Investigative Ophthalmology & Visual Science, Vol. 29, No.

9, September 1988
Copyright © Association for Research in Vision and Ophthalmology

Pigment Adherence as a Measure of Retinal

Adhesion: Dependence on Temperature
Eric G. Endo, Xiao-Ying Yao, and Michael F. Marmor

Retinal adhesion deteriorates rapidly after enucleation. We have developed a new in vitro method for
measuring retinal adhesiveness in the rabbit that is faster than previous models, and can monitor
changes sooner after enucleation. We used the percentage of retina which retained pigment after
peeling from the RPE as a quantitative measure of retinal adhesiveness. We found that the failure of
adhesion after death is more rapid and severe than previously reported, but can be inhibited by cold
temperature. Pigment adherence was also modified by ionic and metabolic factors that have been
found to affect other indices of adhesion. These results emphasize the limitations of in vitro data
relative to the physiologic forces that maintain retinal adhesion in vivo. Invest Ophthalmol Vis Sci
29:1390-1396, 1988

Most previous studies on retinal adhesion have Materials and Methods

measured the force required to peel or pull retina
from the RPE several minutes after enucleation.1"6
The preparation time for such experiments limits
their usefulness, since retinal adhesion deteriorates
markedly and rapidly after enucleation, even when
tissues are maintained in a supportive medium at
37°C.6 By the time these studies have assessed retinal
adhesion, the mechanisms of adhesion may have
changed significantly from those normally involved
in the living eye.
We have developed a faster in vitro method for
measuring retinal adhesiveness in the rabbit, that
overcomes some of the limitations of past experimen-
tal models. In our previous studies, we observed that
strongly adherent retina, when peeled from the RPE,
is heavily pigmented.6 We surmised that this pigment
which remains adherent to the retina after peeling
from the RPE might serve as an efficient method for
measuring retinal adhesiveness. Our initial investiga-
tions with this methodology show that adhesiveness
can be measured within the first minute after enucle-
ation, that the failure of adhesion after death is re-
markably severe, and that the loss of adhesion is sen-
sitive to temperature and other factors.
From the Department of Ophthalmology, Stanford University
Medical Center Stanford, California.
Supported in part by NIH-NEI Research grant EY-01678
(MFM).
Submitted for publication: October 10, 1987; accepted March
11, 1988.
Reprint requests: Michael F. Marmor, MD, Department of Oph-
thalmology, Stanford University Medical Center, Stanford, CA
94305.
These experiments adhere to the ARVO Resolu-
tion on the Use of Animals in Research. Dutch rab-
bits weighing 0.8-1.2 kg were sedated with 2 mg/kg
acepromazine and anesthetized with 20 mg/kg xyla-
zine and 150 mg/kg ketamine, given intramuscularly.
Eyes were rapidly enucleated, hemisected behind the
pars plana, and the vitreous gently removed. Using a
razor blade, the inferior segment of the fundus was
cut into three or four wedges (Fig. 1) approximately
1.2 cm long. The scleral side was glued with instant-
acting cyanoacrylate to glass slides, and the slides
were placed immediately into a Petri dish containing
medium. This preparation required no more than 40
seconds after enucleation.
RPMI (Gibco, Grand Island, NY) served as the
bathing medium except for those experiments exam-
ining the effects of calcium and ouabain (Sigma, St.
Louis, MO), for which Hanks' solution (Gibco) and
Ames' solution (buffered with 95% O2 and 5% CO2)7
were used respectively. Acetazolamide (Lederle, Pearl
River, NY) was used in some experiments.
Following an incubation period, the submerged
retina was gently peeled from the underlying tissue at
a slow steady rate (approximately 120 mm/min),
using jeweler's forceps. The peeled retina was flat
mounted on a glass slide andfixedfor 2 hr with 1.25%
glutaraldehyde and 1% paraformaldehyde in 0.072
mM cacodylate buffer, pH 7.4. The percentage of
retina covered with pigment was estimated by visual
examination on a light box using a magnifier. This
was always done in a blind fashion, without knowl-
edge of experimental conditions. Pigment adherence

