Subplate neurons play an important role in early cortical development. To investigate whether these transient neurons receive synaptic inputs, we performed whole-cell recordings from visually identified and biocytin-labeled subplate cells in somatosensory cortical slices from postnatal day 0–3 rats. Subplate neurons had an average resting membrane potential of –55 mV and input resistance of ~1.1 GΩ. Suprathreshold current injection elicited in 67% of the cells repetitive action potentials at 4–13 Hz and the remaining 33% showed only one spike. Three classes of spontaneous postsynaptic currents (sPSCs) could be identified: (i) Fast sPSCs, with an average amplitude of 14 pA and a decay time of 6.3 ms, which showed a 95% decrease in their frequency during (±)-γ-amino-3-hydroxy-5-methyl-isoxazole-4-propionic acid (AMPA)/kainate receptor blockade. Cyclothiazide caused a 3.5-fold increase in the decay time, indicating that fast sPSCs were mediated by AMPA receptors. (ii) Slow sPSCs, with 18 pA amplitude and 51.2 ms decay time were blocked by the N-methyl-d-aspartate (NMDA) receptor antagonist CPP. (iii) Chloride-driven sPSCs, with 34.4 pA amplitude and 123 ms decay time that were blocked by the γ-amino-butyric acid A (GABAA) receptor antagonist gabazine. While tetrodotoxin citrate (TTX) blocked completely NMDA-mediated slow sPSCs, the frequency of AMPA- and GABAA-mediated sPSCs was reduced in TTX by 55 and 90%, respectively. These results indicate that subplate neurons receive functional synaptic inputs mediated by AMPA, NMDA and GABAA receptors.
The subplate represents a transient layer in the developing cerebral cortex that is located directly under the cortical plate and consists of a heterogeneous neuronal population according to morphology and neurotransmitter identity (Marín-Padilla, 1970; Kostovic and Rakic, 1980, 1990; Luskin and Shatz, 1985). Inverted pyramid-like and horizontal cells as well as polymorphic neurons with different shapes and spiny or smooth dendrites have been classified as subplate neurons (Kostovic and Rakic, 1980; Wahle et al., 1987; Valverde et al., 1989). Due to their earlier onset of maturation, subplate neurons show a relatively large dendritic tree when compared with the pyramidal neurons of the cortical plate (Mrzljak et al., 1992). In rodents, subplate neurons with horizontal dendrites invading the underlying intermediate zone and ascending dendrites extending into the cortical plate have been described (Del Río et al., 2000).
Subplate cells play an important role in the pathfinding of corticopetal and corticofugal axonal projections. Axons arising from subplate neurons pioneer the corticofugal pathway and have been proposed to form a cellular scaffold for guiding thalamocortical axons (McConnell et al., 1989; Ghosh et al., 1990) [for reviews see Allendoerfer and Shatz, and Molnár (Allendoerfer and Shatz, 1994; Molnár, 1998)]. Subplate neurons receive a transient synaptic input from ‘waiting’ thalamic axons (Lund and Mustari, 1977; Rakic, 1977, 1983; Ghosh and Shatz, 1992) and disappear by cell death or transform into other neuronal shapes during postnatal development (Kostovic and Rakic, 1980; Luskin and Shatz, 1985; Valverde and Facal-Valverde, 1988; Valverde et al., 1989; Meyer et al., 1992) [for a review see Supèr et al. (Supèr et al., 1998)]. Early deletion of subplate neurons in kitten visual cortex prevents the segregation of thalamocortical axons within layer IV and the formation of ocular dominance columns (Ghosh and Shatz, 1992). Further evidence for a crucial role of subplate neurons in corticogenesis comes from the recently described Coup-tfI (chicken ovalbumin upstream promoter-transcription factor I) mutant mouse, which shows improper differentiation and premature cell death of subplate neurons (Zhou et al., 1999). In this mutant, the defects in the subplate result in the failure of guidance and innervation of thalamocortical projections and in the absence of layer IV (Zhou et al., 1999).
