This study aims to clarify how endogenous release of cortical acetylcholine (ACh) modulates the balance between excitation and inhibition evoked in visual cortex. We show that electrical stimulation in layer 1 produced a significant release of ACh measured intracortically by chemoluminescence and evoked a composite synaptic response recorded intracellularly in layer 5 pyramidal neurons of rat visual cortex. The pharmacological specificity of the ACh neuromodulation was determined from the continuous whole-cell voltage clamp measurement of stimulation-locked changes of the input conductance during the application of cholinergic agonists and antagonists. Blockade of glutamatergic and γ-aminobutyric acid (GABAergic) receptors suppressed the evoked response, indicating that stimulation-induced release of ACh does not directly activate a cholinergic synaptic conductance in recorded neurons. Comparison of cytisine and mecamylamine effects on nicotinic receptors showed that excitation is enhanced by endogenous evoked release of ACh through the presynaptic activation of α*β4 receptors located on glutamatergic fibers. DHβE, the selective α4β2 nicotinic receptor antagonist, induced a depression of inhibition. Endogenous ACh could also enhance inhibition by acting directly on GABAergic interneurons, presynaptic to the recorded cell. We conclude that endogenous-released ACh amplifies the dominance of the inhibitory drive and thus decreases the excitability and sensory responsiveness of layer 5 pyramidal neurons.
The nucleus basalis magnocellularis (NBM) of the basal forebrain is the primary source of cortical acetylcholine (Ach; Rye et al. 1984) and cholinergic fibers project with a diffuse anatomical innervation pattern across the cortical mantle (Mesulam 1995). The ascending cholinergic pathway is known to play a major role in increasing the “signal/noise ratio” of sensory processing such as top-down control by attention, conscious awareness, and state-dependent learning (Sarter and Bruno 1997; Léna and Changeux 1998; Shulz et al. 2000; Court et al. 2001; Raggenbass and Bertrand 2002).
ACh has been shown to affect the cortical network integration of excitatory glutamatergic and inhibitory γ-aminobutyric acidergic (GABAergic) signals in a complex manner (Woody and Gruen 1987). Pharmacological studies have used applications of cholinergic agonists combined with separate extracellular blockade of either excitatory or inhibitory conductances. Agonists, such as nicotine, increase the release of glutamate (Vidal and Changeux 1993; Gioanni et al. 1999; Wonnacott et al. 2006) and GABA (Porter et al. 1999; Alkondon and Albuquerque 2004) in rat cortex, hippocampus, or corpus striatum through the activation of presynaptic receptors. A small excitatory current induced by the activation of postsynaptic nicotinic receptors has also been described recently in cortical neurons (Xiang et al. 1998; Porter et al. 1999; Christophe et al. 2002). Recent studies showed with more details that focal applications of ACh inhibited layer 5 pyramidal neurons via a direct activation of a potassium conductance when, in contrast, the same application produced no responses in fast spiking interneurons and evoked a large diversity of effects in the different types of non–fast-spiking (non-FS) interneurons: nicotinic depolarization (in particular in layer 1 neurons), hyperpolarizing responses with or without initial nicotinic depolarization, or no response at all (Gulledge and Stuart 2005; Gulledge et al. 2007). Thus, one may conclude that ACh inhibits directly cortical output pyramidal neurons while facilitating or inhibiting specific inhibitory circuits
However, these studies gave only indirect informations about the modulatory roles of ACh affecting the balance between excitation and inhibition. In addition, they did not investigate the effect of the endogenous release of ACh because they relied on the application of exogenous ACh or analogs.
In rat auditory cortex, Metherate and Ashe (1995) examined how spontaneous ACh release acts on synaptic potentials using an anticholinesterase compound. They concluded that ACh tonically depresses synaptic potentials mediated by both glutamate and GABA. A further study by Aramakis et al. (1997) suggested complex interaction effects, the interpretation of which is limited because synaptic effects were derived only from recorded depolarizations and hyperpolarizations using current clamp measurements. In view of the limited conclusions reached by conventional techniques, the aim of the present study was to develop appropriate methods to quantify cortical ACh-mediated modulation of the interaction between synaptic inhibition and excitation during afferent stimulation in rat visual cortex.
For this purpose, we applied in vitro an analysis of synaptic dynamics, developed in vivo (Borg-Graham et al. 1998; Monier et al. 2003, 2008), based on the continuous measurement of the synaptic conductance and its apparent synaptic reversal potential, in response to stimulation of intracortical or white matter (WM) afferents. A decomposition method was used to dissect out the synaptic conductance into excitatory and inhibitory components of fixed known reversal potentials (Monier et al. 2003, Le Roux et al. 2006, 2007). Taking into account that 1) cholinergic fibers have been described to be localized in all cortical layers but more densely in layers 1 and 5 (Butcher et al. 1993) and that 2) a modulation of LTP by the release of endogenous ACh has been observed following stimulation of layer 1 (Hess and Donoghue 1999), we chose to stimulate layer 1 in order to recruit a combined activation of glutamatergic, GABAergic, and cholinergic inputs and record the evoked synaptic responses in identified pyramidal neurons of layer 5. Because the principal aim of the study was to look at changes in the balance between excitation and inhibition in the presence of ACh, a classical dissection of synaptic sources by blocking separately excitation or inhibition was not found appropriate. The measure of the interplay between excitation and inhibition requires 1) to maintain the integrity of the test network, 2) to take into account direct and indirect effects on polysynaptic transmission, and 3) to study possible differential behaviors in the presence of ACh agonists and antagonists.
We focused our study on the release of endogenous ACh following layer 1 stimulation, an approach that has not been yet extensively studied. The bath application of various antagonists of cholinergic receptors indicates that synaptic integration in layer 5 can be up- and downregulated by the activation of nicotinic and muscarinic receptors. These receptors have a specific differential localization on excitatory and inhibitory neurons that are activated presynaptically to the recorded layer 5 cell.
Materials and Methods
Parasagittal slices containing primary visual cortex were obtained from 20- to 25-day-old Wistar rats as described by Edwards et al. (1989). The age of the preparation, which allows stable patch recordings, still fits with the end of maturation of the ascending cholinergic projection (around P19; Aramakis and Metherate 1998). Briefly, rats were decapitated, and brains were quickly removed and placed in cold (5 °C) artificial extracellular solution, in accordance with guidelines of the American Neuroscience Association. Slices of 350 μm thickness were cut on a vibratome and then incubated for at least 1 h at 36 °C in extracellular solution containing (in mM): 126 NaCl, 26 NaHCO3, 10 glucose, 2 CaCl2, 1.5 KCl, 1.5 MgSO4, and 1.25 KH2PO4, which was bubbled with a mixture of 95% O2–5% CO2 (pH 7.5, osmolarity 310/330 mOsm).
