INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Many studies suggest that neural plasticity in the auditory cortex (AC) plays important roles in learning. Tuning curves of neurons and the response map in the AC change following the auditory discrimination test (Edeline et al. 1993; Kilgard and Merzenich 1998; Ohl and Scheich 1996; Recanzone et al. 1993). Synaptic connections are potentiated by correlated activities between AC neurons (Ahissar et al. 1992), and receptive-field plasticity is induced by Hebbian rules (Cruikshank and Weinberger 1996). Discrimination ability between amplitude-modulated tones in rats is enhanced by exposure to sound stimuli, and the enhancement is dependent on N-methyl-D-aspartate receptors (NMDARs) in the AC (Sakai et al. 1999). The cellular mechanisms for the neural plasticity may be explained by long-term potentiation (LTP) in the AC (Kudoh and Shibuki 1994, 1996, 1997; Seki et al. 1999, 2001). However, long-term depression (LTD), as well as LTP, is suggested to play important roles in the neocortex (Artola et al. 1990; Haruta et al. 1994; Kimura et al. 1990; Kirkwood et al. 1993; Sermasi et al. 1999). Because LTD has not been studied in the AC, we investigated LTD in slices obtained from the AC.

LTP and LTD are induced by a marked [Ca²⁺]_i increase and a moderate [Ca²⁺]_i increase, respectively (Lisman 1989). The importance of a moderate [Ca²⁺]_i increase for the induction of LTD is confirmed in the hippocampus (Neveu and Zucker 1996) and the visual cortex (Brocher et al. 1992; Kimura et al. 1990; Yasuda and Tsumoto 1996). However, the mechanisms for the [Ca²⁺]_i increase are controversial. In the cerebral cortex, the LTD induced by low-frequency stimulation (LFS) is dependent on NMDARs (Kirkwood and Bear 1994; Sawtell et al. 1999). It is also dependent on mGluRs (Haruta et al. 1994) and induced by application of an agonist of mGluRs (Kato 1993). Therefore multiple mechanisms for increasing [Ca²⁺]_i appear to play a role in the induction of LTD, and the mechanisms for inducing LTD or LTP may be different among various types of synapses.

Glutamatergic inputs to supragranular pyramidal neurons in the AC mainly originate from MGB neurons or other cortical pyramidal neurons (Kelly and Wong 1981; Vaughan 1983). The mechanisms for increasing [Ca²⁺]_i to induce LTD may be different between the cortical synapses of heterogeneous origin. In this study, therefore we compared the mechanisms for inducing LTD between thalamo-cortical synapses and cortico-cortical synapses.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The experiments were performed according to the guidelines of Niigata University and had the approval of the ethics committee of Niigata University.

Slice preparations

Slice preparations were obtained from Wistar rats of both genders (4-7 wk old). The rats were deeply anesthetized with ether and were immersed in ice-cold water except for the nose for 3 min to reduce the brain temperature. They were decapitated, and a neocortical tissue block including the AC, determined as area 41 (Krieg 1946), was dissected out. Frontal AC slices (thickness, 400 µm) were prepared from the block with a microslicer (DTK-2000, Dosaka, Osaka, Japan) in an ice-cold artificial medium bubbled with 95% O₂-5% CO₂. The composition of the medium (in mM) was 124 NaCl, 5 KCl, 1.24 NaH₂PO₄, 1.3 MgSO₄, 2.4 CaCl₂, 26 NaHCO₃, and 10 glucose. Slices intersecting with a horizontal plane at various angles were also prepared to stimulate synaptic inputs originating from the MGB. The horizontal orientation was estimated from the dorsal surface of the cerebrum (Paxinos and Watson 1986). After incubating the slices at room temperature (21-23°C) for >1 h, recording was performed at 30°C in a small submerged-type chamber (0.3 ml volume) continuously perfused with the oxygenated medium at a flow rate of 1 ml/min.