Despite these reports indicating an important role of the subplate in early cortical development, only few electrophysiological studies on subplate neurons have been performed so far (Friauf et al., 1990; Friauf and Shatz, 1991). In vitro intracellular recordings and current-source density analyses in late embryonic and early postnatal kitten visual cortex demonstrated that subplate neurons show various action potential firing patterns and receive functional excitatory synaptic inputs from axons that course in the developing white matter (Friauf et al., 1990; Friauf and Shatz, 1991). However, the questions of which types of excitatory receptors and whether functional γ-amino-butyric acid (GABA) receptor-mediated synaptic inputs can be demonstrated in subplate neurons have not been addressed. Ultrastructural studies of subplate neurons have demonstrated symmetrical as well as asymmetrical synapses with relatively mature properties in various species (Kostovic and Rakic, 1980; König and Marty, 1981; Chun and Shatz, 1988; Herrmann et al., 1994), indicating that subplate neurons receive inhibitory and excitatory synaptic inputs during early cortical development. As suggested by Kostovic and Rakic (Kostovic and Rakic, 1980), excitatory inputs onto subplate neurons may arise from the thalamus and other cortical areas, whereas inhibitory synaptic inputs may originate from GABAergic subplate neurons. Thalamocortical synaptic contacts with spines and shafts of subplate neuron dendrites have been demonstrated in the neonatal ferret (Herrmann et al., 1994) and a dense network of corticocortical fibers have been reported in the subplate of the embryonic mouse (Crandall and Caviness, 1984). N-Methyl-d-aspartate (NMDA), (±)-γ-amino-3-hydroxy-5-methyl-isoxazole-4-propionic acid (AMPA), kainate receptors (Smith and Thompson, 1999) and the essential subunits for receptor function (Herrmann, 1996; Aoki, 1997; Catalano et al., 1997; Furuta and Martin, 1999) have been demonstrated in the subplate of various species, suggesting the presence of functional glutamatergic synapses in subplate neurons. The expression of benzodiazepine binding sites (Schlumpf et al., 1983), GABAA receptors (Huntley et al., 1990) and GABAA receptor subunits (Meinecke and Rakic, 1992) in the subplate strongly indicate that functional GABAergic synaptic inputs should be present in subplate neurons.
We performed whole-cell recordings from visually identified and biocytin-labeled subplate neurons in neonatal rat cortical slices to study the kinetics and pharmacology of spontaneous synaptic inputs onto subplate neurons. We found three distinct groups of spontaneous postsynaptic currents (sPSCs) mediated by AMPA, NMDA and GABAA receptors. Part of this work appeared in abstract form (Hanganu et al., 2000).
Materials and Methods
Experiments were performed on 400 μm whole brain coronal slices prepared from postnatal day (P) 0–3 (day of birth = P0) male Wistar rats. All experiments were conducted in accordance with the national laws for the use of animals in research and approved by the local ethical committee. The rat pups were anesthetized by hypothermia and decapitated. The brain was rapidly removed and immediately immersed in ice-cold artificial cerebrospinal fluid (ACSF) that contained (in mM): 125 NaCl, 26 NaHCO3, 3 KCl, 1.6 CaCl2, 1.8 MgCl2, 1.3 NaH2PO4 and 20 d-glucose (pH 7.4 after equilibration with 95% O2/5% CO2, osmolarity 333 mOsm). Three to four slices including the primary somatosensory cortex were prepared on a vibroslicer (TPI, St Louis, MO), cut in the middle to divide them into hemispheres and transferred to an incubation chamber containing ACSF at 33°C. After an incubation period of at least 1 h, slices were transferred to a recording chamber (volume <2 ml), where they were continuously superfused at a rate of 2 ml/min with ACSF at 33°C. A thin nylon net was placed on the slice in the recording chamber to increase mechanical stability. The slices were used up to 7 h without any apparent changes in the electrophysiological properties of the cells.
Techniques for performing whole-cell recordings in neonatal rat cortical slices were similar to those described previously (Kilb and Luhmann, 2000; Luhmann et al., 2000). Cells in the subplate were visualized with video-enhanced infrared Nomarski optics (Fig. 1A) in a recording chamber mounted on the fixed stage of an Axioskop microscope (Zeiss, Jena, Germany) and digitized online using a frame grabber card (Screen Machine II, Munich, Germany). Tight-seal whole-cell recordings were obtained in slices similar to those described by Dodt and Zieglgängsberger (Dodt and Zieglgängsberger, 1990). Recording pipettes were pulled from borosilicate glass tubing (CG200F8P, Science Products, Hofheim, Germany) on a vertical puller (PP83, Narishige, Tokyo, Japan). When filled with the intracellular solutions, pipette resistance ranged from 4 to 9 MΩ. Capacitance artifacts and series resistance were minimized using the built-in circuitry of the patch-clamp amplifiers (EPC9, Heka, Lambrecht, Germany and SEC-05L, NPI, Tamm, Germany). In the experiments in which we used the SEC-05L amplifier, signals were amplified and low-pass filtered at 3 kHz, digitized online with an AD/DA-board (ITC-16, Heka), recorded and analyzed using WinTida software (Heka). When using the EPC9 amplifier, the signals were low-pass filtered with a Bessel filter at 2.9 kHz, recorded online and analyzed with WinTida Software.
The standard electrode solution contained (in mM): 117 K-gluconate, 13 KCl, 1 CaCl2, 2 MgCl2, 11 EGTA, 10 K-HEPES, 2 NaATP and 0.5 NaGTP. Electrodes used for the investigation of GABAA receptor-mediated synaptic currents were filled with a solution containing (in mM): 130 KCl, 1 CaCl2, 2 MgCl2, 11 EGTA, 10 K-HEPES, 2 NaATP and 0.5 NaGTP. Both intracellular solutions were adjusted to pH 7.4 with 1 M KOH and to an osmolarity of 306 mOsm with sucrose. In all experiments 0.5% biocytin (Sigma-Aldrich, Steinheim, Germany) was included in the electrode solutions for later cell identification. All potentials were corrected for liquid junction potentials with –10 mV for the gluconate-based electrode solution (Mienville and Pesold, 1999) and –4 mV for the high chloride electrode solution (Marty and Neher, 1995).
