Trends in Neurosciences
Volume 21, Issue 1, 1 January 1998, Pages 8-14
Journal home page for Trends in Neurosciences

Extrasynaptic glutamate spillover in the hippocampus: evidence and implications

https://doi.org/10.1016/S0166-2236(97)01150-8Get rights and content

Abstract

In the mammalian brain most excitatory transmission is mediated by glutamate binding to AMPA and NMDA receptors. These receptors have markedly different biophysical properties, and at synapses in the CA1 region of the hippocampus they play complementary roles in long-term potentiation (LTP): while postsynaptic NMDA receptor activation is necessary for the induction of this form of plasticity, AMPA receptors play a larger role in its expression. Recent studies in hippocampal slices have revealed a further striking difference in the behaviour of the two receptor types: NMDA receptors consistently sense a larger number of quanta of glutamate released from presynaptic terminals than do AMPA receptors. Two alternative explanations for this are either that AMPA receptors are functionally silent at a proportion of synapses (although they can be uncovered by LTP), or that glutamate can spill over from neighbouring synapses and selectively activate NMDA (but not AMPA) receptors. Both of these competing hypotheses have extensive implications for the mechanisms of expression of LTP. Extrasynaptic glutamate diffusion appears to depend critically on the recording temperature, but if excitatory synapses are sufficiently close for cross-talk to occur under physiological conditions, it could have profound implications for the specificity of synaptic communication in the brain.

Section snippets

How does LTP shed light on AMPA-silent signalling?

Pairing presynaptic stimulation of apparently AMPA-silent synapses with postsynaptic depolarization to around 0 mV, to maximize Ca2+ influx via NMDA receptors, can uncover an AMPA receptor-mediated component6, 7. This coincides nicely with several previous observations on the mechanisms of expression of NMDA receptor-dependent LTP: several groups have reported LTP to be expressed preferentially, if not exclusively, by an increase in the size of the AMPA receptor-mediated signal, and to be

Glutamate spillover explains LTP of NMDA receptor-mediated signals

We have recently presented two lines of argument in favour of the spillover hypothesis for excitatory transmission. The first approach relies on a closer examination of the changes in the AMPA and NMDA receptor-mediated signals with LTP ([20]). We found that LTP of AMPA-receptor-mediated signals, elicited by pairing afferent stimulation with depolarization, is consistently associated with a small potentiation of NMDA-receptor-mediated signals, and that this can be accounted for by an increase

Effects of recording temperature on `AMPA-silent' signalling

The second body of evidence against the `latent AMPA-receptor cluster' hypothesis, and in favour of glutamate spillover, comes from an examination of the effect of altering the recording temperature[22]. The studies described above were almost entirely performed at room temperature, which might have widespread consequences for the regulatory processes affecting synaptic transmission. Repeating the comparison of quantal contents mediated by AMPA and NMDA receptors at physiological temperature

Implications of the spillover hypothesis

The spillover hypothesis has extensive consequences, both for the interpretation of experiments on glutamatergic transmission, and for the understanding of the normal functioning of the brain. Consider the possible arrangements of excitatory synapses illustrated in Fig. 3: these differ according to whether or not the presynaptic bouton is invaded by an action potential, transmitter release takes place, and the bouton is in synaptic contact with the recorded postsynaptic cell. Both AMPA and NMDA

Discrepancies in the literature

Thus far, the spillover model is consistent with the experimental data described above. There is, however, considerable disagreement about whether, and to what extent, potentiation of the NMDA receptor-mediated component accompanies conventional LTP: while some groups have reported no change23, 24, others have described either a relatively modest potentiation[25] or even an increase equivalent to that of the AMPA receptor-mediated component[26]. Some of this disagreement can be explained by the

Postsynaptic changes in AMPA receptors

An additional factor that determines the relative potentiation of AMPA- and NMDA-receptor-mediated signals is the degree to which changes in quantal amplitude accompany the increase in quantal content. Many groups have indeed reported increases in the quantal amplitude of AMPA receptor-mediated signals with LTP (10, 11, 35, 36). Although this is not a unanimous finding37, 38, it provides some of the most compelling evidence that LTP expression involves a postsynaptic change. A relatively simple

Problems with the spillover hypothesis

How can one reconcile the pathway specificity of LTP with the spillover hypothesis? It is generally accepted that if two separate populations of afferents converge on the same postsynaptic neurons, only the pathway that is paired with depolarization undergoes potentiation[42]. If, however, the immediately apposed presynaptic bouton in synapse D is activated by a second stimulus pathway, what is to prevent pairing of the first pathway with postsynaptic depolarization from eliciting LTP at this

Further implications of spillover

Another consequence of the spillover hypothesis is that it could complicate the use of the NMDA-receptor open-channel blocker MK-801 as a quantitative tool to study presynaptic transmitter release probability52, 53: the opening probability of a given NMDA receptor could be a complicated function of the probability not only at the immediately apposed release site, but also at all the neighbouring sites that are close enough for significant spillover. This does not necessarily invalidate its use

Could glutamate spillover occur in the living brain?

