Carta Acesso aberto Revisado por pares

The Two Sides of Hippocampal Mossy Fiber Plasticity

2008; Cell Press; Volume: 57; Issue: 1 Linguagem: Inglês

10.1016/j.neuron.2007.12.015

ISSN

1097-4199

Autores

Angharad M. Kerr, Péter Jónás,

Tópico(s)

Neurogenesis and neuroplasticity mechanisms

Resumo

Two studies in this issue of Neuron (Kwon and Castillo and Rebola et al.) show that the mossy fiber-CA3 pyramidal neuron synapse, a hippocampal synapse well known for its presynaptic plasticity, exhibits a novel form of long-term potentiation of NMDAR-mediated currents, which is induced and expressed postsynaptically. Two studies in this issue of Neuron (Kwon and Castillo and Rebola et al.) show that the mossy fiber-CA3 pyramidal neuron synapse, a hippocampal synapse well known for its presynaptic plasticity, exhibits a novel form of long-term potentiation of NMDAR-mediated currents, which is induced and expressed postsynaptically. The mossy fiber synapse onto CA3 pyramidal neurons is a key synapse in the trisynaptic circuit of the hippocampus, providing the major synaptic connection from the dentate gyrus to the CA3 region. This synapse is believed to be highly efficient, acting as a "conditional detonator" that is able to evoke action potentials in CA3 pyramidal neurons if activated repetitively (Henze et al., 2002Henze D.A. Wittner L. Buzsáki G. Nat. Neurosci. 2002; 5: 790-795PubMed Google Scholar). At the network level, the hippocampal mossy fiber synapse may operate as a teacher synapse that triggers spike-timing-dependent plasticity at associational-commissural synapses between CA3 pyramidal neurons (CA3-CA3 cell synapses; Kobayashi and Poo, 2004Kobayashi K. Poo M.M. Neuron. 2004; 41: 445-454Abstract Full Text Full Text PDF PubMed Scopus (83) Google Scholar). Thus, the mossy fiber synapse may play an important role in complex network functions, such as the storage and recall of information and pattern completion (Nakazawa et al., 2002Nakazawa K. Quirk M.C. Chitwood R.A. Watanabe M. Yeckel M.F. Sun L.D. Kato A. Carr C.A. Johnston D. Wilson M.A. Tonegawa S. Science. 2002; 297: 211-218Crossref PubMed Scopus (796) Google Scholar, Bischofberger et al., 2006Bischofberger J. Engel D. Frotscher M. Jonas P. Pflügers Arch.Eur. J. Phys. 2006; 453: 361-372Crossref Scopus (93) Google Scholar). The mossy fiber synapse not only triggers heterosynaptic plasticity, but is also highly plastic by itself. Several forms of homosynaptic short- and long-term plasticity have been reported, including paired-pulse facilitation, frequency facilitation, posttetanic potentiation, and long-term potentiation (LTP; Salin et al., 1996Salin P.A. Scanziani M. Malenka R.C. Nicoll R.A. Proc. Natl. Acad. Sci. USA. 1996; 93: 13304-13309Crossref PubMed Scopus (329) Google Scholar, Toth et al., 2000Toth K. Suares G. Lawrence J.J. Philips-Tansey E. McBain C.J. J. Neurosci. 2000; 20: 8279-8289PubMed Google Scholar). At Schaffer collateral-CA1 pyramidal cell synapses, LTP is N-methyl-D-aspartate receptor (NMDAR) dependent and is currently thought to be induced and expressed postsynaptically (Malinow and Malenka, 2002Malinow R. Malenka R.C. Annu. Rev. Neurosci. 2002; 25: 103-126Crossref PubMed Scopus (2007) Google Scholar). In contrast, the induction of mossy fiber LTP at mossy fiber-CA3 pyramidal neuron synapses is believed to be NMDAR independent and mainly presynaptic, although a postsynaptic component may be present under some conditions (Harris and Cotman, 1986Harris E.W. Cotman C.W. Neurosci. Lett. 1986; 70: 132-137Crossref PubMed Scopus (426) Google Scholar, Zalutsky and Nicoll, 1990Zalutsky R.A. Nicoll R.A. Science. 