Monocular deprivation leads to clear physiological and anatomical changes in the visual cortex known as ocular dominance plasticity. Protein kinase A (PKA) is necessary for ocular dominance plasticity, while protein kinase G (PKG) is not. We have now tested the role of PKA and PKG in long-term potentiation (LTP) and long-term depression (LTD). We have shown that PKA inhibitors have a major effect on both LTP and LTD in the visual cortical slices, whereas a PKG inhibitor affects LTP but not LTD. The PKA activator, 8-chloroadenosine-3′,5′-monophosphorothioate, Sp-isomer (Sp-8-Cl-cAMPS), by itself induces a slowly rising form of LTP, which is occluded by theta-burst stimulation (TBS)-induced LTP. These results support the point that the PKA signaling pathway is crucial for neuronal plasticity in visual cortex, and the dissociation of the role of PKA and PKG in long-term synaptic plasticity in the visual cortex suggests that LTP alone is not sufficient to support ocular dominance plasticity, or LTD plays a more fundamental role than LTP in ocular dominance plasticity.
The study of plasticity in the visual cortex is of particular interest because it is easy to examine the relationship between synaptic plasticity, such as occurs with long-term potentiation (LTP) and long-term depression (LTD), and ocular dominance plasticity, where the anatomical, physiological and perceptual changes are well understood (Wiesel, 1982; Daw, 1995; Katz and Shatz, 1996). Closure of one eye, termed as monocular deprivation, in young animals leads to a shift in the ocular dominance of cells in the visual cortex such that, after a few days, very few cells can be driven through the deprived eye (Wiesel and Hubel, 1963). The animal becomes blind in that eye (Dews and Wiesel, 1970), and the endings in the visual cortex from the deprived eye retract (Shatz and Stryker, 1978). Our previous study demonstrated that both basal and metabotropic glutamate-stimulated cAMP levels in the visual cortex correlate well with the critical period for the physiological effects of monocular deprivation (Reid et al., 1996b). Pharmacological activation of the cAMP/PKA signaling pathway rapidly restores ocular dominance plasticity in the adult visual cortex (Imamura et al., 1999). Normally, this kind of plasticity in the visual cortex only occurs following monocular deprivation early in the critical period. Infusion of a protein kinase A (PKA) inhibitor into the visual cortex blocks the physiological manifestation of this ocular dominance plasticity, whereas infusion of a protein kinase G (PKG) inhibitor (Beaver et al., 2001) or a nitric oxide synthase inhibitor, which blocks the activation of PKG by nitric oxide (Reid et al., 1996a; Ruthazer et al., 1996) does not. Why should PKA be necessary for ocular dominance plasticity, whereas PKG and nitric oxide are not, even though both the soluble guanylyl cyclase, a key enzyme for cGMP synthesis, and type I PKG are expressed in the visual cortex of young animals (Roy and Barnstable, 2001)? In particular, are there any synaptic mechanisms underlying this difference?
To investigate these questions, we examined the effects of the PKA inhibitors (9R,10S,12S)-2,3,9,10,11,12-hexahydro-10-hydroxy-9-methyl-1-oxo-9,12-epoxy-1H-dindolo[1,2,3-fg:3′,2′,1′kl]pyrrolo[3,4-1][1,6]benzodiazocine-10-carboxylic acid, hexyl ester (KT5720) and 8-chloroadenosine-3′,5′-monophosphoro-thioate, Rp-isomer (Rp-8-Cl-cAMPS) on LTP and LTD, the most commonly accepted models for neuronal plasticity in the visual cortex, and compared them to the effects of the PKG inhibitor β-phenyl-1,N2-etheno-8-bromoguanosine-3′,5′-monophosphoro-thioate, Rp-isomer (Rp-8-Br-PET-cGMPS). We also examined the effects of the PKA activator 8-chloroadenosine-3′,5′-mono-phosphoro-thioate, Sp-isomer (Sp-8-Cl-cAMPS). While the role of PKA (Frey et al., 1993; Huang et al., 1994; Weisskopf et al., 1994; Blitzer et al., 1995; Brandon et al., 1995; Salin et al., 1996; Tzounopoulos et al., 1998; Otmakhova et al., 2000) and PKG (Zhuo et al., 1994; Arancio et al., 2001; Kleppisch et al., 1999) in LTP and LTD has been studied extensively in hippocampus and other tissues, this is the first time that the effect of PKA and PKG inhibitors on LTP and LTD and the role of activators have been studied in the visual cortex. It is also the first time that this set of comparisons has been made in the same series of experiments on the same tissue.