We have argued that the discrepancy in quantal content at physiological temperature is less than at room temperature, although not completely abolished[22]. It is difficult to extrapolate from hippocampal slices in vitro to the intact brain, but if spillover does occur, it raises the possibility that excitatory synapses do not function as entirely private communication channels between individual neurones. Another consequence is that a given cell could sample a different population of

Concluding remarks

In this review we have considered how glutamate spillover could explain the mismatch in the behaviour of the two components of fast excitatory synaptic signalling at some hippocampal and neocortical synapses. We cannot, of course, exclude the alternative interpretation, that functional AMPA receptors are absent at a population of synapses, and the evidence that distinguishes between these hypotheses is indirect. It remains to be determined whether the discrepancy between AMPA and NMDA

References (57)

  • K.A. Jones et al.

    Neuron

    (1991)
  • S. Hestrin et al.

    Neuron

    (1990)
  • J.T. Isaac et al.

    Neuron

    (1995)
  • R.C. Malenka et al.

    Neuron

    (1997)
  • D.M. Kullmann et al.

    Neuron

    (1995)
  • G. Tong et al.

    Neuron

    (1994)
  • O. Garaschuk et al.

    Neurosci. Lett.

    (1992)
  • D.M. Kullmann et al.

    Neuron

    (1996)
  • F. Asztely et al.

    Neuron

    (1997)
  • J.A. Kauer et al.

    Neuron

    (1988)
  • D.K. Selig

    Neuron

    (1995)
  • K.J. Collard et al.

    Dev. Brain Res.

    (1993)
  • A. Lehmenkühler

    Neuroscience

    (1993)
  • J. Lisman

    Trends Neurosci.

    (1994)
  • B. Gustafsson et al.

    Trends Neurosci.

    (1988)
  • R.J. Cormier et al.

    Neuron

    (1993)
  • D.M. Kullmann

    Neuron

    (1992)
  • K.M. Harris

    Trends Neurosci.

    (1995)
  • V.V. Uteshev et al.

    Biophys. J.

    (1997)
  • J.S. Isaacson et al.

    Neuron

    (1993)
  • J.M. Bekkers et al.

    Nature

    (1989)
  • R.A. Silver et al.

    Nature

    (1992)
  • D.M. Kullmann

    Neuron,

    (1994)
  • D. Liao et al.

    Nature

    (1995)
  • B. Barbour et al.

    Trends Neurosci.

    (1997)
  • T.V.P. Bliss et al.

    Nature

    (1993)
  • G.M. Durand et al.

    Nature

    (1996)
  • J.T.R. Isaac

    Neuron

    (1996)
  • Cited by (283)

    • Coexisting multiple firing behaviors of fractional-order memristor-coupled HR neuron considering synaptic crosstalk and its ARM-based implementation

      2022, Chaos, Solitons and Fractals
      Citation Excerpt :

      Wang et al. constructed a simple locally active mHR neuron model by using the locally active memristor as an autapse and found an interesting phenomenon of coexisting position symmetry for different attractors [26]. Synaptic crosstalk occurs when components of one synapse overflow into another to affect its signal transmission, or when receptors in one spine release that affect neighboring spines [27–29]. Synaptic crosstalk, which destroys the body's ability to specifically control the strength of a single synapse, is a phenomenon that cannot be ignored and has a profound impact on the specificity of synaptic communication in the human brain [30,31].

    • Cholinergic and Adenosinergic Modulation of Synaptic Release

      2021, Neuroscience
      Citation Excerpt :

      They will activate mostly ionotropic receptors in the postsynaptic membrane. These neurotransmitter can escape from the synaptic cleft – a mechanism termed ‘spill-over’ and may either activate GPCR autoreceptors in the presynaptic bouton of the same synapse or heteroreceptors in the presynaptic bouton of adjacent synapses (Fig. 5B, left; Kullmann and Asztely, 1998; Chen and Diamond, 2002; Bekkers, 2003). For example, ACh, GABA, and glutamate may bind to perisynaptic mAChRs, GABAB receptors or the metabotropic glutamate receptors (mGluRs), respectively, resulting in a feedback regulation of neurotransmitter release probability.

    View all citing articles on Scopus
    View full text