INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Schizophrenia afflicts one-half to one percent of the world's population, including as many as two million Americans. There are several lines of evidence implicating the prefrontal cortex (PFC) in this disorder. Perhaps the most convincing evidence is the linkage of the PFC to many executive functions disrupted in people with schizophrenia, including working memory, abstract thinking, attention, and language coherency (Goldman-Rakic 1991; Goldman-Rakic and Selemon 1997; Levin 1984). These and other symptoms are alleviated in many patients by atypical neuroleptic drugs that share a high affinity for serotonin 5-HT₂ receptors (Leysen et al. 1994; Meltzer et al. 1989; Schotte et al. 1993, 1996). The PFC receives robust serotonergic projections from the raphe nuclei (Lindvall and Björklund 1984) and expresses high levels of 5-HT₂ receptor, particularly in layer V prefrontal pyramidal neurons (PPNs) (Pazos et al. 1985, 1987). In these deep-layer PPNs, 5-HT_2A receptors are enriched in the apical dendrites that extend into more superficial layers (Goldman-Rakic 1999; Willins et al. 1997).

How activation of these PFC 5-HT₂ receptors modulates the excitability of deep-layer (V-VI) PPNs is controversial. 5-HT₂ receptor activation depolarizes rat layer V PPNs (Araneda and Andrade 1991), rat cingulate cortex pyramidal neurons (Tanaka and North 1993), and human layer III neocortical neurons (Newberry et al. 1999). 5-HT₂ receptor activation also increases excitatory postsynaptic currents in rat layer V PPNs (Aghajanian and Marek 1997, 1999). In contrast, 5-HT₂ receptor activation inhibits N-methyl-D-aspartate (NMDA) receptor-mediated currents (Arvanov and Wang 1998; Arvanov et al. 1999a; Arvanov et al. 1999b) and spontaneous activity (Bergqvist et al. 1999; Godbout et al. 1991; Mantz et al. 1990). Furthermore, 5-HT₂ receptor blockade with atypical neuroleptics induces PFC immediate early genes (Robertson and Fibiger 1996; Robertson et al. 1994; Wiesel et al. 1994). These seemingly disparate findings may have a unifying explanation, but if so, it will only come from a better understanding of the way in which 5-HT₂ receptor signaling cascades modulate voltage-dependent ion channels.

In an attempt to isolate the postsynaptic mechanisms by which 5-HT₂ receptors modulate excitability, pyramidal neurons were acutely isolated from the deeper layers of rat prelimbic and infralimbic cortex. Because voltage-dependent Ca²⁺ channels are important targets of monoamine receptors in other cell types (Cepeda et al. 1998; Hernandez-Lopez et al. 2000; Surmeier et al. 1995), our initial studies focused on the ability of 5-HT₂ receptors to modulate whole cell Ca²⁺ channel currents. Voltage-clamp and single-cell RT-PCR (scRT-PCR) studies in these neurons show that 5-HT₂ receptor activation reduces L-type Ca_v1.2 Ca²⁺ channel currents. This modulation is accomplished by a signaling cascade initiated by G_q protein stimulation of phospholipase C (PLC), leading to the mobilization of inositol trisphosphate (IP3)-sensitive intracellular Ca²⁺ stores and activation of calcineurin.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Acute dissociation procedure

Prelimbic and infralimbic PFC was dissected from young adult Sprague-Dawley rats (Charles River, Wilmington, MA), and the neurons were acutely dissociated using methods similar to those previously described (Benson et al. 1992). The rats were anesthetized with methoxyflurane (Schering-Plough, Union, NJ), isoflurane, or pentobarbital (Abbot Laboratories, Chicago, IL) and decapitated. The brains were rapidly removed and plunged into an ice-cold, oxygenated, sucrose solution containing the following (in mM): 250 sucrose, 4 glucose, 2.5 KCl, 1 Na₂HPO₄, 4 MgSO₄, 15 HEPES, 1 kynurenic acid, 0.1 N-nitro-L -arginine, and 0.005 glutathione (pH = 7.4, 300 mosM/l). Unless otherwise noted, all chemicals and reagents were obtained from Sigma (Saint Louis, MO). After cooling, the brains were blocked and sectioned coronally (400 µM) in the ice-cold sucrose solution using a DSK microslicer (Ted Pella, Redding, CA). The slices were then transferred to a chamber where they were held at room temperature in NaHCO₃-buffered Earl's balanced salt solution (EBSS) until use (1-5 h). The EBSS holding solution was bubbled with 95% O₂-5% CO₂ and supplemented with (in mM) 1 kynurenic acid, 0.1 N-nitro-L-arginine, and 0.005 glutathione (pH = 7.4, 300 mosM/1).

