The regulation of hippocampal LTP by the molecular switch, a form of metaplasticity, requires mGlu5 receptors

Neuropharmacology 49 (2005) 13e25 www.elsevier.com/locate/neuropharm

The regulation of hippocampal LTP by the molecular switch, a form of metaplasticity, requires mGlu5 receptors Zuner A. Bortolotto a, Valerie J. Collett a, Francois Conquet b, Zhengping Jia c, Herman van der Putten d, Graham L. Collingridge a,* a

MRC Centre for Synaptic Plasticity, Department of Anatomy, School of Medical Sciences, University of Bristol, University Walk, Bristol BS8 1TD, UK b Institut de Biologie Cellulaire et de Morphologie, Universite´ de Lausanne, 1005 Lausanne, Switzerland c Brain and Behavior, The Hospital for Sick Children, 555 University Avenue, Toronto, Ontario, Canada, M5G 1X8 d Neuroscience Research, Novartis Institutes for BioMedical Research, Novartis Pharma AG, 4002 Basel, Switzerland Received 11 March 2005; received in revised form 20 May 2005; accepted 25 May 2005

Abstract The role of metabotropic glutamate (mGlu) receptors in long-term potentiation (LTP) in the hippocampus is controversial. In the present study, we have used mice in which the mGlu1, mGlu5 or mGlu7 receptor has been deleted, by homologous recombination, to study the role of these receptor subtypes in LTP at CA1 synapses. We investigated the effects of the knockouts on both LTP and the molecular switch, a form of metaplasticity that renders LTP insensitive to the actions of the mGlu receptor antagonist MCPG ((S )a-methyl-4-carboxyphenylglycine). We find that LTP is readily induced in the three knockouts and in an mGlu1 and mGlu5 double knockout. In addition, the molecular switch operates normally in either the mGlu1 or mGlu7 knockout. In contrast, the molecular switch is completely non-functional in the mGlu5 knockout, such that MCPG invariably blocks the induction of additional LTP in an input where LTP has already been induced. The effect of the mGlu5 receptor knockout was replicated in wildtype mouse slices perfused with the specific mGlu5 receptor antagonist MPEP (2-methyl-6-(phenylethynyl)-pyridine). In addition, the mGlu5 selective agonist CHPG ((RS )-2-chloro-5-hydroxyphenylglycine) sets the molecular switch. These data demonstrate that the operation of the molecular switch requires activation of mGlu5 receptors. Ó 2005 Elsevier Ltd. All rights reserved. Keywords: LTP; MCPG; Synaptic plasticity; Hippocampal slice; Metabotropic

1. Introduction NMDA receptor-dependent long-term potentiation (LTP) of AMPA receptor-mediated synaptic transmission in the CA1 region of the hippocampus is the most widely studied experimental system for exploring the molecular basis of memory (Bliss and Collingridge,

* Corresponding author. Tel.: C44 117 928 7420; fax: C44 117 929 1687. E-mail address: [email protected] (G.L. Collingridge). 0028-3908/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.neuropharm.2005.05.020

1993). The ability of synapses to undergo changes in synaptic efficacy is also subject to modification, via a process termed metaplasticity e the plasticity of synaptic plasticity (Abraham and Bear, 1996). An early example of metaplasticity is the observation that activation of NMDA receptors can inhibit the subsequent induction of NMDA receptor-dependent LTP at CA1 synapses (Coan et al., 1989; Huang et al., 1992). A major family of G protein-coupled receptors, the metabotropic glutamate (mGlu) receptors, have been implicated in the induction of LTP in this pathway but an understanding of their precise roles has proved

14

Z.A. Bortolotto et al. / Neuropharmacology 49 (2005) 13e25

elusive (Anwyl, 1999). There is, however, growing acceptance that mGlu receptors are important in various forms of metaplasticity. For example, activation of metabotropic glutamate (mGlu) receptors can facilitate the subsequent induction of NMDA receptor-dependent LTP (Cohen and Abraham, 1996) and inhibits the induction of mGlu receptor-dependent LTD (Rush et al., 2002; Wu et al., 2004). We have previously described a different form of mGlu receptor-mediated metaplasticity at these synapses. We found that the synaptic activation of mGlu receptors renders the subsequent induction of LTP in the same input insensitive to the mGlu receptor antagonist (S )-a-methyl-4-carboxyphenylglycine (MCPG) (Bortolotto et al., 1994). In contrast, under the conditions of our experiments, the induction of LTP in experimentally naı¨ ve inputs is always inhibited by MCPG (Bashir et al., 1993; Bortolotto et al., 1994; Fitzjohn et al., 1998). To explain these results, we proposed the existence of an mGluR-mediated molecular switch that has to be set, either during or prior to the tetanus, for LTP to be induced. Once set, the switch stays on for hours where it negates the need for the synaptic activation of mGlu receptors during the induction of additional LTP. The switch can, however, be actively reset using low frequency stimulation (Bortolotto et al., 1994). Initially, we proposed a model that involved the activation of a single class of MCPG-sensitive mGlu receptor in both switch setting and the induction of LTP (Watkins and Collingridge, 1994). However, when we tested the effects of the potent broad spectrum (mGlu1e8) mGlu receptor antagonist LY341495 on both the setting of the molecular switch and the induction of LTP the results were not compatible with this simple explanation (Fitzjohn et al., 1998). We found that LY341495 blocked the setting of the molecular switch but not the induction of LTP per se. This suggests that there are two MCPG-sensitive mGlu receptors involved in the process; one that is required for the induction of LTP, and has a novel pharmacology, and one that is required for setting the molecular switch, and could be one of the known mGlu receptor subtypes. Of the eight mGlu receptors (i.e., mGlu1e8) identified by molecular cloning, mGlu5 and mGlu7 are highly expressed in the CA1 region of adult rat hippocampus (Shigemoto et al., 1997). Although expressed at lower levels, mGlu1 receptors are also involved in certain functions within the CA1 region of the hippocampus (Mannaioni et al., 2001; Thuault et al., 2002; Stoop et al., 2003). In the present study, we investigated whether any of these three mGlu receptor subtypes could mediate the setting of the molecular switch. We present evidence, from the use of both knockout animals and pharmacology, that the setting of the molecular switch requires the synaptic activation of mGlu5 receptors.

