Download Changes in glucose level affect rod function more than cone

Investigative Ophthalmology & Visual Science, Vol. 33, No. 10, September 1992
Copyright © Association for Research in Vision and Ophthalmology
Changes in Glucose Level Affect Rod Function More
Than Cone Function in the Isolated, Perfused Cat Eye
Claudio Macaluso,* Shoken Onoe,t and Gunter Niemeyer
The glucose concentration (gl) in mammalian serum incorporates a normal range of variation of several
millimoles. We studied the effects of such variations on light-evoked electrical signals in the in vitro
arterially perfused cat eye, avoiding extraocular regulatory mechanisms that might confound data
interpretation. Changes in gl from the nominal control value of 5 mmol/1 were maintained for 5-40 min.
Stimuli of near rod threshold intensity were presented in full dark adaptation, and stimuli of higher
intensity were presented in the presence of a white background for cone responses. We recorded the
dc-electroretinogram (ERG), the scotopic threshold response (STR), the optic nerve response (ONR),
and the transretinal slow P-III and transepithelial retinal pigment epithelium c-wave from the subretinal space. The ocular standing potential changed by up to ±2 mV in parallel with an increase and
decrease in gl, independent of the adaptation condition. Our results show that the rod-ERG, STR, and
rod-driven optic nerve response (ONR) have a marked sensitivity to small changes in gl (±1 to 3
mmol/1). The field potentials increased and decreased in parallel with changes in gl. The cone ERG and
cone ONR, in contrast, failed to respond consistently to increases in gl and revealed decreases in
amplitudes only with an extreme decrease in gl. Decrease in gl, down to 2 mmol/1 and less, is known to
induce drastic behavioral and electrophysiologic phenomena in the central nervous system. Our results
imply that the "normal" glucose level, at least in the cat, could be marginal for rod-mediated retinal
function. The results also suggest a marked difference in metabolic mechanisms for cone versus rod
photoreceptors. Invest Ophthalmol Vis Sci 33:2798-2808,1992
for maintenance of retina or central nervous system
preparations in vitro contain 10-20 mmol/1 glucose.14
Arterial perfusion of isolated cat eyes15"19 usually is
performed using oxygenated tissue culture medium
containing 4.5-5.5 mmol/1 glucose, corresponding to
the range of glucose concentration (or 'gl') of cat
serum. The glucose serum level can change in various
metabolic conditions. The supply of glucose to central
nervous system tissue can be reduced because of ischemia and hypoxia. It also can be enhanced locally (eg,
by adenosine-induced glycogenolysis).20"21
We sought to systematically study the effects of
transient, stepwise increases and decreases in the arterial supply of glucose in the isolated perfused cat eye.
The preparation is well suited for changing constituents of the perfusing medium and observing the resulting effects without the multiple regulatory responses that counteract such changes in situ (ie, insulin and glucagon secretion). We monitored the effects
on light-evoked and light-independent field potentials
that reflect multiple levels of retinal information processing: the standing potential, the ERG b- and cwave, the scotopic threshold response, and the optic
nerve response. The signals were recorded under selective stimulation of the rod system and the cone
system. We found a remarkable sensitivity to discrete
Little is known about the metabolic requirements
of the fully dark adapted mammalian retina. However, retinal oxygen profiles,1 glucose uptake distribution,2 intraretinal pH assessments,3"5 and biochemical
data6"10 indicate there is a significant difference from
the light adapted state. The supply of glucose as the
major source of energy may, like that of oxygen, be
marginal in the fully dark adapted state.""13 Although
the serum levels of glucose vary in humans (3.9-5.6
mmol/1), oxen (2.3-4.1 mmol/1), sheep (2.4-4.5
mmol/1), horses (3.5-6.3 mmol/1), pigs (3.7-6.4
mmol/1), dogs (3.4-6.0 mmol/1), cats (3.4-6.9 mmol/
1), and rats (5.5-11 mmol/1), several standard media
From the Neurophysiology Laboratory, Department of Ophthalmology, University Hospital, Zurich, Switzerland.
This work was supported in part by the Kusch-Neumann Fonds,
University of Zurich.
* Present address: Istituto di Oftalmologia, University of Parma,
Italy.
f Present address: Department of Ophthalmology, Iwate Medical
University, Morioka, Japan.
Submitted for publication: November 15, 1991; accepted March
10, 1992.
Reprint requests: G. Niemeyer, Neurophysiology Laboratory,
Dept. of Ophthalmology, University Hospital, CH 8091 Zurich,
Switzerland.
2798
Downloaded From: http://iovs.arvojournals.org/ on 08/03/2017
No. 10
EFFECTS OF GLUCOSE ON RETINAL FUNCTION / Mocoluso er ol
changes in gl for rod-driven signals, but not for conedriven signals.