Materials and Methods
All procedures were performed according to Yale University Animal Care and Use Committee approved protocols.
Coronal slices were made from rat visual cortex. Sprague–Dawley rats (18–30 days old) were anesthetized with halothane and decapitated soon after the disappearance of a tail reflex. The brain was rapidly removed and immediately placed in fresh, ice-cold, oxygenated dissection buffer containing (mM) sucrose 215; KCl 2.5; NaH2PO4 1.25; NaHCO3 26; dextrose 10; MgCl2 2.8; CaCl2 1.0. After the brain stayed in the dissection buffer for 2 min, a block of visual cortex was dissected and glued to the microslicer tray. Coronal slices were sectioned at 400 μm in a fresh, ice-cold oxygenated dissection buffer using a DSK 1000 microslicer. The slices were quickly and carefully transferred to a submersion holding chamber that contained fresh artificial cerebrospinal fluid (ACSF, in mM: NaCl 124; KCl 5.0; NaH2PO4 1.25; NaHCO3 26; dextrose 10; MgCl2 1.3; CaCl2 2.5) bubbled with 95% O2/5% CO2 at room temperature. The slices recovered in the holding chamber for at least 1 h before the experiments. For recording, slices were then transferred to a submersion-type recording chamber mounted on the stage of an upright microscope (BX50WI, Olympus) and held fixed by a grid of parallel nylon threads and perfused with ACSF at a rate of 3–4 ml/min. All experiments were carried out at room temperature (22–25°C). Drugs were applied by gravity.
Field potentials were recorded in layer II/III with glass electrodes (<1 MΩ) filled with ACSF by electrically stimulating layer IV with a bipolar matrix stimulating electrode (no. MX21XEP, Frederick Haer and Co.) placed in the center of the cortical thickness. Changes in the amplitude of the maximal negative field potential (FP) were used to measure the magnitude of LTP and LTD.
Slices were stimulated at 0.05 Hz with a constant current pulse (200 μs duration) at 15–35 μA, which yielded 50–60% of the maximal response. Stimulation for LTP or LTD was applied after 20 min of stable baseline recording. LTP was induced by theta-burst stimulation (TBS) consisting of 10 bursts at 5 Hz, each burst containing five pulses at 100 Hz, repeated for five times at 0.1 Hz. LTD was induced by low-frequency stimulation (LFS) for 15 min at 1 Hz. Every six or nine samples were averaged and field potential amplitudes were then normalized to baseline and expressed as the mean ± SEM. An unpaired Student’s t-test was used to compare FP amplitudes measured in any two different groups, and a paired Student’s t-test was used to compare responses obtained before and after drug treatments in the same group. Differences were considered significant at the level of P < 0.05. The P value for LTP or LTD was calculated by comparing FP amplitude at the time point 30 min after LTP or LTD induction with that at 2 min before LTP or LTD induction, unless indicated otherwise. Slices were interleaved for control and drug-treated groups.
All the drugs were applied by bath perfusion. Sp-8-Cl-cAMPS and Rp-8-Br-PET-cGMPS were purchased from BioLog (La Jolla, CA) and prepared as 1000× stocks in distilled water and aliquoted and stored at −20°C. KT5720 (BioMol, Plymouth Meeting, PA) was prepared as a 1000× stock solution in DMSO and also aliquoted and stored at −20°C. Rp-8-Cl-cAMPS (BioLog) was directly dissolved in ACSF to its final concentration before experiments. The stocks were diluted in ACSF to achieve their final concentrations. Concentrations were chosen to be 16–100 times greater than the Ki for the protein being antagonized, and two to five times smaller than the Ki for other relevant proteins that might be affected. Other chemicals were purchased from Sigma (St Louis, MO).
In an initial set of experiments we examined the effects of PKA inhibitors and a PKG inhibitor on LTP induction. To minimize the effects of age on LTP, we induced LTP by stimulating layer IV and recording field potentials in layers II/III (Kirkwood et al., 1993, 1995, 1996). Induction of LTP from white matter to layer II/III is age-dependent and requires bicuculline or picrotoxin, which at the concentration usually used, often promotes wide-spread epileptic activity of cortical slices.
