Dye coupling between pyramidal neurons in developing rat prefrontal and frontal cortex is reduced by protein kinase A activation and dopamine

During early postnatal development, lamina II/III pyramidal cells in rat neocortex are extensively coupled via gap junctions. The factors regulating gap junction permeability, as well as the mechanisms underlying the developmental uncoupling process are not understood. To investigate the influence of protein kinase A-mediated phosphorylation on dye coupling in the developing neocortex, pyramidal cells in slices of rat frontal and prefrontal cortex were injected intracellularly with the tracer neurobiotin. Control injections revealed clusters of about 30 dye-coupled neurons. Preincubation with forskolin or direct activation of protein kinase A with Sp-cAMPS reduced the number of coupled cells by about 70%. A significant reduction in dye coupling was also observed following incubation with dopamine. Application of receptor selective agonists and antagonists revealed that the uncoupling was mediated by both dopamine D1 and D2 receptors. The protein kinase A inhibitor Rp-cAMPS reduced the effect of dopamine, suggesting that the neurotransmitter regulates gap junction permeability via protein kinase A activation. In the presence of either forskolin, Sp-cAMPS, or dopamine, neurons displayed a significantly higher input resistance compared to control conditions. During the second postnatal week, transient application of forskolin to single neurons reversibly increased input resistance. At later developmental stages when coupling incidence had declined, this action of forskolin was no longer observed. Our data demonstrate a dependence of gap junction permeability on protein kinase A activity and on dopamine receptor activation in developing rat neocortical neurons. These mechanisms may modulate junctional permeability during the period of circuit formation.

During the early postnatal development of the mammalian neocortex, pyramidal neurons are extensively coupled via gap junctions (Connors et al., 1983;Cepeda et al., 1993;Peinado et al., 1993). It has been suggested that this transitory communication system plays an important role during the period of circuit formation in the cerebral cortex (Yuste et al., 1992;1995: Peinado et al., 1993. Gap junctions in the developing neocortex probably mediate the intercellular transfer of both metabolites (Kandler et al., 1994) and electrical signals (Lo Turco and Kriegstein. 1991;Yuste et al., 1995). Furthermore, they significantly influence electrotonic cell properties (Rorig et al., 1995). All these parameters might affect the efficacy and stability of developing chemical synapses. Thus, one important question is how junctional conductance is regulated by neuromodulators and other trophic factors. Gap junction permeability has been shown to be modulated by a number of physiological factors including transmembrane voltage, intracellular pH, Ca", and phosphorylation of connexin subunits. However, apart from the sensitivity to intracellular acidification (Connors et al., 1984;Rorig et al., 1995). none of these regulatory mechanisms has been investigated so far in immature neocortical neurons.
Phosphorylation via cyclic adenosine monophosphate (CAMP)-dependent protein kinase (PKA) has been shown to affect junctional conductance in a number of neuronal and nonneuronal preparations. In rat hepatocytes (Saez et al.. 1986) and canine heart cells (DeMello 1983(DeMello , 1989Burt and Spray, 1988). CAMP increases gap junction permeability. In teleost, turtle, and mammalian retinal neurons, PKA-mediated phosphorylation decreases junctional communication (Teranishi et al., 1983;Piccolino et al., 1984;Lasater and Dowling, 1985;Lasater, 1987;Hampson et al., 1992;McMahon, 1994). Activation of a number of G-protein-coupled neurotransmitter receptors increase cAMP production, resulting in an enhancement of PKA activity. Dopamine reduces gap junction coupling via DI receptor stimulation in retinal horizontal and amacrine cells of different species (Lasater and Piccolino et al., 1984;Dowling, 1985;DeVries and Schwartz, 1992;Hampson et al., 1992;McMahon, 1994;Mc-Mahon and Brown, 1994), as well as in the rat neostriatum (Cepeda et al., 1989) and nucleus accumbens (O'Donnell and Grace, 1993). The most likely mechanism of transmitter action is connexin phosphorylation, resulting in a short-term modulation of gap junctional conductance (Moreno et al., 1992). Already at prenatal stages, the rat neocortex is invaded by a number of neuromodulatory afferents (Schmidt et al., 1982;Verney et al., 1982;Berger et al., 1985Berger et al., , 1991Kalsbeek et al., 1988). In rodents, two classes of dopaminergic afferents to the prefrontal cortex have been described (Berger et al., 1985). The first arises from the Al0 dopaminergic cell group of the ventral mesencephalon, reaches the prefrontal cortex by embryonic day I6 (Verney et al., 1982) and innervates all cortical layers. Dopaminergic fibers of the second class originate in the A9 group and reach their targets during the first and second postnatal weeks. These afferents predominantly innervate the superficial cortical layers (Berger et al., 1985). Autoradiographic binding studies have revealed an enhanced expression of dopamine Dl receptors during the second and third postnatal weeks in rat prefrontal cortex (Leslie et al., 1991). A recent in situ hybridization and receptor binding study (Schambra et al., 1994) demonstrated the presence of the PKA-coupled dopamine DI receptor (DIA) in rat ventricular cells as early as gestational day 14.
