An ambiguous fast synapse: a new twist in the tale of two transmitters
Michael W. Salter1
& Yves De Koninck2
1 Michael Salter is at the Department of Physiology,
University of Toronto, Hospital for Sick Children, 555 University
Ave., Toronto, Ontario, M5G 1X8,
Canada. mike.salter@utoronto.ca
2 Yves De Koninck is at the Department of Pharmacology
& Therapeutics, McGill University, Montréal,
Québec, H3G 1Y6, Canada.
ydk@spinalcord.mcgill.ca
Jo and Schlichter describe GABA and ATP corelease from the same presynaptic
cell, which suggests that a dorsal horn synapse could be excitatory or inhibitory
depending on its postsynaptic receptors or the amount of transmitter released.
Science, at least the kind of science that makes us sit up and take notice,
is revolution—controlled and hopefully nonviolent, but revolution nonetheless.
For example, it was not until just after the midpoint of this century that
the idea of chemical transmission, in a coup d'état, usurped the main
competing idea of electrical transmission as the dominant form of neuron−neuron
communication in the CNS1. With this insurrection came the concept,
termed Dale's principle, that each neuron releases the same transmitter from
all of its terminals. Sir Henry Dale never explicitly discussed the possibility
of a neuron releasing more than one transmitter2,
3,
4,
5, and
debate has raged about 'what is' versus 'what should have been' Dale's principle.
In any event, the concept of 'one neuron, one transmitter' became entrenched
only long enough for several transmitters to be identified before it was found
that neurons may release more than one transmitter3. Semantic
clashes then erupted as to what was a 'transmitter', 'modulator' or 'mediator'
as it became apparent that, as an endpoint, synaptic transmission may have
diverse time courses, mediation and modulation.
Through an enormous amount of work on cotransmission (see 6), two themes emerged. First, there are differences in the
time domain of responses, such that a given neuron may release one fast transmitter
and then one or more slower transmitter(s); second, different firing patterns
may cause differential release, particularly of the slowly acting transmitters
like peptides. As the peptide phenotype of a neuron is highly modifiable,
the idea 'one neuron, one transmitter' became in essence 'one neuron, one
fast transmitter'—until the finding that two fast transmitters, GABA
and glycine, were coreleased7. In that case, however, both transmitters
are inhibitory, and thus one could preserve the concept of 'one synapse, one
fast synaptic action'. Now even this concept comes under attack in a compelling
series of experiments reported by Jo and Schlichter on page 241 of this issue.
Fundamental to the common understanding of fast synaptic transmission in
the CNS is the concept that a given synapse is either excitatory or inhibitory.
So firmly entrenched is the idea that the sign of a synapse—excitation
or inhibition—is a basic property that most of us never even give it
a second thought. This concept is consistent with the apparently sensible
and world-simplifying notion that a synapse should transduce information unambiguously.
In cases where two opposing transmitters are released, the two components
of the response are separated temporally. So a synapse where there is corelease
of fast transmitters with opposing actions of similar kinetics would seem
illogical.
The results of Jo and Schlichter, however, are quite clear. Using cultures
of superficial spinal cord dorsal horn, they stimulated individual neurons
and recorded postsynaptic currents (PSCs) from neighboring cells with monosynaptic
connections. When GABA, glycine and glutamate receptors were blocked pharmacologically,
nearly half of the presynaptic neurons released ATP, an excitatory transmitter,
as inferred from PSCs mediated by ionotropic P2X receptors in the postsynaptic
cells. Washing out CNQX and AP-5, drugs that block ionotropic glutamate receptors,
did not reveal an additional component of the synaptic current. Surprisingly,
though, when bicuculline, a GABAA receptor antagonist, was washed
out, all of the neurons releasing ATP also elicited GABAA-receptor-mediated
PSCs, and importantly the latency and stimulation thresholds for the GABA-mediated
responses were the same as those for the ATP-mediated responses. Thus, the
logical conclusion is that ATP and GABA are released from the same presynaptic
neuron and possibly at the same synapses.
As it happens, ATP was one of the first culprits implicated as a cotransmitter3. In the peripheral nervous system, ATP is released together with
classical transmitters like noradrenaline or acetylcholine. In these cases,
however, ATP acts in the same direction as the other transmitter. In the dorsal
horn, however, P2X and GABAA receptor PSCs have opposite polarities.
The kinetics of the PSCs are virtually identical and therefore tend to cancel
each other at the resting membrane potential.