Methods
Details of the experimental method of arterial perfusion of the isolated mammalian eye have been published.15"1722 In the present study, eyes were enucleated from deeply anesthetized cats, in accordance
with the ARVO Resolution on the Use of Animals in
Research and with the regulations of the cantonal veterinary authority office of Zurich. The ophthalmociliary artery was cannulated, and the eye was perfused
at 37°C at a flow rate of 1-2 ml/min. The perfusate
consisted of tissue culture medium (TC 199, Earle's
salts; Amimed, Basel, Switzerland) and 30% newborn
calf serum, buffered with HEPES buffer and 26
mmol/1 bicarbonate, adjusted to a pH of 7.4 (at
37°C), and oxygenated by bubbling with 95% O2 and
5% CO2 for 15 min at a rate of 150 ml/min. The final
PO2 was 400-500 mmHg. The gl of "standard" perfusate was 5.5 mmol/1, identical with perfusate used in
previous studies from our laboratory. Glucose concentration could be increased by continuously injecting additional glucose from a stock solution into the
perfusion system. To decrease the gl to below 5.5
mmol/1, perfusate containing only 1.3 mmol/1 glucose was used. In such cases, additional glucose (4.2
mmol/1) was continuously infused to keep gl at a level
of 5.5 mmol/1 in control and recovery phases. The
rate of this injection was reduced for the "hypoglycemic" test periods.
Photic stimuli were provided by a 150 W xenon
lamp (11.54 log quanta [507 nm] deg~2 sec"1) at the
cornea, and attenuated by neutral density and monochromatic filters to achieve rod-matched conditions
in full dark adaptation. The eye was stimulated from
this source in Maxwellian view via a modified fundus
camera.23 The stimulus duration was 400 msec and
the interval was at least 30 sec for recordings of rodmatched signals (50 msec duration for scotopic threshold response recordings). For cone stimulating conditions, an adapting beam (white; 8.6 log quanta [507
nm] deg~2 sec"1) was used to suppress the rod contribution.
The light-evoked potentials, ERG, and scotopic
threshold response (STR) were recorded by salt bridge
Ag-AgCl electrodes24 positioned in the vitreous and
on the posterior scleral surface. A pair of Ag-AgCl
electrodes were positioned on the surface and on the
proximal end of the optic nerve to record the compound action potential of the optic nerve (optic nerve
response, or ONR). For recordings from the subretinal space, glass microelectrodes were made using a
Brown-Flaming puller (Suiter Instrument Co., Novato, CA) and werefilledwith 2 mol/1 potassium-ace-
Downloaded From: http://iovs.arvojournals.org/ on 08/03/2017
2799
tate, resulting in an impedance of 15-70 mohm. By
referencing the microelectrode in the subretinal space
to the vitreous and to the sclera, the transretinal potential (TRP) and the transepithelial potential (TEP),
respectively, were monitored. For these recordings,
rod-matched stimuli, 4 sec in duration, were used.
After cannulation, 20 msec light pulses (narrow
band filter, Xmax 620 nm) were used to monitor the
condition of the eye from the beginning of perfusion
for at least 60 min of dark adaptation, adjusting the
flow rate for a standard amplitude of the vitreal bwave of 700-800 fiV. The standing potential (SP) and
the light-evoked ERG and ONR were recorded at one
low, near threshold intensity, or over a range of intensities before, during, and after changing gl. The duration of changes in gl was 3-15 min, but was 20-60
min for recording of amplitude/intensity functions
(V/log I).
Results
Changing the concentration of glucose in the perfusate affected the electrical activity of the pigment epithelium, retina, and optic nerve in different ways. The
rod-driven light-evoked signals—ERG b-wave and cwave, ONR, and STR—increased or decreased parallel to small changes in gl. Cone-driven light-evoked
signals (b-wave and ONR) failed to respond to small
changes in gl, but decreased during extreme hypoglycemia. The light-independent SP changed in parallel
to changes in gl, independent of dark- or light-adaptation.
ERG b-Wave
The ERG b-wave, recorded under rod-matched
conditions, showed remarkable, immediate, and reversible responses to the changes in gl. The amplitude
of the rod b-wave clearly was enhanced during elevation of the gl above the "standard." The time course
of the changes in b-wave amplitude is shown in Figure
1. The enhancing effect started within 1 min from the
beginning of gl elevation. During the return to control
gl, the b-wave amplitude showed a slight undershoot
prior to recovery. The enhancement varied between 8
and 125% (n = 25), with a tendency to saturate at
8-10 mmol/1 glucose (Fig. 2). There was more variability between preparations than within trials on the
same isolated eye. Dose-dependency was tested and
found to be present in five out of eight preparations.
When we exposed preparations a priori to a higher gl
of 8 or 10 mmol/1, the enhancing effect of additional
glucose was smaller or absent. In three out offiveexperiments of this type, an undershoot was seen after a
return to 8 or 10 mmol/1 gl before recovery, much as
described for the changes starting with standard perfusate.14
2800
A
35
recovered completely upon return to the control level
within 5 min.
An example of the V/log I function for the rod bwave near threshold is shown in Figure 3. During increased gl, the gain of the b-wave was increased in the
rod range (Fig. 3, left), without revealing a clear
change in threshold. Decreasing gl induced a slight
reduction in sensitivity (shift of the V/log I curve
along the abscissa; Fig. 3, right).
Under cone-stimulating conditions, in contrast, the
ERG b-wave was not or only minimally affected by
increasing (n = 7) or decreasing (n = 4) gl (Fig. 2,
empty symbols). Only extensive decreases in gl of
about 4.2 mmol/1 induced marked depression of the
cone-driven b-wave.