The present study was intended to determine whether dye coupling between developing neocortical pyramidal cells is regulated by PKA activation. Following application of either forskolin, membrane-permeable CAMP analogs, or dopamine, we observed a reduction in tracer coupling between superficial pyramidal neurons. The dopamine effect was at least partially mediated via the PKA pathway, suggesting that dopaminergic afferents might be involved in the regulation of electrical and metabolic coupling of neurons during the period of formation of synaptic circuits in the neocortex. Injections were performed in the medial precentral area of the prefrontal cortex, as well as in adjacent frontal areas, and were confined to the superficial layers. Since dye coupling gradually decreases during the developmental period under investigation (Connors et al., 1983; Peinado et al., 1993), data from 2-3 con-  days have been pooled. Control injections made on postnatal days 8-10 (PS-10) resulted in a mean number of 30.4 5 5.7 (n = 14) cells coupled to the injected neuron (range 8-67 cells). After preincubation with forskolin, a marked reduction in cluster size became evident (Fig.1). On P7-9, the mean number of neurons dye coupled to the core cell was reduced to 7.X + 3.3 (n = 19) after forskolin treatment (Fig. 3). The difference between the mean values was statistically significant (17 < 0.0001, Mann-Whitney U test). In the presence of forskolin, tracer injection resulted in a single stained neuron in one case, and in only 2 out of I9 slices were clusters of more than 20 coupled neurons observed. In comparison, under control conditions (at P7-IO), single stained, i.e., uncoupled neurons were never detected, and 9 out of I4 injections revealed clusters of more than 20 cells.

Materials and Methods
A similar result was obtained when slices were incubated with the direct PKA activator Sp-CAMPS (IO0 FM). The mean cluster size on P9/lO was 24. I ? 5.9 (n = IO) under control conditions and 7.7 ? 6.3 (n = 7) in the presence of Sp-cAMPS (Figs. 2 and 3). The mean values differ significantly 0, < 0.01). Com-was observed in three slic-plete Sp-CAMPS-induced uncoupling es.
To investigate whether dopaminergic mechanisms might be involved in the control of gap junction coupling during cortical development, intracellular injections of neurobiotin were made in the presence of dopamine. The injections were performed in slices obtained from 7-l5-d-old animals and were made predominantly in an area of the cerebral cortex receiving strong dopaminergic input, i.e., the medial precentral part of the prefrontal cortex. Figure 4 shows a neuron injected in the presence of nomifensine and sodium metabisulfite (B) as well as neurobiotin-labeled cells after incubation with 100 pM dopamine (A, C). Due to the strong developmental decline in cluster size, neurons were split into three age groups in order to quantify the dopamine effect (Fig. 5). On P8/9, the cluster size was signilicantly reduced (17 < 0.04) from a mean of 37.4 +-8.0 (II = 7) under control conditions to a mean of 14.0 -t 4.4 (n = 7) in the presence of dopamine. A statistically signilicant reduction from a mean of 17.6 + 3.9 (n = 13) coupled neurons to a mean of 5.0 2 1.3 (n = 8) was also observed on PI011 I. Even between PI2 and Pl5, when a strong developmental reduction in cluster size had already occured, dopamine still produced a statistically significant decline in dye coupling from a mean of 6.4 -C 2. I (n = 24) coupled cells under control conditions, to 2.0 2 0.4 (n = 18) in the presence of dopamine. Thus, a dopaminergic modulation of gap junction coupling between layer II/III pyramidal cells is likely to exist during the entire postnatal period.