All extracellular drug applications were added through perfusion to the bathing solution for at least 15 min before recording. Methyllycaconitine (MLA) and QX314 bromide were obtained from Tocris (Bristol, UK). ACh chloride, 2-amino-5-phosphonovalerianic acid (DL-APV), bicuculline, 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX), cytisine, dihydroβerythroidine (DHβE) hydrobromide, mecamylamine, and tetrodotoxin (TTX) were obtained from Sigma (St Louis, MO).
Chemiluminescence Detection of Choline
After setting up the brain slice for electrophysiological recording, a 20 μl volume of the external medium just covering a predefined layer 5 region in the vicinity of the stimulation electrode was sampled before and after electrical stimulation of layer 1 or WM, with an intensity level which induced a subthreshold synaptic response. Two protocols of stimulation were used: 1) 5–10 stimulations at 0.05 Hz, 2) a series of 9 episodes (repeated at 0.1 Hz) of theta burst stimulation (13 high-frequency trains of 4 pulses at 100 Hz, repeated at 5 Hz). Because released ACh is promptly hydrolyzed in the slice, this should lead to a detectable synthesis of free choline. ACh release was measured by first establishing the endogenous choline level prior to the application of electrical stimulation protocols. Thus, any enhancement of choline concentration after the electrical stimulation was attributed to the hydrolysis of released ACh (Szerb 1975; Morot-Gaudry et al. 1985). The method described by Israel and Lesbats (1982) was used to further oxidize choline in betaine and hydrogen peroxide. The final reaction stage between hydrogen peroxide and luminol was then monitored by measuring light emission. The integral value of the emitted light (area under the recorded trace), proportional to the choline content of the sample, was calibrated with convenient standards. One should note that it is impossible to do a direct measurement of ACh in the brain slice with such chemiluminescent assays because the anticholinesterasic compound modifies light emission by luminol and high concentrations are required to fully inactivate acetylcholinesterase. It is the reason why we chose to use the measured changes in the concentration of choline as a quantitative, but relative, indicator of ACh release.
Electrophysiological Recordings and Cell Identification
Slices were perfused continuously and viewed with standard optics using a 40× long working–distance water immersion lens of a Zeiss microscope on an X–Y translation stage with a video-enhanced differential interference contrast system. Pyramidal neurons, identified on the basis of the shape of their soma and the proximal part of the apical dendrite, were recorded in layer 5 using the whole-cell configuration of patch-clamp techniques. Somatic whole-cell recordings were performed at room temperature using borosilicate glass pipettes (3–5 MΩ in the bath) filled with an internal solution (in mM): 140 K-gluconate, 10 N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid, 4 ATP, 2 MgCl2, 0.4 GTP, and 0.5 ethyleneglycol-bis(aminoethylether)-tetraacetic acid, pH adjusted to 7.3 with KOH and the osmolarity adjusted to 285 mOsm.
For some experiments (n = 14), the quaternary lidocaine derivative QX-314 (3 mM) was included in the internal pipette solution to block sodium action potentials and GABAB receptors (Nathan et al. 1990). In a subsample of cells (n = 31), 1% neurobiotin (Vector labs, Burlington, CA) was used to label pyramidal neurons (Lapper and Bolam 1991) and reconstruct their morphology.
Current-clamp and voltage-clamp mode recordings were performed using an Axopatch 1D (Axon Instruments, Molecular Devices, Sunnyvale, CA). Intracellular records were filtered by a low-pass Bessel filter with a cutoff frequency set at 2 kHz and digitally sampled at 4 kHz. Analysis was done off-line with specialized software (Acquis1 and Elphy: written by Gérard Sadoc, Biologic UNIC–CNRS, France). All membrane potential values obtained with this filling solution were corrected off-line by −10 mV in order to subtract the junction potential (Neher 1992). After capacitance neutralization, bridge balancing was done online under current clamp to make initial estimates of the access resistance (Rs). These values were checked and revised as necessary off-line by fitting subthreshold hyperpolarizing current clamp responses to the sum of 2 exponentials (Rs = 13.1 ± 5 MΩ (4–25 MΩ), n = 177). The firing behavior of neurons was determined in response to depolarizing current pulses ranging from −100 to 200 pA in current-clamp mode. The access resistance was compensated off-line in voltage clamp mode.
Electrical stimulations (1–10 μA—0.2 ms duration) were delivered using 1 MΩ impedance bipolar tungsten electrodes (TST33A10KT, WPI) with a tip separation of 125 μm. Electrodes were positioned in WM, either at around 500 μm from the recording site and along the same radial columnar axis or in layer 1 (in the top of layers 2/3) at a lateral distance of around 800 μm from the recording site. We applied paired-pulse stimulation with an ISI delay of 300 ms (3.3 Hz) every 20 s (0.05 Hz), and 5–10 trials were repeated for a given holding current or potential.
Continuous Estimation of the Synaptic Conductances in Voltage Clamp
Data were analyzed by measuring continuously conductance changes during the full time course of the stimulus-evoked synaptic response. The measurement method has been described previously with in vivo preparation (Borg-Graham et al. 1998; Monier et al. 2003) and in vitro preparation (Le Roux et al. 2006, 2007). Methods of conductance estimation in voltage clamp, based on the same principle but with few differences (for an extensive review, see Monier et al. 2008), have been validated in the rat auditory cortex (Wehr and Zador 2003, 2005; Tan et al. 2004), in the ferret prefrontal cortex (Shu et al. 2003, Haider et al. 2006), and in the mouse somatosensory thalamocortical slice (Cruikshank et al. 2007).
I/V plots are commonly used to characterize input conductance and cellular excitability in a static way and can be characterized in voltage clamp or current clamp mode. The present study is performed in voltage clamp mode, which minimizes distortion of synaptic events by transient voltage-dependent channels and capacitance near to the recording site (the derivative of the voltage is consider to be zero). Our method aims at a dynamic measure of input conductance, phase locked to the time of the electrical stimulation, and relies on raw stimulus-locked I/V measurements made at each point in time: the current value includes both evoked and resting components, and the holding potential is corrected for the ohmic drop through the access resistance (V hc(t) = Vh(t) − I(t) × Rs). In this situation, the slope of the best linear I/V fit gives the total input conductance of the cell (Gin) at time t.
The synaptically evoked component (Gsyn(t)) is then measured by subtracting the resting conductance observed in the absence of stimulation (on a time window of 100 ms before the electrical stimulation) from the total conductance. Gsyn(t) can be expressed directly in absolute measurement units (nS) or in relative units (ΔGsyn(t) in %) when compared with the conductance at rest. The synaptic reversal potential of the synaptic conductance increase (Esyn(t)) is taken as the voltage of the intersection between the I/V-curve during the synaptic response and the I/V-curve during the resting condition. Assuming that the evoked conductance change measured at the soma reflects the composite synaptic input effective in driving the cell (because visible at the soma and presumably at the axon hillock), Esyn(t) characterizes the effective balance between excitation and inhibition over time. The global synaptic conductance was further decomposed into 2 conductance components corresponding to the activation of excitatory synapse and inhibitory synapses, each associated with known and fixed reversal potentials.
