L-type Ca Channels

B R I E F C O M M U N I C AT I O N S

L-type voltage-dependent Ca2+ channels mediate expression of presynaptic LTP in amygdala

© 2009 Nature America, Inc. All rights reserved.

Elodie Fourcaudot1–3, Frederic Gambino2,3, Guillaume Casassus1, Bernard Poulain2, Yann Humeau1,2 & Andreas Lu¨thi1 The molecular mechanisms underlying the expression of postsynaptic long-term potentiation (LTP) at glutamatergic synapses are well understood. However, little is known about those that mediate the expression of presynaptic LTP. We found that presynaptic LTP at cortical inputs to the mouse lateral amygdala was blocked and reversed by L-type voltagedependent Ca21 channel (L-VDCC) blockers. Thus, a persistent increase in L-VDCC–mediated glutamate release underlies the expression of presynaptic LTP in the amygdala. Activity-dependent long-term changes in the efficacy of glutamatergic synaptic transmission are important in many forms of learning. Studies on LTP of glutamatergic synapses in different brain areas, including cortex, hippocampus and amygdala, have revealed that synapse strength can be enhanced by two principal mechanisms. Although a rise in intracellular Ca2+ is the trigger in both cases, one form of LTP is expressed by a persistent increase in the postsynaptic response to a fixed amount of glutamate released1, whereas the other major form of LTP is mediated by an increase in the presynaptic probability of release2. The molecular mechanisms underlying postsynaptic LTP expression have been studied in great detail and involve changes in the trafficking or in the properties of postsynaptic glutamate receptors1. In contrast, the mechanisms underlying the expression of presynaptic LTP remain poorly understood2. Because neurotransmitter release is a highly Ca2+-dependent process, long-term modifications in presynaptic Ca2+ influx or sensing could possibly result in a persistent increase in release probability. Action-potential driven presynaptic Ca2+ influx is mediated by voltagedependent Ca2+ channels (VDCCs), a diverse family of molecularly and pharmacologically distinct ion channels3 underlying various forms of synaptic plasticity3–6. We investigated the role of distinct VDCC subtypes in the expression of LTP of presynaptic release in the lateral nucleus of the amygdala (LA), a brain area that is necessary for the acquisition of conditioned fear responses7. In LA, presynaptic LTP is most easily induced at synapses made by cortical afferents onto principal neurons8. Induction of presynaptic cortico-LA LTP requires activation of presynaptic NMDA receptors by glutamate released from thalamo-LA afferents8. Thus, analysis of the role of presynaptic VDCCs in LTP is complicated by the fact that VDCC antagonists may interfere