Some of these results have appeared in abstract form (Bortolotto and Collingridge, 2000).

2. Methods Experiments were performed on transverse hippocampal slices obtained from 3 to 8-week-old mice, using the knockout strains and genotyping as described previously (Conquet et al., 1994; Jia et al., 1998; Sansig et al., 2001). Mice were anaesthetised with halothane and decapitated. Slices (400 mm thick) were prepared by using a mechanical tissue slicer and maintained in a recording chamber at a temperature of 30 G 1  C. They were perfused continuously at 2 ml/min with a solution comprising the following: 124 mM NaCl, 3 mM KCl, 26 mM NaHCO3, 2 mM CaCl2, 1 mM MgSO4, 1.25 mM Na2HPO4, and 10 mM D-glucose. The solution was equilibrated with 95% O2/5% CO2. Two fully independent inputs were stimulated alternately, with 15 s separating the stimulation of each input. The stimulus intensity was set so as to obtain field EPSP slopes of approximately 50% of that at which a population spike was just detectable. All slices were naı¨ ve (i.e., had not received any prior experimental manipulation) such that, whenever tested, the induction of LTP was invariably blocked reversibly by 200 mM MCPG. Synaptic responses were displayed, and field EPSP slopes were measured and plotted online, using software written in-house (Anderson and Collingridge, 2001); available from http://www/ltp-program.com. Each experiment was conducted on a separate slice, each having been obtained from a separate mouse on a different day. The n values refer to the number of times a given result was obtained, which is the same as the number of times the experiment was performed. Data are presented as mean G s.e.m. LTP was defined as a statistically significant (P ! 0.05) stable increase in field EPSP slope, measured at least 60 min following a single tetanus (100 Hz, 1 s, test intensity). Potentiation is expressed as percentage baseline (i.e., 100% equals no change). All experiments were performed blind, with slices from knockout mice and their littermate controls being interleaved in a randomised manner. There were no apparent differences in the ability to obtain synaptic responses between the various groups of mice. The following protocol was adopted to determine whether the molecular switch had been set or not. Two inputs were stimulated alternately and input-specific LTP induced in one of them (this input was then designated as input 1). After a period of between 60 and 100 min the stimulus intensity was reduced to obtain a new baseline of responses of similar size to the original baseline. After 20 min, MCPG was applied and after a further 20 min tetani were delivered to each input in turn. Responses were followed for, at least, a further

Z.A. Bortolotto et al. / Neuropharmacology 49 (2005) 13e25

60 min to determine whether stable LTP was induced. If LTP was not induced in the presence of MCPG then a third tetanus was applied 60 min following the second to ensure that it was a reversible block of LTP induction (rather than a failure of the input to support additional LTP). MPEP, CHPG, LY367385 and MCPG were obtained from Tocris Cookson (Bristol, UK).

3. Results 3.1. The molecular switch in slices of mouse hippocampus Experiments to examine the involvement of mGlu1, mGlu5 and mGlu7 receptors in setting of the molecular switch were performed by replicating in mice the protocol developed using rats (Bortolotto et al., 1994). The three sets of wildtype mice yielded similar results and so the data sets have been pooled (Fig. 1). The results were the same as those previously described in rats (Bortolotto et al., 1994). Thus, the mGlu receptor antagonist MCPG invariably blocked, in a fully reversible manner, the induction of LTP in the experimentally naı¨ ve input. Simultaneously, MCPG failed to block the induction of additional LTP in the input in which LTP was pre-established, by a conditioning tetanus, before addition of MCPG (n Z 18). The level of potentiation in input 1, calculated 60 min following a tetanus, was 190 G 5% of baseline following the first tetanus (i.e., control) and 144 G 4% following the second tetanus (i.e., MCPG treated; conditioned). The corresponding change in input 2 was 97 G 3% baseline in response to the first tetanus (i.e., MCPG treated: naı¨ ve) and 140 G 5% baseline following the second tetanus (i.e., MCPG wash). The level of LTP following washout of MCPG in input 2 was significantly less than the level of LTP induced in input 1 prior to the application of MCPG (P ! 0.05).