What could be the point of generating two such similar yet opposing signals?
There are many potential implications, and we will limit our discussion to
some possible consequences for this apparently heretical synapse. Jo and Schlichter
show that ATP and GABA are coreleased from the same cell, and for simplicity
our speculative models (Fig. 1) show ATP and
GABA coreleased from the same presynaptic terminal. Although this could not
be determined directly, the functional implications would be much the same
if ATP and GABA are released from separate terminals on the same postsynaptic
cell. Indeed, in such a case, the results would be even more provocative and
would drive the final nail into the cherished concept that the transmitter
repertoire is the same at all of a cell's synapses.
Figure 1. Possible mechanisms to explain how the balance of P2X-receptor-mediated
excitation versus GABAA-receptor-mediated inhibition could
be altered to set the net 'sign' of this ambiguous synapse, which releases
both ATP and GABA as fast neurotransmitters.
(a) The postsynaptic cell may control what it hears by altering
receptor expression at the synapse. (b) Alternatively, as transmitter
release increases, more ATP is able to reach its distant receptor before being
intercepted by extracellular nucleotidase. Thus the balance may shift to convert
an inhibitory synapse to an excitatory one.
Given the corelease of ATP and GABA, the sign of the synapse may be ambiguous.
This arrangement provides the opportunity to switch between excitation and
inhibition simply by changing one of the dominant parameters, either the balance
of the responsiveness of the postsynaptic receptors or how much transmitter
is released. In other words, this synaptic arrangement depends not only on
how loudly the presynaptic cell speaks but also on what the postsynaptic cell
wants to hear.
One plausible model for switching sign is that the postsynaptic cell controls
what it hears by altering the cohort of receptors expressed at the synapse
(Fig. 1a). This could occur, for example, via
rapid translocation of receptors to the membrane8 and/or their
clustering and declustering at synapses, or by posttranslational modifications
of the receptors9 that would render cell-surface receptors either
more or less responsive to released transmitter.
Another way of changing the sign of this synapse stems from the peculiar
behavior of synapses that use ATP, due to its rapid extracellular degradation
to adenosine10. A change in sign is conceivable based on the
amount of ATP being released under given conditions. Interestingly, in the
dorsal horn neurons, P2X-receptor-mediated miniature excitatory PSCs (mEPSCs),
attributed to spontaneous release of single vesicles, were not observed, whereas
action potentials in the presynaptic neurons did elicit P2X-mediated EPSCs.
The lack of P2X mEPSCs could be explained if the amount of ATP released by
a single vesicle was sufficiently small that most of it was intercepted by
degradation enzymes before reaching the receptors. This would be especially
plausible if P2X receptors are located at a distance from the release site.
On the other hand, raising the amount of ATP released (for example through
synchronous release of multiple vesicles, which could be evoked by action
potentials) may allow ATP to escape and reach the receptors. From that, we
envisage that increasing the frequency of action potentials may allow progressively
more ATP to escape, and eventually the balance may shift, effectively converting
an inhibitory synapse to an excitatory one (Fig. 1b).
Another interesting observation was that GABAA receptors are
inhibited postsynaptically by adenosine generated from ATP, whereas P2X receptors
were unaffected. Thus, the balance of sensitivity of the postsynaptic cell
to ATP or to GABA could be set by adenosine, presumably acting through metabotropic
adenosine receptors. Thus, increasing firing frequency could increase extracellular
generation of adenosine, which could feedforward to accentuate switching from
excitation to inhibition. This seemingly paradoxical effect of adenosine might
be further complicated if adenosine were to activate postsynaptic potassium
channels11. Thus, the sign of ATP-GABA synapses may have a complex
relationship to the level of activity, but this remains to be shown experimentally.
Could there be other functions of ATP and GABA corelease? One possibility
is that P2X receptor activation may provide a local means to regulate GABA
A receptors, for example by allowing influx of calcium, which has been
shown to modulate the function of GABAA receptors12.
Alternatively, ATP might act on other P2 receptor subtypes14
or even modulate GABAA receptors directly.
In summary, the past year has seen two striking examples of the corelease
of fast synaptic transmitters at central synapses. These will undoubtedly
provoke further exciting investigations and perhaps a renewed revolution in
some basic views on synaptic transmission. What would Sir Henry Dale have
thought of all of this? We guess that, like us, he would remain astonished
by the richness, subtlety and diversity of communication between neurons.