°
300
^
250
"§
200
"5.
o
I
150
100
50
increased [glucose] ( + 3 mM)
i
-2
0
2
i
i
4
6
8
10
12
min
B
35
° T
300
ampli tud
Vol. 33
INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / September 1992
Optic Nerve Response
Changes in gl affected all parts of the ONR, the
initial ON, the following plateau, and the OFF components, corresponding to the time course of the light
stimulus, under rod-matched conditions. Figure 4
shows the time course of the effect on the ON-component of the ONR, revealing an increase in amplitude, similar to the effect on the b-wave. After termination of the step increase in gl, there was a slight
250
350
200
300
r
decreased [glucose] (-4.2 mM)
150
•S
250
b—wa\
3
a.
E
100
200
D
50 -
150
oo
100
-5
10
15
20
25
30
CO
C
D
_C
o
Fig. 1. Time course of typical changes in normalized amplitudes
of the rod ERG b-wave in response to elevation (A) and decrease (B)
in concentration of glucose gl. The b-wave was enhanced by additional glucose and depressed by decreased glucose. Transient increase in gl was followed by a slight undershoot after termination of
the injections. The insets show corresponding original traces of rod
ERGs before and during (thicker line) the changes in gl.
Figure 1 shows the effect of reducing gl on the rod
b-wave (n = 4). The amplitude decreased as soon as gl
decreased, with the maximal effect after 2 min. The
attenuation was dose dependent (Fig. 2, filled symbols). After gl was decreased for 17 min, the b-wave
Downloaded From: http://iovs.arvojournals.org/ on 08/03/2017
50
-5
-3
-1
change in [glucose] (mM)
Fig. 2. Changes in normalized amplitudes of the rod (filled symbols) and cone (empty symbols) ERG b-wave in response to
changes in gl. The intersection of the reference lines indicates the
control amplitude and control gl in the perfusate (5.5 mmol/1).
Whereas the rod signals are sensitive to both increase and decrease
in gl with tendency to saturation (+2.5 to +4.5 mmol/1 change), the
cone b-wave was affected only by extensive decrease in gl by -4.2
mmol/1.
No. 10
2801
EFFECTS OF GLUCOSE ON RETINAL FUNCTION / MQCQIUSO er ol
60+ 4.5mM glue.
X>
|
40
Q.
Fig. 3. V/log I function
for rod ERG b-wave near
threshold, under standard gl
(circles), under elevated gl
(triangles), and under lowered gl (squares).
E
O
20
•o
o
-2.8mM glue.
E610A
J
-6
E611D
-7
-6
log relative intensity
undershoot before recovery. The enhancing effect
also was observed in the OFF-component, whereas
the plateau changed to a lesser extent.
By "loading" the preparation with 8-10 mmol/1 gl
in the perfusate, the enhancement of the ONR with
additional glucose was small or absent.14
The effect of decreasing gl from the "standard"
concentration on the ON-component is shown in Figure 4b. Immediately, the amplitude was markedly attenuated, and 6 min after it returned to the control gl,
the response had recovered. A dose-response relation
was evident for changes in the range of —4.2 to +4.5
mmol/1 glucose referred to control (Fig. 5,filledsymbols). Enhancement or attenuation of the ONR amplitude was parallel to increasing (n = 18) or decreasing (n =6) gl. Figure 6 shows V/log I functions for the
ONR ON-component in the rod-matched range, near
threshold. By raising gl 4.5 mmol/1, the sensitivity was
increased (Fig. 6, left). At a gl reduced to 2.7 mmol/1,
it was decreased (Fig. 6, right).
Under cone-stimulating conditions, the ONR was
found to be almost unresponsive to increasing (n = 7)
and decreasing (n = 2) gl in the perfusate. However, a
substantial decrease of about 4.2 mmol/1 greatly attenuated the signal amplitude (Fig. 5, empty symbols).
Scotopic Threshold Response
Much like the rod-driven ERG b-wave and optic
nerve response, the STR, the most sensitive field po-
Downloaded From: http://iovs.arvojournals.org/ on 08/03/2017
tential of the rod system below and at b-wave threshold, showed clear enhancement or attenuation of its
amplitude parallel to increasing (n .= 5) or decreasing
(n = 3) gl, as illustrated in Figure 7. Experimental
series involving the effects of changes in gl on the V/
log I functions of the STR consistently showed an
increase and decrease in amplitude paralleltp corresponding changes in gl. In some series (two out of five)
there was evidence for an increase in sensitivity—that
is, shift of the V/log I curve to the left during an increase in gl.
Standing Potential
The SP was consistently affected by the changing of
the gl in rod- and cone-stimulating conditions. Original traces of changes in standing potential in response
to changes in gl are shown in Figure 8 (top). An increase in gl (n = 45) consistently induced a small transient decrease in SP, followed by a marked and maintained increase. Decreasing gl (n = 18) induced similar changes in SP with reversed polarity compared to
those seen under an increase in gl. These changes
reached the maximal and maintained effect after 10
min, in contrast to the faster changes in the lightevoked b-wave, ONR, and STR. No difference in
changes in SP was observed between rod- and conestimulating light conditions. The changes in SP were
dose dependent (Fig. 8, bottom) as most clearly recognized within the same preparation. Recordings from
the subretinal space (n = 9) showed that the changes
2802
INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / September 1992
350
A
because changing gl may trigger a response resulting
from modified osmolarity alone. Mannitol induced a
decrease in SP (50-300 ixV) that reached a plateau,
representing the hyperosmolarity response.25 This response is observed consistently as an initial, transient
decrease in SP during increases in gl and as an initial
increase in SP during decreases in gl (Fig. 8).