To investigate reversibility of the uncoupling action of dopamine a sample of PI0 neurons was preincubated and injected in the presence of dopamine. Then slices were superfused for a perid of I.5 hr with dopamine-free solution containing only the antioxidant and uptake inhibitor. Under these conditions we have observed only reduced recovery: four out of five neurons were coupled to one to seven neighboring cells, whereas a large cluster containing 26 coupled neurons was observed in only one case. This suggests that the long presence of the transmitter (up to I hr) might activate secondary processes leading to prolonged or even permanent uncoupling.
To exclude effects of the added antioxidant and the uptake inhibitor on dye coupling, control injections after incubation in Na&O, (100 pM) and nomifensine (I FM) without addition of dopamine were performed in slices from rats aged between P7 and PIO. These substances had no significant effect on cluster size (Figs. 4 and 5). The mean number of coupled cells under control conditions was 30.4 -C 5.7 (II = 14) and 37.0 +-9.9 (11 = 7) after incubation in antioxidant and uptake inhibitor. Thus, the reduction in cluster size observed after incubation in dopamine can be considered as a specific effect of the neurotransmitter.
Since  Under these conditions, dopamine still produced a significant reduction 0, < 0.005) in the number of cells coupled to the injected neuron from a mean of 30.4 ?z 5.7 (n = 14) to a mean of 8.9 ? 5.5 (n = 8, Fig. 7). Thus, despite the relatively high concentration used, we can exclude that the dopamine-induced decrease in gap junction coupling was mediated via p-adrenoceptors. Our results demonstrate a marked reduction in dye coupling following activation of protein kinase A and following activation of dopamine receptors, respectively. Some dopamine receptor subtypes are linked to the CAMP-PKA system. We therefore investigated whether the dopamine effect was mediated by this intracellular signal transduction pathway. Slices were incubated in dopamine and the PKA inhibitor Rp-CAMPS (100 FM).
Under these conditions, we observed reduced cluster sizes on P8-IO (18.1 t 5.7, n = 7). However, this mean value was not statistically different from those obtained under control conditions (31.6 ? 6.9, n = 11, p = 0.2109, Figs. 6 and 7). Thus, direct inhibition of protein kinase A at least partially antagonized the uncoupling effect of dopamine, suggesting that PKA-coupled dopamine receptors were involved in this action.
In order to determine which types of dopamine receptors modulate gap junction permeability, the effects of receptor selective agonists and antagonists were investigated. A rise in has been shown to be triggered by a subtype of the D I receptor, the D I A receptor (Dearry et al., 1990;Monsma et al., 1990;Zhou et al., 1990). We therefore tested the effect of the Dl receptor agonist SKF 38393 on tracer coupling. Preincubation in 100 FM SKF 38393 resulted in a significant reduction in dye coupling both on P7-9 (p < 0.02) and on PI I @ < 0.001, Figs. 8A,C and 9A), suggesting strongly that Dl receptors contribute to the uncoupling action of dopamine. However, the D2/D3 receptor agonist quinpirole (I 0 FM), although less effective, also caused a statistically significant 0, < 0.04) reduction in cluster size (Figs. 8B and 9A) on P7-9.
In the presence of dopamine and the selective Dl receptor antagonist SCH 23390 (5 PM), the cluster size was found to be slightly reduced, but not significantly different to controls (Fig.  9B). A weak antagonism of the dopamine effect was also observed upon application of the competive D2 receptor antagonist sulpiride (IO pM, Fig. 9B). The analysis of variance (nonparametric Kruskal-Wallis ANOVA) revealed that the antagonistic effects of both SCH 23390 and sulpiride were statistically significant (p < 0.03). However, ANOVA posttests (Dunn's Multiple Comparison Test) showed that the cluster sizes observed in the presence of the antagonists were significantly different from controls, but not significantly different from the cluster size obtained in the presence of dopamine alone. After incubation in dopamine and haloperidol (30 PM), a combined Dl/D2 receptor antagonist, the cluster size was found to be not statistically different from control values 0~ = 0.19) and significantly larger @ < 0.04) than the mean cluster size found in slices incubated The Journal of Neuroscience, November 1995. 15(11) 7393 exclusively in dopamine (Fig. 9B). Thus, the most effective antagonism of the dopamine effect on dye coupling between superficial layer pyramidal cells in the prefrontal cortex was achieved by blocking both Dl and D2 dopamine receptors.