All substances were purchased from Merck with the exception of 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX) which was from Tocris (Ballwin, MO), (±)- or R(–)-3-(2-carboxypiperazin-4-yl)-propyl-1-phosphonic acid (CPP), SR-95531 (gabazine), cyclothiazide (CYZ) and tetrodotoxin citrate (TTX) from RBI (Natick, MA). Stock solutions of these drugs were prepared as follows: TTX and CPP in distilled water, CNQX, gabazine and CYZ in dimethylsulfoxide (DMSO). Stock solutions were stored at –20°C and diluted in ACSF on the day of experiment. Maximal concentration of DMSO in the superfusate was 0.1%. The drugs were bath applied after stable control recordings were obtained for at least 6 min. Drugs reached synaptic location within the first 5 min of application. Generally, a drug-free ACSF wash was applied before and after each drug application. A partial or a complete washout was obtained after 5–45 min. During an experiment, the series resistance was continuously controlled. Typical recordings lasted 30–60 min and up to four drugs were tested on the same cell.
The passive electrophysiological properties of the cells and their firing patterns were characterized under current-clamp conditions. The resting membrane potential was measured immediately after obtaining the whole-cell configuration. For the determination of the input resistance and membrane time constant, hyperpolarizing current pulses of 300 ms duration and 20 pA amplitude were applied from a holding potential of –70 mV. The input resistance was calculated according to Ohm's law by dividing the maximal resulting potential changes by the amplitude of the injected current. The membrane time constant (t) was calculated by fitting a monoexponential function of the type a + b × exp(time/t) to the induced potential deflection. Active membrane properties and current voltage relationships were assessed by recording responses to a series of 2 s long hyper- and depolarizing current pulses from a holding potential between –60 and –70 mV. Spike amplitude was measured from action potential threshold to the peak.
Analysis of spontaneous postsynaptic currents (sPSCs) was performed using the Mini Analysis Software (Synaptosoft, Leonia, NJ). The sPSCs were captured using a threshold-crossing detector set above the noise level. Events that did not show a typical sPSC waveform were rejected manually and by optimal settings of the program parameters rise time, decay time, area and baseline. Peak amplitude, inter-event interval, rise time and decay time of the sPSCs were determined. Rise time was measured as the time between points at 0.5% and peak current amplitude and the decay time was determined between time points at 100 and 37% of the peak. Decay time constants (τ) were calculated by fitting single or double exponential functions to averaged sPSCs using a simplex algorithm. For statistical analyses, sPSCs measurements were calculated as amplitude or decay time distribution histograms using Origin 6 (Microcal Software, Northampton, MA). Correlation coefficients (r) were calculated using a least-squares linear regression analysis. Data are presented as mean ± SEM. For statistical analyses the two-tailed Student's t-test, one-way ANOVA test and χ2-test were used. Probability levels of P < 0.01 and P < 0.001 were considered significant.
For staining putative subplate neurons, all recordings were made with biocytin-containing solutions. Intracellular diffusion of biocytin from the electrode followed by active transportation resulted in adequate filling of the recorded neurons (Fig. 1B–D). At the end of the recording session the electrode was slowly withdrawn, the slice was fixed overnight in 4% paraformaldehyde containing phosphate buffer and processed by a modification of the protocol described by Schröder and Luhmann (Schröder and Luhmann, 1997). Slices were pre-incubated for 60 min in phosphate-buffered saline (PBS) containing 0.5% H2O2 to saturate endogenous peroxidase, followed by incubation overnight in avidin- coupled peroxidase (ABC kit, Vectorlabs, Burlingame, CA). After being washed in PBS and TRIS the slices were incubated for 30 min in 20% 3,3-diaminobenzidine (DAB, Sigma-Aldrich, Taufkirchen, Germany) and for 10 min in DAB containing 0.01% H2O2. The reaction product was intensified by 2–3 min incubation in 0.5% OsO4. Finally, the slices were rinsed in TRIS and distilled water, dehydrated in ethanol and propylenoxide and embedded in Durcopan (Fluka, Buchs, Switzerland). In general, one or two cells per slice were labeled with little or no background staining. Biocytin stained cells were photographed with a Zeiss Axioskop and some cells were reconstructed by camera lucida drawings.
Whole-cell recordings were performed from 152 subplate neurons in somatosensory cortical slices from P0–3 rats. Since no significant age-dependent differences in the membrane properties and synaptic inputs could be observed between cells recorded in P0–3 rats, data from this age group were pooled. The subplate neurons analyzed in this study showed a morphology as previously described (Valverde et al., 1989) in perinatal rat cerebral cortex. The location of the neurons within the subplate directly under the cortical plate, their appearance under video-assisted Nomarski microscopy and the morphology of the biocytin-stained cells (n = 75) served as criteria to identify the subplate neurons. The majority of the subplate neurons considered in this study showed a more horizontal orientation, a ramified dendritic tree and ascending and descending axonal collaterals (Fig. 1B–D).
