experimental eyes were analyzed based on the extent of
their optic nerve injury, they also segregated into three groups. Six
showed no injury, four had partial optic nerve injury (focal lesions),
and seven had degeneration of nearly all optic nerve axons (global
lesions).
Table 2 compares the light- and dark-phase mean IOP of each group with the mean
IOP of the control eyes.
For experimental nerves with no lesions (n = 6), the
mean light-phase IOP for the experimental eyes was 21 ± 1 mm Hg,
identical with that of the control group. The dark-phase mean was
33 ± 3 mm Hg. Although this was slightly higher than that of the
control eyes, the difference was not statistically significant
(P = 0.08).
Four nerves showed a partial optic nerve lesion, ranging in grade from
1.7 to 2.6. IOP in all these eyes showed persistent wide fluctuations
between light- and dark-phase IOPs. The light-phase IOP for this group
overall was 22 ± 1 mm Hg and that for the dark phase was 39 ± 3 mm Hg. Only the dark-phase value was significantly greater than
that in the corresponding fellow eye value (P <
0.001).
In the final group, all seven eyes with optic nerve injury involving
nearly all axons had sustained mean IOP elevations during both the
light and dark phases. Overall mean pressures for the light and dark
phases were 37 ± 4 and 42 ± 2 mm Hg, respectively. Both
were significantly higher than in their respective control eyes
(P < 0.001).
Agreement among observers regarding optic nerve injury grades using
this evaluation system was almost perfect (kappa 0.90 ± 0.07 SE,
Z [Test Statistic] 12.8).
Accurate, noninvasive, repeatable measurements of IOP provide the
cornerstone for acquiring a thorough understanding of the relationship
between IOP and optic nerve damage. This relationship must be
understood if we are to use animal models of pressure-induced optic
nerve damage to study mechanisms of injury and to test the efficacy of
potential neuroprotective strategies.
We initially showed that readings with the tonometer in rat eyes
correlate linearly with transducer measurements of actual
IOP.
4 Subsequently, we found that these measurements could
be repeated over months without ocular side effects.
5 We
finally determined that IOP measured in awake animals using only
topical anesthesia could record subtle, physiologic variations in IOP
in normal rat eyes.
10 This strongly
suggested that measurement of IOP without general anesthesia would
provide the most accurate assessment of the IOP responsible for optic
nerve damage in eyes with experimentally elevated IOP.
In the present study, we have made frequent observations of IOP in
awake animals during both the light and dark phases of a strictly
controlled circadian cycle. This points to important refinements of our
techniques to measure IOP in rats, particularly when applied to our
model of pressure-induced optic nerve damage. The results confirm our
initial observations on the circadian fluctuation of IOP in normal rat
eyes and offer several similarities in IOP behavior between our model
and human glaucoma. The results also provide important insights into
the mechanism of pressure elevation in this model, the relationship
between this pressure elevation and optic nerve damage, and methods of
measuring IOP in rats housed in a light–dark environment.
Measurements in the uninjected fellow eyes confirmed our prior
observation that IOP in the normal rat eye varies in a distinct
circadian fashion, between 21 mm Hg during the light phase and 31 mm Hg
in the dark.
10 Rhythmic oscillations of IOP, centered
around the 24-hour biological clock, represent a common cause of IOP
fluctuation both in humans and in laboratory animals. In humans, some
studies report that peak IOP occurs during the day,
16 17 whereas others have found that peak IOP in normal humans occurs during
the dark, or sleep, portion of the cycle.
18 19 20 In
contrast, animal studies in both rabbits and rats have consistently
demonstrated IOP elevation during the dark phase.
10 13 21 Fluorophotometric studies in rabbits indicate that this increase
results from increased rates of aqueous humor production during the
dark.
22 Fluctuations of aqueous humor formation and IOP
are both inhibited by removing sympathetic input to the
eye.
23 24 25
In the initial description of our model, we demonstrated that
episcleral vein injection of hypertonic saline produced variable
degrees of trabecular meshwork sclerosis and angle
closure.
3 From this, we hypothesized that the mechanism of
IOP elevation was due to obstruction of aqueous humor outflow. This
received strong support by our subsequent demonstration that the
topical application of glaucoma drops that suppress aqueous humor
formation significantly lowers IOP in this model.
9
Because increased IOP during the dark in normal eyes probably depends,
at least in part, on an increase in the rate of aqueous humor
formation,
22 25 we would expect aqueous outflow
obstruction to accentuate the dark-phase IOP increase. This is exactly
what we observed in the present study. Elevations in IOP were much more
common during the dark phase of the circadian cycle. Overall, six
animals (
Table 1 , group 2) demonstrated persistent, wide fluctuations
in IOP, due in every case to a greater than normal mean IOP increase
during the dark. Thus, this study provides a further physiologic
correlate to our prior observations of anatomic angle closure in this
model and suggests strongly that injection of hypertonic saline
accomplishes our original intent: to sclerose aqueous outflow pathways
and increase the resistance to outflow.