Since deep cortical layers receive the first dopaminergic afferents during development and the strongest dopaminergic innervation in the adult rat, additional neurobiotin injections were performed in layer V and VI neurons, respectively, to examine layer specificity of the dopamine effect. Control injections in layers V/VI on P8-IO resulted in a much smaller number of coupled cells per injected neuron compared to superficial superficial layers (2.3 ? 1.32, n = 7, Figs. 10 and I I). Injections performed in the presence of dopamine (100 pM) did not significantly 0, = 0.37) change the number of dye-coupled neurons (3.2 + 0.8, n = 6, Fig. I I).
Effects of dopumine, $-CAMPS, md ,forskolin on rnernhrar~e potential and input resistance In order to analyze to what extent dopamine, PKA activating agents, or receptor-selective agonist affect neuronal properties, resting membrane potential and input resistance were measured in control and test solutions prior to intracellular injection of neurobiotin. Figure 13 summarizes data obtained from neurons ranging from P7 to PI 1. None of the substances tested (dopamine, forskolin, Sp-CAMPS) significantly altered the average resting membrane potential of the neurons. In contrast, the average neuronal input resistance was found to be significantly increased after preincubation in each of these substances (Fig.  13).
Since a reduction in gap junction permeability should change electrotonic cell properties, we investigated the effects of forskolin on electrotonic parameters. Application of 20 FM forskolin reversibly increased neuronal input resistance by 29-165s in five out of nine neurons tested (Fig. l2A). This enhancement in membrane resistance should result in a potentiation of the amplitudes of excitatory postsynaptic potentials (EPSPs). However, since forskolin markedly augments GABA,-mediated inhibitory postsynaptic potentials (IPSPs, Penit-Soria et al., 1987;Sutor and Mayr, 1991) and since GABA, receptor blockade induces epileptiform activity in the neocortex (Gutnick et al., 1982;Lee and Hablitz, 1991) the effect of forskolin on EPSPs could not be studied directly. We therefore injected transient currents resembling glutamatergic synaptic currents (100 pA amplitude, 1 msec time to peak, 5 msec decay time) and recorded the corresponding voltage responses. These voltage deflections were reversibly potentiated by 33-575s in seven out nine cells (Fig. l2B). Effects of forskolin on current-induced electrotonic potentials were not observed in neurons on PI7118 (n = 6, Fig.  12C,D), i.e., at a time when the extent of dye coupling had declined to a minimum. This finding strongly suggests that the forskolin-induced alterations in electrotonic potentials detected in younger neurons were due to a reduction in gap junctional conductance rather than to an action of forskolin on calciumdependent potassium currents (Hiramatsu et al., 1994), or voltage-gated sodium conductances (Ono et al., 1995).

Discussion
The present study provides evidence for a regulation of gap junction coupling between superficial pyramidal neurons in rat prefrontal and frontal cortex by protein kinase A. A significant reduction in neuronal dye coupling, as well as a prominent effect on input resistance was observed after adenylyl cyclase activa- Figure 8. Micrographs of layer II/III cells injected in the presence of the DI receptor selective agonist SKF 38393 (100 PM; A and C) and the D2/3 receptor selective agonist quinpirole (10 pM; B), respectively. SKF 38393 induced significant uncoupling on both Pl 1 (A) and P 7 (C). Quinpirole also significantly reduced dye coupling. The neuron shown in B was injected on P7. Scale bar: 50 km.
tion by forskolin and direct PKA activation by Sp-CAMPS. Do-We have used dye coupling as our major assay system to pamine also induced a marked reduction in gap junction couelucidate the mechanisms regulating gap junctional communipling. The transmitter effect was partially mediated by Dl re-cation. The tracer neurobiotin has been repeatedly used as an ceptor-stimulated PKA activation. Thus, we propose that dopaindicator of dye coupling between cortical neurons (Peinado et minergic afferents to the neocortex might be involved in the al., 1993). To allow for comparison between individual slices regulation of electrotonic and metabolic coupling between py-from different animals, we standardized injection and survival ramidal neurons during early postnatal development.
times of slices after injection. Only one neuron was injected in . Effects of dopamine receptor selective agonists and antagonists on dye coupling. A, Effects of receptor selective agonists. The Dl receptor agonist SKP 38393 reduced dye coupling as efficiently as dopamine both between P7 and P9 and on Pl 1. The D2/3 receptor agonist quinpirole was less effective, but the reduction in the number of coupled neurons per injection was also statistically significant (Mann-Whitney U test). B, Effects of receptor selective antagonists. After incubation in dopamine and either the Dl receptor antagonist SCH 23390 or the D2 receptor antagonist sulpiride, the size of neurobiotinlabeled cell clusters was no longer statistically significant from controls (ANOVA, P < 0.03). The most effective antagonism of the uncoupling action of dopamine was achieved by incubation in dopamine and the Dl/D2 receptor antagonist haloperidol (for P values, see text). each slice and only pyramidal-shaped neurons were included into the analysis. The extent of coupling was assessed by counting the coupled somata in the vicinity of the injected neuron.
