RAPID COMMUNICATION
Sodium Action Potentials Are Not Required for Light-Evoked
Release of GABA or Glycine From Retinal Amacrine Cells
MARK C. BIEDA1,2 AND DAVID R. COPENHAGEN1
1
Departments of Ophthalmology and Physiology, UCSF School of Medicine, San Francisco 94143; and 2Stanford
Neuroscience Program, Stanford, California 94309
INTRODUCTION
Most vertebrate CNS neurons require sodium action potential (Na-AP) production for stimulus-evoked fast classical neurotransmitter release. Many classes of neurons in sensory structures release fast classical neurotransmitters via graded
potentials instead of Na-APs.
The use of Na-APs versus graded potentials divides neurons
into two separate categories. Previous studies have demonstrated that these two categories of neurons have consistently
different physiological properties and also probably have different information processing properties (Juusola et al. 1996).
Therefore knowledge of a neuron’s use of Na-APs or graded
potentials gives important clues to physiological and information processing aspects of that neuron.
The costs of publication of this article were defrayed in part by the payment
of page charges. The article must therefore be hereby marked “advertisement”
in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
3092
In the retina, photoreceptors, horizontal cells, and bipolar
cells are graded potential neurons, whereas ganglion cells
require Na-AP production for evoked neurotransmitter release.
The position of amacrine cells in this classification is still unclear.
Although many amacrine cells produce Na-APs (Barnes and
Werblin 1986; Roska et al. 1998), many of these cells have
anatomic features compatible with graded release, including lack
of an axon and reciprocal synapses with bipolar cells (Dowling
1987). Paradoxically, cultured GABAergic rat amacrine cells require Na-AP production for evoked release (Taschenberger and
Grantyn 1995), whereas cultured GABAergic chick amacrine
cells probably do not (Gleason et al. 1993). Although a recent
report (Cook et al. 1998) concluded that glycinergic amacrine
cells use both Na-AP–dependent and -independent modes of
light-evoked neurotransmitter release, there is no information on
whether GABAergic transmission requires Na-APs in the intact
retina.
The goal of this study was to assess the role of Na-APs in
mediating light-evoked release of classical neurotransmitters
from amacrine cells. Because amacrine and ganglion cells are
the only retinal cell classes that have active sodium channels
during light stimulation, only these cells will be directly affected by TTX. To measure the release of inhibitory neurotransmitters by amacrine cells, we recorded the ON inhibitory
postsynaptic current (IPSC) in ganglion cells by voltage clamping ganglion cells near 0 mV, the reversal potential for glutamate-gated synaptic currents (Mittman et al. 1990). To block
Na-APs in the circuit, we applied 1 mm TTX, which has been
shown to block all voltage-gated sodium currents in amacrine
cells (Barnes and Werblin 1986; Bieda and Copenhagen, unpublished observations). In brief, we found that release of
GABA or glycine from amacrine cells onto ganglion cells uses
both Na-AP– dependent and -independent processes. A brief
report of this work has previously been made (Bieda and
Copenhagen 1998).
METHODS
Retinal slices (200 –300 mm) were prepared from tiger salamander
(Ambystoma tigrinum) eyes following UCSF and NIH guidelines.
After forming the eyecup (see Coleman and Miller 1989), we removed
the retina and applied hyaluronidase (Sigma type IV-S, 300 – 400
units/ml; 2–5 min) to eliminate vitreous. Similar hyaluronidase treatments do not affect retinal response properties (Coleman and Miller
1989; Winkler and Cohn 1985). All procedures were performed with
infrared illumination.
The extracellular solution consisted of (in mM) 112 NaCl, 2 KCl,
2 CaCl2, 1 MgCl2, 5 hemisodium HEPES, and 25 glucose, pH 7.6
0022-3077/99 $5.00 Copyright © 1999 The American Physiological Society
Downloaded from http://jn.physiology.org/ by 10.220.33.3 on July 31, 2017
Bieda, Mark C. and David R. Copenhagen. Sodium action potentials are not required for light-evoked release of GABA or glycine
from retinal amacrine cells. J. Neurophysiol. 81: 3092–3095, 1999.