Field potential recording

Field potentials were recorded via an electrolytically polished silver wire (200 µm diam) insulated by polyvinyl chloride except for 60 µm from the tip (<50 k). A platinum wire (125 µm diam), sharpened and insulated by polyvinyl chloride except for 100 µm from the tip, was used as a stimulation electrode (10-15 k). Focal stimulation was applied to layer IV below the recording sites or layer III 300 µm apart from the recording sites. Biphasic stimulus current pulses (duration of each pulse: 100 µs) were generated with a stimulator (SEN-7203, Nihon Kohden, Tokyo, Japan) to reduce stimulus artifacts. Field potentials were elicited every 1 min or two consecutive traces elicited every 30 s were averaged. Signals were amplified 10-fold with a circuit using an operational amplifier (OPA128LM, Burr-Brown, Tucson, AZ). The output passed through a band-pass filter (0.2 Hz to 10 kHz) was stored in a computer (PC-9801BA2, NEC, Tokyo, Japan) at a sampling rate of 100 kHz via an analog-digital converter (ADXM-98A, Canopus, Kobe, Japan) for later analysis. The BASIC programs for the recording and later analysis were developed using the software library supplied by Canopus. The magnitude of the postsynaptic field potentials was measured from the peak amplitude of the second negativity (Kudoh and Shibuki 1996). Only when the maximal amplitude of the postsynaptic potentials was larger than 1 mV, the slice was used for the LTD recording. The intensity of the test stimuli was adjusted so that 90-80% of the maximal responses were recorded. The stimulus intensity was 50-200 µA for stimulating layer III or IV, and 0.5-1 mA for stimulating the white matter 600-1,000 µm apart from the recording sites. To evoke LTD, LFS (1 Hz, 900 pulses) was applied at the same intensity of the test stimulation, unless otherwise specified. The magnitude of the LTD was estimated as the relative amplitude of the average field potentials recorded 38-40 min after LFS normalized by those recorded 1-3 min before LFS. LTD and the effects of various drugs were statistically evaluated by Mann-Whitney U test.

MGB lesion

The MGB was destroyed in 25 rats 1 wk before slicing. Rats were anesthetized with fluothane (1.5% in O₂ gas). After the skin covering the skull was disinfected, a small hole was opened in the skull above the MGB. An electrolytically polished steel electrode (300 µm diam, <10 k), which was insulated with polyvinyl chloride except for 300 µm from the tip, was inserted stereotaxically into the MGB. The position of the tip was located 5.4 mm caudal and 3.4 mm lateral from the Bregma and 5.4-6.0 mm deep from the surface of the brain (Paxinos and Watson 1986). Characteristic biphasic potentials (Hall and Borbely 1970) were recorded in the MGB after click stimulation. At the depth where the maximal positive potentials (>50 µV) were recorded, a positive current of 5 mA was applied for 10 s three times via the electrode to destroy the MGB. Fradiomycin was locally applied to the incision, and the incised skin was sutured.

We could observe the lesion in the MGB in the brain block during preparation of slices. However, trimming of the brain block frequently distorted or destroyed the structures around the MGB lesion. Therefore we examined the extent of the lesion in nine rats not used for slices. A part of the MBG was not destroyed in one rat, in which the field potentials in the MGB were <50 µV. In the other eight of the nine rats, the potentials >50 µV were recorded in the MGB. In these rats, the lesion covered the MBG (Fig. 5C). Therefore it was confirmed that the conditions used for MGB lesions in slice experiments were sufficient to destroy most of the MGB.