Intrinsic Membrane Properties
Subplate neurons exhibited relatively uniform passive membrane properties. Using gluconate-based electrode solution, the average resting membrane potential (RMP) and input resistance (Rin) was –55 ± 0.5 mV and 1173 ± 47 MΩ (n = 131). Similar values were measured when using a chloride-based electrode solution (RMP = –52.7 ± 1.2 mV, Rin = 1143 ± 150 MΩ, n = 21). These values are similar to those reported for Cajal–Retzius cells (Mienville and Pesold, 1999) and other types of neurons in the neonatal cortex (LoTurco et al., 1991; Kim et al., 1995). All subplate neurons were capable of firing overshooting action potentials in response to sustained depolarization by intracellular current injection. At the resting membrane potential, six of the 11 cells tested showed spontaneous action potentials. In agreement with previous observations (Friauf et al., 1990), subplate cells revealed different firing patterns in response to injection of suprathreshold current pulses. Repetitive firing at frequencies between 4.5 and 13 Hz could be observed in 67% of the 131 neurons tested. In accordance with Friauf et al. (Friauf et al., 1990), these cells were classified as regular spiking subplate neurons (Fig. 2A). The remaining 33% of the cells, which were classified as single spiking neurons (Friauf et al., 1990), responded to suprathreshold current injection with a single action potential followed by membrane oscillation (Fig. 2C). The firing patterns of the cells were independent of the electrode solution used. Furthermore, the proportion of cells showing single spiking (P0: 34%, n = 32; P1: 42%, n = 55; P2–3: 30%, n = 30) and regular spiking (P0: 66%; P1: 58%; P2–3: 70%) discharge was not dependent on the age. Neither the passive membrane properties nor the action potential amplitudes were significantly different between regular spiking (RMP = –56 ± 0.8 mV, Rin = 1181 ± 59 MΩ, n = 88; 52 ± 1.6 mV, n = 35) and single spiking subplate cells (RMP = –54.3 ± 0.8 mV, Rin = 1146 ± 57 MΩ, n = 43; 49.9 ± 2.9 mV, n = 10). In both groups the voltage deflection showed either linear dependence or time-independent inward rectification from the injected current (Fig. 2B,D). No significant correlation could be noted between the firing type and the presence of inward rectification.
Characterization of sPSCs
Voltage-clamp recordings at a holding potential of –70 mV revealed in 89% of the cells sPSCs (Fig. 3). The frequencies of these sPSCs varied considerably among the cells, ranging from 0.1 to 0.9 Hz (0.27 ± 0.02 Hz, n = 117 cells). The average amplitude of the sPSCs was 15 ± 0.6 pA (n = 117 cells). The peak-to-peak noise level varied between 2 and 7 pA, and was below the smallest sPSCs amplitude. In 68% of the cells that showed sPSCs, the decay times of the sPSCs were either fast or slow, and across those cells gave rise to a bimodal distribution (Fig. 3C). In these neurons, decay time distribution histograms served to determine the limit between fast and slow sPSCs, which was between 10 and 20 ms for all cells studied. The properties of the fast and slow sPSCs are summarized in Table 1. In comparison to the slow sPSCs, fast sPSCs occurred approximately twice as often and revealed a significantly smaller average amplitude and faster rise time. The fast sPSCs decay could be best fitted with a monoexponential function, while the slow sPSCs decay was fitted with a biexponential function (Fig. 3B). Fast and slow sPSCs showed an unimodal amplitude distribution (Fig. 3D) and could be observed in regular spiking as well as in single spiking subplate neurons.
The remaining 32% of the subplate neurons gave rise to a unimodal decay time distribution of fast sPSCs only (data not shown). In these cells, all sPSCs decayed with a single exponential time constant and revealed a decay time of <20 ms. This, in addition to their average frequency (0.22 ± 0.02 Hz), amplitude (13.7 ± 1 pA), rise time (2.3 ± 0.2 ms) and decay time (6.3 ± 0.5 ms, n = 37) led us to the conclusion that these sPSCs belong to the fast population.
A possible effect of dendritic cable filtering on fast and slow sPSCs was examined in 21 neurons by plotting the rise time versus amplitude (Fig. 3E) and the rise time versus decay time for fast and slow sPSCs (Fig. 3F). These plots show neither a significant correlation between the distribution of sPSC rise time and amplitude, nor between rise time and decay time, suggesting that dendritic filtering does not profoundly influence the kinetics of the sPSCs.