1990; 248: 1619-1624Crossref PubMed Scopus (622) Google Scholar, Yeckel et al., 1999Yeckel M.F. Kapur A. Johnston D. Nat. Neurosci. 1999; 2: 625-633Crossref PubMed Scopus (272) Google Scholar). Furthermore, the expression of mossy fiber LTP appears to be entirely presynaptic. Despite some controversy, there has been "unanimous agreement […] that mossy fiber LTP is independent of NMDARs and that mossy fiber LTP is expressed presynaptically" (Nicoll and Schmitz, 2005Nicoll R.A. Schmitz D. Nat. Rev. Neurosci. 2005; 6: 863-876Crossref PubMed Scopus (452) Google Scholar). However, two papers in this issue of Neuron (Kwon and Castillo, 2008Kwon H.-B. Castillo P.E. Neuron. 2008; 57 (this issue): 108-120Abstract Full Text Full Text PDF PubMed Scopus (112) Google Scholar, Rebola et al., 2008Rebola N. Lujan R. Cunha R.A. Mulle C. Neuron. 2008; 57 (this issue): 121-134Abstract Full Text Full Text PDF PubMed Scopus (261) Google Scholar) reveal a substantially more complex picture. The fact that conventional mossy fiber LTP is independent of NMDAR activation (Harris and Cotman, 1986Harris E.W. Cotman C.W. Neurosci. Lett. 1986; 70: 132-137Crossref PubMed Scopus (426) Google Scholar) does not necessarily mean that NMDARs do not contribute to mossy fiber synaptic transmission. NMDAR-mediated current components were found in excitatory postsynaptic currents (EPSCs) evoked by stimulation of both mossy fiber axons and dentate gyrus granule cells (Jonas et al., 1993Jonas P. Major G. Sakmann B. J. Physiol. 1993; 472: 615-663Crossref PubMed Scopus (518) Google Scholar, Weisskopf and Nicoll, 1995Weisskopf M.G. Nicoll R.A. Nature. 1995; 376: 256-259Crossref PubMed Scopus (156) Google Scholar). Furthermore, NMDAR-mediated components were evident in glutamate-activated currents recorded from outside-out patches isolated from the proximal apical dendrites of CA3 pyramidal neurons (Spruston et al., 1995Spruston N. Jonas P. Sakmann B. J. Physiol. 1995; 482: 325-352Crossref PubMed Scopus (371) Google Scholar). Although NMDARs will be largely blocked by extracellular Mg2+ at rest, the large amplitude of unitary EPSPs at the mossy fiber-CA3 cell synapse may relieve the Mg2+ block, either directly or indirectly by triggering action potentials in CA3 cells (Henze et al., 2002Henze D.A. Wittner L. Buzsáki G. Nat. Neurosci. 2002; 5: 790-795PubMed Google Scholar). Taken together, these results indicate that NMDARs are present at hippocampal mossy fiber-CA3 pyramidal neuron synapses and could make an important contribution to mossy fiber synaptic transmission. Kwon and Castillo, 2008Kwon H.-B. Castillo P.E. Neuron. 2008; 57 (this issue): 108-120Abstract Full Text Full Text PDF PubMed Scopus (112) Google Scholar and Rebola et al., 2008Rebola N. Lujan R. Cunha R.A. Mulle C. Neuron. 2008; 57 (this issue): 121-134Abstract Full Text Full Text PDF PubMed Scopus (261) Google Scholar examine the NMDAR-mediated component of EPSCs at mossy fiber-CA3 pyramidal neuron synapses in more detail, using whole-cell patch-clamp recordings in acute slices. The main finding is that moderate repetitive stimulation (typically 24 stimuli at 25 Hz in Kwon and Castillo, 2008Kwon H.-B. Castillo P.E. Neuron. 2008; 57 (this issue): 108-120Abstract Full Text Full Text PDF PubMed Scopus (112) Google Scholar; 6 bursts of 6 stimuli at 50 Hz in Rebola et al., 2008Rebola N. Lujan R. Cunha R.A. Mulle C. Neuron. 