The increased circadian pressure variation shown in several of
these eyes has strong parallels with observations in human glaucoma.
Although IOP fluctuation is generally less than 5 mm Hg in normal
individuals, glaucoma patients can have an exaggerated phasic variation
of IOP at all stages of the disease.
26 27 28 29 One study
reported a mean fluctuation of approximately 11 mm Hg, and more than 20
mm Hg for some glaucoma patients, whereas another showed a mean
variation of 18 mm Hg and more than 30 mm Hg in two individuals. As in
our model, the increased IOP fluctuations in glaucoma patients most
likely result from obstruction of aqueous humor outflow superimposed on
the natural circadian rhythm of IOP. In humans with elevated IOP,
aqueous humor outflow obstruction can be confirmed by tonography, as
well as by the effectiveness of glaucoma drops, which reduce aqueous
humor formation.
With this in mind, the complex array of pressure histories seen in our
study probably resulted from differing degrees of angle closure among
these eyes. We suspect that those eyes with only mild elevation of mean
IOP, or none at all, had the least amount of outflow obstruction. Eyes
with more striking mean IOP elevation only during the dark phase had
greater obstruction, enough to accentuate elevation during the periods
of highest aqueous production. Extensive angle closure apparently
reduced aqueous outflow enough to produce elevated IOP during both the
light and the dark phases, at both rates of aqueous humor formation.
Although sustained IOP elevations during both the light and dark phases
clearly result in devastating optic nerve injury (
Table 1 , group 3),
persistent dark-phase elevations in IOP with intervening normal
light-phase values can also produce axonal degeneration. Of the six
animals in group 2, four demonstrated partial, but unequivocal evidence
of optic nerve damage. The absence of histologic evidence of nerve
damage in the other two probably reflects individual variation in optic
nerve sensitivity to IOP and the relatively short duration of this
experiment. More prolonged dark-phase elevations would probably have
resulted in identifiable injury in these animals as well.
In humans, exaggerated IOP fluctuations are also thought to be
associated with optic nerve damage and may have important implications
for the prognosis of glaucoma.
30 31 Undetected increases
in pressure may precede the development of disc cupping and field loss.
Other investigators have noted that patients with progressive field
defects tend to have a greater magnitude of IOP circadian variation
than do patients with stable visual fields.
32
In this study, large fluctuations in IOP usually appeared as
repetitive, periodic, sharp increases in IOP during the dark that
nearly doubled the normal circadian oscillation, as illustrated in
Figure 2 . Because four of these eyes also demonstrated partial optic
nerve injury without experiencing sustained IOP elevation during both
phases, this shows that repeated, periodic, sharp increases, or spikes,
in IOP are capable of producing optic nerve damage.
Similarly, undetected intermittent IOP elevations in humans have
been identified as a potential mechanism for optic nerve damage in eyes
with otherwise normal IOP.
33 34 35 Both situations suggest
that the mechanisms by which elevated IOP damages the optic nerve may
be cumulative, with a limited capacity to recover after restoration of
normal IOP. How or why this occurs is currently unknown, but our model
may provide the means to study this problem and lead to important
insights into understanding this puzzling facet of glaucomatous optic
neuropathy.
Our findings validate the reliability of using the tonometer (TonoPen)
to measure IOP in awake rats.
10 They also confirm that the
timing of pressure measurement is critical to the accurate
documentation of IOP history.
4 5 Because IOP is typically
higher in the dark phase, measurements at the same time of day should
yield more reproducible values for monitoring IOP over time or for
comparisons among groups of animals or treatments. However, several of
our animals demonstrated large dark-phase IOP elevations, without a
significant mean elevation in the light. This demonstrates that
light-phase measurements alone did not accurately reflect the true IOP
exposure of the eye. In these eyes, pressure measurements obtained
during the light phase only would have yielded normal pressure
readings. The optic nerve damage observed in several of these eyes
would have gone unexplained or resulted in mistakenly attributing this
damage to abnormal sensitivity to normal IOP or to the injection
procedure itself.
Conversely, these data suggest that animals with IOP elevated during
the light phase are likely to have extensive, near total optic nerve
damage. Thus, analyzing only optic nerves from eyes that manifest
elevated IOP, determined by light-phase measurements alone, could give
the erroneous impression that IOP-induced damage in rats is an
all-or-none phenomenon. Rats housed in a light–dark cycle after
aqueous humor outflow obstruction should have IOP monitored during both
the light and dark phases of the circadian cycle.