with LTP induction by reducing release at thalamo-LA synapses. Instead of electrically inducing LTP, we analyzed LTP that was induced by the adenylyl cyclase activator forskolin (FSK, referred to as LTPFSK), which enhances cortico-LA synaptic release by directly activating cAMP/protein kinase A–dependent signaling mechanisms9, engaging the same expression mechanisms as those involved in electrically induced LTP9. First, to determine the contribution of different VDCC subtypes to baseline synaptic release at cortico-LA synapses, we applied blockers of L-type (verapamil, 50 mM), P/Q-type (o-agatoxin IVA (o-agaIVA), 0.5 mM), N-type (o-conotoxin GVIa (o-CgTx), 1 mM) and R-type (Ni2+, 10 mM) VDCCs (Fig. 1a). Although the proportion of synaptic release mediated by L-VDCCs (15 ± 4%, n ¼ 7, P o 0.01) and R-VDCCs (18 ± 8%, n ¼ 6, P 4 0.05) was relatively minor, both P/Q-VDCCs (43 ± 6%, n ¼ 5, P o 0.01) and N-VDCCs (55 ± 6%, n ¼ 5, P o 0.001) substantially contributed to release at cortico-LA synapses (Fig. 1a,b). In contrast with baseline synaptic transmission, comparing blockers of specific VDCC subtypes revealed a specific role of L-VDCCs in LTPFSK. Application of FSK (50 mM) induced normal LTP in the presence of N-, P/Q- and R-VDCC blockers (control: 144 ± 13% of baseline, n ¼ 17, P o 0.05; o-CgTx: 146 ± 14% of baseline, n ¼ 10, P o 0.05; o-agaIVA: 145 ± 13% of baseline, n ¼ 4, P o 0.05; Ni2+: 145 ± 17% of baseline, n ¼ 5, P o 0.05; Fig. 1c). However, LTPFSK was completely abolished by the L-VDCC blockers verapamil (78 ± 13% of baseline, n ¼ 6, P 4 0.05 versus pre-FSK baseline, P o 0.05 versus control; Fig. 1c,d) and nimodipine (10 mM; 90 ± 15% of baseline, n ¼ 8, P 4 0.05 versus pre-FSK baseline, P o 0.05 versus control; Fig. 1d). L-VDCCs were not only required for LTPFSK, but were also necessary for electrically induced LTP (control: 146 ± 12%, n ¼ 19; verapamil: 95 ± 11%, n ¼ 9, P o 0.05; Fig. 1e). Thus, although L-VDCCs only weakly contributed to baseline synaptic release, they were specifically required for presynaptic LTP. To examine whether L-VDCCs contribute to the expression of LTPFSK, we applied L-VDCC blockers after stable LTPFSK had been induced. Both verapamil (n ¼ 5) and nimodipine (n ¼ 4) completely reversed previously established LTPFSK (Fig. 2a). In the absence of blockers, LTPFSK remained stable for the duration of the experiment (Fig. 2a). Comparing the effect of L-VDCC blockers on synaptic transmission at naive synapses and at synapses at which LTPFSK had been induced revealed a marked increase in the L-VDCC–mediated component after induction of LTPFSK (verapamil: naive, 15 ± 4% inhibition, n ¼ 7; LTPFSK, 41 ± 11%, n ¼ 5, P o 0.05 versus naive; nimodipine: naive, 1 ± 8%, n ¼ 9; LTPFSK, 34 ± 15%, n ¼ 4, P o 0.05 versus naive; Fig. 2b). Conversely, after induction of LTPFSK, the fraction of synaptic transmission mediated by other VDCC subtypes decreased (o-agaIVA: naive, 43 ± 6%, n ¼ 5; LTPFSK, 13 ± 13%, n ¼ 4, P o 0.05 versus naive; o-CgTx: naive, 55 ± 6%, n ¼ 5; LTPFSK, 34 ± 8%, n ¼ 5, P o 0.05 versus naive) or remained constant (Ni2+: naive,

1Friedrich Miescher Institute for Biomedical Research, Basel, Switzerland. 2Centre National de la Recherche Scientifique, UPR3212, Strasbourg, France. 3These authors contributed equally to this work. Correspondence should be addressed to A.L. ([email protected]).

Received 8 April; accepted 30 June; published online 2 August 2009; doi:10.1038/nn.2378

NATURE NEUROSCIENCE VOLUME 12

[

NUMBER 9

[

SEPTEMBER 2009

1093

B R I E F C O M M U N I C AT I O N S

–5 0 5 10 Time (min)

–5 0 5 10 Time (min)

100

0

***

60

** 40 20

**

Ve ra ω- pa ag mi a l ω- IVA C gT x N 2 i +

0

d

* 150

*

*

BL FSK

100 Control

0 0

e

50

C Ve on N rap tro im a l ω- odip mil ag in a e ω- IV C A gT x N 2 i +

0

BL FSK

FSK

200

Verapamil

10 20 30 Time (min)

Control

* * *

100

–5 0 5 10 Time (min) Verapamil

Control

c EPSC amplitude (%)

EPSC amplitude (% inhibition)

b

FSK potentiation (% of baseline)

Ni

Post Pre

40

Verapamil Post Pre

200 100 Pairing

Ctrl

0 –5

0

5

10

Verap.

15

20

Time (min)