15

mGlu1 receptors are not required for setting of the molecular switch. Thus, in all respects examined the mGlu1ÿ/ÿ mouse appears indistinguishable from its wildtype littermate or from control rats. 3.3. The molecular switch can be set in mGlu7ÿ/ÿ mice Results from the mGlu7ÿ/ÿ mice are illustrated in Fig. 3. In agreement with our earlier study, LTP was readily induced in the mGlu7ÿ/ÿ mouse (Bushell et al., 2002). The level of potentiation, 60 min following a tetanus was 196 G 18% (n Z 3), which was not significantly different from the value for the interleaved wildtype littermates of 188 G 9% (n Z 4; P O 0.05). In the non-conditioned input, MCPG fully blocked LTP, but not STP, in a reversible manner. In the conditioned input, MCPG was completely ineffective at blocking the induction of LTP. This suggests that mGlu7 receptors are not required for setting of the molecular switch. Thus, in this respect, the mGlu7ÿ/ÿ mouse appears similar to its wildtype littermate and also similar to control rats. 3.4. The molecular switch is absent in mGlu5ÿ/ÿ mice Results from the mGlu5ÿ/ÿ mice are illustrated in Fig. 4. LTP was readily induced in the mGlu5ÿ/ÿ mouse. However, the level of potentiation, 60 min following a tetanus (174 G 4%; n Z 6), was significantly less than that the interleaved wildtype littermates of 189 G 3% (n Z 4; P ! 0.05). In the non-conditioned input, MCPG fully blocked LTP, but not STP, in a reversible manner. This shows that the ability of MCPG to block the induction of LTP in naı¨ ve inputs is not due to its ability to inhibit mGlu5 receptors. Interestingly, in the conditioned input, MCPG was fully effective in blocking the induction of LTP. Thus, the conditioning tetanus was unable to set the molecular switch. This, therefore, suggests that mGlu5 receptors are necessary for setting of the molecular switch.

3.2. The molecular switch can be set in mGlu1ÿ/ÿ mice

3.5. The mGlu5 receptor antagonist MPEP prevents setting of the molecular switch

Results from the mGlu1ÿ/ÿ mice are illustrated in Fig. 2. In agreement with our earlier study, LTP was readily induced in the mGlu1ÿ/ÿ mouse. The level of potentiation, 60 min following a tetanus was 188 G 5% of control (n Z 6), which was not significantly different from the value for the interleaved wildtype littermates of 197 G 22% (n Z 4; P O 0.05). In the non-conditioned input, MCPG fully blocked LTP, but not STP, in a reversible manner. This shows that the ability of MCPG to block the induction of LTP in naı¨ ve inputs is not due to its ability to inhibit mGlu1 receptors. In the conditioned input, MCPG was completely ineffective at blocking the induction of LTP. This suggests that

Although these results demonstrate that mGlu5 receptors are needed to be able to set the molecular switch it is not possible to tell whether this is an acute effect or an indirect effect due to the absence of mGlu5 receptors throughout development. We, therefore, used the highly specific mGlu5 receptor antagonist MPEP (1 mM). These experiments were also performed blind using mGlu5C/C mice and mGlu5ÿ/ÿ mice. In the wildtype group, MPEP mimicked the effects of the knockout; thus, MPEP had no effect on the induction of LTP (control input, 188 G 7%; MPEP-treated input, 177 G 8%; P O 0.05) but prevented the setting of the molecular switch, such that MCPG invariably and

16

Z.A. Bortolotto et al. / Neuropharmacology 49 (2005) 13e25

A a

b

c

d

e

f

g

h

0.5 mV

wildtype

10 ms

3

MCPG b

Normalised EPSP slope

2

d

∇

a

c

1

input 1 0 3

2

h e

f

g

1 input 2 0

0

1

2

3

4

Time (h)

B

3

MCPG

Normalised EPSP slope

2

∇ 1 input 1 0 3 2

1 input 2 0

0

1

2

3

4

Time (h) Fig. 1. Setting the molecular switch in slices from wildtype mice: MCPG blocks the induction of LTP in experimentally naı¨ ve inputs but not in inputs that have experienced prior LTP. The graphs show a single representative experiment (A) and pooled data from 18 experiments (B). The traces are averages of four successive responses obtained at the times indicated (aeh) for the slice illustrated in (A). The following protocol was adopted throughout: in input 1, a tetanus (100 Hz, 1 s, baseline intensity) was delivered (arrow). Sixty minutes later, the baseline was reset by reducing the stimulus intensity (to exclude ceiling effects; open triangle). After 20 min of collecting a new baseline, MCPG (200 mM) was applied for 20 min. A second tetanus (100 Hz, 1 s, new baseline intensity) was delivered immediately prior to the washout of MCPG. A first tetanus (100 Hz, 1 s, baseline intensity) was also delivered to input 2 at this time. After 60 min of wash a second identical tetanus was delivered to ensure that the input was capable of exhibiting LTP. Note that MCPG invariably blocked LTP in input 2 (naı¨ ve input) but never affected LTP in input 1 (conditioned input).

reversibly blocked the induction of LTP in the conditioned input (n Z 6; Fig. 5A). For MPEP to be effective it needed to be present during the conditioning tetanus, since the application of MPEP following a tetanus did not enable MCPG to block the induction

of additional LTP. The addition of MPEP to the mGlu5ÿ/ÿ knockout had no additional effect than that of the knockout alone; results consistent with the only effect of MPEP in these experiments being to inhibit mGlu5 receptors (n Z 4; Fig. 5B). For example, MPEP