1390

Downloaded from iovs.arvojournals.org on 03/11/2019

 No. 9 PIGMENT ADHERENCE AS A MEASURE OF RETINAL ADHESION / Endo er ol 1391

NASAL TEMPORAL

Fig. 2. Peeled retinae with RPE fragments adherent to (left) 90%,

Fig. 1. Diagram of the rabbit fundus showing the regions used in (middle) 50%, and (right) 10% of the retinal surface.
pigment adherence experiments.

was scored from 0-100%, in increments of 10% There was no significant difference in pigment ad-
(Fig. 2). herence between right and left eyes of the same ani-
For scanning electron microscopy, peeled retina mal when exposed to identical experimental condi-
and underlying tissue were pinned to thin sheets of tions. Nevertheless, whenever possible, individual
wax with insect pins and fixed as described above. experiments were designed to compare tissues from
The tissues were dehydrated with ethyl alcohol and within a single eye.
critical point-dried before routine coating and exami- Peeling rate has been shown to influence measure-
nation with an ISI scanning electron microscope. ments of retinal adhesiveness within a range of very
slow mechanically driven speeds.4 To determine
Results whether pigment adherence was sensitive to varia-
tions in the faster range of peeling rates that occur
Validity of the Model with manual peeling, we incubated tissue in RPMI at
25 °C for 1, 5 or 10 min to obtain three levels of
In this method, the peeling is manual and the scor-
ing subjective. Several experiments were performed
to insure that human variables did not alter the valid-
ity or reliability of our measurements. 25°C
To determine the reliability of estimating the 100
amount of adherent pigment, one observer (EE)
scored the same set of 20 specimens twice, 2 weeks
apart. There was no significant difference between
the two scores for each specimen; the average differ-
ence was only 0.5% ± 1.1% (mean ± SEM) and no
difference exceeded 10%. Two observers (EE and
XY) independently ranked another set of 20 speci-
mens, but again the average difference was only 0.5%
± 1.1% and ho difference exceeded 10%.
To test whether there are regional differences in the
retina, with respect to pigment adherence, four sec-
tions of the eye (see Fig. 1) were compared with one
another. A total of 26 eyes were incubated in RPMI
for various durations and at various temperatures,
but all four sections from each eye were peeled under 50 100 150 200
the same conditions. Pigment adherence was 11-16%
Peeling rate (mm/min)
lower in regions A and D, relative to B and C (P
< 0.01). However, there was no significant difference Fig. 3. Effect of manual peeling rates on pigment adherence. The
between regions B and C. Therefore, retina sectioning figure shows results after a L min incubation at 25°C; regression
analysis shows no significant effect. After a 5 or 10 min incubation,
was modified so that only the central regions of the the line was shifted downward but there was still no relationship
retina were used in later experiments. In addition, between peeling rate and pigment adherence, at rates up to 330
retina sections were rotated within each experiment. mm/min.

Downloaded from iovs.arvojournals.org on 03/11/2019

 1392 INVESTIGATIVE OPHTHALMOLOGY & VI5UAL SCIENCE / September 1988 Vol. 29

Fig. 4. Scanning electron micrographs of the photoreceptor surface (left) and RPE surface (right) after peeling the layers apart. Top: After 5
min incubation at 4°C, retina was strongly adherent and separated only with RPE cell cleavage and rupture. Most of the outer segments are
obscured by adherent fragments of RPE cells, and the RPE surface shows several levels of cleavage through the cells. Bottom: After 10 min
incubation at 37°C, retina was only weakly adherent and separated cleanly. The larger magnification bars = 10 pim.