In preparation for mechanical dissociation, the deep layers V and VI of the prelimbic and infralimbic cortex were dissected and transferred to oxygenated HEPES-buffered (HBSS) containing 1 mg/ml protease (type XIV, bacterial), where they were incubated for 25 min at 37°C. The HBSS enzyme solution was supplemented with (in mM) 1 pyruvic acid, 1 kynurenic acid, 0.1 N-nitro-L-arginine, and 0.005 glutathione (pH = 7.4, 300 mosM/1). After enzyme treatment, the neurons were dissociated by triturating with a series of progressively smaller fire-polished Pasteur pipettes. The dissociation solution was oxygenated and contained the following (in mM): 140 sodium isethionate, 2 KCl, 4 MgCl₂, 23 glucose, 15 HEPES, 1 kynurenic acid, 0.1 N-nitro-L -arginine, and 0.005 glutathione (pH = 7.4, 300 mosM/1). The tissue suspension was then placed in a 35 mm petri dish (Nunc, Naperville, IL) positioned on the stage of an inverted microscope. As soon as the neurons settled to the bottom of the dish, they were continuously perfused with physiological saline. The saline solution contained the following (in mM): 140 NaCl, 2 KCl, 2 MgCl₂, 1 CaCl₂, 23 glucose, and 15 HEPES (pH = 7.4, 300 mosM/1).

Whole cell recording

Whole cell recordings were made using standard techniques (Hamill et al. 1981). Electrodes were made by pulling 7052 glass (A-M Systems, Carlsborg, WA) on a Sutter P-97 Flaming/Brown micropipette puller (Sutter Instrument, Novato, CA) and fire-polished with a MF-83 (Narishige Scientific Instrument Lab, Tokyo). The internal pipette solution contained the following (in mM): 190 N-methyl-D-glucamine, 4 MgCl₂, 2 Na₂ATP, 0.2 Na₂GTP (Boehringer Mannheim, Indianapolis, IN), 12 phosphocreatine, 0.1 leupeptin (Bachem, Torrance, CA), and 40 HEPES (pH = 7.2 with H₂SO₄, 270 mosM/1). Some internal solutions also contained 15 mM bis-(o-aminophenoxy)-N,N,N',N'-tetraacetic acid (BAPTA, Molecular Probes, Eugene, OR). The external recording solution contained the following (in mM): 130 tetraethylamonium (TEA) chloride, 1 MgCl₂, 2 BaCl₂, 10 glucose, 10 HEPES, and 0.001 tetrodotoxin (TTX; pH = 7.4 with TEA-OH, 300 mosM/1). In some recordings, 135 mM NaCl and 20 mM CsCl replaced TEA-Cl. The external recording solutions were locally applied to the neurons through an array of glass capillary tubes 500 µM in diameter. The flow of external solutions was gravity-fed controlled by a system of electronic valves (Lee, Essex, CT). Exchanging the external solution was completed in 1 s by altering the position of the array with a PMC200-P2 motion controller and an 850B high-speed actuator (Newport, Irvine, CA). The valve perfusion system and the Newport motion controller were synchronously driven by a PC running Labview software (Version 4.0.1, National Instruments, Austin, TX).

Depending on their solubility, the drugs and toxins used in the experiments were first dissolved in either H₂O or dimethyl sulfoxide (DMSO) to make stock solutions at concentrations of 1,000 to 10,000×. When drugs were dissolved in DMSO, equivalent amounts were added to the external control solutions. If the manufacturer indicated the drug was stable in solution, the stocks may have been frozen for 1 wk before use; otherwise, all were freshly prepared on the day of use. The 5-HT₂ agonists (±)-2,5-dimethoxy-4-iodoamphetamine hydrochloride (DOI), and -methyl-5-hydroxytryptamine maleate (-m-5-HT) and 5-HT₁ agonist (+)-8-hydroxy-2-(dipropylamino)-tetralin (R)-(+)-2-dipropylamino-8-hydroxy-1,2,3,4-tetrahydronaphthalene (8-OH-DPAT), which may have some small potential to oxidize, were dissolved in 1% sodium metabisulfite stock solution. All three were obtained from Sigma. The final concentration of metabisulfite (0.001%) was never seen to affect baseline currents. Other drugs used include ritanserin, ketanserin, risperidone, clozapine, LY-53,857, (±)-Bay K 8644, chelerythrine chloride, U-73122, U-73343, xestospongin C, thapsigargin, calcineurin autoinhibitory peptide, FK-506 (Calbiochem, San Diego, CA), nifedipine (Alexis, San Diego, CA), -conotoxin GVIA (CgTx; Bachem), and -agatoxin TK (AgTx; Peninsula Laboratories, San Carlos, CA). Cytochrome C (0.05%) was added to AgTx-containing solutions. To control for any effects of cytochrome C, it was included at equal concentrations in all external solutions in experiments using AgTx. The guanine nucleotide-binding protein peptide (G_q peptide) was designed to mimic C-terminal residues 343-353 of the alpha subunit (GenBank Accession No. P82471). The sequence was LQLNLKEYNLV.