Glutamate Receptor-mediated sPSCs
To identify the nature of the receptors which mediate the fast and slow sPSCs, we examined the effects of the AMPA/kainate receptor antagonist CNQX (10 μM, n = 19 cells) and of the selective NMDA receptor antagonist CPP (10–20 μM, n = 9 cells). Whereas the slow sPSCs were not affected significantly by CNQX, the frequency of fast sPSCs was reduced by 95 ± 2.5% (P < 0.001) (Fig. 4A,B). The effects of CNQX were not accompanied by any changes in the holding current or input resistance and were reversible. These results suggest that the majority of fast sPSCs were mediated by AMPA/kainate receptors. In order to discriminate between AMPA and kainate receptors, we investigated in six cells the sPSCs modulation by cyclothiazide, a drug known to slow the desensitization of AMPA, but not of kainate receptors (Yamada and Tang, 1993). Bath application of 100 μM cyclothiazide in the presence of CPP (20 μM) and the GABAA receptor antagonist gabazine (50 μM) caused a significant (P < 0.001) and reversible prolongation in the decay time of the fast sPSCs to 348 ± 43% of the control value (Fig. 4C,D). Cyclothiazide had no significant effect on the frequency of the fast sPSCs (control: 0.095 ± 0.02 Hz, CYZ: 0.19 ± 0.05 Hz, wash: 0.18 ± 0.05), but unexpectedly increased significantly (P < 0.05) the amplitude from 7.4 ± 0.6 to 9 ± 0.8 pA. This effect was reversible after washout of cyclothiazide (7.8 ± 1 pA) (n = 6 cells). These results indicate an almost exclusive participation of AMPA receptors on the generation of fast sPSCs.
Activation of NMDA receptors appears to be involved in the generation of both slow and fast sPSCs. Bath application of CPP reduced the frequency of the fast sPSCs by 34 ± 10.2% (P < 0.01), a result likely due to blockade of presynaptic neuron activation since amplitude (as well as rise and decay time) remained unchanged. The frequency of the slow sPSCs decreased in CPP by 96 ± 1.5% (P < 0.001) (Fig. 4E,F). In seven out of nine investigated cells no slow sPSCs could be observed after 5–10 min application of CPP, indicating that these events were mediated by NMDA receptors. This conclusion is also supported by the long decay time constants of the slow sPSCs (Table 1), as described previously for NMDA receptor-mediated synaptic events (Forsythe and Westbrook, 1988; Bellingham et al., 1998). Since CPP dissociates slowly (Benveniste and Mayer, 1991), only a partial recovery could be observed upon washout of CPP. In eight cells combined application of CPP and CNQX caused a complete blockade of the fast sPSCs and a 98 ± 1.5% reduction in the frequency of the slow sPSCs, indicating that the spontaneous activity observed in these neurons using gluconate-based solution was exclusively mediated by AMPA and NMDA receptors (Fig. 4G). After washout of CPP and CNQX the frequency of the fast and slow sPSCs recovered to 45 ± 14.6 and 37.3 ± 15.9% of the control values (n = 8 cells).
Effects of TTX on the Fast and Slow sPSCs
In order to investigate the role of action potential-dependent synaptic activity on the generation of sPSCs we studied the effects of TTX in eight subplate neurons (Fig. 5). Bath application of 1 μM TTX reduced the frequency of the fast sPSCs by 46 ± 7% (P < 0.001), indicating that about half of the population of the fast sPSCs were mediated by action potential-dependent presynaptic transmitter release. TTX had no significant effect on the amplitude distribution, rise or decay time of the remaining fast sPSCs. Similar results were obtained when the fast sPSCs were pharmacologically isolated by application of CPP. Under this condition, TTX reduced the frequency of the fast sPSCs significantly (P < 0.01) by 55 ± 4.2% (n = 5 cells) of the control values. As illustrated in Figure 5B, in all subplate neurons (n = 8) the slow sPSCs were completely blocked by TTX, suggesting that these events required action potential-dependent transmitter release. For both populations of currents, the TTX effects were completely reversible within 5 min.
GABAA Receptor-mediated sPSCs
Since previous studies have shown the expression of benzodiazepine binding sites (Schlumpf et al., 1983), GABAA receptors (Huntley et al., 1990) and GABAA receptor subunits (Meinecke and Rakic, 1992) in the subplate, we were interested in the question of whether GABAA receptor-mediated sPSCs could be also demonstrated in subplate neurons. In three cells recorded with gluconate-based electrode solution, bath application of gabazine (50 μM) had no effect on the slow or fast sPSCs. The lack of GABAA receptor-mediated sPSCs under this experimental condition results from the small driving force at an estimated equilibrium potential for chloride of about –50 mV. Increasing the chloride concentration in the patch pipette to 133 mM shifts the chloride equilibrium potential to ~0 mV, thereby displaying GABAA receptor-mediated currents at the holding potential. Using a high chloride electrode solution, the frequency of the sPSCs was reduced by addition of 10 μM CNQX and 20 μM CPP to 47.9 ± 7.5% (n = 8 cells), and the remaining sPSCs were blocked by bath application of gabazine (Fig. 6A). These GABAA-mediated sPSCs were characterized by their low frequency (0.07 ± 0.009 Hz, n = 8 cells), variable amplitudes (34.4 ± 4.7 pA) and long rise (6.1 ± 0.9 ms) and decay (123.4 ± 13 ms) times (Fig. 6B). In agreement with previous studies (Edwards et al., 1990), the decay was fitted with two exponentials with average time constants of 40.5 ± 4.8 and 211.5 ± 22.5 ms (n = 8 cells). All properties of the GABAergic sPSCs were significantly different (P < 0.001) from those of the fast and slow sPSCs when using the one-way ANOVA test. Gabazine showed a very slow washout rate and a complete recovery could be observed in only three cells after 45 min washout. Blocking the GABAA receptor-mediated sPSCs by bath application of gabazine (n = 5 cells), the NMDA and AMPA receptor-mediated sPSCs were isolated and displayed similar kinetics as those described in ACSF (Fig. 3). A combined application of gabazine, CPP and CNQX completely blocked the spontaneous synaptic activity in four out of four subplate neurons tested.