pigment adherence} and for each we plotted the effect
of peeling at different speeds (Fig. 3). Peeling rates
well above and below our usual rate (approximately
120 mm/min) had no significant effect on pigment
adherence.
Our gross observations showed RPE pigment on
the retina, but could not rule out the possibility that
photoreceptor outer segments were also being rup-
tured with adherence to the RPE, an event which
might confound judgments of adhesiveness based on
RPE pigment alone. To determine whether outer
segment adherence was occurring, we examined the
peeled surfaces of both retina and RPE by scanning
electron microscopy. We found that separation of the
two layers occurred at the subretinal space, except
under conditions of strong adhesion when many RPE
cells ruptured, leaving apical fragments adherent to
the retina (Fig. 4). We did not find sheared outer
segments on the RPE surface.
Modifying Factors
Previous work measuring peeling force has shown
that retinal adhesion diminishes steadily after enucle-
ation,6 but the earliest measurements were made 4-5
min after enucleation. We measured pigment adher-
ence in a large number of eyes at 37°C, beginning 30
seconds after enucleation (Fig. 5), and found that
most of the loss of adhesion occured within the first
2-3 min. During the first minute, pigment adherence
was nearly 100%.
Our previous experiments with retinal peeling were
performed only at 37 °C in an effort to simulate phys-
iologic conditions. To determine whether the rapid

Downloaded from iovs.arvojournals.org on 03/11/2019

 No. 9 PIGMENT ADHERENCE AS A MEASURE OF RETINAL ADHESION / Endo er ol
37°C
iooH
0>
E 50-
Time after enucleation (min)
Fig. 5. Deterioration of pigment adherence after enucleation for
tissue incubated at 37°C. The curve represents a logarithmic re-
gression.
fall-off in adhesion after enucleation was temperature
sensitive, we compared the course of pigment adher-
ence at three different temperatures. Figure 6 shows
that cooler temperatures delay the deterioration of
retinal adhesion; in fact, in some experiments, strong
adhesion was maintained at 4°C for 18 hr.
We also studied other factors which have been
shown with different methodology to weaken (acidic
pH and calcium depletion)8 or strengthen (ouabain
and systemic acetazolamide)6'9 retinal adhesion, as
measured 5-20 min after enucleation. Figure 7 shows
that low pH markedly lowered pigment adherence,
not only at 37°C as found previously, but also at
37°C
Fig. 7. Effects of pH on
O)
pigment adherence at two ~
different temperatures.
Downloaded from iovs.arvojournals.org on 03/11/2019
1393
E
CL
Incubation time (min)
Fig. 6. Effects of temperature on the loss of pigment adherence
after enucleation. Pigment adherence is shown as a function of
incubation time at 4°C, 25°C and 37°C. On this and subsequent
Figures, bars indicate the standard error of the mean.
25°C, while an elevated pH had little or no effect at
either temperature. The effect of low pH was more
striking 1 min after enucleation than at 5 or 10 min,
when pigment adherence was low and thus was no
longer a sensitive indicator of adhesiveness. Pigment
adherence was lowered similarly when the tissue was
incubated in calcium and magnesium-free Hanks'
solution with 1 mM EDTA (Fig. 8).
Incubating eye cup tissue in 25°C or 37°C Ames'
solution with 0.5 mM ouabain increased pigment ad-
herence (Fig. 9), an effect which lasted for at least 20
min. The administration of intravenous acetazol-
amide (15 mg/kg) 30 min before enucleation also
25°C
10 0 10
Incubation time (min)
 1394 INVESTIGATIVE OPHTHALMOLOGY G VISUAL SCIENCE / September 1988 Vol. 29