Electrophysiological recordings were obtained using an Axopatch 200B amplifier (Axon Instruments, Foster City, CA). Data were acquired though an ITC-16 interface linked to an Apple computer running Pulse and Pulse Fit software (Version 8.1, Instrutech, Elmont, NY). Pyramidal neurons were voltage clamped at room temperature using patch electrodes with series resistances of 1.5-2.0 M. After G seal formation and rupture, series resistance was compensated 75% and periodically monitored. A sample of pyramidal neurons showed a mean whole cell capacitance of 16.3 ± 0.6pF (n = 54). Pyramidal neurons were selected by their morphology and all had a large apical dendrite. Voltage control was determined by examining the decay of the tail currents after depolarization from 80 to 0 mV. If, under control conditions, the tails did not decay smoothly and completely within 10 ms, the neurons were discarded. Current trace data were analyzed with Igor Pro software (Version 3.12, WaveMetrics, Lake Oswego, OR).

scRT-PCR

scRT-PCR was performed using protocols previously described (Baranauskas et al. 1999; Tkatch et al. 2000). Individual neurons were patched and aspirated into micropipettes while being continuously perfused by the control solution. The micropipettes contained 1 µl of diethylpyrocarbonate (DEPC)-treated water. To minimize RNase activity, the micropipettes were autoclaved, sterile gloves were worn at all times during the procedures, and the external control solutions were prepared with essentially RNase-free water. After aspirating the cell into the tip of the micropipette, the tip was broken off and the contents expelled into an Eppendorf tube containing DEPC-treated water (1.6 µl), RNasin (0.7 µl, 40U/µl), BSA (0.7 µl), and random hexamers (2 µl, 50 ng/µl; Promega, Madison, WI).

Single-stranded cDNA was generated by reverse transcription. First, the neuron-containing mixture was heated to 70°C for 10 min to denature the nucleic acids, then cooled on ice for 1 min. To this mixture was added 10× PCR buffer (2 µl), MgCl₂ (2 µl, 25 mM), DTT (2 µl, 0.1 M), dNTPs (1 µl, 10 mM each; Promega), and DEPC-treated water to bring the final volume to 20 µl. The reaction mixtures were then heated to 42°C, at which point SuperScript II (1 µl, 200 U/µl; Life Technologies, Grand Island, NY) was added. Next, these RT reactions were run at 42°C for 50 min. The temperature was then increased to 70°C for 15 min to terminate the reactions. Finally, to eliminate any residual RNA, RNase H (0.5 µl, 2 U/µl; Life Technologies) was added and the reaction mixtures were held at 37°C for 20 min.

PCR amplification of the resulting cDNA was done by standard methods using a thermal cycler (P-200, MJ Research, Watertown, MA). The PCR reaction was carried out using 4 µl of cDNA template, 10× PCR buffer (2.6 µl), MgCl2 (2.6 µl, 25 mM), dNTPs (1 µl, 10 mM), primers (1.5 µl each, 20 µM), and DEPC-treated water for a final volume of 30 µl. Taq polymerase (0.5 µl, 5 U/µl; Promega) was added immediately before the first cycling step. "Touch-down" protocols were implemented for more efficient amplification of single-cell cDNA. This was done by modifying a standard 45-cycle PCR protocol in the following way: first, 35 cycles were run at the optimal annealing temperature for each primer set, then the annealing temperature was decreased by 1°C for two cyclesfive timesfor a total of 45 cycles. The optimal annealing temperatures for the primers were determined using Oligo software (Version 6.4, National Biosciences, Plymouth, MN).