Perforated-patch recording from pyramidal neurons

The LTD of EPSPs was measured in supragranular pyramidal neurons using a blind slice patch technique (Kudoh and Shibuki 1997, Seki et al. 2001). A glass micropipette was filled with medium containing (in mM) 92 K₂SO₄, 31 KCl, 10 HEPES, and 1 MgCl₂ adjusted to pH 7.4 with KOH. Amphotericine B dissolved in dimethyl sulfoxide (10 mg/ml) was added to the patch medium (final concentration, 50 µg/ml). The tip was filled with an amphotericine B-free medium to facilitate the sealing between the pipette and cell membrane. The resistance of the electrode was 5-10 M. The electrode was inserted into the supragranular layers, and a positive pressure of 10-30 mmHg was applied in the pipette during insertion. After a sudden increase in the access resistance of the electrode, the pressure on the electrode was removed. In successful cases, the access resistance and resting membrane potential were continuously decreased for ~10 min before these parameters were stabilized. The intracellular potentials were recorded through an amplifier for intracellular recording (MEZ-8300, Nihon Kohden) and were processed by the same system used for field potential recordings. Neurons showing a resting membrane potential more negative than 60 mV [69 ± 1 (SE) mV, n = 36] and a membrane resistance >30 M (64 ± 5 M) were selected as in our previous studies (Kudoh and Shibuki 1997, Seki et al. 2001). If the series resistance was clearly changed by >10% during recording, the findings were discarded. Pyramidal neurons were identified by their antidromic spikes following white matter stimulation (<1 mA) (Kudoh and Shibuki 1996). The LTD was estimated as the changes in the initial slopes of excitatory postsynaptic potentials (EPSPs) because this parameter was not very likely to be affected by subsequent inhibitory synaptic potentials. The intensity of the test stimuli and LFS, applied at layer IV, was adjusted between 90 and 300 µA, so that the initial slope was ~3 mV/ms. Test stimuli were applied every 30 s, and the average slopes of EPSPs were measured from two consecutive traces.

Drugs

D-()-2-amino-5-phosphonovalerate (APV) and (RS)-1-aminoindan-1,5- dicarboxylic acid (AIDA) were obtained from Tocris Cookson (Bristol, UK). Nifedipine and (RS)--cyclopropyl-4-phosphonophenylglycine (CPPG) were purchased from Wako (Osaka, Japan). (+)--methyl-4-carboxyphenylglycine (MCPG) and L(+)-2-amino-3-phosphonopropionic acid (L-AP3) were obtained from Research Biochemicals (Natick, MA). ()-Bicuculline methiodide was purchased from Sigma (St. Louis, MO). These drugs were added to the perfusing medium for applying to the slices. Fradiomycin and amphotericin B were purchased from Nippon Kayaku (Tokyo) and Wako, respectively.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

LTD in supragranular field potentials

Supragranular field potentials following layer IV stimulation had early and late negative components, and only the late one was sensitive to 6-cyano-7-nitroquinoxalene-2,3-dione (CNQX, Fig. 1, A and B). As the stimulus intensity was increased, the latency of the early component was constant while that of the late component was shortened (Fig. 1C). These findings, similar to those of our previous study (Kudoh and Shibuki 1996), suggest that the late component represents postsynaptic population spikes in pyramidal neurons.

View larger version (18K): [in this window] [in a new window]   Fig. 1. A: scheme of experiments. Stimulation was applied at layer IV and responses were recorded in the supragranular layers. B: effects of 10 µM 6-cyano-7-nitroquinoxalene-2,3-dione (CNQX) on field potentials. Traces recorded before and during application of CNQX are superimposed. C: field potentials evoked by the stimulation of various intensities (25-120 µA).

We used LFS of layer IV for inducing LTD in AC slices because it is used for inducing LTD in the visual cortex (Artola et al. 1990; Kirkwood et al. 1993; Haruta et al. 1994). The postsynaptic potentials were depressed by LFS, while the early component was unchanged (Fig. 2, A and C). The amplitude of the postsynaptic potentials immediately after LFS was 37 ± 6% (n = 8). The depression gradually recovered during 15-20 min and became stable thereafter. The amplitude of LTD, estimated as the relative amplitude of postsynaptic potentials 38-40 min after LFS, was 74 ± 3%, (n = 8, Fig. 2C). To test the contribution of inhibitory synapses in LTD, depression was induced in the medium containing 1 µM bicuculline, an antagonist of GABA_A receptors. Although the amplitude of field potentials before LFS was increased by bicuculline to 151 ± 12% of the control, the amplitude of LTD (78 ± 7%, n = 5) was comparable with that in the normal medium.