37°C 25°C previous techniques; measurements are possible

100'
Control
Ca-Mg-free
c as a quantitative index of retinal adhesiveness.
Q)
E ject to human error and require comment. First, pig-
Incubation time (min)
Fig. 8. Effect of low calcium and magnesium on pigment adher-
ence at two temperatures. The tissues were incubated in calcium
and magnesium-free medium with EDTA.
increased pigment adherence at 37°C (Fig. 10), while
1 mM acetazolamide in the bathing medium at 25°C
had no effect.
Discussion
We have described a new in vitro method for evalu-
ating retinal adhesiveness in the pigmented rabbit
that allows much more rapid experimentation than
within 30 seconds of enucleation. The method is
based oh the finding that strongly adherent retina
cannot be separated cleanly from the RPE because
RPE cells give way and cleave before cellular separa-
tion occurs. This leaves pigmented cell fragments ad-
herent to the retinal surface. We have shown that the
percentage of retina covered with pigment can serve
Some of the procedures used in this model are sub-
ment adherence was estimated by direct visual exam-
ination. We considered the use of photometric or
densitometric analysis, but retinae vary in translu-
cency and color from rabbit to rabbit and sometimes
from region to region. The assessment of pigment
adherence must account for this variation, and this
would require very sophisticated correction (and pos-
sibly error) in an automated system. Our experimen-
tal measurements show excellent reproducibility and
reliability, despite the subjectivity of scoring, and vi-
sual estimation furthers our goal of establishing a
simple and efficient procedure.
Secondly, the peeling rate in our procedure is sub-
ject to human variability. DeGuillebon and Zauber-
man showed that there was a semilogarithmic rela-
tionship between peeling force and peeling velocity
between rates of 2-42 mm/min. 4 We considered
using a mechanical device to peel the retina at a con-
stant rate, but this would have significantly length-
ened the preparation time. Furthermore, we found
that it was very difficult to peel retina manually and
steadily at a rate less than 40 mm/min, but in the
range of 40-330 mm/min, there was no relationship
between pigment adherence and manual peeling rate.

37°C 25°C

100 100

Ouabain
Ouabain

Qi Fig. 9. Effect of ouabain

on pigment adherence at
O) 50- 50-
two temperatures. The con-
a. trol curve at 37°C is repro-
duced from Figure 6.
Control Control

10 15 20 10

Incubation time (min)