In order to precisely determine the GABAA reversal potential under our experimental conditions, we pharmacologically blocked the GABAB (with QX314 in the internal solution) and 2-amino-3-hydroxy-5-methyllisoazol-4-propionic acid (AMPA) and N-methyl-D-aspartic acid (NMDA) excitatory components (with bath application of CNQX [25 μM] and APV [20 μM], respectively). The apparent synaptic reversal potential of the peak conductance evoked by an electrical stimulation applied to layer 4 was −80.1 ± 3 mV (n = 19, data not shown; see details in Monier et al. 2008). The reversal potential of the excitation could not be tested within our experimental paradigms because the pharmacological blockade of inhibition gives rise to uncontrolled bursts in the spontaneous activity. Using the same preparation, the same experimental setup and analysis software as in the present study, Le Roux et al. (2006) performed a related control (see figure D in their supplementary data) where voltage-clamped ramps were applied between −80 and +30 mV in the absence of exogenous application of glutamate. The crossover point between the 2 linear I–V relationships (control vs. glutamate) established that the reversal potential for the excitatory equals 0 mV for our experimental conditions.
Accordingly, the reversal potentials used for the decomposition of the global synaptic conductance were set at 0 mV for excitatory (Eexc) and −80 mV for inhibitory conductance (Einh). The value of Einh is compatible with the Nernst potential calculated for the intracellular chloride concentration in our experimental conditions (about 5 mM). In addition, these values for the reversal potentials are classically accepted and used in other studies (Wehr and Zador 2003, 2005). The synaptic conductance was equated to the excitatory or to the inhibitory component, in the cases when the reversal potential was found, respectively, either above Eexc (0 mV) or below Einh (−80 mV). This ensures that all the global synaptic conductance terms in the decomposition are positive or null. A decomposition method based on 3 synaptic components (Excitatory, GABAA and GABAB) has been previously applied in vivo (Monier et al. 2003). Our method here is similar mathematically and considers that the GABAA and the GABAB components can be added and replaced by a unique inhibitory component. The fact that we are not dissociating GABAA and GABAB components (for sake of simplification) does not imply that we are underestimating one component relatively to the other one.
In the present stimulation and slice preparation conditions, no change in the evoked synaptic test response was found when the putative NMDA component was blocked with an antagonist and when a low frequency of stimulation (0.05 Hz) was applied. It is however well established that NMDA responses can be recruited in visual cortex, in vitro (Carmignoto and Vicini 1992; Nase et al. 1999) as well as in vivo (Kleinschmidt et al. 1987), for afferent stimulation regimes (electrical or visual) susceptible to elicit a strong postsynaptic depolarization. Indeed, high frequency of stimulation protocols (i.e., theta burst) induced NMDA receptors–dependent potentiations of excitatory and inhibitory inputs on layer 5 pyramidal neurons in our experimental conditions (Le Roux et al. 2006, 2007). In addition, in the present study, we used a 1.5 mM Mg2+ concentration in the artificial cerebrospinal fluid (ACSF), whereas, in many studies, a free Mg2+-perfusing solution is used to enhance the visibility of NMDA receptor–mediated excitatory postsynaptic potentials (Agmon and O'Dowd 1992; Gil and Amitai 1996). Furthermore, we did not block monosynaptic and polysynaptic inhibition (GABAA and GABAB) that are known to strongly downregulate the NMDA response component (Luhmann and Prince 1990).
Because our recordings were done at the somatic level, the conductance measurement method that we have devised does not overcome the loss of visibility of dendritic inputs due to local nonlinear synaptic interactions or strong cable attenuation or exclude the contamination by currents from poorly clamped voltage-dependent dendritic membranes. The motivation of using the voltage-clamp method is that distortion of synaptic events by transient voltage-dependent channels and capacitance near to the recording site are minimized (Borg-Graham et al. 1998).
For each component, excitatory and inhibitory, we calculated 2 parameters to quantify the conductance changes: the peak value and the mean averaged over a time window of 200 ms. The contribution of each synaptic component was expressed by the ratio between its mean value and the mean composite global synaptic conductance change. Onset latencies of synaptic conductance changes were measured from the time of stimulation onset by determining the earliest delay at which the conductance waveform began to deviate significantly from the average baseline value.
Reconstruction of the Membrane Potential Trajectory
The computation of the excitatory and inhibitory conductances on the basis of the voltage clamp measurements allows to predict the membrane potential trajectory (Vrec(t)) that would have been observed in current clamp. This is done by solving numerically the following differential equation:
In order to check the coherence between the different estimates, we calculated the root mean square error (“rms error”) between the actual mean voltage measured in current clamp () and that predicted (Vrec (t)) from the excitatory and inhibitory conductances estimated in voltage clamp or current clamp (Wehr and Zador 2003):
Standard error of the mean bars and mean values averaged across cell population for a given experimental condition are reported in the graphs. Differences between means were evaluated for statistical significance using the nonparametric Wilcoxon statistical test and the parametrical t-test for paired samples. In this latter case, data were normalized as “percent of control value.” Differences were considered statistically significant if: (*) the probability of occurrence by chance was 0.05 or less (P ≤ 0.05), (**) for P < 0.01, and (***) for P < 0.001.
ACh Release by the Cortical Cholinergic Network
Cholinergic fibers innervating the cortex originate from the CH4 cell group of the NBM (Johnston et al. 1981; Mesulam et al. 1992). Cholinergic fibers can be found in all cortical areas and layers (Mesulam et al. 1986, 1992) with a density of cholinergic fibers differing between cortical areas and across layers (Lysakowski et al. 1989). Layer 1 has been shown to contain the highest laminar densities of ACh axons and varicosities (Mechawar et al. 2000). Accordingly, the immunocytochemical labeling of cholinergic fibers with the specific marker of the vesicular transporter of ACh indicated a strong labeling in layers 1/2 and a more discrete labeling from layer 3 to layer 5 (data not shown).