17

Z.A. Bortolotto et al. / Neuropharmacology 49 (2005) 13e25

A

mGlu1 -/3

MCPG

Normalised EPSP slope

2

∇

1

input 1 0 3 2

1 input 2 0

0

1

2

3

4

Time (h)

B

3

MCPG

Normalised EPSP slope

2

∇ 1 input 1

0 3

2

1

0

input 2 0

1

2

3

4

Time (h) Fig. 2. Setting the molecular switch in slices from mGlu1ÿ/ÿ mice. Data are presented as in Fig. 1 and show a single representative experiment (A) and pooled data from six experiments (B).

had no effect on the induction of LTP (control input, 171 G 5%; MPEP-treated input, 177 G 4%; P O 0.05). 3.6. Pharmacological activation of mGlu5 receptors sets the molecular switch The findings from both the knockout animals and pharmacological experiments implicate mGlu5 receptors as an essential trigger for the molecular switch. If activation of mGlu5 receptors is sufficient for this process then it should be possible to set the molecular switch by selective pharmacological activation of mGlu5 receptors. To test this possibility we applied CHPG (1 mM), which activates mGlu5 but not mGlu1 receptors (Doherty et al., 1997). In each slice tested, transient application of CHPG resulted in MCPG failing to block

the induction of LTP when tested 40 or 140 min later (Fig. 6). The effect of CHPG was via activation of mGlu5 receptors since MCPG was fully effective when the equivalent protocol was performed using the one available mGlu5ÿ/ÿ mouse (data not shown). 3.7. LTP can be generated in the mGlu1 and mGlu5 double knockout mouse We considered the possibility that mGlu1 and mGlu5 receptors may compensate for one another, such that in the mGlu1 knockout the mGlu5 receptor mediates LTP and vice versa. To test this possibility we bred the mGlu1 and mGlu5 knockouts to generate mGlu1 and mGlu5ÿ/ÿ mice. These animals invariably exhibited LTP. The level of LTP was 173 G 15% (n Z 5) which

18

Z.A. Bortolotto et al. / Neuropharmacology 49 (2005) 13e25

A

mGlu7-/3

MCPG

Normalised EPSP slope

2

∇

1

input 1 0 3

2

1 input 2 0

0

1

2

3

4

Time (h)

B

MCPG

3

Normalised EPSP slope

2

∇

1

input 1 0 3

2 1 input 2

0

0

1

2

3

4

Time (h) Fig. 3. Setting the molecular switch in slices from mGlu7ÿ/ÿ mice. Data are presented as in Fig. 1 and show a single representative experiment (A) and pooled data from three experiments (B).

was significantly less than that in the interleaved wildtype controls of 198 G 3% (n Z 4; P ! 0.05; Fig. 7) but similar to that observed in the mGlu5ÿ/ÿ mouse. Interestingly, MCPG was still able to inhibit the induction of LTP in naı¨ ve inputs, in a fully reversible manner. Thus, the ability of MCPG to inhibit LTP in wildtype animals is unlikely to be due to an action on either mGlu1 or mGlu5 receptors. Interestingly, we were not able to induce more than one episode of LTP in any given input. To determine whether this was a developmental effect of the double knockout or related to the need to activate either mGlu1 or mGlu5 receptors to be able to induce multiple episodes of LTP, we performed a new set of pharmacological experiments. Using wildtype mice, we co-applied the mGlu1 receptor antagonist (S )-(C)-2methyl-4-carboxyphenylglycine (LY367385; 30 mM) and

MPEP (30 mM). This treatment failed to block LTP, confirming that LTP can be induced without the need to activate mGlu1 or mGlu5 receptors. In contrast to the double knockout, however, a second tetanus in the presence of these antagonists invariably induced additional LTP (Fig. 8). Thus the failure of the double knockout to support more than one episode of LTP relates to some developmental change.

4. Discussion 4.1. The role of mGlu receptors in the induction of LTP at CA1 synapses The present finding that LTP was unaffected in mGlu1ÿ/ÿ mice is consistent with our previous report

19

Z.A. Bortolotto et al. / Neuropharmacology 49 (2005) 13e25

mGlu5-/a

b

c

f

g

h

d

e

0.5 mV

A

10 ms

3

MCPG

Normalised EPSP slope

2

b

∇

a

1

e c

d

input 1

0 3

2

h

0

g

f

1

input 2 0

1

2

3

4

Time (h)

B

3

MCPG

Normalised EPSP slope

2

∇

1

input 1

0 3

2 1 0

input 2

0

1

2

3

4

Time (h) Fig. 4. The molecular switch is absent in slices from mGlu5ÿ/ÿ mice. Data are presented similarly to Fig. 1 and show a single representative experiment (A) and pooled data from six experiments (B). The traces are averages of four successive responses obtained at the times indicated (aeh) for the slice illustrated in (A). In these experiments a third tetanus was delivered to input 1 to determine whether MCPG had blocked the induction of LTP or whether the slice was incapable of exhibiting a second LTP. In all cases there was a reversible effect of MCPG.