Since the long decay and rise times could indicate that GABAA receptor-mediated sPSCs originate from electrotonically distant sites and are thus influenced by cable filtering, we investigated the relationship between decay versus rise time and between amplitude versus rise time. When plotted for all GABAA receptor-mediated sPSCs recorded in 12 cells, neither decay time versus rise time nor amplitude versus rise time were correlated. However, the relatively small number of GABAergic events recorded per neuron precluded a complete analysis for each cell. In order to investigate the involvement of action potential-dependent transmitter release, we tested the effects of TTX on these GABAA receptor-mediated sPSCs in five cells. Bath application of 1 μM TTX reduced the frequency of GABAA receptor-mediated sPSCs by 89.8 ± 1.5%. The remaining sPSCs were blocked by gabazine. This result suggests that the majority of the GABAA receptor-mediated sPSCs were dependent on presynaptic action potentials.
The main results of the present in vitro electrophysiological study on morphologically identified subplate neurons in neonatal rat somatosensory cortex are as follows. (i) All subplate neurons were capable of firing action potentials in response to injection of suprathreshold depolarizing current pulses. The majority of the cells discharged repetitively between 4.5 and 13 Hz, whereas one-third of the subplate neurons responded with only a single spike. (ii) Almost 90% of the subplate cells showed sPSCs with an average amplitude of 15 pA and frequency of 0.27 Hz. (iii) The majority of these cells revealed fast and slow sPSCs, which differed in their kinetics, frequency, amplitude and pharmacology. In 32% of the subplate neurons, only fast sPSCs could be observed. (iv) The fast sPSCs were mediated by AMPA receptors and about half of this population was sensitive to TTX. (v) The slow sPSCs were action potential-dependent and mediated by NMDA receptors. (vi) A third population of chloride-driven sPSCs was mediated by GABAA receptors and mostly action potential-dependent.
Electrophysiological Properties of Subplate Cells
Our electrophysiological data support the hypothesis that subplate neurons may play an important and active role in early cortical development [for a review see Allendoerfer and Shatz (Allendoerfer and Shatz, 1994)]. In agreement with previous observations in fetal and early postnatal cat visual cortex (Friauf et al., 1990), subplate neurons showed different firing patterns to injection of depolarizing current pulses and were classified as regular spiking or single spiking cells. Single spiking cells did not differ in their active and passive membrane properties from regular spiking subplate neurons, indicating that poor cell health or recording quality does not account for the lack of repetitive firing in single spiking cells. It is also unlikely that single spiking neurons represent a population of immature cells, which during further development will transform into regular spiking cells or vice versa, since the relative proportion of cells showing single or regular spiking patterns remained constant during our observation period (P0–3). We suggest that two distinct classes of subplate neurons with different firing patterns but without obvious differences in their morphology and synaptic inputs may exist in the subplate of the newborn rat. The absence of a correlation between the firing pattern and morphology has been previously also reported for cat subplate neurons (Friauf et al., 1990).
As reported in previous studies on rat cortical neurons (Kim et al., 1995; Luhmann et al., 1999), subplate neurons reveal a relatively depolarized resting membrane potential. It seems unlikely that this property results from poor recording quality, since gigaohm seals were achieved before obtaining the whole-cell configuration and cells exhibited a high input resistance throughout the measurement. Although a depolarized resting membrane potential may be a typical feature of immature cortical neurons, it cannot be excluded that the depolarized resting membrane potential of subplate neurons contributes to the induction or progression of programmed cell death, as it has been suggested for Cajal–Retzius cells (Mienville and Pesold, 1999). Spontaneous action potentials were described previously (Blanton and Kriegstein, 1991) in the embryonic turtle cerebral cortex. Our experiments show that subplate neurons in the rat neocortex also reveal spontaneous action potentials, which may also contribute to the spontaneous synaptic activity observed in these cells.