300
ampli tude
250
200
ERGc-Wave
150
ONR-
z
o
100
50 -
increased [glucose] ( + 3 m M )
0
-2
0
2
4
6
8
10
12
min
B
Vol. 33
35 r
°
300
Changes in glucose level markedly affected the amplitude of the ERG c-wave. Decreasing (n - 6) or
increasing (n = 11) gl consistently induced a decrease
or an increase, respectively, in the amplitude of the
vitreal c-wave (Fig. 9). The changes in c-wave amplitude were dose related (Fig. 9B) and closely followed
the changes in SP (Fig. 9A; only the effects of decreased gl are shown).
To identify which component contributing to the
c-wave was affected by glucose, we recorded from the
subretinal space. Referring the subretinally positioned microelectrode to the sclera permitted assessment of the transepithelial potential and the RPE cwave. Referring to the vitreous allowed the transretinal potential and slow P-III to be assessed. Analysis
250
350
"°
200
~a.
o
r
300
decreased [glucose] (-4.2 mM)
vS
150
0)
250
100
I
50
200
o
i
-5
0
5
10
15
20
25
30
min
Fig. 4. Time course of typical changes in normalized amplitudes
of the rod ONR-ON response during increase (A) and decrease (B)
in gl. The response was enhanced by additional glucose and depressed by decreased glucose. There was a slight undershoot after
termination of the injections. The insets show the corresponding
original traces of the rod ONR-ON before and during (thicker line)
the changes in gl.
in SP induced by glucose were due solely to a change
in the TEP, not to changes in the TRP.
In one eye, the effect on SP of injections of hyperosmotic solutions (4.5 and 10 mmol/1 mannitol, not
shown) was tested. This experiment was performed
Downloaded From: http://iovs.arvojournals.org/ on 08/03/2017
0>
CO
c
o
_c
o
150
O
100
_Q
50
-
5
-
3
-
1
1
3
5
7
9
change in [glucose] (mM)
Fig. 5. Changes in normalized amplitudes of the rod (filled symbols) and cone (empty symbols) ONR-ON component in response
to changes in gl from the standard (about 5.5 mmol/1) perfusate.
Whereas the rod signals are sensitive to both increase and decrease
of gl with a tendency to saturation at +2.5 to +4.5 mmol/1 change,
the cone ONR-ON component was affected only at extensive decreases in gl by -4.2 mmol/1.
EFFECTS OF GLUCOSE ON RETINAL FUNCTION / Mocoluso er ol
No. 10
2803
40
+ 4.5mM glue.
Q.
Fig. 6. V/log I function of
the rod ONR-ON component under standard (circles), elevated (triangles),
and under lowered gl
(squares).
en
-2.8mM glue.
o 20
10
CL
o
-8
-7
-8
-7
log relative intensity
of these light-evoked signals revealed that the changes
in the amplitude of the c-wave during increased (n
= 6) or decreased (n = 3) gl resulted from a change in
the amplitude of the transepithelial c-wave rather
than change in the amplitude of slow P-III (Fig. 10,
only the effects of decreased gl are shown).
control
Discussion
In this study, we found that changing the supply of
glucose, even within the "normal" range, influences
light-evoked signals and the standing potential, which
all arise from different retinal layers. That the various
retinal layers are affected differently in dark and light
adaptation supports and extends the concept of their
different metabolic characteristics. The external retina is recognized as being the most sensitive to the
light adapting condition, probably because the
photoreceptors reach their highest activity in the
d a r k
effect +3.6 mM glucose
effect - 2.7 mM glucose
control
50 msec
Fig. 7. Original STR traces under increased gl (+3.6 mmol/l,
top), and decreased gl (-2.7 mmol/l, bottom). The thicker line
shows the effects. (Calibration bars: 2 ^V [top] and 10 M V [bottom].)
Downloaded From: http://iovs.arvojournals.org/ on 08/03/2017
1-10,26-28
b-Wave
Small increases and decreases in gl increased and
decreased, respectively, the rod-driven b-wave, but
did not or only slightly increased and decreased the
cone-driven b-wave. The effect in the dark adapted
condition appears to be a change in gain rather than a
change in sensitivity, as derived from the V/log I
curves. The b-wave is thought to be generated by the
movement of K+ along the Miiller cell,29'30 as a result
of the light-evoked depolarization of the ON bipolar
cells.31 Our results regarding the b-wave could be due
to an effect of glucose on the neuronal activity or the
Miiller cell resting potential, as hypothesized by Masland and Ames.32 The finding of a clear-cut difference
of effects in rod- and cone-stimulating conditions
tends to exclude the latter possibility and points to a
2804
INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / Seprember 1992
change in standing potential
— — glucose
glucose
:
3 -
2/
2 -
1
+4.5mM
-4.2mM
•
-
1
A
•
-1
-2
* :
#
• °*
o o
- 5 - 3 - 1
1
3
5
the guinea pig in low glucose, suggesting a decrease in
the number of activated neurons. Creutzfeldt35 reported a decrease in the firing rate in cortical recordings from the anesthetized cat under hypoglycemia.