(Conquet et al., 1994) although a deficit was reported in a different mGlu1 knockout (Aiba et al., 1994). Given that LTP is resistant to the actions of the broad spectrum antagonist LY341495, applied in much higher concentrations than the Ki for inhibition of mGlu1 receptors, (Fitzjohn et al., 1998) and given that mGlu1 receptor protein is expressed at low levels in stratum radiatum (Shigemoto et al., 1997) we suspect that the

reported deficit is an indirect consequence of the knockout of this receptor. In particular, mGlu1ÿ/ÿ mice have a motor deficit of cerebellar origin and appear more agitated than their wildtype littermates. They also appear to be more prone to stress, which itself can lead to LTP deficits under certain circumstances (McEwen, 1994). Of course, our results do not exclude a role for mGlu1 receptors under different experimental conditions

20

Z.A. Bortolotto et al. / Neuropharmacology 49 (2005) 13e25

A

wildtype 3

MPEP

MCPG

Normalised EPSP slope

2

∇

1

input 1 0 3

2

∇

1

input 2 0

0

1

2

3

4

Time (h)

B

mGlu5-/3

MPEP

MCPG

Normalised EPSP slope

2

∇

1

input 1 0 3

2

∇

1

input 2 0

0

1

2

3

4

Time (h) Fig. 5. MPEP blocks the setting of the molecular switch. Data are presented similarly to Fig. 1 except that MPEP (1 mM) was applied following the induction of LTP in input 1 and a tetanus delivered to input 2 after 20 min of perfusion with MPEP. Sixty minutes later, the baseline was reset by reducing the stimulus intensity (to exclude ceiling effects; open triangle). The graphs are pooled data from six wildtype slices (A) and four mGlu5ÿ/ÿ (B) slices.

or in LTP in other brain regions. For example, recently it has been reported that intraventricular application of the mGlu1 receptor selective antagonist LY367385 inhibits LTP in the dentate gyrus of freely moving rats (Naie and Manahan-Vaughan, 2005). Our results do not agree with the conclusion of one pharmacological study that mGlu1 receptors are required for the induction of LTP at this synapse (Francesconi et al., 2004). These authors found that LY367385 blocked LTP induced by a single tetanus; however, they used a concentration of 200 mM which is well in excess of the IC50 value for inhibition of mGlu1 receptors (Schoepp et al., 1999) and they did not demonstrate reversibility of the block. Furthermore, the antagonist was ineffective when three

tetani were delivered suggesting that their LY367385sensitive component was masked under these conditions. The observation that the stable component of LTP was also unaffected in mGlu7ÿ/ÿ mice is consistent with our previous report (Bushell et al., 2002). These mice did show some qualitative differences from the other mice tested; the level of LTP induced in the presence of MCPG in the conditioned input was larger than in the other groups. Conversely, the duration of STP observed in the presence of MCPG in the naı¨ ve input was less than in wildtype, mGlu1ÿ/ÿ and mGlu5ÿ/ÿ mice, consistent with a role of mGlu7 receptors in STP (Bushell et al., 2002). However, the small number of

21

Z.A. Bortolotto et al. / Neuropharmacology 49 (2005) 13e25

A

wildtype 3

CHPG

MCPG

MCPG

Normalised EPSP slope

2

∇

1

input 1 0 3

2

1 input 2 0

0

1

3

2

4

Time (h)

B

3

MCPG

MCPG

CHPG

Normalised EPSP slope

2

∇

1

input 1 0 3

2

1 input 2 0

0

1

2

3

4

Time (h) Fig. 6. Pharmacological activation of mGlu5 receptors sets the molecular switch. The graph shows a single example (A) and pooled data from four slices (B), using wildtype mice. A tetanus was delivered in the presence of MCPG to input 1 and input 2 following washout of CHPG (1 mM) for 40 and 140 min, respectively. Note that LTP was invariably induced.

available mice precluded us from drawing any firm conclusions. The finding that LTP of AMPA receptor-mediated synaptic transmission is slightly, but significantly, affected in mGlu5ÿ/ÿ mice is consistent with a previous report (Lu et al., 1997). However, the finding that MPEP did not significantly affect LTP suggests that this is a developmental consequence of the absence of mGlu5 receptors. Therefore, our experiments argue against an acute role for mGlu5 receptors in LTP, at least under the conditions of these experiments (e.g., in response to a single tetanus, delivered at 100 Hz). In addition, we previously found that a competitive mGlu5 receptor antagonist, LY344545, also had no effect on LTP at concentrations that were selective for this subtype of receptor (Doherty et al., 2000). In contrast, it has been reported that MPEP blocks

the induction of LTP, but not short-term potentiation, at CA1 synapses (Sanna et al., 2002). The reason for this difference is not known. Our results also cannot preclude a role of mGlu5 receptors in the induction of LTP in other brain regions or under different experimental conditions. For example, it has been reported that intraventricular application of MPEP inhibits LTP in the dentate gyrus of freely moving rats (Balschun and Wetzel, 2002; Naie and Manahan-Vaughan, 2004). In addition, it has been shown that MPEP reduces LTP at corticostriatal fibres by approximately 50%. Interestingly, in this study the mGlu1 receptor selective antagonist LY367385 also decreases this LTP by approximately 50% and the combined inhibition of mGlu1 and mGlu5 receptors completely eliminated the induction of LTP (Gubellini et al., 2003).