Downloaded from iovs.arvojournals.org on 03/11/2019

 No. 9 PIGMENT ADHERENCE AS A MEASURE OF RETINAL ADHESION / Endo er ol
If pigment adherence is a measure of retinal adhe-
siveness, it should be dependent on factors that have
been shown previously to influence the force required
to peel retina from the RPE. Comparing our results
with those obtained using a direct measurement of
adhesive force (peeling force)',6'819 we found that incu-
bation in low pH or low calcium reduced pigment
adherence, just as it reduces peeling force. Ouabain
and intravenous administration of acetazolamide
comparably increased pigment adherence, while acet-
azolamide in the bathing medium had no effect, as
was the case in peeling experiments. However, the
new model allows us to show that experimental ef-
fects occur more rapidly, with greater intensity and
over a broader range of temperatures than previously
realized.
Although we have documented that pigment ad-
herence serves well as a rapid and semiquantitative
index of retinal adhesiveness, the amount of adherent
pigment must be a function of RPE and photorecep-
tor cell integrity and strength, as well as the strength
of the bond to the outer segments. If the RPE cells are
severely weakened, they will fragment and adhere to
the retina irrespective of the strength of the bond to
retina. Similarly, conditions which fragment the
outer segments could cause an erroneous conclusion
that adhesion is weak, by allowing separation through
the outer segments (which we ordinarily do not ob-
serve) rather than at the subretinal space. We have,
for example, found there is often marked pigment
adherence after poisoning the RPE with sodium io-
date, even though the force required to peel off the
retina is low by direct measurement (Yoon, Yao and
Marmor, unpublished experiments). Retinal adhe-
sion should be measured with caution (using any
methodology) under conditions that are likely to
damage the integrity of cells.
We found that the deterioration in retinal adhesion
after enucleation occurred more rapidly and with
greater severity than previously recognized with the
peeling model. The earliest peeling measurements of
retinal adhesion were made 4-5 min after enucle-
ation, after which the peeling force decreased steadily
in a linear fashion. Using the pigment adherence
model, retinal adhesion deteriorated precipitously
within the first 2-3 min of enucleation and pigment
retention was reduced to approximately 15% after 5
min. These results emphasize the importance of
measuring retinal adhesion as soon as possible after
enucleation, and raise concerns about evaluating
physiologic mechanisms of retinal adhesion with in
vitro techniques that require long preparation times.
The latter are not invalid but may be measuring
Downloaded from iovs.arvojournals.org on 03/11/2019
1395
37°C
100
Acetazolamide
c
<D
50-
E
O)
Control
0 5
Incubation time (min)
Fig. 10. Effect of intravenous acetazolamide on pigment adher-
ence at 37°C.
forces that are not primary in the maintenance of
adhesion during life.
Another point to consider with respect to our data
is that we may not be measuring the full range of
physiologic adhesive forces within the first 2 min of
enucleation. Once pigment adherence nears 100%,
we no longer can judge the magnitude of adhesion. It
could be barely sufficient to cleave the RPE, or con-
siderably stronger, without causing any further effect.
Our new methodology gets closer to physiologic ad-
hesive forces than techniques which require more
preparation time, but in vivo methods will be needed
to monitor variations within a truly physiologic
range. Similarly, our methodology cannot monitor
very small adhesive forces. Even when pigment ad-
herence was 0%, the retina did not necessarily sepa-
rate spontaneously from the RPE. Even these mini-
mal adhesive forces could have relevance in some
clinical situations.
Our results show that the failure of retinal adhesion
after death can be inhibited by cold temperatures. At
25 °C, pigment adherence was greater than 50% after
5 min compared to 15% in 37°C. The inhibition was
even more striking at 4°C where pigment adherence
remained at 85% or more for 18 hr. These data might
seem to suggest that retinal adhesion is lost after
 1396 INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / Seprember 1988
death by a failure of temperature-dependent meta-
bolic or enzymatic processes. However, later experi-
ments with this system (unpublished data) have
shown that temperature effects are rapidly reversible,
which is inconsistent with the hypothesis that adhe-
sive failure involves irreversible metabolic or enzy-
matic decay.
Key words: retinal adhesion, retinal pigment epithelium
(RPE), temperature, ouabain, acetazolamide
References
1. Zauberman H and Berman ER: Measurement of adhesive
force between the sensory retina and pigment epithelium. Exp
Eye Res 8:276, 1969.
2. Bloch D, O'Connor P, and Lincoff H: The mechanism of the
cryosurgical adhesion: III. Statistical analysis. Am J Ophthal-
mol 71:666, 1971.
Downloaded from iovs.arvojournals.org on 03/11/2019
Vol. 29
3. De Guillebon H, DeLaTribonniere MM, and Pomerantzeff O:
Adhesion between retinal and pigment epithelium: Measure-
ment by peeling. Arch Ophthalmol 86:679, 1971.
4. De Guillebon H and Zauberman H: Experimental retinal de-
tachment: Biophysical aspects of retinal peeling and stretching.
Arch Ophthalmol 87:545, 1972.
5. Owczarek FR, Marak GE, and Pilkerton AR: Retinal adhesion
in light-and-dark-adapted rabbits. Invest Ophthalmol 14:353,
1975.
6. Marmor MF, Abdul-Rahim AS, and Cohen DS: The effect of
metabolic inhibitors on retinal adhesion and subretinal fluid
absorption. Invest Ophthalmol Vis Sci 19:893, 1980.
7. Ames A III, Isukada Y, and Nesbett FB: Intracellular Cl", Na+,
K+, Ca++, Mg++ and P in nervous tissue: Response to gluta-
mate and to changes in extracellular calcium. J Neurochem
14:145, 1967.
8. Marmor MF and Maack T: Local environmental factors and
retinal adhesion in the rabbit. Exp Eye Res 34:727, 1982.
9. Marmor MF and Maack T: Enhancement of retinal adhesion
and subretinal fluid absorption by acetazolamide. Invest Oph-
thalmol Vis Sci 23:121, 1982.