Cholinergic fibers innervating the occipital cortex originate from the diagonal band and enter the cortex by initially running within layer 6 before ascending toward the pial surface and terminating mainly in layers 5 and 1. Thus, it seems rather unlikely that the sole stimulation of WM afferents contributes significantly to the direct activation of cholinergic fibers in visual cortex. An appropriate electrical stimulation of layer 1 should more readily activate cholinergic fibers in addition to glutamatergic and GABAergic ones (Patil and Hasselmo 1999). To ascertain that electrical stimulation of layer 1 was effectively inducing the release of endogenous ACh, a chemiluminescent characterization of released ACh was carried out. We checked the stability over time of the concentration of choline in the supernatant of control slices in the absence of any electrical stimulation. The concentration of choline was stable (0.393 ± 0.034 μM, n = 26) for 20 min (Fig. 1A). Five stimulations of the layer 1 at 0.05 Hz increased the concentration of choline from 0.316 ± 0.096 μM to 0.668 ± 0.129 μM (n = 5; t-test P < 0.05; Fig. 1B). In order to obtain a maximal stimulation of cholinergic fibers, we used a theta burst stimulation (see Methods) of the layer 1, a protocol which is largely recognized as mimicking physiological patterns of afferent activity. With this protocol, the concentration of choline is significantly (P < 0.01) increased to 1.100 ± 0.211 μM (n = 8, Fig. 1B). The concentration of choline remained significantly above control level (0.550 ± 0.077 μM, n = 7; P < 0.001) when action potential initiation was prevented by the bath application of 1 μM TTX (Fig. 1B). Consequently, we ascertained that electrical stimulation of layer 1 induces ACh release. An identical stimulation of WM (Fig. 1B) also increased the amount of choline from 0.350 ± 0.070 μM to 0.921 ± 0.790 μM (n = 4, P < 0.05) as expected from the histological course of cholinergic fibers in the cortex.
Conductance Decomposition and Reconstruction of Voltage Dynamics
Stable patch-clamp recordings were obtained from neurons (n = 183), the soma of which were exclusively located in layer 5 of rat visual cortex. The excitability behavior of each neuron was characterized by the discharge pattern in response to test depolarizing current pulses. Recorded neurons had a resting potential of −71.1 ± 0.7 mV (n = 118, for recordings with standard pipette solution) and showed the typical regular adaptation discharge pattern of pyramidal neurons as described by McCormick et al. (1985). Membranes of these neurons had an input resistance of 224.5 ± 11.4 MΩ (n = 118) and a time constant of 30.8 ± 1.3 ms (n = 118). To ascertain identification, some of the recorded cells were also labeled with neurobiotin 1% (n = 31), and all indeed exhibited the typical morphology of pyramidal neurons with thick-tufted dendrites, suggesting that all recordings refer to a morphologically and electrophysiologically identified class of cortical neuron.
We recorded electrically evoked synaptic responses in current clamp and voltage clamp modes. In order to deliver a focal stimulus, tips of the stimulating tungsten bipolar electrode were kept separated by a distance of 125 μm; electrodes were positioned at different distances from the recording site to stimulate different cortical layers and afferent circuits: in the layer 1, in order to stimulate more distal inputs, and in WM, in order to directly stimulate thalamocortical fibers. The intensity of the stimulation was adjusted in current clamp mode to induce a subthreshold postsynaptic response (Fig. 2B). This intensity was around 2–3 times the amplitude of the stimulation necessary to induce a detectable response (minimal threshold stimulation) and weak enough to avoid recruiting dominant nonlinear processes, linked for instance to NMDA receptor activation. In current-clamp mode, starting from a resting state slightly more depolarized than the reversal potential of GABAergic synapses, evoked responses mainly exhibited 2 components: a fast depolarization followed by a hyperpolarization (Fig. 2B). Occasionally, a slower phase of hyperpolarization was observed after the transient initial hyperpolarization.
To estimate the synaptic conductances, the membrane current (Im) of the cell was recorded in voltage-clamp mode, at 4–5 holding potentials between −90 and −50 mV, and the stimulus-locked waveform was averaged over 5–8 trials (see Fig. 2C). The total input conductance (Gin(t)) was computed on the basis of current and corrected voltage measurements by calculating at every point in time the slope of the best linear fit of the IV-curve (see Methods). The composite synaptic conductance change (Gsyn(t)) was then derived by subtracting the resting conductance (Grest(t)) estimated before the electrical stimulation and decomposed into excitatory (Gexc(t)) and inhibitory (Ginh(t)) conductances using the apparent reversal potential of the synaptic conductance (Esyn(t)) (Fig. 2C; for details, see Methods).
The synaptic conductance change evoked by a layer 1 stimulation starts at 6.4 ± 0.2 ms (onset latency) and peaks at 15.3 ± 0.4 ms (peak latency) after the electrical shock, with an amplitude of 12.4 ± 0.7 nS (n = 183), corresponding to a mean increase of 290 ± 19% from the resting conductance value. The apparent reversal potential at the peak of the response was −56.2 ± 1.2 mV (n = 183). This value indicates a dominance of inhibition at the peak of the evoked conductance change. The decomposition of the evoked synaptic conductance allows us to further evaluate the relative contribution of excitation and inhibition (Fig. 2C; for details, see Methods). When integrated over the full time course of the synaptic response, the average balance between the excitatory and inhibitory conductance components, representing, respectively, 17 ± 1% and 83 ± 1% of the global synaptic conductance (n = 183). The onset latencies of excitation and inhibition differed significantly and were, respectively, 6.3 ± 0.2 ms and 7.6 ± 0.2 ms (difference 1.2 ± 0.2 ms). The excitatory conductance peaked at 12.1 ± 0.4 ms on average, when the inhibition peaked 4.1 ± 0.3 ms after at 16.2 ± 4 ms. Thus, in all cells, excitation was elicited before inhibition for a short temporal window, after which the balance shifted strongly in favor of inhibition.
The continuous decomposition at any point in time of the total conductance into excitatory and inhibitory components allows us to recalculate the dynamics of the membrane potential trajectory during activation on the basis of the measurements of the membrane time constant (τm). This reconstituted potential response is the prediction based on voltage clamp recordings of the behavior that normally should be observed during current clamp measurements if the contribution of voltage-dependent nonlinearities remains limited. The direct comparison between the reconstructed and CC-recorded voltages is illustrated for different cells (Fig. 2D). We see clearly that both traces fit relatively well together. In order to quantify the difference between VC-reconstructed voltages and raw CC-records, we calculated the rms error (see Methods) for each response. We observed good agreement between the 2 data sets (rms error = 0.96 ± 0.3 mV [n = 50]).
Pharmacological Dissection of the Synaptic Response
The respective contribution of excitation and inhibition was dissected out pharmacologically. In order to block a putative GABAB component, the internal solution of the recording pipette was supplemented with QX314 (3 mM). The application, by perfusion in the extracellular medium, of APV (NMDA receptor blocker) and CNQX (AMPA/kainate receptor blocker) suppressed the inward current, the excitatory conductance and the fast depolarization in the voltage trace recorded in current clamp (see example on Fig. 3A). In addition, the application of CNQX/APV produced a strong concomitant reduction of the inhibitory component due to the suppression of the polysynaptic inhibition. The percentage of monosynaptic inhibition (ratio between the inhibitory conductance increase in the control situation and that measured during blockade of excitation) was only 14 ± 7% (n = 29), and the apparent reversal potential of this synaptic response during application of CNQX/APV was −80.8 ± 1.3 mV (n = 29). This value which is slightly more negative than that found for the layer 4 stimulation (see Methods) is likely due to the distal inhibitory input. Finally, the application of bicuculline (GABAA receptor antagonist) in the extracellular medium fully blocked the monosynaptic inhibitory conductance and the fast hyperpolarization in the voltage trace recorded in current clamp (example Fig. 3A).