Changes in gl may directly affect the mechanisms that
control firing frequency. We favor the hypothesis that
in our ONR data the modulation induced by changes
in glucose is simply carried along the information processing pathway within the retina, resulting in
changes in the number or firing rate of activated ganglion celis.
Scotopic Threshold Response
o
n
\j
Vol. 33
7
9
change in [glucose] (mM)
Fig. 8. Changes in standing potential (SP) induced by changing
gl. Top: original traces; the heavy line indicates the duration of the
change in gl, which lasted 30 min and 25 min for the increase and
decrease, respectively. Bottom: changes in amplitude (mV) of the
SP induced by changes in gl in rod (filled symbols) and cone (empty
symbols) adapting conditions. The SP consistently increased during
increases, and decreased during reduction in gl, independently of
the light-adapting condition.
primary change in the neural response, mainly of rod
ON bipolar cells, underlying the changes in b-wave
amplitude.
Optic Nerve Response
The ONR reflects the light-evoked retinal ganglion
cells activity. Its configuration, which makes up the
ON, plateau, and OFF components, is understood to
be the summation of all the on-going excitation and
inhibition of the axons underlying the electrodes.17'33
The changes seen in the ONR, induced by small
changes in gl, are similar to those for the ERG b-wave.
They have a virtually identical time course, good
correlation of change in amplitude for a given change
in gl and, finally, marked effects in rod-matched, but
not cone-stimulating conditions.
Cox and Bachelard34 reported attenuation of the
population spike in superfused hippocampal slices of
Downloaded From: http://iovs.arvojournals.org/ on 08/03/2017
The STR is a newly characterized rod-driven component of the ERG, elicited by very dim illumination
below the threshold of the rod b-wave and originating
in the proximal retina.36 The mechanism for its generation possibly involves an increase in K+37 induced by
activation of inner retinal neurons.38 The glucose-induced changes in amplitude of the STR, parallel to
those observed for b-wave and ONR, provide clear
evidence for a remarkable sensitivity of the rod system to small changes in gl. Based on this view, it is not
surprising that the STR reacts similarly to the optic
nerve response.
Standing Potential and c-Wave
The increase and decrease in SP parallel to changes
in gl were consistent and reversible. They differed
from changes in b-wave, ONR, and STR in two ways:
(1) the changes in SP were independent of light and
dark conditions; and (2) they revealed a slower time
course. The c-wave amplitude also increased and decreased in parallel to corresponding changes in gl and
with a time course identical to that of the SP. Covariation of the two voltages39 has been described for several conditions, such as the enhancement of the cwave during the light peak.40 Our recordings from the
subretinal space identified the transepithelial component of the c-wave as the source of these changes, indicating a probable change in RPE resistance, rather
than a change in the light-induced decrease in [K+]o in
the subretinal space.40 Therefore, glucose changes
would not primarily affect the light-evoked electrical
activity—ie, hyperpolarization—of photoreceptors.
In preliminary experiments, the aspartate-isolated receptor potential (fast P-III) was not modified by changing gl. Similarly, Winkler7 reported that the fast P-III
amplitude remained constant during changes in gl, in
the range 1-10 mmol/1.
These features point to a different mechanism un-
No. 10
EFFECTS OF GLUCOSE ON RETINAL FUNCTION / MQCOIUSO er ol
2805
c-waves
0.5 mV I
©w
> a)
100
6§.
50
e-e
Glucose -1.7 mM
cc
Fig. 9. (A) DC recording of the standing
potential (upper trace) and amplitude of
the ERG c-waves (lower trace) during a decrease in gl plotted against the same time
scale. The c-wave amplitude appears to covary with the standing potential. (B)
Changes in normalized c-wave amplitudes
in response to changes in gl from the standard (about 5 mmol/1). The c-wave increased during increases and decreased
during reduction in gl. The inset shows
original traces of c-waves before (thin
trace) and during (heavy trace) an increase
in gl.
10
20
15
B
effect: + 3.5 mM glucose
300
Mi
'••
250
I control
T3
I
\
1 mV
200
4 sec
Q.
CO
6
I
100
c
CO
"
50
- 5 - 3 - 1
1
3
5
7
change in [glucose] (mM)
derlying the glucose-induced change in SP and c-wave
compared to the mechanism of the effects on b-wave,
ONR, and STR. Harik et al41 demonstrated an abundance of glucose transporters in the RPE, in addition
to their presence in capillaries of retina and optic
nerve—that is, at the blood-eye barriers. These glucose transporters are of the "brain-type," also referred
to as "erythrocyte-type" or GLUT-1,42 They perform
a facilitated transport, without cotransport with ions,
as demonstrated in the RPE.43"44 Modifications of the
rate of this nonelectrogenic transport of glucose
across the RPE membrane cannot explain the observed glucose-induced changes in SP. Rather, these
Downloaded From: http://iovs.arvojournals.org/ on 08/03/2017
might depend on metabolism of glucose in the RPE
cell, leading to change in electrogenic transport mechanisms.