The primer sequences for Ca²⁺- and calmodulin-dependent kinase type-II (CaMKII) and 5-HT_2A and 5-HT_2C have been published (Vysokanov et al. 1998) and have predicted product lengths of 354, 465, and 545 bp respectively. The primers for PLC1-4 have also been published (Hernandez-Lopez et al. 2000) and have predicted lengths of 253, 547, and 420 bp for PLC1-3, respectively, with PLC4 having two products with predicted lengths of 209 and 246 bp.

In each set of single-cell reactions, both positive and negative controls were performed. Positive controls were run with pooled PFC cDNA. To control for the possible contribution of genomic DNA in the single-cell reactions, reverse transcriptase was omitted from one of the samples in each set. To control for cDNA contamination of reagents, the cell was omitted from one reaction mixture in each set. These controls were consistently negative.

Statistical procedures

Statistical analyses were performed with Systat software (Version 5.2, Evanston, IL). Significant differences between small unmatched samples were determined using Kruskal-Wallis one-way ANOVA. Box plots were generated for graphic presentation of the results. Sample statistics are given as medians because they provide the most robust estimate of central tendency. This feature is important in cases where small sample sizes do not allow the properties of their underlying sampling distributions to be determined.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

PPNs express serotonin 5-HT_2A- and 5-HT_2C-class mRNA

Layers V and VI were dissected from rat prelimbic and infralimbic cortex and the PPNs were acutely isolated (Fig. 1A). These neurons have one large apical dendrite and several finer basal dendrites that can be preserved with careful dissociation (Fig. 1B). As expected from previous studies (Pompeiano et al. 1994; Wright et al. 1995), RT-PCR analysis of pooled PFC RNA revealed the presence of 5-HT_2A, 5-HT_2C, and calmodulin-dependent protein kinase II (CaMKII) mRNAs (Fig. 1C, top left). To determine the cellular expression of 5-HT₂-class mRNAs, scRT-PCR experiments were conducted on large, visually identified pyramidal neurons (Fig. 1C, top, right). The identification of these pyramidal neurons was confirmed by their expression of CaMKII mRNA (Benson et al. 1992). This nonquantitative profiling detected 5-HT_2A mRNA in 67% of all PPNs sampled (n = 18). Ten of these neurons were also examined for their expression of 5-HT_2C mRNA; 5-HT_2C was detected in one of the neurons in this sample (this neuron did not have detectable levels of 5-HT_2A mRNA). These data are summarized in the bar graph shown in Fig. 1C, bottom. The simplest interpretation of these results is that the majority of layer V-VI PPNs express 5-HT_2A receptor mRNA.

View larger version (52K): [in this window] [in a new window]   Fig. 1. Pyramidal neurons in the rat PFC express 5-HT2-class mRNAs. A: a section of rat brain containing the putative rat homologue of human PFC. The outlined area contains the deep layers (V-VI) of prelimbic and infralimbic frontal cortex. B: a single dissociated pyramidal neuron from the area outlined in A with a section of the proximal apical dendrite intact. C: PCR analysis of prefrontal cortex (PFC) tissue and prefrontal pyramidal neuron (PPN) expression of 5-HT2-class receptor mRNAs. The 1st set of amplicons (top, left) shows that CaMKII, 5-HT2A, and 5-HT2C mRNA is expressed in tissue from rat PFC. The 2nd panel of amplicons is from individual PPNs [single-cell RT-PCR (scRT-PCR)] and shows cells that had detectable levels of 5-HT2A or 5-HT2C mRNA, and CaMKII mRNA (a marker of pyramidal neurons). The frequency distribution results indicate that from this sample of PPNs, 67% had detectable levels of 5-HT2A mRNA, and 10% had detectable levels of 5-HT2C mRNA, with none of the cells having detectable levels of both mRNAs. All of the PPNs screened expressed CaMKII mRNA.