2008; 57 (this issue): 121-134Abstract Full Text Full Text PDF PubMed Scopus (261) Google Scholar) leads to a substantial (∼200%), long-lasting enhancement of the NMDAR component of mossy fiber EPSCs. In both studies, mossy fiber inputs were rigorously identified using application of the group 2 mGluR agonists DCG-4 or LCCG-1, which selectively reduce transmitter release from hippocampal mossy fiber synapses. Furthermore, in the study by Kwon and Castillo, 2008Kwon H.-B. Castillo P.E. Neuron. 2008; 57 (this issue): 108-120Abstract Full Text Full Text PDF PubMed Scopus (112) Google Scholar polysynaptic events from CA3-CA3 cell synapses were minimized using an elevated Ca2+ and Mg2+ concentration in the bath solution. Thus, the new studies convincingly demonstrate a novel form of NMDAR-LTP at hippocampal mossy fiber synapses. Although the two papers converge on similar final conclusions, the starting points of the studies are very different. Kwon and Castillo, 2008Kwon H.-B. Castillo P.E. Neuron. 2008; 57 (this issue): 108-120Abstract Full Text Full Text PDF PubMed Scopus (112) Google Scholar start out from the well-known properties of LTP at glutamatergic synapses. They examine the effects of previously used LTP induction protocols on the amplitude of NMDAR-mediated EPSCs and find a substantial long-lasting potentiation. They identify several factors necessary for NMDAR-LTP induction: NMDAR activation, metabotropic glutamate receptor 5 (mGluR5) activation, G protein activation, a rise in postsynaptic Ca2+ concentration, Ca2+ release from intracellular stores, and protein kinase C (PKC) activation. Furthermore, the authors provide evidence that the potentiation of the NMDAR-mediated EPSC is dependent on soluble N-ethylmaleimide-sensitive factor attachment protein receptors (SNAREs), suggesting that it requires insertion of previously assembled NMDARs into the postsynaptic membrane. In contrast, Rebola et al., 2008Rebola N. Lujan R. Cunha R.A. Mulle C. Neuron. 2008; 57 (this issue): 121-134Abstract Full Text Full Text PDF PubMed Scopus (261) Google Scholar start out by investigating the role of the neuromodulator adenosine, which reduces the probability of transmitter release at mossy fiber-CA3 pyramidal neuron synapses via activation of A1 receptors (Moore et al., 2003Moore K.A. Nicoll R.A. Schmitz D. Proc. Natl. Acad. Sci. USA. 2003; 100: 14397-14402Crossref PubMed Scopus (117) Google Scholar). They test the effects of activation of A2A receptors, but do not detect any change in the non-NMDAR component of mossy fiber EPSCs. They go on to examine the location of A2A receptors using immunoelectron microscopy and find that A2A receptors are located postsynaptically. Consequently, they test whether activation of A2A receptors affects NMDAR-mediated EPSCs at mossy fiber synapses and find that, under physiological conditions, repetitive mossy fiber stimulation induces a long-lasting, A2A receptor-dependent potentiation of NMDAR-mediated currents. In addition to A2A receptors, NMDAR-LTP at hippocampal mossy fiber synapses is also dependent on NMDARs, mGluR5 receptors, G proteins, a rise in postsynaptic Ca2+ concentration, and Src tyrosine kinase. Thus, both studies show that NMDAR activation, mGluR5 activation, G proteins, and a rise in postsynaptic Ca2+ are required for NMDAR-LTP. Kwon and Castillo, 2008Kwon H.-B. Castillo P.E. Neuron. 2008; 57 (this issue): 108-120Abstract Full Text Full Text PDF PubMed Scopus (112) Google Scholar demonstrate the involvement of PKC and SNARE-dependent mechanisms, whereas Rebola et al., 2008Rebola N. Lujan R. Cunha R.A. Mulle C. Neuron. 