Although most CNS neurons require sodium action potentials (NaAPs) for normal stimulus-evoked release of classical neurotransmitters, many types of retinal and other sensory neurons instead use only
graded potentials for neurotransmitter release. The physiological
properties and information processing capacity of Na-AP–producing
neurons appear significantly different from those of graded potential
neurons. To classify amacrine cells in this dichotomy, we investigated
whether Na-APs, which are often observed in these cells, are required
for functional light-evoked release of inhibitory neurotransmitters
from these cells. We recorded light-evoked inhibitory postsynaptic
currents (IPSCs) from retinal ganglion cells, neurons directly postsynaptic to amacrine cells, and applied TTX to block Na-APs. In control
solution, TTX application always led to partial suppression of the
light-evoked IPSC. To isolate release from glycinergic amacrine cells,
we used either bicuculline, a GABAA receptor antagonist, or picrotoxin, a GABAA and GABAC receptor antagonist. TTX application
only partially suppressed the glycinergic IPSC. To isolate release from
GABAergic amacrine cells, we used the glycine receptor blocker
strychnine. TTX application only partially suppressed the lightevoked GABAergic IPSC. Glycinergic and GABAergic amacrine
cells did not obviously differ in the usage of Na-APs for release.
These observations, in conjunction with previous studies of other
retinal neurons, indicate that amacrine cells, taken as a class, are the
only type of retinal neuron that uses both Na-AP– dependent and
-independent modes for light-evoked release of neurotransmitters.
These results also provide evidence for another parallel between the
properties of retinal amacrine cells and olfactory bulb granule cells.
ACTION POTENTIALS AND AMACRINE CELL RELEASE
3093
Efficacy of TTX blockade
Because TTX is an open channel blocker, we applied relatively
large numbers of successive light flashes (typically ;20) during the
TTX application period. In all cases, the TTX-induced suppression of
the ON IPSC quickly stabilized after the initial approximately five
responses. Our results demonstrate a wide range of variability in the
magnitude of TTX suppression (Figs. 1B; 2, B and D, and 3B). To
verify that this variability was not due to problems with the stock of
TTX or the perfusion system, we typically monitored TTX block of
the ganglion cell sodium current during experiments (data not shown).
In all cases, TTX (1 mm) induced complete, partially reversible block
of the sodium current. Therefore the observed variability in TTXinduced ON IPSC suppression was not due to problems with TTX
efficacy or the perfusion system.
RESULTS
Figure 1 shows the suppression of the ON IPSC induced by
TTX (1 mm) application in control saline. In the experiment
shown in Fig. 1A, TTX induced a significant, mostly reversible
suppression of the amplitude and charge of the ON IPSC. Block
of this IPSC by coapplication of bicuculline (100 mm) and
strychnine (1 mm) indicates that it was mediated by GABA
and/or glycine receptor activation (Fig. 1A). The summary of
experiments demonstrates a wide distribution in the amount of
TTX suppression (Fig. 1B). These experiments demonstrate
that light-evoked release of inhibitory transmitters from at least
some amacrine cells does not require Na-APs.
In tiger salamander, light-evoked ON IPSCs in ganglion cells
usually have both GABAergic and glycinergic components
(Mittman et al. 1990; Zhang et al. 1997). Therefore, to explain
the variability seen in Fig. 1B, we hypothesized that glycinergic and GABAergic transmission might have different dependencies on Na-APs and that different ganglion cells possess
different ratios of glycinergic to GABAergic input. Hence we
studied the role of Na-APs in mediating release of GABA and
glycine separately. A recent study with picrotoxin to block
GABA receptors indicates that release from glycinergic amacrine cells does not require Na-APs (Cook et al. 1998). However, because application of picrotoxin has been found to
greatly increase the release of glutamate by bipolar cells (Dong
FIG. 1. Light-evoked release of inhibitory neurotransmitters from amacrine
cells occurs by both sodium action potential (Na-AP)-independent and -dependent mechanisms. A: representative experiment. ON inhibitory postsynaptic
currents (IPSCs), reflecting release from amacrine cells, were recorded from a
ganglion cell. Responses are averages of 5 consecutive traces. Bar indicates
period of light ON. Scale bars: 40 pA, 500 ms. B: summary results from 9 cells
demonstrates wide range of suppression of ON IPSC charge by TTX application.
and Werblin 1998; Zhang et al. 1997), the relative role of
Na-APs in mediating glycine release may be distorted under
this condition. In contrast, the GABAA receptor antagonists
bicuculline and SR95531 have been reported to have weak
effects on the release of glutamate (Dong and Werblin 1998;
Zhang et al. 1997) and therefore may provide more easily
interpretable results. Hence to extend and confirm the results of
Cook et al. (1998), we tested the effect of TTX on the glycinergic ON IPSC by employing either picrotoxin or bicuculline to
block GABAergic inputs.