Functional Synaptic Inputs of Subplate Neurons
Our study demonstrates for the first time that subplate neurons receive functional synaptic inputs mediated by AMPA, NMDA and GABAA receptors. Although the low frequency and moderate amplitude of the sPSCs indicate a small number of synapses and a low density of channels devoted to them, the high input resistance of subplate neurons in the range of 1.1 GΩ will amplify the efficiency of these synapses. Already a small synaptic current of 9 pA will depolarize the membrane by ~10 mV and may elicit an action potential in the cell. Since the average amplitude of the AMPA and NMDA receptor-mediated sPSCs is >14 pA, glutamatergic inputs arising from the thalamus (Rakic, 1977; Friauf et al., 1990; Herrmann et al., 1994), other cortical areas (Innocenti, 1981; Crandall and Caviness, 1984) and from intra-areal sources (Assal and Innocenti, 1993; Galuske and Singer, 1996) may easily elicit action potentials in subplate neurons following activation of AMPA and NMDA receptors. The GABAA-mediated sPSCs may have a similar depolarizing action, since previous studies have shown that GABA acts as an excitatory transmitter during early cortical development (Luhmann and Prince, 1991; Yuste and Katz, 1991; LoTurco et al., 1995; Owens et al., 1999) [for a review see Cherubini et al. (Cherubini et al., 1991)]. This GABAergic input may arise from other subplate cells via local and long-distance axonal connections (Wahle et al., 1987; Antonini and Shatz, 1990; Del Río et al., 2000). Compared with previous studies in immature rat hippocampus (Edwards et al., 1990; Taketo and Yoshioka, 2000), the GABAAergic sPSCs in subplate neurons are relatively slow. Since the slow kinetics are not the result of dendritic filtering, we suggest that developmental differences (Hollrigel and Soltesz, 1997) related to a certain subunit composition of the GABAA receptor during the perinatal period (Laurie et al., 1992) may account for the slow rise and decay times of the GABAAergic sPSCs. Beside their slow kinetics, GABAAergic sPSCs in subplate neurons appear at a relatively low rate, as previously also reported in neonatal dentate granule cells (Hollrigel and Soltesz, 1997) and neocortical neurons (Owens et al., 1999), suggesting that immature cortical networks do not express spontaneous activity at high frequencies.
The large majority of the subplate cells exhibited a mixture of AMPA and NMDA or pure AMPA-mediated sPSCs, indicating functional glutamatergic synaptic transmission at this early age. These electrophysiological data confirm previous immunohistochemical and receptor autoradiographical analyses, which showed a high expression of AMPA and NMDA receptors and their essential subunits in the subplate (Herrmann, 1996; Aoki, 1997; Catalano et al., 1997; Furuta and Martin, 1999). Previous studies have demonstrated in hippocampal CA1 (Isaac et al., 1995; Durand et al., 1996) and neocortical layer V/VI pyramidal neurons (Rumpel et al., 1998) of the immature rat so-called silent synapses, i.e. glutamatergic synapses that express only NMDA receptors but no functional AMPA receptors. Since subplate cells are generated earlier than these pyramidal neurons, it may well be that subplate neurons express synapses with more mature molecular, structural and functional properties. Using the in vitro preparation of the intact hippocampal formation, Diabira et al. recently demonstrated AMPA receptor-mediated EPSPs in the CA1 region of the rat as early as embryonic day 19 (Diabira et al., 1999). However, these AMPA-mediated PSCs could be only found in a small percentage (10%) of the CA1 pyramidal neurons and only in those cells that showed a more mature arborized dendritic tree (Gozlan et al., 1999). Interestingly, the complex and large dendritic tree is also one of the typical morphological properties of subplate neurons (Mrzljak et al., 1992; Del Río et al., 2000). In agreement with our observations in subplate neurons, Diabira et al. (Diabira et al., 1999) also observed a contribution of AMPA and GABAA receptors, but not kainate receptors, to the field EPSPs recorded in the neonatal rat CA1 region. Our experiments with cyclothiazide strongly indicate that the fast sPSCs were predominantly mediated by AMPA receptors. This suggestion is also supported by the high levels of the AMPA receptor assembling subunits GluR2/3 that are present in the subplate during early corticogenesis (Furuta and Martin, 1998). Since it has been shown that AMPA receptors carrying the flip forms exhibit greater sensitivity to cyclothiazide than receptors assembled from flop variants (Johansen et al., 1995; Partin et al., 1995), we also suggest a dominant presence of flip splice variants in the AMPA receptors of subplate neurons.
A large percentage of the sPSCs recorded in subplate neurons at a holding potential of –70 mV were mediated by NMDA receptors. Functional NMDA receptors have been already described in cortical plate neurons of the late embryonic rat (LoTurco et al., 1991), indicating that NMDA receptors may be involved in the very early development of the cerebral cortex. Although space clamp problems due to remote dendritic localization of these synapses cannot be excluded, it is likely that NMDA receptors located on subplate neurons exhibit a reduced Mg2+ block, as previously described in the hippocampus (Morrisett et al., 1990) and visual cortex (Kato and Yoshimura, 1993) of the immature rat. Furthermore, the slow kinetics of the NMDA receptor-mediated sPSCs may result from a subunit composition incorporating the NR2B subunit, which is present during early development [for a review see Feldmeyer and Cull-Candy (Feldmeyer and Cull-Candy, 1996)]. NMDA receptors are thought to play an important role in synaptic plasticity of the juvenile cerebral cortex (Kleinschmidt et al., 1987). Our data indicate that NMDA receptors may mediate activity-dependent synaptic plasticity already in the neonatal cortex.