18 ± 8%, n ¼ 6; LTPFSK, 14 ± 17%, n ¼ 5, not significant versus naive; Fig. 2b and Supplementary Fig. 1). Because induction of LTPFSK results in an increase in the presynaptic probability of release9, it was associated with a faster depletion of the readily releasable pool of synaptic vesicles as reflected by a shift toward a more pronounced synaptic depression during repetitive 20-Hz stimulation (n ¼ 10, P o 0.05; Fig. 2c). We therefore examined whether subsequent application of an L-VDCC blocker reverses not only excitatory postsynaptic current (EPSC) amplitude, but also short-term plasticity. Blockade of L-VDCCs completely reversed FSKinduced changes in short-term plasticity (n ¼ 7, P o 0.05 versus FSK, P 4 0.05 versus control; Fig. 2c) with no effect on axonal excitability (Supplementary Fig. 2), indicating a direct effect on synaptic release. Finally, we applied verapamil after induction of presynaptic LTP by co-stimulation of thalamo- and cortico-LA afferents. Consistent with its effect on LTPFSK, verapamil blocked a larger fraction of transmission at potentiated synapses compared with naive synapses (naive: 15 ± 4%, n ¼ 7; LTP: 49 ± 10%, n ¼ 6, P o 0.05) and abolished previously established LTP (Fig. 2d). In contrast, the relative contribution of P/Q-VDCCs and N-VDCCs was reduced after electrical LTP induction (o-agaIVA: naive, 43 ± 6%, n ¼ 5; LTP, 22 ± 11%, n ¼ 5; o-CgTx: naive, 55 ± 6%, n ¼ 5; LTP, 35 ± 8%, n ¼ 3).

These data indicate that, unlike in hippocampal mossy fiber potentiation10, L-VDCCs are necessary for the persistent expression, and possibly the induction, of presynaptic cortico-LA LTP. Consistent with previous reports3,10, we found that L-VDCCs only weakly contributed to baseline release. After induction of LTP, however, the contribution of L-VDCCs to synaptic transmission was markedly increased and previously established LTP was completely reversed by L-VDCC blockade. In principle, changes in channel number, subcellular localization or in their functional properties could mediate this effect. In an attempt to address this question, we increased the efficacy of L-VDCCs by applying the positive allosteric modulator BayK 8644 (10 mM)11. Application of BayK 8644 strongly enhanced EPSC amplitude and decreased the paired-pulse ratio, indicating a presynaptic mechanism of action (Supplementary Fig. 3). Although LTP and BayK 8644 may act on L-VDCCs by distinct molecular mechanisms, this finding demonstrates that it is possible to increase synaptic strength by pharmacologically enhancing activation of pre-existing L-VDCCs. Thus, LTP expression could be mediated by a persistent enhancement of L-VDCC efficacy, possibly involving a protein kinase A–dependent mechanism3,9 and/or proteins interacting with L-VDCCs such as the presynaptic active zone component RIM1a9,12,13. Our data are consistent with in vivo pharmacological experiments demonstrating a role for L-VDCCs in fear conditioning14,15. Systemic and intra-LA

a

FSK

Baseline

300

Verapamil

Verapamil

FSK

200 100 FSK

0

FSK+Verap.

FSK

300

Nimodipine

FSK 0

FSK+Nimo.

10 20 30 40 50 60 Time (min)

b

Baseline LTPFSK 60

*

*

*

40

* *

20

FSK+Verap.

150 100 50 Ctrl

0

FSK

0.0

FSK+Verap.

0.5

1.0

d

Baseline

LTP

Verap. Verapamil

200 100 Pairing

0 0

10

20

30

Time (min)

VOLUME 12

[

x

N 2 i +

gT

A

C

a ag

ω-

ω-

IV

in ip od

im

N

Ve

ra p

am

e

0 il

EPSC amplitude (% inhibition)

FSK

Time (s)

0

1094

Control

200 100

Figure 2 L-VDCC mediate expression of presynaptic LTP. (a) L-VDCC blockers reversed previously established LTPFSK. In the absence of L-VDCC blockers, LTPFSK remained stable. Scale bars represent 50 pA and 5 ms. (b) Induction of LTPFSK selectively increased the contribution of L-VDCCs to synaptic transmission. VDCC blockers were applied to naive synapses or after induction of stable LTPFSK. (c) LTPFSK was associated with altered short-term synaptic plasticity. FSK-induced changes in 20-Hz depression were fully reversed by L-VDCC blockade. Scale bars represent 50 pA and 250 ms. (d) L-VDCC blockade reversed previously established LTP induced by afferent stimulation. Scale bars represent 50 pA and 5 ms. * P o 0.05. Error bars are s.e.m.

c EPSC amplitude (%)