Figure 2 shows experiments examining the effect of TTX on
glycinergic ON IPSCs. In Fig. 2, A and B, picrotoxin (200 mm)
was continuously present to block GABAA and GABAC receptors. In a representative experiment, TTX led to a mostly
reversible decrease in the charge and amplitude of the strychnine-sensitive ON IPSC (Fig. 2A). Figure 2B shows significant
variability from cell to cell in the amount of TTX-induced
suppression. These results are consistent with those of Cook et
al. (1998). In Fig. 2, C and D, bicuculline (100 mm) was
continuously present to block GABAA receptors. Application
of TTX induced a small, reversible decrease in the charge and
amplitude of the strychnine-sensitive ON IPSC in a representa-
Downloaded from http://jn.physiology.org/ by 10.220.33.3 on July 31, 2017
with NaOH. The internal solution consisted of (in mM) 107 CsMES,
3 MgCl2, 5 EGTA, 5 HEPES, 2 Na2ATP, and 0.5 NaGTP, pH 7.3–7.4
with CsOH. Whole-cell currents were recorded with an Axopatch 1D
and acquired with PULSE (HEKA electronik GmbH; Lambrecht,
Germany). Most ganglion cells had transient EPSCs in response to
light, with transient ON–OFF cells predominating. Voltages were corrected for junction potential and electrode offset. After initiation of
whole cell recording, we let cells dialyze for ;10 min before recording the baseline for experiments. We used whole-bath perfusion of a
;300-ml chamber with solution flow at ;2 ml/min to exchange
solutions.
A red LED was used to provide full-field light stimulus (1- or 2-s
duration; 12- to 20-s interstimulus interval). These interstimulus intervals generated stable baseline responses with no evidence of any
slow light-adaptational processes. We typically recorded ;10 –15
baseline responses before applying TTX.
All compounds were exclusively purchased from Sigma, except
TTX, which was purchased from Calbiochem and Sigma.
All analyses were conducted with custom written software in the
IGOR PRO (Wavemetrics, Lake Oswego, OR) environment (by
Bieda).
3094
M. C. BIEDA AND D. R. COPENHAGEN
FIG. 2. Both Na-AP– dependent and –independent
modes contribute to light-evoked glycine release. A:
representative experiment; 200 mm picrotoxin present in
all solutions. Markings as in Fig. 1. Scale bars: 80 pA,
500 ms. B: summary results of experiments with picrotoxin. Note wide range of suppression. C: representative
experiment with 100 mm bicuculline present in all solutions. Markings as in Fig. 1. Scale bars: 100 pA, 500
ms. D: results from 4 cells with bicuculline present
continuously. All responses (A and C) are averages of 5
consecutive traces. Measurements in B and D are of
suppression of ON IPSC charge.
cells synapsing onto the ganglion cell. Taken together, the
results of Fig. 2 demonstrate that light-evoked release of glycine can occur from amacrine cells in the absence of Na-APs.
In Fig. 3, we selectively measured GABA release from
amacrine cells by performing experiments in strychnine (10
mm). In some experiments, GABAergic ON IPSCs were quite
transient (Fig. 3A) and in others relatively sustained (data not
shown). Application of TTX induced a small decrease in the
picrotoxin-sensitive ON IPSC in the experiment shown in Fig.
3A. Figure 3B displays summary results indicating a wide
distribution in the magnitude of TTX-induced suppression. In
total, these results demonstrate that light-evoked GABA release from amacrine cells does not require Na-AP generation.
The ranges of TTX suppression of glycinergic (Fig. 2, B and
D) and GABAergic (Fig. 3B) transmission are overlapping and
not obviously distinguishable. Therefore these results do not
support a model with differential reliance of GABAergic and
glycinergic amacrine cells on Na-APs for release of neurotransmitters.
DISCUSSION
FIG. 3. Both Na-AP– dependent and -independent modes contribute to
light-evoked GABA release; 10 mm strychnine present in all solutions. A:
representative experiment. Responses are averages of 5 consecutive traces.