View larger version (24K): [in this window] [in a new window]   Fig. 2. Long-term depression (LTD) in supragranular field potentials induced by low-frequency stimulation (LFS) of layer IV. A: sample traces of field potentials before and 40 min after (*) LFS. B: traces recorded in the presence of 50 µM ()-2-amino-5-phosphonovalerate (APV). C: LTD in control medium or in the presence of APV. D: traces before and after (*) LFS during application of 500 µM (+)--methyl-4-carboxyphenylglycine (MCPG). E: traces during application of APV plus MCPG. F: LTD during application of MCPG alone or APV plus MCPG. Control LTD is also shown. G: traces before and after (*) LFS during application of APV plus nifedipine. H: traces in the presence of APV plus Ni2+. I: LTD during application of APV plus nifedipine or APV plus Ni2+.

We investigated the role of NMDARs in the induction of LTD. Application of 50 µM APV, an antagonist of NMDARs, had no apparent effect on the field potentials. APV slightly suppressed LTD (81 ± 3%, n = 7, Fig. 2, B and C), but the effect of APV was not significant. Roles of NMDARs might appear more clearly in LTD induced by strong stimuli because Mg²⁺ blockage of NMDARs is removed only by strong depolarization (Nowak et al. 1984). However, the LTD induced by LFS at twice the intensity of the test stimuli (amplitude: 68 ± 7%, n = 7) was only slightly suppressed by APV (74 ± 2%, n = 5), and the effect of APV was not significant. These findings suggest only partial roles of NMDARs, if any, for inducing LTD.

We investigated the possible roles of mGluRs in the induction of LTD. Application of 500 µM MCPG, an antagonist of mGluRs, had no apparent effect on the field potentials. MCPG suppressed the amplitude of LTD only slightly (84 ± 6%, n = 7, Fig. 2, D and F), and the effect was not significant. Therefore we tested the effect of a mixture of APV and MCPG. In the presence of APV plus MCPG, no significant depression was observed 38-40 min after LFS (101 ± 5%, n = 7, Fig. 2, E and F), and the effect of APV plus MCPG on LTD was significant (P < 0.01).

We examined the effects of 200 µM Ni²⁺, a blocker of T-type Ca²⁺ channels, or 20 µM nifedipine, a blocker of L-type Ca²⁺ channels. The amplitude of field potentials was significantly decreased by Ni²⁺ (66 ± 6%, n = 4, P < 0.05), while nifedipine had no clear effect on field potentials. We recorded LTD in Ni²⁺ or nifedipine together with APV. However, these cocktails were not more effective than APV alone (Fig. 2, G-I). These findings suggest the importance of NMDARs and mGluRs but not that of T- or L-type Ca²⁺ channels for the induction of LTD.

LTD in EPSPs of supragranular pyramidal neurons

We investigated LTD in supragranular pyramidal neurons using a perforated patch recording. Monosynaptic EPSPs after layer IV stimulation were identified with their constant latency regardless of the stimulus intensity. The initial rising slope of monosynaptic EPSPs was not affected by bicuculline, while the amplitude of EPSP was increased (data not shown). LTD was estimated as the changes in the rising slope of EPSPs. The amplitude of LTD was 61 ± 6% (n = 11) 25 min after LFS (Fig. 3, A and C). LTD was only partially suppressed by APV alone (75 ± 9%, n = 9, Fig. 3, B and C) or MCPG alone (73 ± 7%, n = 8, Fig. 3, D and F), and these effects were not significant. However, the cocktail of APV plus MCPG suppressed the depression significantly (104 ± 10%, n = 8, P < 0.005, Fig. 3, E and F). These findings are in good agreement with those obtained from field potential recordings.

View larger version (23K): [in this window] [in a new window]   Fig. 3. LTD in excitatory postsynaptic potentials (EPSPs) of supragranular pyramidal neurons. A: sample traces of EPSPs before and 25 min after (*) LFS. B: traces during application of APV. C: LTD in control slices or those incubated with APV. D: LTD during application of MCPG alone. E: LTD in the presence of APV plus MCPG. F: LTD during application of MCPG alone or APV plus MCPG. Control LTD is also shown.