PPNs have L-, N-, P/Q-, and R-type Ca²⁺ currents

Whole cell voltage-clamp recordings were made from PPNs using techniques that allowed the isolation of currents through voltage-dependent Ca²⁺ channels (Bargas et al. 1994). To determine the contribution of N-, P/Q-, L-, and R-type currents, channel-specific blockers were applied while activating currents with depolarizing voltage steps. Currents were evoked every 5 s to minimize rundown. Blockers of L-type current (nifedipine, 5 µM), N-type current (-CgTx GVIA, 1 µM), and P/Q-type current (-AgTx TK, 1 µM) consistently blocked a portion of the whole cell current (Fig. 2A). The current that remained after application of nifedipine, -CgTx GVIA, and -AgTx TK was operationally defined as R-type current. The R-type current was completely blocked by cadmium (200 µM). Figure 2B shows individual current traces taken at each stage in the addition of the channel blockers. The percent contribution of each high-voltage-activated (HVA) Ca²⁺ current type was calculated by subtraction. The median percent L-type current was 21%, N-type was 18%, P/Q-type was 41%, and R-type was 25%. Kruskal-Wallis ANOVA revealed that PPNs display significantly more P/Q-type current than L-, N-, or R-type current (P < 0.01, n = 4; Fig. 2C).

View larger version (15K): [in this window] [in a new window]   Fig. 2. PPNs have L-, N-, P/Q-, and R-type Ca2+ currents. A: high-voltage-activated (HVA) current (2 mM Ba2+) is evoked by stepping from 80 to 10 mV. The time course shows that successive and concurrent application of nifedipine (5 µM), -conotoxin GVIA (CgTx, 1 µM), and -agatoxin TK (AgTx, 1 µM) blocks a portion of the current. Cadmium (200 µM) blocks all HVA current. B: individual current traces show that application of nifedipine (5 µM), CgTx (1 µM), and AgTx (1 µM) each block a portion of the HVA current. These traces are all cadmium (200 µM) subtracted to eliminate any contaminating current. C: the box plot summarizes the percent of total HVA current carried by each Ca2+ channel type (n = 4).

5-HT₂ receptor stimulation reduces L-type Ca²⁺ currents

An initial set of experiments revealed that the 5-HT₂ agonist DOI consistently reduced whole cell Ca²⁺ channel currents (data not shown) and L-type currents in particular. L-type currents were isolated by applying nifedipine (5 µM) and subtracting the resulting currents from those evoked in its absence. Previous work with cortical pyramidal neurons (Foehring et al. 2000; Furukawa et al. 1999) and heterologous expression systems (Furukawa et al. 1999) had shown that this concentration of nifedipine has no detectable effect on N-, P/Q-, or R-type HVA currents using similar voltage protocols. To determine the effect of 5-HT₂ receptor stimulation on L-type current, recordings were made in the presence and absence of DOI (10 µM) and -m-5-HT (10 µM; Fig. 3A). Nifedipine-subtracted traces show that both DOI (Fig. 3B) and -m-5-HT inhibited peak L-type current. The median percent inhibition by 5-HT₂ agonist was 24% for DOI and 20% for -m-5-HT. The box plot of the data shows no significant difference in inhibition of peak L-type HVA current by DOI or -m-5-HT at nominally saturating equimolar concentrations (10 µM) (Kruskal-Wallis ANOVA, n = 7; Fig. 3C).

View larger version (17K): [in this window] [in a new window]   Fig. 3. 5-HT2 receptor agonists inhibit L-type currents. A: the time course demonstrates the method for calculating the 5-HT2-mediated effects on peak L-type currents. Initially, control -methyl-5-hydroxytryptamine maleate (-m-5-HT, 10 µM) and (±)-2,5-dimethoxy-4-iodoamphetamine hydrochloride (DOI, 10 µM) records are taken. Nifedipine (5 µM) is then applied to block L-type current, and the 5-HT2 agonist applications are repeated. The difference between the control records and the records in nifedipine reflects the L-type current component. B: individual records of nifedipine-subtracted (5 µM) traces shown before and after the application of 5-HT2 agonist demonstrate that DOI (10 µM) inhibits peak L-type Ca2+ current. C: the data summarized in the box plot show the percent inhibition of peak L-type current by DOI (10 µM) or -m-5-HT (10 µM) (n = 7).

