AMPA receptors mediate much of the rapid synaptic excitatory
neurotransmission. The functional properties of the receptor, as in the
nicotinic receptors, are dependent on the subunit composition, with the
receptor composed of a combination of five of these
subunits.
6 Four closely related subunits have been cloned
thus far, named GluR1 through GluR4
17 18 19 (also termed
GluR-A through GluR-D).
20 Furthermore, each of these
subunits exists in two different forms—termed “flip” and“
flop,”—due to alternate splicing of a 115-bp region immediately
preceding one of the transmembrane regions. Each of the eight possible
splice variants of the four subunit types shows different expression
patterns within the brain, both spatially
21 and
temporally.
22 The alternate forms confer different
properties to the receptors: flip channels continue to open in the face
of repeated binding of glutamate, whereas the flop channel shows a
gradual decrease in response. It has been suggested that this
difference in desensitization and that change from flop to flip may
play a role in long-term potentiation,
21 the relatively
long-lived strengthening of synaptic connectivity believed to be
associated with memory.
23
AMPA receptors, again depending on the composition of subunits, have a
variable selectivity to different ions, specifically
Ca
2+ and Na
+.
24 The GluR2 subunit is
responsible for much of the difference: inclusion of this subunit in
the channel substantially reduces the ability of the channel to pass
Ca
2+ ions.
25 AMPA channels are found
throughout the brain—including the retina and specifically on
RGCs.
18 20
KA receptors can be further subdivided into two classes based on the
subunits cloned thus far: GluR5 through GluR7
19 26 27 and
KA-1 to -2
28 29 (also termed gamma 1 and gamma
2).
30 Channels (which also have five subunits) that are
composed of the GluR5 through GluR7 subunits are often referred to as
low-affinity KA channels, with binding constants for KA approximately
10 times lower than for those channels containing KA-1 and -2
subunits.
20 28
GluR5 to 7 subunits are expressed in many regions of the brain,
including the retina and retinal ganglion cells, but in comparison to
GluR1 to 4 are more restricted, and the distribution appears to be
developmentally regulated.
19 26 The KA-1 and -2 subunits
are not found as functional homomeric KA channels but rather are found
in combination with the GluR5 through GluR7 subunits.
29 KA-1 subunits are found in the CA3 and dentate gyrus of the
hippocampus, which is the classical KA high-affinity-binding site in
the CNS. Both these subunits are found in the retina.
31 No
splice variants have yet been reported for these subunits.
The NMDA receptors are the most widely studied of the three subtypes of
glutamate receptors, partly because they have been implicated in many
CNS functions and dysfunctions, which are discussed below in the
context of excitotoxicity. NMDA channels, unlike certain AMPA-KA
channels, show very high selective permeability to
Ca
2+ compared with that of other
cations.
6
Five subunits have been cloned and are named NMDAR1, and NMDAR2A
through NMDAR2D.
32 33 Functional channels can be formed
completely from NMDAR1. This is not true of the 2A-2D subunits, which
must be expressed in concert NMDAR1 to make a functioning
channel.
33 Inclusion of the 2A-2D subunits in functional
NMDA channels alters the pharmacokinetics of the channel considerably.
In particular, heteromeric channels containing these subunits increase
the amplitude of the Ca
2+ flow through the
receptor by from 5- to 60-fold.
34 35 36 NMDAR1 and 2A
subunits are found throughout the brain,
32 36 37 38 whereas
2B is expressed selectively within the forebrain,
33 2C is
found predominately in the cerebellum,
35 and 2D is most
prominently expressed in the brain stem, cerebellum, and olfactory
bulb.
37 All the subunits have been found in the
retina.
31
The major interest in limiting excitotoxic damage in the past several
decades has been directed at blockade of the NMDA receptor. We have
previously shown (manuscript submitted) that optic nerve crush leads to
release of glutamate into the vitreous of a rat eye, and that NMDA
antagonists can limit damage from crush. Yoles and
Schwartz
3 39 have shown similar results, exploring
elevation of glutamate in the vitreous. They and others have proposed
the concept of secondary degeneration, whereby the initial insult of
crush leads to loss of a population of ganglion cells; however, this
primary insult also triggers additional death, perhaps through the
release of toxic levels of glutamate. Although, as note above, NMDA
antagonists are partially protective against this insult, we
demonstrate here that AMPA-KA antagonists are if anything more
protective, although not additive in this model system.
In summary, these data suggest that activation of the AMPA-KA receptors
in the face of optic nerve crush may be critical in regulating neuronal
death; their blockade may deserve additional consideration in limiting
ganglion cell loss from glutamate-mediated damage.