LTD in the synapses between pyramidal neurons connected with horizontal axon collaterals

To investigate LTD in the synapses between pyramidal neurons connected with horizontal axon collaterals, supragranular field potentials were evoked by stimulating layer III 300 µm horizontally apart from the recording sites. LTD (67 ± 8%, n = 6) was induced by LFS. APV alone was sufficient to suppress the induction of LTD significantly (P < 0.05, Fig. 4, A, B, and D), while MCPG had almost no effect (Fig. 4, C and E). These findings suggest that LTD in the synapses between pyramidal neurons connected with horizontal axon collaterals is dependent on NMDARs but not on mGluRs.

View larger version (33K): [in this window] [in a new window]   Fig. 4. LTD in synapses between horizontal connections. A: sample traces of supragranular field potentials before and after (*) LFS applied at 300 µm horizontally apart from the recording site. B: traces in the presence of APV. C: traces during application of MCPG. D: LTD in control medium or in the presence of APV. E: LTD in control medium or during application of MCPG.

LTD in thalamo-cortical synapses

MGB inputs come into the AC from the anterior direction (Metherate and Cruikshank 1999; Webster 1985). To preserve the MGB inputs in the white matter, slices intersecting with a horizontal plane at various angles () were prepared (Fig. 5A). When  was 10-20°, the postsynaptic potentials elicited by stimulating the anterior white matter (800-1,000 µm from the recording site, Fig. 5B) was significantly larger than those in the slices prepared at  of 0 or 30° (P < 0.05, Fig. 5D). Therefore the potentials dependent on  could be attributed to the MGB inputs. To test this possibility, we prepared slices ( = 10-20°) from the rats with MGB lesions. The postsynaptic potentials after stimulating the anterior white matter (800-1,000 µm from the recording site) was significantly smaller than that recorded in normal slices (P < 0.005, Fig. 5E), suggesting a significant contribution of MGB inputs for these potentials.

View larger version (38K): [in this window] [in a new window]   Fig. 5. LTD in thalamo-cortical synapses. A: lateral view of the rat brain. The auditory cortex (AC) and the slicing plane are shown. B: schematic illustration of the experiments for recording the field potentials elicited by the medial geniculate body (MGB) inputs. C: lesions produced in the MGB of 3 rats. D: postsynaptic potentials evoked by white matter stimulation at various sites in slices cut at different angles. The amplitude is plotted against the antero-posterior position of the stimulated sites. E: postsynaptic potentials in the slices obtained from the rats with MGB lesions. F: sample traces of field potentials elicited by the stimulation of the anterior white matter before and 30 min after (*) LFS. G: traces during application of MCPG. H: LTD in control medium or in the presence of APV or MCPG.

We investigated LTD in the field potentials, in which the contribution of MGB inputs was expected. LTD (81 ± 5%, n = 15) was comparable with that after LFS of layer IV (Fig. 5, F and H). While LTD in these experiments was not clearly affected by 50 µM APV (Fig. 5H), it was significantly suppressed by 500 µM MCPG alone (P < 0.05, Fig. 5, G and H). These findings suggest that the induction of LTD in thalamo-cortical synapses is dependent on mGluRs but not on NMDARs.

LTD in cortico-cortical synapses

To investigate LTD induced by LFS of cortico-cortical inputs, we prepared frontal slices, and stimulated the white matter locating 600-1,000 µm medial and dorsal from the recording site (Fig. 6A), where the MGB inputs to the recording site are unlikely to be stimulated (Metherate and Cruikshank 1999; Webster 1985). The amplitude of LTD was 80 ± 5% (n = 7, Fig. 6, B and D) and not significantly different from that after LFS of layer IV. The induction of LTD was significantly suppressed by 50 µM APV alone (P < 0.05, Fig. 6, C and D), whereas 500 µM MCPG had almost no effect (Fig. 6D). These findings suggest that the induction of LTD in cortico-cortical synapses is dependent on NMDARs but not on mGluRs.

View larger version (20K): [in this window] [in a new window]   Fig. 6. LTD in cortico-cortical synapses. A: schematic illustration of experiments for recording the field potentials evoked by cortico-cortical inputs. B: sample traces of field potentials before and 30 min after (*) LFS. C: traces in the presence of APV. D: LTD in control medium or in the presence of APV or MCPG.