One of two different selective and membrane-permeant PKA inhibitors, KT5720 (Kase et al., 1987) or Rp-8-Cl-cAMPS (Gjertsen et al., 1995), was bath applied for 20 min starting at 15 min before LTP induction by TBS. Although short-term potentiation (STP) was observed, both drugs resulted in the blockade of LTP, whereas the interleaved control slices exhibited normal LTP (Fig. 1A). Unexpectedly, Rp-8-Br-PET-cGMPS, a selective and membrane-permeant PKG inhibitor (Butt et al., 1995), also completely blocked LTP (Fig. 1B). However, neither of these two protein kinase inhibitors affected basal synaptic transmission at the concentrations used in our experiments. Perfusing slices with either KT5720 or Rp-8-Br-PET-cGMPS (Fig. 1C,D) for 30 min did not significantly affect the amplitude of the field potential in the absence of stimuli to induce LTP or LTD. Thus, both PKA and PKG are required for LTP induction in the visual cortex, but do not have an obvious effect on basal synaptic transmission.
Under similar conditions, we examined the effects of a PKA activator on synaptic transmission in layer II/III from stimulation of layer IV. As previously reported for other PKA activators in hippocampus (Bolshakov et al., 1997; Pockett et al., 1993), exposing slices for 15 min to Sp-8-Cl-cAMPS, a membrane permeant PKA activator, caused a potentiation that gradually developed after the removal of the drug (Fig. 2A,C). The results of this experiment show that activation of PKA starts a reaction that continues for some time after the activator is removed. To determine whether this kind of synaptic potentiation and TBS-induced LTP share common mechanisms, a set of two-way occlusion experiments was carried out. Application of Sp-8-Cl-cAMPS even at 50 μM did not produce potentiation following saturated LTP by twice TBS (Fig. 2B). Surprisingly, Sp-8-Cl-cAMPS-induced potentiation did not occlude TBS-induced LTP (Fig. 2C). These results indicate that PKA activity is required for LTP induction, but LTP induction depends on factors in addition to PKA.
The preceding experiments showing that a PKA activator produced a LTP after its removal indicated that activation of PKA initiated an intracellular biochemical cascade that no longer needed PKA activity. Transient activation of PKA is detected during hippocampal LTP (Roberson and Sweatt, 1996). Thus we now tested the effect of a PKA inhibitor on the potentiated synaptic transmission. Expectedly, the PKA inhibitor KT5720 showed no significant effect on the potentiated synaptic transmission when it was bath applied 30 min after TBS (Fig. 3A). The PKG inhibitor Rp-8-Br-PET-cGMPS did not affect the potentiated synaptic transmission either (Fig. 3B). Taken together with the results from the initial experiments, these results indicate that transient PKA or PKG activity is required for LTP induction, but sustained PKA or PKG activity is not necessary for its maintenance.
Because of the unexpected results of the PKG inhibitor on LTP induction, we examined the effect of PKA and PKG inhibition on LTD, another commonly used form of activity-dependent synaptic plasticity. After 5 min of stable baseline recording, slices were perfused with KT5720 for 35 min, to cover the whole LTD induction period. Only a small change in field potential amplitude was observed following the standard induction protocol of stimulation at 1 Hz for 15 min, whereas interleaved control slices generated robust LTD (Fig. 4A). Similarly, Rp-8-Cl-cAMPS (250 μM) blocked LTD. In contrast, application of the PKG inhibitor Rp-8-Br-PET-cGMPS (3 μM) had no effect on LTD (Fig. 4B).
There are two primary findings in this study. First, both PKA and PKG signaling pathways are required for the early phase of LTP in the visual cortex. Second, PKA is required for the induction of LTD, but PKG is not. These results demonstrate that PKA plays a major role in synaptic plasticity in the visual cortex while the contribution of PKG appears to be of less importance, at least under these conditions. Compared with our previous results from whole animal experiments, this study suggests that LTP alone is not sufficient to support ocular dominance plasticity, or LTD may be more important than LTP for the ocular dominance plasticity in the visual cortex.