Activity-dependent and -independent Processes
Subplate neurons receive synaptic inputs from various cortical and subcortical sources [for a review see Allendoerfer and Shatz (Allendoerfer and Shatz, 1994)]. These inputs are mediated by AMPA, NMDA and GABAA receptors (present study). Subplate cells relay their output to neurons in the developing cortical plate and layer I (Friauf and Shatz, 1991; Galuske and Singer, 1996; Finney et al., 1998) and project back to the thalamus and other cortical regions (McConnell et al., 1994; Del Río et al., 2000). Therefore, during a restricted period of late embryonic and early postnatal ontogenesis, subplate cells act as a crucial link in the developing cortical circuit. A large proportion of the sPSCs were TTX sensitive, indicating that they were dependent on presynaptic action potentials. Catalano and Shatz (Catalano and Shatz, 1998) have shown in cats that thalamocortical connections require action potential-dependent activity for their correct topographic organization. Since our study indicates that the slow NMDA receptor-mediated sPSCs are TTX sensitive, we propose that blockade of NMDA receptors in the subplate may also cause a disorganized thalamocortical projection. However, activity-dependent processes are probably more important for the refinement of connections during later stages of embryonic and early postnatal development. Recent studies on various mutant mice demonstrate that activity-independent processes, e.g. molecular cues, specify the basic organization of the cerebral cortex. The regional specification of the neocortex develops normally in the Gbx-2 mutant, whose thalamus and thalamocortical projection is disrupted (Miyashita-Lin et al., 1999), indicating that molecular markers intrinsic to the cortex specify the functional subdivisions (Rakic, 1988). The two regulatory genes Emx2 and Pax6 may act as such molecular markers, since they regulate the arealization of the cerebral cortex in a cooperative manner (Bishop et al., 2000). Verhage et al. (Verhage et al., 2000) recently demonstrated in the munc18-1 mutant mouse that the cerebral cortex develops normally even in the absence of synaptic transmission. Further evidence for the need of molecular cues comes from a number of organotypic slice culture studies on the development of the thalamocortical projection in the absence of subplate neurons [for a review see Bolz et al. (Bolz et al., 1993)]. The limbic system-associated membrane protein (LAMP) is one of the molecules that fulfills the role of a specific axonal guidance factor (Mann et al., 1998).
In conclusion, although the early activity-independent assembly of the cerebral cortex depends primarily on molecular markers, during further development neuronal activity plays an important role in the refinement of the initial crude connectivity. Subplate neurons receive functional glutamatergic and GABAergic synaptic inputs and most likely participate in thalamocortical, corticothalamic and intracortical neuronal circuits during early development. By integrating synaptic inputs from various cortical and subcortical sources, subplate cells may influence profoundly the functional status of these circuits and the activity-dependent maturation of the cerebral cortex.
This work was supported by DFG grant Lu 375/3 and the neuroscience graduate program GRK 320. The authors thank B. Hellmuth for technical assistance.
Address correspondence to Heiko J. Luhmann, Institute of Neurophysiology, Heinrich-Heine-University, POB 101007, D-40001 Düsseldorf, Germany. Email: firstname.lastname@example.org.
|Property||Fast sPSCs||Slow sPSCs||Significance|
|Significant differences between fast and slow sPSCs are indicated by ** (P < 0.01) and *** (P < 0.001).|
|Decay time (ms)||6.3 ± 0.25||51.2 ± 3||***|
|Frequency (Hz)||0.22 ± 0.02||0.1 ± 0.01||***|
|Amplitude (pA)||14 ± 0.6||18 ± 1.4||**|
|Rise time (ms)||2.6 ± 0.1||3.9 ± 0.1||***|
|Decay time constants (ms)||4.9 ± 0.2||17.5 ± 1.2||***|
|104 ± 6.1|
|Property||Fast sPSCs||Slow sPSCs||Significance|
|Significant differences between fast and slow sPSCs are indicated by ** (P < 0.01) and *** (P < 0.001).|
|Decay time (ms)||6.3 ± 0.25||51.2 ± 3||***|
|Frequency (Hz)||0.22 ± 0.02||0.1 ± 0.01||***|
|Amplitude (pA)||14 ± 0.6||18 ± 1.4||**|
|Rise time (ms)||2.6 ± 0.1||3.9 ± 0.1||***|
|Decay time constants (ms)||4.9 ± 0.2||17.5 ± 1.2||***|
|104 ± 6.1|
- action potentials
- pulmonary arteriogram
- pulmonary artery
- antimicrobial chemoprophylaxis
- gamma-aminobutyric acid
- alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid
- physician assistant
- plasminogen activator
- gaba-a receptor
- somatosensory cortex
- whole-cell recording