The lack of a consensus concerning specialized cholinergic synapses in the cortex has led to the hypothesis that released ACh could diffuse and act at distance from its releasing site (Descarries et al. 1997). The fact that in the presence of QX314, CNQX, APV, and bicuculline, the evoked synaptic response was totally suppressed (n = 8, example on Fig. 3A) indicates that the cholinergic effects are not due to a direct action of released ACh on pyramidal neurons of layer 5 following stimulation of layer 1.
A paired-pulse stimulation (ISI delay 300 ms, 0.33 Hz) of layer 1 induced only a weak reduction of the second response (Fig. 3A). The peak and the mean of the synaptic conductance increase were, respectively, decreased by 19 ± 2% and 13 ± 2% (n = 153, P ≤ 0.001, see distribution on Fig. 3B). The decomposition of synaptic conductances showed that the peak and the mean of the inhibitory conductance were, respectively, decreased by 22 ± 2% and 15 ± 2% (n = 153, P ≤ 0.001), whereas the peak of the excitatory conductance was decreased only by 7 ± 2% and the mean increased by 12 ± 4% (n = 153, P ≤ 0.001, Fig. 3B). In consequence during the synaptic response to the second stimulation, the balance between the excitation and the inhibition was shifted in average from 17/83% to 21/78% (n = 153, P ≤ 0.001, see example Fig. 3A, the balance shifts from 19/81% to 25/75%). At the dendritic level, the depression of excitation would be partly compensated and late excitatory inputs might be unmasked as indicated by the increase of the mean excitatory conductance.
In order to distinguish between the monosynaptic and polysynaptic nature of the paired-pulse depression in inhibition, we blocked excitation and polysynaptic inhibition (mediated by an excitatory drive) with CNQX/APV (n = 29). The remaining monosynaptic inhibition was only slightly depressed in the second response compared with the first response in the example of the Figure 3A, this depression being nonsignificant at the population level (n = 29, P = 0.12). Thus, in the control situation, the slight paired-pulse depression of the inhibition results from the depression of the polysynaptically mediated inhibitory component.
Presence of Functional Nicotinic Receptors in the Network
Our bioluminescence experiments suggest that the stimulation of layer 1 recruits cholinergic fibers releasing ACh, allowing a potential activation of cholinergic receptors. In order to study more selectively the nature of the cholinergic modulation, we have studied the pharmacological effects of bath application of agonists and antagonists of ACh receptors.
First, we checked the possible involvement of nicotinic receptors with application in the external medium of cytisine, an agonist of the nicotinic receptors (Paterson and Nordberg 2000). A typical experiment and pooled data (n = 8) are illustrated in Figure 4. Cytisine (10 μM) increased the peak of the synaptic conductance evoked by layer 1 stimulation by 32 ± 17% compared with the control condition (P < 0.05) (example in Fig. 4A and pooled data in Fig. 4B). The reversal potential corresponding to the peak of the synaptic conductance was shifted toward a more depolarized state by 11 ± 4 mV (−49 ± 5 mV vs. −60 ± 3 mV, P < 0.05), reflecting a change in the balance between excitation and inhibition in favor of the excitation.
The conductance decomposition analysis confirmed this observation and showed a selective increase in the peak (by 52 ± 22% [P < 0.05]) and in the mean (by 91 ± 35% [P < 0.05]) of the excitatory conductance compared with the control situation. There was no significant change in the inhibitory conductance component (Fig. 4B; P > 0.43). Consequently, the fraction of the synaptic drive due to the excitatory component was significantly increased to 24 ± 3% (P < 0.05) in the presence of cytisine. This change in the balance between evoked excitation and inhibition was clearly visible in the reconstructed current clamp voltage trace: The amplitude of the depolarization increased and the amplitude of the hyperpolarization decreased (see example in the lower row of Fig. 4A).
Paired-pulse protocols gave identical levels of modulatory effects of cytisine on synaptic responses to the first and to the second stimulation. As seen before, the depression of the inhibitory conductance to the second stimulation pulse is attributed to a decrease in polysynaptic and not monosynaptic inhibition, the efficacy of glutamatergic excitatory synapses afferent to GABAergic inhibitory neurons being very likely reduced. The similarity in the modulatory effects of cytisine on the first and on the second test responses suggests that nicotinic receptors activated by cytisine are not involved in the mechanism responsible of the paired-pulse depression. We conclude that a selective activation of nicotinic receptors produces a significant modulation of the response evoked by layer 1 stimulation. In addition, the increase in the excitatory component suggests a preferential location of nicotinic receptors situated on glutamatergic neurons presynaptic to the pyramidal layer 5 recorded cell, where their activation could increase the release of glutamate and, thus, the excitatory drive of the pyramidal neuron.
Recruitment of Nicotinic Receptors by Stimulation of the Endogenous Cholinergic System
The role of endogenously released ACh on nicotinic receptors was further analyzed by application of mecamylamine (5 μM), a noncompetitive antagonist of nicotinic receptors (Chavez-Noriega et al. 1997). A typical experiment and results on the whole neuronal population (n = 14) are illustrated in Figure 5A. During application of mecamylamine, no significant change was found in the peak value of the global composite synaptic conductance. However, the decomposition method showed that both the peak and the mean values of the inhibitory conductance decreased slightly but that, due to a large dispersion, these results were not statistically significant (P > 0.2). Nevertheless, a significant decrease of the excitatory conductance, visible in the example of the Figure 5A, was found at the population level (−13 ± 5%, n = 14, P < 0.01, Fig. 5B). In the illustrated cell, this decrease in the excitatory conductance and, to a lesser extent, in the inhibitory conductance, produced a reduction in the depolarization phase observed in the reconstructed voltage trace. The paired-pulse protocol gave identical effects of mecamylamine on synaptic responses to the first and to the second stimulation pulse.
These results suggest that endogenous ACh might modulate the excitatory conductance component. This effect is relatively weak but in contrast with that produced by cytisine application, it concerns only the activation of receptors by endogenous ACh. It suggests that cytisine or endogenous ACh were both able to activate the same subtype of nicotinic receptor.