The role of the PKA signaling pathway in the late phase of LTP in hippocampus and other tissues is well documented (Frey et al., 1993; Huang et al., 1994; Nguyen and Kandel, 1996; Abel et al., 1997; Bolshakov et al., 1997). Our results support the point that PKA is also required for the early phase of LTP (Blitzer et al., 1995; Otmakhova et al., 2000), since PKA inhibitors block the induction of LTP, and a PKA activator induces LTP by itself. In an attempt to determine whether TBS-induced LTP and potentiation induced by the PKA activator share common mechanisms, we found that LTP induced by repeated TBS occluded Sp-8-Cl-cAMPs-induced potentiation, but application of Sp-8-Cl-cAMPs did not occlude TBS-induced LTP. One explanation for this phenomenon is that PKA activity is not saturated by Sp-8-Cl-cAMPs, even at 50 μM. An alternative explanation is that more than one signaling pathway is required for LTP induction in the visual cortex, so that TBS can still induce LTP even when the cAMP/PKA pathway is saturated. In fact, both CaMKII and p42/44 mitogen-activated protein kinase (MAPK) are reported to contribute to LTP in the visual cortex (Kirkwood et al., 1997; Cristo et al., 2001).
Our results agree with the report of Hensch et al. (Hensch et al., 1998) who found that LTP of extracellular field potentials is absent in visual cortical slices from PKA RIβ-deficient mice. The present study also indicates that PKA activity is not required for the maintenance of LTP, since application of the PKA inhibitor KT5720 at 30 min after LTP induction does not reduce the magnitude of LTP. This is consistent with the study in hippocampus showing that PKA is transiently activated during LTP (Roberson and Sweatt, 1996). Our result that a PKA activator produces a potentiation of synaptic transmission that develops gradually after the removal of the activator also suggests that transient activation of PKA initiates a cascade of events. Thus the activity of PKA is not necessary to be kept at high level during the maintenance of LTP.
In contrast to PKA, the role of the cGMP/PKG signaling pathway in LTP is controversial. The discrepancies between the previous reports might be attributed to differences in the experimental conditions. One important variable is the selectivity and membrane permeability of those cGMP analogs and inhibitors used in those studies. Son et al. found that perfusion with 8-Br-cGMP, a cGMP analog, for 5 or 10 min before weak tetanic stimulation resulted in long-lasting potentiation, while perfusion with 8-Br-cGMP for 15 min before weak tetanic stimulation resulted in no long-lasting potentiation at all. However, perfusion with 8-pCPT-cGMP, a different cGMP analog that is more membrane permeable and more selective for activation of PKG, for 15 or even 30 min before weak tetanic stimulation produced robust LTP (Son et al., 1998). Another important variable might be the protocol for LTP induction. Parent et al. reported that activation of cyclic nucleotide-gated ion channel (CNG) 1 contributed to LTP induction by TBS but not to that by tetanic stimulation (Parent et al., 1998). Most cGMP analogs have effects on CNGs to a different extent. Thus we chose Rp-8-Br-PET-cGMPs, a more selective and more membrane permeable PKG inhibitor that does not affect CNG at the concentration used in this study, to determine the role of PKG in LTP in the present experiments. The results show that the PKG inhibitor can completely block TBS-induced LTP. These results agree with some studies in the hippocampus (Zhuo et al., 1994; Arancio et al., 1995; Malen and Chapman, 1997) and clearly demonstrate that the cGMP/PKG signaling pathway is required for the early phase of LTP in visual cortex. In hippocampus, it was reported that cGMP acted directly in the presynaptic neuron to produce LTP (Son et al., 1998; Arancio et al., 2001). But in the visual cortex, cGMP has postsynaptic facilitating effects on both excitatory synaptic transmission and NMDA response (Wei et al., 2002), so it will be intriguing to determine whether PKG contributes to LTP in the visual cortex by presynaptic or postsynaptic mechanisms.