The initial increase and decrease in SP opposite to
changes in gl were due to the osmolarity change induced by injection of glucose. Mannitol injections in
the same osmolarity range as glucose allowed us to
distinguish between simple osmotic and glucose-specific effects. Reasons why the change induced by glucose is smaller than that induced by mannitol might
be: (1) that glucose enters cells, whereas mannitol
does not; thus, the effective osmotic change pressure
of glucose is smaller; or (2) because of algebraic sum-
2806
INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / September 1992
Vol. 33
mV
1.5
Fig. 10. Changes in amplitude of the cwave recorded from the vitreous, and of its
components, transepithelial c-wave and
slow P-III, recorded from the subretinal
space, during a decrease in gl. Left column:
amplitude vs. time changes. Right column:
original traces; thicker line designates effects. The changes in the amplitude of the
c-wave during decreased gl resulted from a
change in the transepithelial c-wave rather
than in slow P-III.
7.5
Q.
a>
w
c
5.5
Glucose -1.7 mM
10
4 sec
15
min
mation with the nonosmotic, opposite, and more
prominent effect of gl on the SP.
The Photoreceptor Synapse as a Possible Site for the
High Sensitivity of the Retina to "Mild
Hypoglycemia" in the Dark
The effect of changing gl on b-wave and ONR
under cone stimulating conditions, showing a decrease only for extremely low arterial gl, is consistent
with data obtained in other parts of the central nervous system under hypoglycemia. Only severe hypoglycemia (1.2 mmol/1 and less) induces convulsions,
coma, and, finally, an isoelectric electroencephalogram in association with a significantly decreased
level of adenosine triphosphate and creatine phosphate.45"48 Central nervous system functions (monitored as electrical or as behavioral activity) are affected when the arterial gl falls below approximately 2
mmol/1, sometimes referred to as "mild hypoglycemia."34'46"50 The metabolic correlate of this condition
is very interesting. Different from what might be expected, the energy state of the tissue—eg, the concentration of high energy phosphates—is not lowered.34-45"49 Rather, failure of neurotransmission is
thought to be the underlying mechanism, because
glucose is metabolically coupled to molecules involved in the release of transmitters. Levels of gluta-
Downloaded From: http://iovs.arvojournals.org/ on 08/03/2017
mate, glutamine, alanine, 7 aminobutyric acid, and
acetylcholine have been found to be decreased, and
levels of aspartate, lysine, and NH4+ have been found
to be increased in mild hypoglycemia.45"51 The damage to and death of cells occurring in severe hypoglycemia have been demonstrated to be a result of an increase in excitotoxins—the aspartate/glutamate proportion reaches a new steady-state in hypoglycemia.51
In the present study, thefindingthat the rod-driven
b-wave, the ONR, and the STR respond to decreases
in gl to about 3 mmol/1 reveals an exquisite sensitivity
to hypoglycemia of the dark adapted retina. Parallels
to the high sensitivity to mild hypoxia of the dark
adapted cat retina in vivo are evident in the work of
Linsenmeier1 and Steinberg.52 This is attributed to the
high Na+/K+ ATPase activity of photoreceptors, necessary for keeping the light-sensitive cation channels
open.52 Because "mild hypoglycemia" does not inhibit energy metabolism, the highly energy-dependent Na+/K+ ATPase activity in photoreceptors
should not be influenced by lowering gl. Our finding
that the light-evoked electrical activity of photoreceptors is not affected by changing gl would confirm this
view. The possibility arises that the neurotransmission between photoreceptors and second order neurons is the site primarily affected by change in gl. This
hypothesis is consistent with three facts. First, in dark
adaptation, not only does the Na+/K+ ATPase pump
No. 10
EFFECTS OF GLUCOSE ON RETINAL FUNCTION / Mocoluso er ol
exert maximal activity, but neurotransmission to second order neurons reaches its highest rate. Second,
glutamate is considered most likely to be the neurotransmitter at those synapses.53'54 Third, glutamate
levels are significantly lowered in hypoglycemia.45"51
Conclusion
11.
12.
13.
These observations, particularly that small changes
in gl induced marked changes in the amplitudes of
rod-driven responses, raise the question whether the
supply of glucose in fully dark-adapted conditions is
marginal or suboptimal for rod-driven retinal function. From recording field potentials and from single
cell responses in this preparation, which are comparable to and exhibit the same sensitivity as in the in vivo
cat retina, we consider the isolated eye to be nourished with an adequate amount of oxygen. It remains
to be tested whether the light-evoked electrical signals
of the rod system in vivo also are sensitive to small
changes in gl.
14.
15.
16.
17.
18.
19.
Key words: glucose, retina, optic nerve, electrophysiology,
perfused eye
20.
Acknowledgments
We thank Dr. Urs Gerber, Zurich, for valuable comments
on the manuscript, and Ms. F. Uldry and Ms. F. Werren for
excellent technical assistance.
21.
References
22.
1. Linsenmeier RA: Effects of light and darkness on oxygen distribution and consumption in the cat retina. J Gen Physiol
88:521, 1986.
2. Bill A and Sperber GO: Aspects of oxygen and glucose consumption in the retina: Effects of high intraocular pressure and
light. Graefes Arch Clin Exp Ophthalmol 228:124, 1990.