LTD in the slices obtained from the rats with MGB lesion

The findings obtained thus far suggest that the induction of LTD in the slices lacking the MGB inputs might be solely dependent on NMDARs. Therefore LTD was tested in the slices obtained from the rat with MGB lesion. The amplitude of the LTD after LFS of layer IV was 67 ± 5% (n = 5, Fig. 7, A and C), and significantly larger than that in normal slices (P < 0.05). As expected, the induction of the LTD was not affected by 500 µM MCPG (Fig. 7, B and C). LTP was suppressed by 50 µM APV significantly but only partially (P < 0.05, Fig. 7, D and F). APV plus MCPG produced no further suppression of LTD (Fig. 7, E and F), suggesting involvement of other mechanisms for the induction of LTD. We tested the effects of 20 µM nifedipine, which was not effective in normal slices (Fig. 2I). Nifedipine suppressed LTD partially, although the effect was not significant (Fig. 7, G and I). The cocktail of APV, MCPG and nifedipine suppressed LTD significantly, and almost no depression was observed 38-40 min after LFS (P < 0.05, Fig. 7, H and I). These findings are explained by the possibility that L-type Ca²⁺ channels, which have only a minor role in LTD in normal rats, could be important for inducing LTD in the rats with MGB lesions.

View larger version (26K): [in this window] [in a new window]   Fig. 7. LTD in the slices prepared from the rats with MGB lesion. A: sample traces of field potentials before and 40 min after (*) LFS. B: traces in the presence of MCPG. C: LTD in control medium or during application of MCPG. D: traces during application of APV. E: traces in the presence of APV plus MCPG. F: LTD in control medium or during application of APV or APV plus MCPG. G: traces during application of nifedipine. H: traces during application of APV, MCPG plus nifedipine. I: LTD during application of nifedipine alone, or APV, MCPG plus nifedipine.

Subtypes of mGluRs required for the induction of LTD

MCPG is a nonspecific antagonist for mGluRs. Eight subtypes of mGluRs (mGluR1-8) are divided into group I (mGluR1 and mGluR5) coupled positively with phospholipase C, and group II/III coupled negatively with adenylyl cyclase (Abe et al. 1992). We investigated the effects of various antagonists of mGluRs on LTD induced by LFS of layer IV in the presence of APV. APV plus 300 µM L-AP3, an antagonist of group I mGluRs (Mistry et al. 1996), suppressed LTD significantly (P < 0.05, Fig. 8). However, the cocktail of APV plus 300 µM AIDA, an antagonist selective to mGluR1 (Pelliciari et al. 1995) and that of APV plus 1 mM CPPG, an antagonist of group II/III, were not more effective than APV alone (Fig. 8). These findings suggest that mGluR5 may be the responsible subtype for the induction of LTD in thalamo-cortical synapses.

View larger version (17K): [in this window] [in a new window]   Fig. 8. Subtype of mGluRs responsible for the LTD of field potentials induced by LFS of layer IV.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Thalamo-cortical and cortico-cortical synapses in the slices obtained from adult rats

To investigate the functions of thalamo-cortical synapses in small brains such as those of mice or juvenile rats, slice preparations including the sensory cortex and thalamic nucleus were developed for the somatosensory cortex (Agmon and Connors 1991) and AC (Metherate and Cruikshank 1999). In the present study, we adjusted the angle of the slicing plane, so that MGB afferents to the AC were attached to the slices obtained from adult rats. Although these preparations did not contain the MGB, thalamo-cortical fibers were specifically stimulated by adjusting the stimulating sites in the white matter. Pharmacological properties were quite different between the thalamo-cortical and cortico-cortical synapses regarding the induction of LTD. Therefore it is suggested that thalamo-cortical and cortico-cortical synapses can be investigated separately even in the slices obtained from adult rats.