A second way to assess modulation of L-type current is to use the L-type channel agonist Bay K 8644 (1 µM) to enhance L-type currents and slow their deactivation (Hernandez-Lopez et al. 2000; Howe and Surmeier 1995; Tsien et al. 1986). As shown in Fig. 4A, Bay K 8644 greatly slows the deactivation of L-type channels and effectively isolates L-type channel tail currents late in the response to a repolarizing step. Plotting the inverted tail currents on a semi-log scale makes the biexponential decay in the presence of Bay K 8644 apparent and allows the determination of a time in the decay when the contribution of non-L-type currents should be minimal (more than 2-3 ms, Fig. 4B). This temporal separation was used to assess the impact of neuromodulators on L-type channel function. Tail current amplitude (at 6 ms following the repolarizing step) is plotted as a function of time in Fig. 4C; the plot shows the enhancement of tail currents by Bay K 8644 and their reversible inhibition by DOI (10 µM). Similar results were seen in nearly every PPN tested (92%, n = 59). Representative voltage-clamp current records taken before, during, and after DOI application are shown in Fig. 4D. Records taken before the application of Bay K 8644 also illustrate the change induced in current amplitude and kinetics. The point in the tail current decay that was used for measurement is identified by a gray bar in Fig. 4D.

View larger version (35K): [in this window] [in a new window]   Fig. 4. Bay K 8644 reveals that DOI suppresses L-type channel tail currents. A: the deactivation of the L-type tail current is dramatically slowed by the application of the L-channel agonist Bay K 8644 (1 µM). Currents were evoked by stepping the voltage from the holding potential of 80 to 10 mV for 20 ms and then to 60 mV (see inset protocol). B: semi-log plot of absolute tail currents from before (A) and after Bay K 8644 application. Prior to Bay K 8644, the tail currents were fit with a single exponential having a time constant of 0.46 ms; after application, a second time constant was evident having a value of 6.3 ms. C: the amplitude of the Bay K 8644-induced L-type tail current, sampled every 5 s, is plotted to show the kinetics of the DOI-mediated (10 µM) inhibition. D: the control trace shows the current evoked by stepping the voltage from the holding potential of 80 to 10 mV for 20 ms and then to 60 mV. Measurements of tail current amplitude are taken at the point where the control current completely decays (vertical bar). The DOI (10 µM) trace shows the 5-HT2-mediated inhibition of the Bay K 8644-induced L-type tail current. This inhibition is clearly seen in the magnification box to the right. E: semi-log plot of absolute tail currents before and after DOI application. Both currents were well-fit with a bi-exponential function having a fast time constant of 0.60 ms and a slow time constant of 6.87 ms. The slow component is plotted for both tail currents. The amplitude of the slow component was reduced from 2.56 to 2.11 nA by DOI (18%), whereas the fast component was reduced by only 7%. F: the data summarized in the box plot show the percent inhibition of L-type tail current by DOI (10 µM) or -m-5-HT (10 µM; n = 10).

In all of the neurons chosen for analysis, tail currents in the presence and absence of DOI decayed smoothly back to baselineindicating that DOI had not altered current amplitudes by influencing voltage-clamp quality. But to provide a more rigorous test of this possibility, tail current kinetics were estimated by fitting with a biexponential function. If the quality of the voltage control fluctuated, it should be reflected in altered decay kinetics. However, in all of the cases examined (n = 6), the deactivation time constants in the absence of DOI were indistinguishable from those in the presence of DOI. In particular, the amplitude of the slowly deactivating current was diminished by DOI application, but the decay time constant was unaltered (Fig. 4E). Furthermore, there was excellent concordance between the percent inhibition of the slowly deactivating component estimated by curve fitting and that estimated by simply measuring the tail current amplitude at 6 ms (no more than 4% difference between the 2 measures in the same cell). As a consequence, the DOI effects on tail currents in subsequent experiments were determined by the simple amplitude measurement shown in Fig. 4D. These experiments are summarized in box plot format in Fig. 4F. The median inhibition of L-type tail current by DOI was 19.5 and 20% by -m-5-HT. These effects were not significantly different (Kruskal-Wallis ANOVA, P > 0.05, n = 10).

In contrast to the effects of the 5-HT₂ receptor agonists, application of the 5-HT₁ receptor-preferring agonist, 8-OH-DPAT, failed to significantly modulate Bay K 8644-enhanced L-type tail currents. As reported previously in sensorimotor cortex pyramidal neurons, 8-OH-DPAT reduced whole cell Ca²⁺ channel currents (Foehring 1996). A plot of peak current evoked by a slowly repeated test step to 10 mV is shown in Fig. 5A (). 8-OH-DPAT at both 1 and 10 µM reduced the amplitude of these currents as did subsequent application of DOI. However, the slow tail currents were not reduced by either concentration of 8-OH-DPAT, whereas application of DOI resulted in a reversible modulation (Fig. 5A, ). Representative current traces are shown in Fig. 5B. In a sample of seven pyramidal neurons, the median 8-OH-DPAT (10 µM) reduction of currents evoked by a step to 10 mV was 9%, but the reduction in slow tail current amplitude was not significantly different from control solution changes (median reduction = 3%).