Of greater interest were the results from LTD experiments. LTD has been suggested to be simply a reversal of LTP, specifically that LTD and depotentiation share a common mechanism. An attractive hypothesis is that LTP and LTD reflect the phosphorylation and dephosphorylation, respectively, of a common set of synaptic proteins (Lisman, 1989; Bear and Malenka, 1994). If this is true, inhibitors of the relevant protein kinases should block LTP and may facilitate or even produce LTD by themselves (Kameyama et al., 1998), but at least should not block LTD. In our case, neither PKA nor PKG inhibitors produced LTD by themselves, though they blocked LTP. To the contrary, inhibitors of PKA blocked LTD, which is consistent with the results from visual cortical (Hensch et al., 1998) and hippocampal slices of PKA RIβ-deficient mice (Brandon et al., 1995) and results from the hippocampus of mice lacking the Cβ1 catalytic sub-unit of PKA (Qi et al., 1996). So far little is known about the mechanisms of PKA action in LTD. If basal synaptic transmission is governed by PKA, down-regulation of PKA activity would induce LTD and subsequent application of electrical LTD-induction protocols would produce no further depression (Kameyama et al., 1998). But this is not true in our case, because the PKA inhibitor KT5720 did not significantly change basal synaptic transmission when it was applied in the same manner as in the LTD experiments. Kato et al. reported that intracellular calcium release through inositol-1,4,5-trisphosphate (IP3) facilitates induction of LTD in visual cortex (Kato et al., 2000), and PKA is well demonstrated to be a major regulator of intracellular calcium release by phosphorylation of IP3 receptors (Haug et al., 1999; Bugrim, 1999). Thus one possibility is that PKA takes part in LTD induction through regulation of intracellular Ca2+ release by IP3 receptors. An alternative possibility that could account for our results is that PKA plays a role in both LTD and LTP by modulating NMDA receptors. Because inhibition of PKA blocks both LTP and LTD, the simplest hypothesis is that PKA inhibitors affect a common component of mechanisms for LTP and LTD. Both LTP and LTD in the pathway from layer IV to layer II/III in visual cortex are NMDA receptor-dependent (Kirkwood and Bear, 1994a,b). So NMDA receptors would be an attractive target for PKA in this hypothesis. It has been demonstrated that the NMDA receptor 1 subunit (NR1) can be phosphorylated by PKA on serine 890 and 897 (Tingley et al., 1997). In addition, NMDA channel activity can be enhanced by PKA but is limited by PP1, and PKA is bound to PP1 by Yotiao, an NMDA receptor-associated protein (Westphal et al., 1999). Thus once PKA is inhibited, the function of NMDA channel will be depressed by PP1 so that LTP or LTD is blocked.
The current study also allows us to relate our results on LTP and LTD with previous results on ocular dominance plasticity (Beaver et al., 2001), and to ask what role LTP and LTD may play in plasticity in the whole animal. PKA inhibitors abolish both LTP and LTD and ocular dominance plasticity. PKG inhibitors abolish LTP but not LTD, and do not affect ocular dominance plasticity. The simple suggestion from this is that LTP is not sufficient to support ocular dominance plasticity, or LTD is more important for ocular dominance plasticity than LTP. Antonini and Stryker (Antonini and Stryker, 1993, 1996) looked at the anatomical correlates of ocular dominance plasticity, and found that terminals from the deprived eye retract over the first week, well before the terminals of the non-deprived eye expand. If LTD is associated with retraction of terminals and LTP with expansion, and retraction has to occur before expansion to liberate space for the expansion, then blockage of LTD will prevent ocular dominance plasticity, but blockage of LTP may not. A similar result has been found in the process of retraction of connections from the retina within the lateral geniculate nucleus to form eye-specific layers. This retraction does not occur in mice mutant for class I major histocompatibility complex, and LTD is absent, while LTP is enlarged (Huh et al., 2000). However, as we noted, all the experiments in the present study were carried out at room temperature. Reservations might be necessary for us to draw conclusions regarding in vivo conditions, since synaptic transmission in the visual cortex, like other mammalian physiological functions, can be affected by changes in temperature (Volgushev et al., 2000).
The authors thank Mark Yeckel, Colin Barnstable, Reiko Fitzsimmonds and Marina Piccioto for comments on the manuscript. This work was supported by Public Health Service Grant RO1 EY11353. Nigel Daw is a Senior Science Investigator of Research to Prevent Blindness.
Address correspondence to: Nigel W. Daw, Department of Ophthalmology, Yale University Medical School, New Haven, CT 06520-8061, USA. Email: firstname.lastname@example.org.