Effect of MLA, an α7 Selective Antagonist
The implication of α7 nicotinic receptors in the modulatory effects of ACh was checked by the perfusion of the selective antagonist, MLA (Ward et al. 1990). Whatever the concentration used (10−6, 10−7, and 10−8 M, n = 7, 7, and 9, respectively), no significant effects (Fig. 6) on the amplitude of the synaptic conductance (P > 0.65, 0.94 and 1 respectively) and either on the inhibitory (P > 0.98, 0.69, 0.91, respectively) or the excitatory conductances (P > 0.51, 0.93, 0.46, respectively) were observed. Consequently, in our preparation, α7 nicotinic receptors may be weakly or not implicated in the modulatory effects of ACh.
Effect of DHβE, an α4β2 Receptor Antagonist
DHβE (500 nM), an antagonist of α4β2 receptors (Alkondon and Albuquerque 1993) which are the most numerous receptors in the cortex (Zoli et al. 1998), was applied by perfusion. A typical experiment and pooled data (after electrical stimulations in layer 1 or in WM) are illustrated in Figure 7. In the cell illustrated in Figure 7A where the stimulus was applied in layer 1, DHβE induced a decrease in the global synaptic conductance. This effect was replicated for the whole population of tested pyramidal neurons (n = 7, Fig. 7B), and the peak value of the global synaptic conductance was decreased by 26 ± 7% (P < 0.05). The reversal potential corresponding to the peak of the synaptic conductance was shifted significantly toward a more depolarized potential by a mean value of 2.4 ± 0.5 mV (P < 0.05).
The conductance decomposition analysis further shows a significant decrease in the peak and in the mean values of the inhibitory conductance component (reduction by 29 ± 6% [P < 0.01] and 24 ± 4% [P < 0.05], respectively), whereas the excitatory conductance component remained unchanged (Fig. 7B). The relative part taken by the excitatory component in the global change in conductance increased significantly by 22 ± 3% (P < 0.05). This differential modulation of excitatory and inhibitory conductance components induced an increase in the mean depolarization and a decrease in the amplitude of the hyperpolarization in the reconstructed current clamp voltage trace (compare the 2 traces in the lower row in Fig. 7A).
DHβE had a weaker effect on the response to the second stimulation when the paired-pulse protocol was used (Fig. 7A). Over the whole population of recorded cells, the parameters that characterized the response to the second stimulation remained statistically unchanged in the presence or the absence of DHβE in the bath. Consequently, the paired-pulse ratio of the inhibitory conductances was increased in the presence of DHβE by 24 ± 10% (n = 7, P < 0.05, control = 69 ± 5% vs. DHβE = 82 ± 10%) compared with the control situation.
DHβE did not change the global synaptic conductance in layer 5 pyramidal neurons when the electrical stimulation was applied in WM. For the whole population of tested pyramidal neurons (n = 9,) neither the peak value of the global conductance, the excitatory and the inhibitory conductance components, nor the mean values of the excitatory and inhibitory conductance components were significantly modified (Fig. 7C).
We conclude that the apparent dominant effect of DHβE could be mediated by α4β2 nicotinic receptors located on GABAergic interneurons projecting from superficial cortical layers on layer 5 pyramidal neurons. The activation of these α4β2 nicotinic receptors by endogenous ACh may enhance the release of GABA and its functional impact at the level of the recorded layer 5 pyramidal cell. On the contrary, GABAergic interneurons probably located in layers 5 and 6 are probably devoid of α4β2 nicotinic receptors because DHβE had no effect on responses of layer 5 pyramidal neurons evoked by electrical stimulation in WM assuming that this stimulation can also activate cholinergic fibers (Butcher et al. 1993).
Involvement of Functional Muscarinic Receptors
To check the possible involvement of muscarinic receptors in the recruited network by layer 1 stimulation, muscarine, an agonist of the muscarinic receptors, was added in the external medium. A typical experiment and the data on the whole neuronal population (n = 9) are illustrated in Figure 8. Muscarine (5 μM) decreased the peak of the synaptic conductance by 62 ± 16% compared with the control condition (P < 0.01; Fig. 8A,B). The analysis of this response showed that the decrease was due to the modulation of both excitatory and inhibitory conductances. Their amplitude decreased, respectively, by 57 ± 14% (P < 0.01) and 69 ± 10% (P < 0.01) and their mean decreased by 51 ± 9% (P < 0.001) and 70 ± 6% (P < 0.001) in regard with the control situation. The reconstructed current clamp voltage trace showed that the amplitude of the depolarizing phase of the response largely decreased (by 59 ± 10%, P < 0.001), whereas the amplitude of the hyperpolarization slightly decreased (Fig. 8A). Paired-pulse protocols gave identical levels of modulatory effects of muscarine on synaptic responses to the first and to the second stimulation pulses, which suggests a lack of action of muscarine on the polysynaptic inhibition responsible of the depression of the second test response.
Atropine (10 μM), the nonselective antagonist of muscarinic receptors, increased the inward current and decreased the outward current (see example Fig. 9A). Surprisingly, at the global population level, the peak of the synaptic conductance was not significantly changed when compared with the control condition (Fig. 9B). This absence of a reliable effect is probably due to a balance between modulatory effects on both inhibitory and excitatory components. Indeed, the inhibitory conductance component amplitude was decreased by 22 ± 8% (P < 0.02, n = 17). In contrast, the amplitude of the excitatory conductance component was increased by 67 ± 30% (P < 0.05) and its mean by 164 ± 70% (P < 0.01). The reconstructed voltage trace showed that the amplitude of the depolarizing phase increased by 147 ± 61% (P < 0.05) and the mean depolarization increased by 190 ± 49% (P < 0.005). The decrease in the conductance changes evoked by the second stimulation pulse (by 31 ± 12%) is mainly due to a decrease of inhibition (polysynaptic inhibition), whereas excitation appears unchanged.
These results suggest a localization of functional muscarinic receptors on both glutamatergic and GABAergic neurons. Muscarinic activation changes the balance in favor of inhibition by concomitantly increasing GABA release and decreasing glutamate release. This dual modulation certainly implies the activation of different types of muscarinic receptors, a point which will be developed in a further study.
The main finding of this paper is that the release of endogenous ACh by the stimulation of layer 1 can modulate the synaptic drive of cortical layer 5 pyramidal neurons through a diversity of regulatory effects affecting in fine the balance between excitation and inhibition. Using a pharmacological approach based on the application of nicotinic antagonists in vitro, we provide evidence for a modulation of glutamatergic and/or GABAergic transmission via the activation of nicotinic receptors by endogenous ACh. The observation of differential effects of DHβE and mecamylamine on the modulation of synaptic integration in layer 5 pyramidal cells are in favor of the involvement of at least 2 different subtypes of nicotinic receptors. Moreover, the activation of muscarinic receptors must be considered although their precise characterization was out of the scope of this paper. Thus, the endogenous release of ACh can at the same time upregulate glutamate and GABA inputs to the recorded cell. However, it can be noticed that the functional balance between these apparently antagonistic effects is in favor of a decrease of layer 5 pyramidal cell excitability. In view of the complexity of the cortical network, several points ought to be discussed to further analyze the neuromodulatory effect of ACh.