To investigate transmitter effects on dye coupling, slices were incubated for 1.5 to 2 hr, including preincubation and survival times, in ACSF containing a rather high agonist concentration. Therefore, desensitization of receptors resulting in an underestimation of the effect cannot be excluded. However, the reduction in dye coupling was of high statistical significance. We used high agonist concentrations mainly for two reasons: (1) dye coupling displays high variability already under control conditions, and (2) receptor densities might be low as compared to the adult cortex. Furthermore, among the various dopamine receptor subtypes, the affinity for dopamine is lowest at the Dl and D2 receptors (Schwartz et al., 1992). The dopamine-induced uncoupling could be mimicked by both the Dl receptor selective agonist SKF 38393 and the D2 receptor selective agonist quinpirole. It was also observed in the presence of the P-adrenoceptor antagonist propranolol. These findings provide strong evidence for a dopamine effect selectively mediated by dopamine receptors. Since we used high agonist concentrations, the observed uncoupling effect of dopamine could not be completely suppressed by receptor selective antagonists applied in the "normal" pharmacological concentration range. Nevertheless, in neurons pretreated with dopaminergic antagonists, the reduction in cluster size induced by subsequent treatment with dopamine failed to reach statistical significance. In retinal amacrine cells, the antagonizing effect of SCH 23390 has been shown to depend strongly on the dopamine/SCH 23390 concentration ratio (Hampson et al., 1992). However, we did not further increase antagonist concentrations, since this might have resulted in a loss of receptor subtype selectivity.
In order to quantify effects on dye coupling, dopamine was bath applied, and coupling within the entire dendritic tree of the injected cell was analyzed. However, local uncoupling due to transmitter release at single modulatory terminals affecting only single dendrites or even parts of dendrites is likely to occur under physiological conditions. Although dopamine is released predominantly from varicosities and might reach its receptors by diffusion over relatively large distances, dopaminergic terminals have also been shown to be involved in triadic synaptic complexes in both rodent (Verney et al., 1990) as well as primate (Goldman-Rakic et al., 1989) prefrontal cortex, suggesting highly localized effects of the transmitter. The distance between dopaminergic release sites and gap junctions, however, is not known to date, and whether dopamine receptors are located in close proximity to junctional complexes also remains to be shown. Since dopamine receptor-mediated uncoupling of gap junctions might occur via widespread "volume neurotransmission" (i.e., undirected spread of transmitter in a larger volume of cortical tissue; Lidow, 1995) or in a highly localized fashion, it is unclear whether activation of dopaminergic afferents uncouples an assembly of neurons, e.g., an entire column, single neurons, or only local dendritic compartments from the functional syncytium.
We have only analyzed the effects of exogeneous transmitter on dye coupling without attempting to stimulate release from dopaminergic terminals in the slice preparation. Dopaminergic fibers originating in the brainstem are cut during preparation and the diffusely projecting dopaminergic afferents are difficult to stimulate. The dopamine uptake inhibitor nomifensine did not reduce dye coupling, indicating that spontaneously released dopamine does not reach concentrations high enough to significantly affect gap junction coupling. Microelectrodes rather than patch pipettes were used for tracer loading of pyramidal cells to exclude false negative results due to washout of components of intracellular second messenger pathways. However, to obtain higher resolution recordings, we used the patch-clamp technique to study the influence of gap junction closures on electrotonic potentials. Under these conditions, forskolin, which most effectively reduced tracer coupling, exerted effects on electrotonic potentials in five out of nine neurons tested. The lack of effect in four cells might be due partially to a long electrotonic distance between the location of gap junctions and the recording site, although a dialysis of intracellular components might also have taken place by the time PKA-mediated phosphorylation reduces junctional communication.