2008; 57 (this issue): 121-134Abstract Full Text Full Text PDF PubMed Scopus (261) Google Scholar reveal the involvement of A2A receptors and Src tyrosine kinases. Taken together, the results show that NMDAR-LTP is generated by a complex network of intracellular signaling molecules (Figure 1). Intriguingly, the molecular mechanisms underlying NMDAR-LTP are very different from those of conventional LTP at mossy fiber synapses. For example, cAMP and protein kinase A (PKA) are of critical importance in conventional LTP (Huang et al., 1994Huang Y.Y. Li X.C. Kandel E.R. Cell. 1994; 79: 69-79Abstract Full Text PDF PubMed Scopus (435) Google Scholar), whereas PKC is involved in NMDAR-LTP at hippocampal mossy fiber synapses. Furthermore, adenosine plays very different roles in conventional LTP and NMDAR-LTP. In conventional LTP, tonic activation of presynaptic A1 receptors by ambient adenosine reduces release probability, which in turn facilitates plasticity (Moore et al., 2003Moore K.A. Nicoll R.A. Schmitz D. Proc. Natl. Acad. Sci. USA. 2003; 100: 14397-14402Crossref PubMed Scopus (117) Google Scholar). In contrast, in NMDAR-LTP, adenosine appears to be generated during tetanic stimulation, and the subsequent activation of postsynaptic A2A receptors contributes to LTP induction. The cellular source of adenosine is not entirely clear. Rebola et al., 2008Rebola N. Lujan R. Cunha R.A. Mulle C. Neuron. 2008; 57 (this issue): 121-134Abstract Full Text Full Text PDF PubMed Scopus (261) Google Scholar show that blockers of ecto-ATPases and ecto-5′-nucleotidases reduce the extent of NMDAR-LTP, suggesting that adenosine is formed through conversion from previously released ATP. Where is the locus of induction of the new form of NMDAR-LTP at mossy fiber synapses? In conventional mossy fiber LTP, it is thought that the induction is mainly presynaptic, although a postsynaptic component may be present under some conditions (Zalutsky and Nicoll, 1990Zalutsky R.A. Nicoll R.A. Science. 1990; 248: 1619-1624Crossref PubMed Scopus (622) Google Scholar, Yeckel et al., 1999Yeckel M.F. Kapur A. Johnston D. Nat. Neurosci. 1999; 2: 625-633Crossref PubMed Scopus (272) Google Scholar). In the new form of NMDAR-LTP at mossy fiber synapses, both Kwon and Castillo, 2008Kwon H.-B. Castillo P.E. Neuron. 2008; 57 (this issue): 108-120Abstract Full Text Full Text PDF PubMed Scopus (112) Google Scholar and Rebola et al., 2008Rebola N. Lujan R. Cunha R.A. Mulle C. Neuron. 2008; 57 (this issue): 121-134Abstract Full Text Full Text PDF PubMed Scopus (261) Google Scholar report clear evidence for a postsynaptic locus of induction. Kwon and Castillo, 2008Kwon H.-B. Castillo P.E. Neuron. 2008; 57 (this issue): 108-120Abstract Full Text Full Text PDF PubMed Scopus (112) Google Scholar find that the Ca2+ chelator BAPTA, GDP-β-S, and the PKC blocker chelerythrine suppress LTP when included in the solution of the postsynaptic recording pipette. Similarly, Rebola et al., 2008Rebola N. Lujan R. Cunha R.A. Mulle C. Neuron. 2008; 57 (this issue): 121-134Abstract Full Text Full Text PDF PubMed Scopus (261) Google Scholar find that BAPTA, GDP-β-S, the tyrosine kinase inhibitor PP1, and an Src inhibitory peptide block NMDAR-LTP from the postsynaptic side. These results converge toward the conclusion that the induction of NMDAR-LTP at hippocampal mossy fiber synapses is postsynaptic. Where is the locus of expression of the new form of NMDAR-LTP? Both studies provide evidence that the expression of NMDAR-LTP, like its induction, is postsynaptic. First, the potentiation selectively affects the NMDAR component, whereas the non-NMDAR component of the EPSC remains unchanged (Kwon and Castillo, 2008Kwon H.