Evoked Endogenous ACh Release
The presence of cholinergic fibers in layer 1 of the cortex was previously demonstrated by morphological studies in the visual cortex of the cat or of the rat (Parnavelas et al. 1986; Butcher et al. 1993; Siciliano et al. 1999; Mechawar et al. 2000). The implication of released ACh from cholinergic fibers was also proposed on the basis of electrophysiological studies of long-term plasticity in the rat pyriform cortex (Hess and Donoghue 1999). In early development of the rat visual cortex, cholinergic inputs play a critical role by synchronizing oscillatory activity in the alpha–beta frequency range acting as a functional template for the development of early cortical columnar networks (Dupont et al. 2006; Hanganu et al. 2007). We confirm that visual cortex of rat contains cholinergic fibers mostly located in layers 1 and 2. We show in addition that the electrical stimulation of layer 1 produces an effective release of ACh. From morphological observations, Descarries et al. (1997)—but see Turrini et al. (2001)—proposed that spontaneously released ACh could diffuse in cortical layers until reaching its targets. However, although we cannot exclude a “basal” modulatory effect on the cortical network by a spontaneous tonic release of ACh, diffusion of ACh may not be as important in our experimental conditions because slices are continuously perfused. Indeed, we noticed that ACh concentration that we measured just after stopping the perfusion was lower than that we measured a few minutes later when ACh concentration at the surface of the cortical layer reached a steady state in the lack of perfusion. We conclude that it is the ACh release evoked by layer 1 electrical stimulation which is the most likely responsible of the reported event-driven modulatory effects under our experimental conditions (i.e., perfusion of the slices).
The stimulation of layer 1 recruits glutamatergic, GABAergic, and cholinergic fibers. These activations induce a complex synaptic response pattern in pyramidal neurons of layer 5 remarkably homogenous across target cells, in the balance between excitatory (16%) and inhibitory (84 %) conductance components. We indeed reported that we could not detect a direct cholinergic input to layer 5 pyramidal neurons when stimulating electrically layer 1. We checked, however, the presence of cholinergic receptors on these neurons by applying ACh exogenously in the bath as done in numerous previous studies. During bath application of ACh, electrical stimulation in layer 1 or WM induced a depolarization of the pyramidal neurons, without changing their input membrane resistance (E. Lucas-Mennier, C. Monier, M. Amar, G. Bunx, Y. Fregnuc, P. Fossier, unpublished data). This effect could be due to the activation of muscarinic receptors negatively coupled to K+ channels (Murakoshi 1995), leading to the depolarization of the neuron resting potential. If these channels are located far from the soma, an increase in the membrane resistance may not be detected when patch clamping the soma, as discussed by Kimura and Baughman (1997). Several possible schemas might explain why electrical stimulation of layer 1 gives no direct detectable synaptic effects 1) cholinergic afferents may synapse pyramidal neurons on distal dendrites undetectable from the recording site (Casu et al. 2002) and 2) alternately, cholinergic fibers may not contact the pyramidal neurons of layer 5. Consequently, the modulatory effects of ACh may target the cortical network recruited presynaptically to the recorded pyramidal neurons. The comparison with WM stimulation suggests also that the cholinergic receptor distribution is not uniform across the various stages of synaptic integration to be crossed before reaching layer 5 cells.
Modulatory Role of Nicotinic and Muscarinic Receptors
Bath application of an agonist of nicotinic receptors, cytisine, showed that functional nicotinic receptors could still be recruited to facilitate glutamate release (i.e., the excitation) following layer 1 stimulation. Mecamylamine, an antagonist of the nicotinic receptors, decreased the excitatory drive, and this result is consistent with the effect observed under cytisine. However, these drugs did not change the test response to the same extent. This could be due to a difference in the density/distribution of nicotinic receptors involved and/or to a difference in the affinity of cytisine and mecamylamine on the various nicotinic receptor subtypes (Boyd 1997). Cytisine is known to activate nicotinic receptors containing β4 subunits (Sargent 1993; Papke and Heinemann 1994) and particularly α3β4 nicotinic receptors (Albuquerque et al. 1997), whereas it is relatively ineffective in activating nicotinic receptors containing β2 subunit (Luetje and Patrick 1991; Smith et al. 2007). Low concentrations of mecamylamine preferentially block α3β4 receptors (Alkondon and Albuquerque 1993). Thus our results, showing that cytisine and mecamylamine have antagonistic effects, favor the implication of α*β4 subtype and probably α3β4 subtype nicotinic receptors on synaptic terminals of glutamatergic neurons afferent to the recorded pyramidal neuron (Fig. 10).
However, in contrast with the reports by Porter et al. (1999) in motor cortex and by Xiang et al. (1998) in visual cortex, we did not observe any modulation of the GABA release in the presence of cytisine or mecamylamine. These differences could be due to 1) the age of the studied rats in which stabilization of the cholinergic pathway is known to occur after P19 (Aramakis and Metherate 1998); 2) intrinsic differences between the studied cortical areas; 3) a heterogeneity in the distribution of interneurons expressing nicotinic receptors responsible for the enhancement of the GABA release, with a possible absence of recruitment of such neurons by layer 1 stimulation (Xiang et al. 1998; Porter et al. 1999); or 4) to the presence of other subtypes of nicotinic receptors on GABAergic interneurons.
The present study indicates an absence of direct activation of cholinergic receptors located on the recorded pyramidal neurons, thus raising the possibility that nicotinic effects are found at the level of neurons located presynaptically with regard to the recorded pyramidal neuron. The paired-pulse protocol, combined with pharmacological tools, was used here to test the possible implication of a modulation of the release mechanism in the presynaptic terminal by searching for a differential modulation of the first versus the second synaptically evoked test responses. The first and the second synaptically evoked responses were equally modulated by cytisine or mecamylamine. So, it can be assumed that the presynaptic activation of α*β4 nicotinic receptors does not interfere with presynaptic mechanisms leading to the classical depression of the second response in regard to the first under our experimental conditions. Nicotinic receptors activated by cytisine have been localized on presynaptic terminals of glutamatergic neurons (Vidal and Changeux 1993; Gioanni et al. 1999); thus, we propose that α*β4 nicotinic receptor are located on presynaptic terminals of glutamatergic fibers afferents to the recorded pyramidal neuron.
Activation of α7 nicotinic receptors has a strong modulatory effect in the hippocampal network through the activation of GABAergic interneurons (Jones and Yakel 1997) and can also enhance glutamate release facilitating the induction of long-term potentiation (Yamamoto et al. 2005). However, the activation of α7 receptors has no effect on the synaptic response induced by the layer 1 stimulation in the visual cortex. This underlines an absence of recruitment of these receptors by ACh released after layer 1 stimulation or the lack of detection of a modulatory effect. A possible reason might again be linked to the age of the preparation because the presence of α7 receptors is largely decreased in P20-25 old rats compared with earlier stages of postnatal development (Broide et al. 1995, 1996; Aramakis and Metherate 1998).