Roles of mGluRs in thalamo-cortical synapses and those of NMDARs in cortico-cortical synapses for inducing LTD

In the present study, APV alone or MCPG alone had only a small effect on LTD after LFS of layer IV (Fig. 2), while APV plus MCPG had a significant effect on LTD. These findings can be interpreted in two different ways. First, homogenous LTD might be only weakly blocked by APV alone or MCPG alone, whereas it could be completely blocked by the cocktail because the postsynaptic Ca²⁺ rise required for the induction of LTD was more completely blocked by APV plus MCPG. Alternatively, the LTD after LFS of layer IV can be a mixture of APV-sensitive LTD and MCPG-sensitive LTD. To determine which hypothesis is more likely, we conducted two series of additional experiments. First, conditions were adjusted to stimulate thalamo-cortical and cortico-cortical pathways separately (Figs. 4-6). Second, AC slices were prepared from the rats with MGB lesions (Fig. 7). The results of both series suggest the heterogeneity of the LTD after LFS of layer IV in the AC: it is probably a mixture of NMDARs- dependent LTD in cortico-cortical synapses and mGluRs-dependent LTD in thalamo-cortical synapses.

The differential roles of NMDARs and mGluRs for inducing LTD may be attributable to the distribution of NMDARs and mGluRs. MGB inputs terminate in deep layer III and layer IV in the AC, while cortico-cortical fibers and horizontal axon collaterals terminate in deep layer I, layer II, and layer III (Kelly and Wong 1981; Vaughan 1983). Among various subtypes of mGluRs, mGluR5 was suggested to be responsible for the induction of LTD in thalamo-cortical synapses of the AC (Fig. 8). In the visual cortex, mGluR5 is present in layers where thalamic fibers terminate (Reid et al. 1995), while NMDARs or the NR1 subunit are primarily concentrated in superficial layers (Johnson et al. 1996; Rosier et al. 1993). These reports suggest that mGluR5 is present mainly at thalamo-cortical synapses, while NMDARs are at cortico-cortical synapses. Thus the difference in the induction mechanisms of LTD between thalamo-cortical and cortico-cortical synapses can be explained by the distribution of mGluR5 and NMDARs, although this hypothesis needs verification in the future.

It is unclear what the roles of glutamate receptors in thalamo-cortical and cortico-cortical synapses are for inducing LTD. NMDARs are present in postsynaptic pyramidal neurons in the visual cortex, and a mild increase in [Ca²⁺]_i via NMDARs is essential for inducing LTD (Yasuda and Tsumoto 1996). mGluR5 is also present in postsynaptic neurons (Romano et al. 1995; Shigemoto et al. 1993), and activation of postsynaptic mGluRs is required for inducing LTD in the visual cortex (Kato 1993). IP₃ receptors are found in cell bodies and proximal dendrites of pyramidal neurons (Sharp et al. 1993), where thalamo-cortical synapses are present (Peters et al. 1979). Therefore activation of mGluR5 is expected to raise [Ca²⁺]_i probably via IP₃-mediated Ca²⁺ release from the intracellular Ca²⁺ store. Taken together, the postsynaptic mGluR5 and NMDARs located at the thalamo-cortical and cortico-cortical synapses, respectively, can be responsible for a mild increase in postsynaptic [Ca²⁺]_i required for inducing LTD in the AC.

LTP is induced by tetanic stimulation of the white matter, layer IV and layer III in the AC of adult rats, and the induction is almost completely blocked by application of APV alone (Kudoh and Shibuki 1994, 1996, 1997; Seki et al. 1999, 2001). Application of 50 µM Ni²⁺ also suppressed the induction of LTP (Seki et al. 1999), suggesting that not only NMDARs but also Ni²⁺-sensitive Ca²⁺-channels are involved for the induction of LTP. However, the induction of LTD was not affected by Ni²⁺ (Fig. 2I). In thalamo-cortical synapses, LTD was insensitive to APV, whereas LTP insensitive to APV have not been observed in the adult AC. Because the induction of LTP requires a marked increase in postsynaptic [Ca²⁺]_i, multiple mechanisms including NMDARs and Ni²⁺-sensitive Ca²⁺-channels may be required for producing a sufficient postsynaptic [Ca²⁺]_i increase to induce LTP. In contrast, a mild increase in postsynaptic [Ca²⁺]_i required for inducing LTD may be achieved by activation of mGluR5 alone in thalamo-cortical synapses and NMDARs alone in cortico-cortical synapses.