22

Z.A. Bortolotto et al. / Neuropharmacology 49 (2005) 13e25

A

mGlu1&5-/3

MCPG

Normalised EPSP slope

2

∇

1

input 1

0 3 2 1

input 2 0

0

1

2

3

4

Time (h)

B

3

MCPG

Normalised EPSP slope

2

∇

1

∇

input 1

0 3 2

1

input 2 0

0

1

2

3

4

Time (h) Fig. 7. LTP can be induced in the mGlu1 and mGlu5 double knockout mouse. Data are presented similarly to Fig. 1 and show a single representative experiment (A) and pooled data from five experiments (B). Note that, unlike in the mGlu5ÿ/ÿ mouse, a tetanus delivered after washout of MCPG did not induce LTP (input 1). Note also that MCPG reversibly inhibited the induction of LTP (input 2).

We considered the possibility that mGlu1 and mGlu5 receptors could compensate for one another, given that both can mediate similar effects on CA1 neurons (Mannaioni et al., 2001; Thuault et al., 2002; Stoop et al., 2003). However, we consistently observed LTP in the double knockout mouse, and LTP in these mice was sensitive to MCPG. The finding that LY341495 applied at a concentration (100 mM) sufficient to inhibit both mGlu1 and mGlu5 (as well as all other known mGlu receptor subtypes) failed to inhibit the induction of LTP (Fitzjohn et al., 1998) also suggests that neither the mGlu1 nor the mGlu5 receptor is required for the induction of LTP per se. This is further supported by the present observations that co-application of LY367385 and MPEP failed to block the induction of LTP.

Therefore, the identity of the receptor that is blocked by MCPG and needs to be activated for the induction of LTP at CA1 synapses, in the absence of the setting of the molecular switch, remains a mystery. 4.2. The role of mGlu receptors in setting of the molecular switch Based on the analysis of knockout mice we have obtained no evidence that either the mGlu1 or mGlu7 receptor is involved in the molecular switch. However, whilst we find no evidence for a role of mGlu5 receptors in LTP per se, we have found that these receptors are the trigger for the molecular switch. Thus, MCPG was able to inhibit the induction of LTP in conditioned as well as

23

Z.A. Bortolotto et al. / Neuropharmacology 49 (2005) 13e25

A

wildtype 3

LY367385 MPEP

Normalized EPSP slope

2

1 input 1 0 3

2

1 input 2 0

0

1

2

3

Time (h)

B

3

LY367385 MPEP

Normalized EPSP slope

2

1

input 1 0 3

2

1

input 2 0

0

1

2

3

Time (h) Fig. 8. Blockade of mGlu1 and mGlu5 receptors fails to inhibit LTP. The graph shows a single example (A) and pooled data from four slices (B), using wildtype mice. A combination of LY367385 (30 mM) and MPEP (30 mM) were applied immediately following the tetanus delivered to input 1. After 20 min a tetanus was delivered to input 2 in the presence of these antagonists. Subsequently, the baselines were reset, the antagonists re-applied and tetani redelivered to both inputs. In all cases, LTP was induced in response to both tetani. (There is a break in the X-axis of the pooled data since in one experiment an additional tetanus was delivered to each input in the presence of MCPG to test the state of the molecular switch; the switch was set in input 1 but not input 2 as expected; data not shown.)

naı¨ ve inputs in either mGlu5ÿ/ÿ mice or MPEP-treated slices (and both). The finding that similar results were obtained in both knockouts and pharmacological experiments is compelling evidence for a role of mGlu5 receptors in this process. Reliance on the knockout alone could be misleading since the deficit might be developmental in origin or due to a background effect. Similarly, reliance on a single pharmacological agent alone could lead to erroneous conclusions since its activity against all potential targets is unknown. The finding that the mGlu5 receptor selective agonist CHPG sets the molecular switch supports the role of mGlu5 receptors. Indeed, it shows that activation of mGlu5

receptors is sufficient to set the switch. Of course, none of these data exclude a role for other receptors linked to the same signalling cascades but activated by other neurotransmitters in the molecular switch. A consistent observation was that the level of control LTP in input 1 was greater than the level of LTP in input 2 following the washout of MCPG. It is unlikely that this was due to incomplete washout of MCPG, given that 1 h was allowed to washout this low affinity antagonist. This differential effect was also observed in each of the knockout strains. It is possible that this represents another form of metaplasticity induced by the tetanus that was delivered in the presence of MCPG.