3. Borgula GA, Karwoski CJ, and Steinberg RH: Light-evoked
changes in extracellular pH in frog retina. Vision Res 29:1069,
1989.
4. Oakley Bll and Wen R: Extracellular pH in the isolated retina
of the toad in darkness and during illumination. J Physiol
419:353, 1989.
5. Yamamoto F and Steinberg RH: [H+]o in the subretinal space
of cat: Interactions between arterial oxygen tension and intraocular perfusion pressure. Invest Ophthalmol Vis Sci 31(suppl):
388, 1990.
6. Sickel W: Retinal metabolism in dark and light. In: Handbook
of Sensory Physiology, Fuortes MGF, editor. Berlin, SpringerVerlag, vol. VII/2, 1972.
7. Winkler BS: Glycolytic and oxidative metabolism in relation to
retinal function. J Gen Physiol 77:667, 1981.
8. Kimble EA, Svoboda RA, and Ostroy SE: Oxygen consumption and ATP changes of the vertebrate photoreceptor. Exp Eye
Res 31:271, 1980.
9. Ostroy SE, Svoboda RA, and Wilson MJ: A stage in glycolysis
controls the metabolic adjustments of vertebrate rod photoreceptors upon illumination. Biochem Biophys Res Commun
168:155, 1990.
10. Winkler BS, Solomon FJ, and Orselli SM: Glucose metabolism
Downloaded From: http://iovs.arvojournals.org/ on 08/03/2017
23.
24.
25.
26.
27.
28.
29.
30.
31.
2807
in darkness and light as measured in whole and mutant retinas.
Invest Ophthalmol Vis Sci 32(suppl):l 135, 1991.
Niemeyer G, Onoe S, and Macaluso C: Glucose concentration
affects the rod-ERG and optic nerve response in the perfused
cat eye. Invest Ophthalmol Vis Sci 31(suppl):391, 1990.
Macaluso C, Onoe S, and Niemeyer G: Discrete changes in
glucose level affect rod- but not cone-function in the perfused
cat eye. Invest Ophthalmol Vis Sci 32(suppl):903, 1991.
Niemeyer G, Onoe S, and Macaluso C: Effects of glucose concentration on rod-function in the isolated, perfused cat eye.
Klin Monatsbl Augenheilkd 198:406, 1991.
Onoe S and Niemeyer G: Changing glucose concentration affects rod-mediated responses in the perfused cat eye. Acta Soc
Ophthalmol Jpn 96:634, 1992.
Gouras P and Hoff M: Retinal function in an isolated, perfused
mammalian eye. Invest Ophthalmol 9:388, 1970.
Niemeyer G: The function of the retina in the perfused eye.
Doc Ophthalmol 39:53, 1975.
Niemeyer G: Neurobiology of perfused mammalian eyes. J
Neurosci Methods 3:317, 1981.
Alder VA, Niemeyer G, Cringle S, and Brown J: Vitreal oxygen
tension gradients in the isolated perfused cat eye. Curr Eye Res
5:249, 1986.
Schuurmans RP and Zrenner E: The arterially perfused eye:
Colour vision mechanisms and neurotransmitters. In: Color
Vision in Clinical Pharmacology, Rietbrock N and Woodcock
BG, editors. Braunschweig/Wiesbaden, Germany, Friedr. Vieweg & Sohn, 1980, pp. 89-104.
Magistretti PJ, Hof PR, and Martin JL: Adenosine stimulates
glycogenolysis in mouse cerebral cortex: A possible coupling
mechanism between neuronal activity and energy metabolism.
J Neurosci 6:2558, 1986.
Osborne NN: [3H]glycogen hydrolysis elicited by adenosine in
rabbit retina: Involvement of A2 receptors. Neurochem Int
14:419, 1989.
Dawis SM, Hofmann H, and Niemeyer G: The electroretinogram, standing potential and light peak of the in vitro cat eye
during acid-base changes. Vision Res 25:1163, 1985.
Funkhouser A and Niemeyer G: Adaptation of a fundus camera permitting complex stimulation and observation in the visible and the infrared. Doc Ophthal Proc Series 31:145, 1982.
Niemeyer G: Light modulation of the standing potential in the
perfused mammalian eye: Characteristics and responses to acidosis. Doc Ophthal Proc Series 37:41, 1983.
Madachi-Yamamoto S, Yonemura D, and Kawasaki K: Hyperosmolarity response of ocular standing potential as a clinical test for retinal pigment epithelium activity. Normative
data. Doc Ophthalmol 57:153, 1984.
Lowry OH, Roberts NR, Shulz DW, Clow JE, and Clark JR:
Quantitative histochemistry of retina. II. Enzyme of glucose
metabolism. J Biol Chem 236:2813, 1961.
Ames AMI and Gurian BS: Effects of glucose and oxygen deprivation on function of isolated mammalian retina. J Neurophysiol 26:617, 1963.
Tornquist P and Aim A: Retinal and choroidal contribution to
retinal metabolism in vivo. A study in pigs. Acta Physiol Scand
106:351, 1979.
Newman EA and Odette LL: Model of electroretinogram bwave generation: A test of the K+ hypothesis. J Neurophysiol
51:164, 1984.