ω-CgTx

Figure 1 Differential role of distinct VDCC subtypes in synaptic transmission and plasticity. (a) Basal synaptic transmission at cortico-LA synapses was predominantly mediated by P/Q- and N-VDCCs (blocked by o-agaIVA and o-CgTx, respectively), whereas L- and R-VDCCs (blocked by verapamil and Ni2+, respectively) contributed only a minor fraction. (b) Summary plot illustrating the percentage of inhibition of the EPSC amplitude induced by different VDCC blockers. (c) LTPFSK was abolished in the presence of the L-VDCC blocker verapamil (BL, pre-FSK baseline). Scale bars represent 50 pA and 10 ms. (d) Summary plot illustrating the effect of different VDCC blockers on LTPFSK. LTPFSK was selectively impaired by two different L-VDCC blockers (verapamil and nimodipine), but not by blockers of P/Q-, N- or R-VDCCs. (e) LTP induced by afferent stimulation was abolished in the presence of an L-VDCC blocker. Scale bars represent 2 mV and 10 ms. * P o 0.05, ** P o 0.01 and *** P o 0.001. Error bars are s.e.m. All experiments were carried out following a protocol approved by the Veterinary Department of the Canton of Basel-Stadt (Supplementary Methods).

EPSC amplitude (%)

ω-agaIVA

Verapamil

–5 0 5 10 Time (min)

© 2009 Nature America, Inc. All rights reserved.

2+

EPSC amplitude (%)

200

EPSP slope (%)

EPSC amplitude (% of baseline)

a

NUMBER 9

[

SEPTEMBER 2009 NATURE NEUROSCIENCE

B R I E F C O M M U N I C AT I O N S application of an L-VDCC antagonist selectively interferes with longterm fear memory without affecting short-term memory14,15, thus suggesting a role for L-VDCC–dependent presynaptic LTP in long-term memory formation. Note: Supplementary information is available on the Nature Neuroscience website.

ACKNOWLEDGMENTS We thank J. Letzkus and all members of the Lu¨thi laboratory for helpful discussions and comments on the manuscript. This study was supported by the Agence Nationale pour la Recherche, the European Neuroscience Institutes Network, the Eucor Learning and Teaching Mobility Program, the Swiss National Science Foundation, the Austrian Science Fund and the Novartis Research Foundation.

1. 2. 3. 4. 5. 6.

Malinow, R. & Malenka, R.C. Annu. Rev. Neurosci. 25, 103–126 (2002). Nicoll, R.A. & Schmitz, D. Nat. Rev. Neurosci. 6, 863–876 (2005). Catterall, W.A. Annu. Rev. Cell Dev. Biol. 16, 521–555 (2000). Catterall, W.A. & Few, A.P. Neuron 59, 882–901 (2008). Mato, S. et al. Neuropharmacology 54, 87–94 (2008). Pelkey, K.A., Topolnik, L., Lacaille, J.C. & McBain, C.J. Neuron 52, 497–510 (2006). 7. LeDoux, J.E. Annu. Rev. Neurosci. 23, 155–184 (2000). 8. Humeau, Y., Shaban, H., Bissie`re, S. & Lu¨thi, A. Nature 426, 841–845 (2003). 9. Fourcaudot, E. et al. Proc. Natl. Acad. Sci. USA 105, 15130–15135 (2008). 10. Castillo, P.E., Weisskopf, M.G. & Nicoll, R.A. Neuron 12, 261–269 (1994). 11. Fisher, R.E., Gray, R. & Johnston, D. J. Neurophysiol. 64, 91–104 (1990). 12. Coppola, T. et al. J. Biol. Chem. 276, 32756–32762 (2001). 13. Kiyonaka, S. et al. Nat. Neurosci. 10, 691–701 (2007). 14. Bauer, E.P., Schafe, G.E. & LeDoux, J.E. J. Neurosci. 22, 5239–5249 (2002). 15. Shinnick-Gallagher, P., McKernan, M.G., Xie, J. & Zinebi, F. Ann. NY Acad. Sci. 985, 135–149 (2003).

© 2009 Nature America, Inc. All rights reserved.

AUTHOR CONTRIBUTIONS E.F., F.G, G.C., B.P., Y.H. and A.L. designed the study. E.F., F.G., G.C. and Y.H. carried out the experiments. E.F., F.G., G.C., B.P., Y.H. and A.L. analyzed the data. E.F., G.C., B.P., Y.H. and A.L. wrote the manuscript.

Published online at http://www.nature.com/natureneuroscience/. Reprints and permissions information is available online at http://www.nature.com/ reprintsandpermissions/.

NATURE NEUROSCIENCE VOLUME 12

[

NUMBER 9

[

SEPTEMBER 2009

1095