-B. Castillo P.E. Neuron. 2008; 57 (this issue): 108-120Abstract Full Text Full Text PDF PubMed Scopus (112) Google Scholar, Rebola et al., 2008Rebola N. Lujan R. Cunha R.A. Mulle C. Neuron. 2008; 57 (this issue): 121-134Abstract Full Text Full Text PDF PubMed Scopus (261) Google Scholar). This selective potentiation is easier to reconcile with a postsynaptic than with a presynaptic expression of LTP. Second, changes in synaptic efficacy are not associated with changes in the paired-pulse ratio or the coefficient of variation of EPSC peak amplitude, which would be expected to track changes in release probability (Kwon and Castillo, 2008Kwon H.-B. Castillo P.E. Neuron. 2008; 57 (this issue): 108-120Abstract Full Text Full Text PDF PubMed Scopus (112) Google Scholar, Rebola et al., 2008Rebola N. Lujan R. Cunha R.A. Mulle C. Neuron. 2008; 57 (this issue): 121-134Abstract Full Text Full Text PDF PubMed Scopus (261) Google Scholar). Third, the light chain of botulinum toxin B (cleaving synaptobrevin) and a peptide that mimics the C-terminal sequence of SNAP-25 suppress LTP when present in the postsynaptic cell (Kwon and Castillo, 2008Kwon H.-B. Castillo P.E. Neuron. 2008; 57 (this issue): 108-120Abstract Full Text Full Text PDF PubMed Scopus (112) Google Scholar). These results suggest that the expression of NMDAR-LTP involves the insertion of new NMDARs into the postsynaptic membrane. Taken together, these results imply that the expression of NMDAR-LTP is postsynaptic, presumably occurring through insertion of new NMDARs. In parallel, PKC and Src kinase may modulate NMDAR function directly, possibly by channel phosphorylation (Salter and Kalia, 2004Salter M.W. Kalia L.V. Nat. Rev. Neurosci. 2004; 5: 317-328Crossref PubMed Scopus (621) Google Scholar). The two papers considerably change our way of thinking about mossy fiber plasticity. Whereas the conventional mossy fiber LTP is expressed presynaptically, the new NMDAR-LTP is induced and expressed postsynaptically. The two studies further show that both forms of LTP coexist at the same individual synapse. Why should a single mossy fiber synapse coexpress two different forms of LTP? Intriguingly, the two papers demonstrate that the threshold for LTP induction is different for the two forms of LTP. A weak repetitive stimulation is sufficient to induce the NMDAR-LTP, whereas a strong repetitive stimulation is needed to induce conventional LTP at mossy fiber synapses (Salin et al., 1996Salin P.A. Scanziani M. Malenka R.C. Nicoll R.A. Proc. Natl. Acad. Sci. USA. 1996; 93: 13304-13309Crossref PubMed Scopus (329) Google Scholar, Kwon and Castillo, 2008Kwon H.-B. Castillo P.E. Neuron. 2008; 57 (this issue): 108-120Abstract Full Text Full Text PDF PubMed Scopus (112) Google Scholar). Similarly, because of their differential dependence on PKA and PKC, presynaptic LTP and postsynaptic NMDAR-LTP could be differentially sensitive to neuromodulators. What could be the functional significance of NMDAR-LTP? Presynaptic mossy fiber LTP is expected to scale the amplitude of synaptic events, by potentiating both the fast non-NMDAR component and the slow NMDAR component. In contrast, postsynaptic NMDAR-LTP is expected to alter the time course of the EPSP by selectively potentiating the slow NMDAR component. Thus, NMDAR-LTP may facilitate the temporal summation of EPSPs and the initiation of bursts of action potentials in postsynaptic CA3 pyramidal neurons following repetitive activity in granule cells. In this framework, NMDAR-LTP may contribute to the "conditional detonator" properties of mossy fiber synapses (Henze et al., 2002Henze D.A. Wittner L. Buzsáki G. Nat. Neurosci. 