24

Z.A. Bortolotto et al. / Neuropharmacology 49 (2005) 13e25

One possibility is that it is due to the activation of NMDA receptors during the tetanus, which can result in inhibition of subsequent LTP (Coan et al., 1989; Huang et al., 1992). Further experiments are required to establish the basis of this effect. 4.3. The mechanism of the molecular switch Previously, we have shown that the setting of the molecular switch can be prevented by a CaMKII inhibitor, KN-62, applied at a concentration subthreshold for the inhibition of LTP (Bortolotto and Collingridge, 1998). CaMKII is an attractive candidate molecule for metaplasticity since following its activation by Ca2C it can undergo autophosphorylation and enter a persistently active state (Lisman, 1994). Although this autophosphorylation has been implicated in the maintenance of LTP a role in metaplasticity is equally feasible (Abraham and Tate, 1997). We have also found that the molecular switch is blocked by various inhibitors of PKC (Bortolotto and Collingridge, 2000). Interestingly, other forms of mGlu receptor-mediated metaplasticity also involve activation of PKC (Rush et al., 2002; Wu et al., 2004). PKC has also been implicated in metaplasticity on theoretical grounds (Abraham and Tate, 1997). Another form of metaplasticity with many features in common with the molecular switch has also been described at CA1 synapses (Cohen et al., 1998). Transient activation of group I mGlu receptors with DHPG leads to the facilitation of LTP induced by a subsequent weak tetanus. This facilitated component of LTP might equate with the MCPGinsensitive LTP observed in the present study. The priming effect of DHPG involves activation of phospholipase C (Cohen et al., 1998) and protein synthesis (Raymond et al., 2000). One possible mechanism for metaplasticity, therefore, is that activation of mGlu5 receptors leads, via phospholipase C, to the activation of PKC and CaMKII, the latter possibly activated as a consequence of release of Ca2C from intracellular stores. Since both CaMKII and PKC inhibitors are effective in blocking the setting of the molecular switch fully then it can be assumed that both PKC and CaMKII are required for this process to occur. These kinases may, therefore, act in synergy to initiate protein synthesis. Such a mechanism would explain the lack of requirement for mGlu receptor activation for additional LTP, since the necessary protein synthetic machinery would already be engaged by the initial receptor activation. Further experiments are required to establish the role of the molecular switch in mnemonic processing in the brain. However, it is interesting to note that LTP in the dentate gyrus in vivo is associated with an increase in expression of mGlu5 receptors (Manahan-Vaughan et al., 2003). If this increase is of functional receptors,

and assuming that mGlu5 receptors play a similar role in regulating LTP in the dentate gyrus, then their regulation by LTP would constitute a plasticity of metaplasticity e supermetaplasticity.

Acknowledgements This work was supported by the MRC. We are grateful to Guijin Lu for genotyping.

References Abraham, W.C., Bear, M.F., 1996. Metaplasticity: the plasticity of synaptic plasticity. Trends Neurosci. 19 (4), 126e130. Abraham, W.C., Tate, W.P., 1997. Metaplasticity: a new vista across the field of synaptic plasticity. Prog. Neurobiol. 52 (4), 303e323. Aiba, A., Chen, C., et al., 1994. Reduced hippocampal long-term potentiation and context-specific deficit in associative learning in mGluR1 mutant mice. Cell 79 (2), 365e375. Anderson, W.W., Collingridge, G.L., 2001. The LTP program: a data acquisition program for on-line analysis of long-term potentiation and other synaptic events. J. Neurosci. Methods 108 (1), 71e83. Anwyl, R., 1999. Metabotropic glutamate receptors: electrophysiological properties and role in plasticity. Brain Res. Brain Res. Rev. 29 (1), 83e120. Balschun, D., Wetzel, W., 2002. Inhibition of mGluR5 blocks hippocampal LTP in vivo and spatial learning in rats. Pharmacol. Biochem. Behav. 73 (2), 375e380. Bashir, Z.I., Bortolotto, Z.A., et al., 1993. Induction of LTP in the hippocampus needs synaptic activation of glutamate metabotropic receptors. Nature 363 (6427), 347e350. Bliss, T.V., Collingridge, G.L., 1993. A synaptic model of memory: long-term potentiation in the hippocampus. Nature 361 (6407), 31e39. Bortolotto, Z.A., Bashir, Z.I., et al., 1994. A molecular switch activated by metabotropic glutamate receptors regulates induction of long-term potentiation. Nature 368 (6473), 740e743. Bortolotto, Z.A., Collingridge, G.L., 1998. Involvement of calcium/ calmodulin-dependent protein kinases in the setting of a molecular switch involved in hippocampal LTP. Neuropharmacology 37 (4e5), 535e544. Bortolotto, Z.A., Collingridge, G.L., 2000. A role for protein kinase C in a form of metaplasticity that regulates the induction of long-term potentiation at CA1 synapses of the adult rat hippocampus. Eur. J. Neurosci. 12 (11), 4055e4062. Bushell, T.J., Sansig, G., et al., 2002. Altered short-term synaptic plasticity in mice lacking the metabotropic glutamate receptor mGlu7. Sci. World J. 2, 730e737. Coan, E.J., Irving, A.J., et al., 1989. Low-frequency activation of the NMDA receptor system can prevent the induction of LTP. Neurosci. Lett. 105 (1e2), 205e210. Cohen, A.S., Abraham, W.C., 1996. Facilitation of long-term potentiation by prior activation of metabotropic glutamate receptors. J. Neurophysiol. 76 (2), 953e962. Cohen, A.S., Raymond, C.R., et al., 1998. Priming of long-term potentiation induced by activation of metabotropic glutamate receptors coupled to phospholipase C. Hippocampus 8 (2), 160e 170. Conquet, F., Bashir, Z.I., et al., 1994. Motor deficit and impairment of synaptic plasticity in mice lacking mGluR1. Nature 372 (6503), 237e243.