DHβE decreases the inhibitory conductance component and, in some neurons, enhances the duration of the activation of excitatory conductances, although this latter effect was not found to be significant for the whole neuronal population. This diversity might be explained by a variable level of unmasking of excitation by the DHβE-induced depression of inhibition. Assuming that DHβE has a preferential selectivity to the α4β2 nicotinic receptor subtype (Alkondon and Albuquerque 1993; Gioanni et al. 1999), our results are in favor of a preferential localization of α4β2 nicotinic subtype receptors on GABAergic interneurons with a localization at a presynaptic level (Zhu and Chiappinelli 1999) and/or at a somatic level (Christophe et al. 2002). This latter localization is consistent with the activation of a nicotinic current in GABAergic neurons of superficial layers in rat cortex (Xiang et al. 1998; Christophe et al. 2002) and in human cortex (Alkondon et al. 2000). These results are substantiated by our observations showing that in the presence of DHβE, there is no modulation of the synaptic drive in layer 5 pyramidal neurons when the WM is stimulated. They are also corroborated by morphological and functional differences between GABAergic interneurons through the cortical layers (Markram et al. 2004). Moreover, the delay of the inhibition (around 7.5 ms) induced by layer 1 stimulation (same value in the presence of glutamatergic blockers, data not shown) suggests a disynaptic activation of inhibition by nonglutamatergic fibers, for instance cholinergic fibers. DHβE prevents the depression of synaptic responses to the second stimulation in the paired-pulse protocol. However, we assume that the activation of α4β2 nicotinic receptors exclusively located on GABAergic neurons facilitates the release of GABA, that is, the inhibition. The consequence might be an enhanced depression of the synaptic response to the second stimulation in the presence of DHβE. One explanation is linked to the fact that DHβE is a competitive antagonist of α4β2 nicotinic receptor (Dwoskin et al. 2000) which can be displaced from the receptor by accumulation of released ACh in response to the second stimuli. We conclude that, in the network activated by the layer 1 stimulation, GABA release could be modulated by presynaptic α4β2 receptors and/or induced by the activation of somatic α4β2 receptors (Fig. 10).
To summarize, 2 types of nicotinic receptors are activated by endogenous ACh and are involved in the modulation of the layer 1 activated network: one type (noted N1 in Fig. 10, probably the α*β4 subtype) is probably located on some presynaptic terminals of glutamatergic neurons and, as a result, would increase the excitation. The second type (noted N2, probably the α4β2 subtype) would activate the release of GABA possibly via a cholinergic synapse situated on the soma of GABAergic interneurons presynaptic to the recorded cell and possibly via a modulation of GABAergic interneurons activated by glutamatergic fibers (Fig. 10).
Muscarine decreased the synaptic response by depressing excitation and inhibition as already observed (Vidal and Changeux 1993; Kimura and Baughman 1997), but atropine did not induce opposite effects as expected for inhibition. The paired-pulse protocol showed an increase of the depression of the second response in the presence of atropine. One possibility is the presence of several types of muscarinic receptors (Caulfield and Birdsall 1998) in the activated neuronal network with a differential distribution on GABAergic interneurons and on glutamatergic fibers (Gulledge et al. 2007). One type would induce a reduction in the release of GABA and would not be recruited by the electrical stimulation of layer 1 but only by muscarine (Vidal and Changeux 1993). The other type would enhance the release of GABA and would be activated by the release of ACh following layer 1 stimulation. Such a possibility explains the effect of atropine on the paired-pulse experiment. The pharmacological characterization of the different types of muscarinic receptors on glutamate fibers and on GABAergic interneurons will constitute the next step of this work.
The modulation of the thalamocortical pathway by the cholinergic system was studied by Gil et al. (1997) who applied various cholinergic agonists. These authors showed that excitatory postsynaptic potentials induced by stimulation of the thalamocortical pathway are increased by nicotine. According to their study, the action of nicotine would be exerted via the activation of α7 nicotinic receptors. In our experimental situation, the nicotinic modulation that we observed by electrical activation of the superficial layers is not exerted by the activation of α7 nicotinic receptors but rather by the subtypes α4β2 and possibly α*β4. This discrepancy in the structural substrate of the ACh effect may imply 3 possible mechanisms: 1) the sign of the regulation by nicotinic receptors may differ between visual cortex and somatosensory cortex, 2) and/or the sign of the modulatory effect may differ following activation of the feedforward thalamocortical pathway versus the horizontal and recurrent “corticocortical” pathway, or 3) the resulting receptor activation was undiscriminative in terms of subnetwork selection and layer targeting (because bath application of cholinergic agonists were used in the study of Gil et al. ). This situation differs from our protocol, where the study of cholinergic modulation is restricted to the activation of the final stages of the neuronal pathway activated by visual inputs.
When electrical stimulation is applied to the superficial cortical layers, the action of endogenous ACh on the α4β2 and α*β4 nicotinic receptors seems to increase both the release of GABA and of glutamate. In order to orchestrate a coherent sign in the modulatory effect, a differential regulation of both types of nicotinic receptors is to be expected. This hypothesis is supported by looking at the functional impact of the pharmacological activation of these receptor subtypes: 1) the effect of mecamylamine is weak compared with that of DHβE suggesting a weaker implication of α*β4 receptors than α4β2 receptors in the activated network and 2) the localization of α4β2 receptors is found preferentially on GABAergic interneurons whereas that of α*β4 receptors is on glutamatergic cells.
These various lines of argument lead us to conclude that, when endogenous ACh is released in the network following electrical stimulation of layer 1, its dominant effect on nicotinic and muscarinic receptors is to amplify the dominance of the drive of inhibitory inputs and thus decrease the excitability and sensory responsiveness of layer 5 pyramidal neurons. Our observations also illustrate the versatility of the role of the cholinergic input and suggest that the strength and the sign of its functional impact may depend specifically on the preexisting balance between excitation and inhibition in the recruited subnetworks. Thus, in view of the diversity (and possibly antagonism) that exists between cholinergic-based processes upregulating at the same time excitation and inhibition, it is likely that a wide spectrum of functional effects may be expected for local circuits that subserve distinct cortical-based computations.
Conseil Général de l'Essonne (France). The Délégation Générale pour l'Armement (France); the Fondation pour la Recherche Médicale (France); and the Institut Lilly (France) to E.L.M.
We thank Drs Kirsty Grant and Andrew Davison for helpful comments. Conflict of Interest: None declared.
- gamma-aminobutyric acid
- cholinergic agents
- cholinergic agonists
- depressive disorders
- electric stimulation
- muscarinic acetylcholine receptor
- nicotinic receptors
- visual cortex