Markings as in Fig. 1. Scale bars: 80 pA, 500 ms. B: summary results of
suppression of GABAergic ON IPSC charge by TTX. Note wide range of
suppression.
We conclude that both GABAergic and glycinergic amacrine cells can produce light-evoked release of GABA or
glycine without employing Na-APs. Therefore, in the context
of previous studies (Dowling 1987), our results indicate that
amacrine cells are the only class of retinal cells that employs
both Na-AP– dependent and -independent release modes. This
Na-AP–independent release may occur by graded potentials
and/or production of calcium action potentials. Also these
results demonstrate another significant parallel between the
properties of retinal amacrine cells and olfactory bulb granule
cells, which also comprise a class of Na-AP–producing neurons that do not require Na-APs for inhibitory neurotransmitter
release (Isaacson and Strowbridge 1998).
Two aspects of the data are also noteworthy. First, there was
a wide distribution in the amount of TTX-induced suppression
of the ON IPSC. Second, there was no clear difference between
the reliance of GABAergic release on Na-APs and the reliance
of glycinergic release on Na-APs. However, because amacrine
cells form mutually inhibitory networks (Roska et al. 1998;
Downloaded from http://jn.physiology.org/ by 10.220.33.3 on July 31, 2017
tive experiment (Fig. 2C). Figure 2D displays the results of
four experiments of this type. Three of four neurons showed
;30% suppression of the ON IPSC by TTX. However, one
neuron showed an increase in the ON IPSC with TTX. Because
GABAergic amacrine cells are mutually inhibitory (Zhang et
al. 1997), augmentation of release by TTX in this experiment
probably reflects a net disinhibition of the glycinergic amacrine
ACTION POTENTIALS AND AMACRINE CELL RELEASE
al. 1998). Also Na-APs may induce faster, more transient
release under some conditions. Local graded potential-mediated release could allow localized neurotransmitter release,
allowing dendrites to act as separate computational compartments (Shepherd and Koch 1990). These suggestions represent
only a few of the possibilities for this highly flexible system.
We thank Dr. Tania Vu for helpful comments on the manuscript.
M. Bieda was supported by a Howard Hughes Medical Institute predoctoral
fellowship and by a training grant from the National Institutes of Health (NIH).
This research was supported by the NIH. Additional support was provided by
Research to Prevent Blindness and That Man May See, Inc.
Address for reprint requests: D. R. Copenhagen, Dept. of Ophthalmology,
UCSF School of Medicine, Room K-141, Box 0730, San Francisco, CA
94143-0730.
Received 22 December 1998; accepted in final form 23 February 1999.
REFERENCES
BARNES, S. AND WERBLIN, F. Gated currents generate single spike activity in
amacrine cells of the tiger salamander retina. Proc. Natl. Acad. Sci. USA 83:
1509 –1512, 1986.
BIEDA, M. C. AND COPENHAGEN, D. R. Action potentials are not required for
release of GABA or glycine from amacrine cells. Soc. Neurosci. Abstr. 24:
136, 1998.
BLOOMFIELD, S. A. Effect of spike blockade on the receptive-field size of
amacrine and ganglion cells in the rabbit retina. J. Neurophysiol. 75:
1878 –1893, 1996.
COLEMAN, P. A. AND MILLER, R. F. Measurement of passive membrane
parameters with whole-cell recording from neurons in the intact amphibian
retina. J. Neurophysiol. 61: 218 –230, 1989.
COOK, P. B., LUKASIEWICZ, P. D., AND MCREYNOLDS, J. S. Action potentials are
required for the lateral transmission of glycinergic transient inhibition in the
amphibian retina. J. Neurosci. 18: 2301–2308, 1998.
DONG, C. J. AND WERBLIN, F. S. Temporal contrast enhancement via GABAC
feedback at bipolar terminals in the tiger salamander retina. J. Neurophysiol.
79: 2171–2180, 1998.
DOWLING, J. E. The Retina: An Approachable Part of the Brain. Cambridge,
MA: Belknap, 1987.
GLEASON, E., BORGES, S., AND WILSON, M. Synaptic transmission between
pairs of retinal amacrine cells in culture. J. Neurosci. 13: 2359 –2370, 1993.
ISAACSON, J. S. AND STROWBRIDGE, B. W. Olfactory reciprocal synapses:
dendritic signaling in the CNS. Neuron 20: 749 –761, 1998.