Z.A. Bortolotto et al. / Neuropharmacology 49 (2005) 13e25 Doherty, A.J., Palmer, M.J., et al., 2000. A novel, competitive mGlu(5) receptor antagonist (LY344545) blocks DHPG-induced potentiation of NMDA responses but not the induction of LTP in rat hippocampal slices. Br. J. Pharmacol. 131 (2), 239e244. Doherty, A.J., Palmer, M.J., et al., 1997. (RS )-2-chloro-5-hydroxyphenylglycine (CHPG) activates mGlu5, but no mGlu1, receptors expressed in CHO cells and potentiates NMDA responses in the hippocampus. Neuropharmacology 36 (2), 265e267. Fitzjohn, S.M., Bortolotto, Z.A., et al., 1998. The potent mGlu receptor antagonist LY341495 identifies roles for both cloned and novel mGlu receptors in hippocampal synaptic plasticity. Neuropharmacology 37 (12), 1445e1458. Francesconi, W., Cammalleri, M., et al., 2004. The metabotropic glutamate receptor 5 is necessary for late-phase long-term potentiation in the hippocampal CA1 region. Brain Res. 1022 (1e2), 12e18. Gubellini, P., Saulle, E., et al., 2003. Corticostriatal LTP requires combined mGluR1 and mGluR5 activation. Neuropharmacology 44 (1), 8e16. Huang, Y.Y., Colino, A., et al., 1992. The influence of prior synaptic activity on the induction of long-term potentiation. Science 255 (5045), 730e733. Jia, Z., Lu, Y., et al., 1998. Selective abolition of the NMDA component of long-term potentiation in mice lacking mGluR5. Learn. Mem. 5 (4e5), 331e343. Lisman, J., 1994. The CaM kinase II hypothesis for the storage of synaptic memory. Trends Neurosci. 17 (10), 406e412. Lu, Y.M., Jia, Z., et al., 1997. Mice lacking metabotropic glutamate receptor 5 show impaired learning and reduced CA1 long-term potentiation (LTP) but normal CA3 LTP. J. Neurosci. 17 (13), 5196e5205. Manahan-Vaughan, D., Ngomba, R.T., et al., 2003. An increased expression of the mGlu5 receptor protein following LTP induction at the perforant pathedentate gyrus synapse in freely moving rats. Neuropharmacology 44 (1), 17e25. Mannaioni, G., Marino, M.J., et al., 2001. Metabotropic glutamate receptors 1 and 5 differentially regulate CA1 pyramidal cell function. J. Neurosci. 21 (16), 5925e5934. McEwen, B.S., 1994. Corticosteroids and hippocampal plasticity. Ann. N.Y. Acad. Sci. 746, 134e142 discussion 142e144, 178e179. Naie, K., Manahan-Vaughan, D., 2004. Regulation by metabotropic glutamate receptor 5 of LTP in the dentate gyrus of freely moving rats: relevance for learning and memory formation. Cereb. Cortex 14 (2), 189e198.

25

Naie, K., Manahan-Vaughan, D., 2005. Pharmacological antagonism of metabotropic glutamate receptor 1 regulates long-term potentiation and spatial reference memory in the dentate gyrus of freely moving rats via N-methyl-D-aspartate and metabotropic glutamate receptor-dependent mechanisms. Eur. J. Neurosci. 21 (2), 411e421. Raymond, C.R., Thompson, V.L., et al., 2000. Metabotropic glutamate receptors trigger homosynaptic protein synthesis to prolong long-term potentiation. J. Neurosci. 20 (3), 969e976. Rush, A.M., Wu, J., et al., 2002. Group I metabotropic glutamate receptor (mGluR)-dependent long-term depression mediated via p38 mitogen-activated protein kinase is inhibited by previous highfrequency stimulation and activation of mGluRs and protein kinase C in the rat dentate gyrus in vitro. J. Neurosci. 22 (14), 6121e6128. Sanna, P.P., Cammalleri, M., et al., 2002. Phosphatidylinositol 3kinase is required for the expression but not for the induction or the maintenance of long-term potentiation in the hippocampal CA1 region. J. Neurosci. 22 (9), 3359e3365. Sansig, G., Bushell, T.J., et al., 2001. Increased seizure susceptibility in mice lacking metabotropic glutamate receptor 7. J. Neurosci. 21 (22), 8734e8745. Schoepp, D.D., Jane, D.E., et al., 1999. Pharmacological agents acting at subtypes of metabotropic glutamate receptors. Neuropharmacology 38 (10), 1431e1476. Shigemoto, R., Kinoshita, A., et al., 1997. Differential presynaptic localization of metabotropic glutamate receptor subtypes in the rat hippocampus. J. Neurosci. 17 (19), 7503e7522. Stoop, R., Conquet, F., et al., 2003. Determination of group I metabotropic glutamate receptor subtypes involved in the frequency of epileptiform activity in vitro using mGluR1 and mGluR5 mutant mice. Neuropharmacology 44 (2), 157e162. Thuault, S.J., Davies, C.H., et al., 2002. Group I mGluRs modulate the pattern of non-synaptic epileptiform activity in the hippocampus. Neuropharmacology 43 (2), 141e146. Watkins, J., Collingridge, G., 1994. Phenylglycine derivatives as antagonists of metabotropic glutamate receptors. Trends Pharmacol. Sci. 15 (9), 333e342. Wu, J., Rowan, M.J., et al., 2004. Synaptically stimulated induction of group I metabotropic glutamate receptor-dependent long-term depression and depotentiation is inhibited by prior activation of metabotropic glutamate receptors and protein kinase C. Neuroscience 123 (2), 507e514.