Wen R and Oakley BII: K+ -evoked Miiller cell depolarization
generates b-wave of electroretinogram in toad retina. Proc Natl
AcadSci USA 87:2117, 1990.
Stockton RA and Slaughter MM: b-wave of the electroretinogram. A reflection of ON-bipolar cell activity. J Gen Physiol
93:101, 1989.
2808
INVESTIGATIVE OPHTHALMOLOGY 6 VISUAL SCIENCE / September 1992
32. Masland RH and Ames AIII: Dissociation of field potential
from neuronal activity in the isolated retina: Failure of the
b-wave with normal ganglion cell response. J Neurobiol 6:305,
1975.
33. Niemeyer G: The optic nerve action potential: A monitor for
pharmacological effects in the perfused cat eye. In: Le Indagini
Elettrofisiologiche Nelle Affezioni del Nervo Ottico, Cordelia
M and Macaluso C, editors. Universita degli Studi di Parma,
Italy, 1989.
34. Cox DWG and Bachelard HS: Attenuation of evokedfieldpotentials from dentate granule cells by low glucose, pyruvate
+ malate, and sodium fluoride. Brain Res 239:527, 1982.
35. Creutzfeldt OD and Meisch JJ: Changes of cortical neuronal
activity and EEG during hypoglycemia. Electroencephalogr
Clin Neurophysiol 24(suppl):158, 1963.
36. Sieving PA, Frishman LJ, and Steinberg RH: Scotopic threshold response of proximal retina in cat. J Neurophysiol 56:1049,
1986.
37. Frishman LJ and Steinberg RH: Light evoked increases in
[K+]o in proximal portion of dark adapted cat retina. J Neurophysiol 61:1233, 1989.
38. Naarendorp F and Sieving PA: The scotopic threshold response of the cat ERG is suppressed selectively by GABA and
glycine. Vision Res 31:1, 1991.
39. Nilsson SEG and Skoog K: Covariation of the simultaneously
recorded c-wave and standing potential of the human eye. Acta
Ophthalmol 53:721, 1975.
40. Linsenmeier RA and Steinberg RH: A light-evoked interaction
of apical and basal membranes of retinal pigment epithelium:
c-wave and light peak. J Neurophysiol 50:136, 1983.
41. Harik SI, Kalaria RH, Whitey PM, Anderson L, Lundahl P,
Ledbetter SR, and Perry G: Glucose transporters are abundant
in cells with occluding junction at the blood-eye barriers. Proc
Natl Acad Sci USA 87:4261, 1990.
42. Pardridge WM, Boado RJ, and Farrell CR: Brain-type glucose
transporter (GLUT-1) is selectively localized to the bloodbrain barrier. J Biol Chem 265:18035, 1990.
43. Zadunaisky JA and Degnan KJ: Passage of sugars and urea
Downloaded From: http://iovs.arvojournals.org/ on 08/03/2017
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
Vol. 33
across the isolated retina pigment epithelium of the frog. Exp
Eye Res 23:191, 1976.
DiMattio J and Streitman J: Facilitated glucose transport
across the retinal pigment epithelium of the bullfrog (Rana
Catesbiana). Exp Eye Res 43:15, 1986.
Ferrendelli JA: Hypoglycemia in the central nervous system.
In: Brain Work, Ingvar DH and Lassen NA, editors. Alfred
Brenzon Symposium VIII, Munskgaard. New York, Academic
Press, 1975, pp. 298-313.
Lewis LD, Ljunggren B, Norberg K, and Siesjo BK: Changes in
carbohydrate substrates, amino acids, and ammonia in the
brain during insulin-induced hypoglycemia. J Neurochem
23:659, 1974.
Lewis LD, Ljunggren B, Ratcheson RA, and Siesjo BK: Cerebral energy state in insulin-induced hypoglycemia, related to
blood [glucose] and EEG. J Neurochem 23:677, 1974.
Bachelard HS: Cerebral metabolism and hypoglycemia. In: Hypoglycemia, Marks V and Rose F, editors. Oxford, Blackwell,
1981.
Norberg K, Ljunggren B, and Siesjo BK: Cerebral metabolism
in relation to function in insulin-induced hypoglycemia. In:
Brain Work, Ingvar DH and Lassen NA, editors. Alfred Brenzon Symposium, Munskgaard. New York, Academic Press,
1975, pp. 314-319.
Gibson GE and Blass JP: Impaired synthesis of acetylcholine in
brain accompanying mild hypoxia and hypoglycemia. J Neurochem 27:37, 1976.
Auer RN: Progress review: Hypoglycemic brain damage.
Stroke 17:699, 1986.
Steinberg RH: Monitoring communications between photoreceptors and pigment epithelial cells: Effects of "mild" systemic
hypoxia. Friedenwald lecture. Invest Ophthalmol Vis Sci
28:1888, 1987.
Ayoub GS, Korenbrot JI, and Copenhagen DR: Release of endogenous glutamate from isolated cone photoreceptors of the
lizard. Neurosci Res 10(suppl):47, 1989.
Bloomfield SA and Dowling JE: Roles of aspartate and glutamate in synaptic transmission in rabbit retina. I. Outer plexiform layer. J Neurophysiol 53:699, 1985.