2002; 5: 790-795PubMed Google Scholar). Does the new form of NMDAR-LTP play any role for the operation of the hippocampal network in vivo? The two studies do not provide a direct answer to this question. However, there is room for speculation. A large proportion of granule cells in the dentate gyrus are "place cells," showing a firing rate that is modulated by the animal's position in space. If the animal is located outside the place field of a given granule cell, the cell fires at low frequency (<0.5 Hz; Jung and McNaughton, 1993Jung M.W. McNaughton B.L. Hippocampus. 1993; 3: 165-182Crossref PubMed Scopus (559) Google Scholar). However, if the animal passes through the place field center, the granule cell transiently fires with a frequency up to 50 Hz (Jung and McNaughton, 1993Jung M.W. McNaughton B.L. Hippocampus. 1993; 3: 165-182Crossref PubMed Scopus (559) Google Scholar, Henze et al., 2002Henze D.A. Wittner L. Buzsáki G. Nat. Neurosci. 2002; 5: 790-795PubMed Google Scholar). According to the present results (Kwon and Castillo, 2008Kwon H.-B. Castillo P.E. Neuron. 2008; 57 (this issue): 108-120Abstract Full Text Full Text PDF PubMed Scopus (112) Google Scholar, Rebola et al., 2008Rebola N. Lujan R. Cunha R.A. Mulle C. Neuron. 2008; 57 (this issue): 121-134Abstract Full Text Full Text PDF PubMed Scopus (261) Google Scholar), NMDAR-LTP may be triggered under these conditions. In contrast, several or more long-lasting visits of the animal to the same place field may be required to trigger conventional presynaptic mossy fiber LTP. These predictions need to be tested in future experiments. Adenosine A2A Receptors Are Essential for Long-Term Potentiation of NMDA-EPSCs at Hippocampal Mossy Fiber SynapsesRebola et al.NeuronJanuary 10, 2008In BriefThe physiological conditions under which adenosine A2A receptors modulate synaptic transmission are presently unclear. We show that A2A receptors are localized postsynaptically at synapses between mossy fibers and CA3 pyramidal cells and are essential for a form of long-term potentiation (LTP) of NMDA-EPSCs induced by short bursts of mossy fiber stimulation. This LTP spares AMPA-EPSCs and is likely induced and expressed postsynaptically. It depends on a postsynaptic Ca2+ rise, on G protein activation, and on Src kinase. Full-Text PDF Open ArchiveLong-Term Potentiation Selectively Expressed by NMDA Receptors at Hippocampal Mossy Fiber SynapsesKwon et al.NeuronJanuary 10, 2008In BriefThe mossy fiber to CA3 pyramidal cell synapse (mf-CA3) provides a major source of excitation to the hippocampus. Thus far, these glutamatergic synapses are well recognized for showing a presynaptic, NMDA receptor-independent form of LTP that is expressed as a long-lasting increase of transmitter release. Here, we show that in addition to this "classical" LTP, mf-CA3 synapses can undergo a form of LTP characterized by a selective enhancement of NMDA receptor-mediated transmission. This potentiation requires coactivation of NMDA and mGlu5 receptors and a postsynaptic calcium rise. Full-Text PDF Open Archive

Referência(s)
Altmetric
PlumX