JUUSOLA, M., FRENCH, A. S., UUSITALO, R. O., AND WECKSTREOM, M. Information processing by graded-potential transmission through tonically active
synapses. Trends Neurosci. 19: 292–297, 1996.
MITTMAN, S., TAYLOR, W. R., AND COPENHAGEN, D. R. Concomitant activation
of two types of glutamate receptor mediates excitation of salamander retinal
ganglion cells. J. Physiol. (Lond.) 428: 175–197, 1990.
ROSKA, B., NEMETH, E., AND WERBLIN, F. S. Response to change is facilitated
by a three-neuron disinhibitory pathway in the tiger salamander retina.
J. Neurosci. 18: 3451–3459, 1998.
SHEPHERD, G. M. AND KOCH, C. Introduction to synaptic circuits. In: The
Synaptic Organization of the Brain, edited by G. M. Shepherd. New York:
Oxford Univ. Press, 1990, p. 3–31.
TASCHENBERGER, H. AND GRANTYN, R. Several types of Ca21 channels mediate
glutamatergic synaptic responses to activation of single Thy-1-immunolabeled rat retinal ganglion neurons. J. Neurosci. 15: 2240 –2254, 1995.
WINKLER, B. S. AND COHN, E. M. Hyaluronidase and retinal function. Arch.
Ophthal. 103: 1743–1746, 1985.
ZHANG, J., JUNG, C. S., AND SLAUGHTER, M. M. Serial inhibitory synapses in
retina. Vis. Neurosci. 14: 553–563, 1997.
Downloaded from http://jn.physiology.org/ by 10.220.33.3 on July 31, 2017
Zhang et al. 1997), we must be cautious in drawing conclusions
about the relative magnitudes of TTX effects. This is particularly true in our experiments with inhibitory antagonists (Figs.
2 and 3) because these antagonists have been shown to affect
glutamate release (Cook et al. 1998; Zhang et al. 1997) and will
affect the inhibition received by amacrine cells. Therefore the
percentage of release remaining in TTX may not accurately
reflect the actual percentage of light-evoked Na-AP–independent release from amacrine cells under normal conditions, a
conclusion consistent with the TTX-induced increase in release
observed in one cell (Fig. 2D). Nonetheless, the large amount
of TTX-insensitive, light-evoked release implies that Na-AP–
independent release is significant under normal conditions
when Na-APs are present.
Because this study measures aggregate release from presumably many amacrine cells onto a single ganglion cell, there are
several models to account for our results. First, all amacrine
cells may use both Na-AP– dependent and -independent release
modes (model I). Second, some amacrine cells may rely solely
on Na-AP– dependent release, whereas others may rely solely
on Na-AP–independent release (model II). Third, there may be
a continuum of reliance on Na-APs for release (model III).
For GABAergic release from cultured amacrine cells, Taschenberger and Grantyn (1995) find an absolute requirement
for Na-APs, whereas the results of Gleason et al. (1993) imply
that there is no requirement. Our results are not incompatible
with these previous, seemingly contradictory results. Either the
segregation model (model II) or model III could encompass
these disparate conclusions. The finding by Taschenberger and
Grantyn (1995) that some amacrine cells absolutely require
Na-AP production for evoked GABA release would seem to
remove model I from consideration. However, because their
work was performed in a cultured system, it is possible that the
results do not apply to the normally developed retina in vivo.
Therefore these experiments do not eliminate model I.
Future studies with paired recordings of amacrine cells and
ganglion cells should help determine which, if any, of these
models is the best reflection of amacrine cell processing. However, a recently proposed model (Cook et al. 1998) implies that
the ratio of Na-AP– dependent to -independent release for a
single cell may not be fixed. Also Bloomfield (1996) demonstrates that, for wide-field rabbit amacrine cells, TTX-sensitive
channels are critical for propagation of distal inputs to the
soma. These results taken together imply that Na-APs may
play diverse and complex roles in amacrine cell function extending beyond those included in our set of simplified models.
The implications of our finding that amacrine cells, as a
class, use both Na-AP– dependent and -independent release
modes for functional light-evoked release depends on which of
the three models presented previously is closest to the truth.
There are many rationales for why cells might possess each
mode of transmission. For example, Na-APs may be required
for transmission of long-distance surround inhibition (Cook et
3095