Cyclic adenosine monophosphate and dopamine as negative regulators of gap junctional communication during postnatal development of the neocortex In different tissues, PKA stimulation has heterogeneous effects on gap junction coupling. Whereas coupling is enhanced in some non-neuronal tissues (DeMello 1983;Saez et al., 1986;Burt and Spray, 1988;DeMello 1989), an inhibitory effect of PKA activation has been shown in the vertebrate retina (Teranishi et al., 1983;Piccolino et al. 1984;Lasater and Dowling, 198.5;Lasater, 1987;DeVries and Schwartz, 1992;Hampson et al., 1992;Mc-Mahon, 1994). Our results demonstrate a suppressive effect of raising intracellular CAMP concentrations on dye coupling in the developing rat neocortex. This second-messenger pathway is regulated by a number of G-protein<oupled neurotransmitter receptors, suggesting that chemical and electrical communication systems are closely interrelated in the immature neocortex. Whereas resting membrane potentials were not significantly altered, all three substances produced a statistically significant increase in input resistance (two-tailed Student's t test) in lamina II/III neurons as compared to agematched control neurons. In neurons of deep cortical layers, dopamine did not significantly increase input resistance. Statistical significance is indicated by asterisks.
We have not directly demonstrated a PKA-mediated phosphorylation of connexins, i.e., the observed reduction in dye coupling might be indirectly induced via phosphorylation of a different protein. Stimulation of the PKA pathway via Dl receptors has recently been shown to modulate voltage-activated calcium currents in striatal neurons (Surmeier et al., 1995). However, depolarizing pulses applied to inject neurobiotin were below spike threshold, thus preventing calcium entry during the action potential. Since the dopamine-induced reduction in dye coupling was antagonized by the PKA inhibitor Rp-CAMPS, an indirect effect via changes in intra-or extracellular pH can be excluded. It has been shown that activation of adenylyl cyclase and PKA results in a decrease in open probability of gap junction channels in retinal neurons (Lasater, 1987;McMahon and Brown, 1994). Thus, the most likely explanation for the reduc-tion in dye coupling induced by PKA in neocortical neurons is phosphorylation of connexins, which are endowed with several serine phosphorylation sites (see Bayer, 1993, for review).
The uncoupling effect of dopamine in the developing rat neocortex was mediated by both dopamine Dl and D2 receptors. D2 receptors are known to reduce intracellular CAMP concentrations (Onali et al., 1984;Battaglia et al., 1985;Weiss et al., 1985); however, a linkage to other second-messenger pathways has been shown (Piomelli et al., 1987;Kanterman et al., 1991, Jackson andWestlind-Danielsson, 1994). The application of receptor selective agonists demonstrated that activation of either Dl or D2 receptors is sufficient to depress dye coupling. Therefore, an intracellular signal transduction system different from the CAMP pathway might be involved in the action of dopamine on gap junction permeability. Possible candidates are protein kinase C-dependent phosphorylation, calcium ions, or arachidonic acid metabolites (Rose and Loewenstein, 1975;Rao et al., 1987;Miyachi et al., 1994).
Regulation of gap junctional communication-a new role for dopaminergic afferents to the developing neocortex? During the early postnatal period, dopaminergic terminals are present in both superficial and deep cortical layers (Berger et al., 1991). The strongest expression of both Dl and D2 receptors occurs in the deep layers, although these receptors are also existent in the superficial layers of prefrontal areas (Schambra et al., 1994). A significant increase in Dl receptor expression has been observed between the second and third postnatal week when gap junction coupling disappears and synaptogenesis increases (Leslie et al., 1991). Thus, the dopaminergic fiber system might exert a twofold function: at embryonic and early postnatal stages, dopamine might regulate metabolic and electrical signal transfer, and, by regulating electrotonic cell parameters, it might affect the efficacy of developing chemical synapses.
Layer ViVI pyramidal cells were already uncoupled, and dopamine had no significant effect on residual tracer coupling in the deep layers during the second postnatal week, suggesting that uncoupling follows a developmental gradient reflecting the "inside-first outside-last" pattern of neurogenesis in the neocortex.
Dopamine has been suggested to act as a neurotrophic factor during development of the neocortex, since lesioning of dopaminergic afferents reduces cortical thickness (Kalsbeek et al., 1987) and impairs the development of cortical neurons (Kalsbeek et al., 1989). In the adult rat prefrontal cortex, dopamine potentiates GABAergic synaptic transmission via a presynaptic mechanism (Penit-Soria et al. 1987;Sutor and Mayr, 1991) and shifts the balance between long-term potentiation and long-term depression in favor of long-term depression (Law-Tho et al., 1995). Thus, the functions of the dopaminergic projection to the neocortex obviously change during development.