View larger version (19K): [in this window] [in a new window]   Fig. 5. 5-HT1 receptor stimulation inhibits aggregate HVA current but not L-type tail current. A: currents were evoked by stepping the voltage from the holding potential of 80 to 10 mV for 20 ms and then to 60 mV. Aggregate peak current was measured near the end of the step to 10 mV () and compared with Bay K 8644-evoked tail current from the same cell (). The time course shows that while the 5-HT1 agonist 8-OH-DPAT (1 µM and 10 µM) reversibly inhibited peak current it had no effect on the L-type tail current (n = 7) in cells where DOI (10 µM) inhibited L-type tail current (n = 7). B: representative current traces from the experiment shown in A further demonstrate that the effect of 5-HT1 receptor stimulation with (R)-(+)-2-(dipropylamino)-8-hydroxy-1,2,3,4-tetrahydronaphthalene (8-OH-DPAT, 10 µM) is limited to a reduction in peak current.

To provide an additional verification of 5-HT₂ receptor mediation of the DOI and -m-5-HT effects, several 5-HT₂ receptor antagonists were examined. Ritanserin, ketanserin, LY-53,857, clozapine, and risperidone were tested. All of the antagonists inhibited HVA current at low micromolar concentrations in the absence of agonist (data not shown). At 10 µM, ritanserin reduced currents dramatically (median =76%, n = 3), whereas ketanserin (20%, n = 6), clozapine (19%, n = 3), and LY-53,857 (16%, n = 6) had smaller but significant effects. Risperidone (10 µM) only modestly reduced currents (median = 5%, n = 3). This pattern of effects is similar to that reported by Sah and Bean in response to neuroleptics (Sah and Bean 1994). As described in these earlier studies, a direct interaction of the antagonists with the Ca²⁺ channel pore was tested by examining unblocking kinetics in the presence and absence of the pore blocker Cd²⁺ (Chow 1991; Donaldson and Beam 1983). If Cd²⁺ and the receptor antagonist bind within the pore to block channel currents, then allowing Cd²⁺ access to its binding site first should prevent the antagonist from gaining access to its binding site, leading to rapid current recovery kinetics resembling those of Cd²⁺ alone. Ketanserin and ritanserin had slow enough recovery kinetics relative to those of Cd²⁺ to make this type of analysis feasible. Kitanserin recovery kinetics were dramatically accelerated by preexposure to Cd²⁺ (200 µM), changing from a median of 9.5 to 4.2 s (n = 7)a rate that was not different from that seen with Cd²⁺ alone (median = 3.7 s, n = 14). Washing kinetics following exposure to ritanserin were also accelerated by preexposure to Cd²⁺. However, the recovery occurred with a fast phase (resembling that of Cd²⁺) and a slower phase. The time constant of the slower phase (median = 26 s, n = 7) was much faster than the mono-exponential recovery seen in the absence of Cd²⁺ (median = 45 s, n = 3). These results are consistent with the hypothesis that both ketanserin and ritanserin act as Ca²⁺ channel pore blockers in a subset of Ca²⁺ channels. Further studies will be necessary to fully characterize this phenomenon, but it clearly makes the use of these antagonists problematic in the analysis of 5-HT₂ receptor modulation of Ca²⁺ channels.

A powerful alternative pharmacological strategy is to target elements in the putative signaling cascade linking the receptor to the ion channel, in this case the L-type Ca²⁺ channel. In the brain, as well as other tissues, 5-HT₂ receptor activation stimulates PLC isoforms and phosphoinositide (PI) hydrolysis (Berridge et al. 1983; Conn and Sanders-Bush 1984, 1986; Roth et al. 1984, 1986; Sanders-Bush and Conn 1986). Both 5-HT_2A and 5-HT_2C receptor activation of PLC is thought to be mediated by the G protein, G_q (Goppelt-Struebe and Stroebel 1998; Jope et al. 1994; Shah et al. 1999; Wang and Friedman 1999). The experiments described in the following text were designed to test the hypothesis that 5-HT₂ receptors modulate L-type Ca²⁺ channels through this signaling pathway.

Disruption of G_q interactions blocks 5-HT₂-mediated inhibition of L-type Ca²⁺ current

To determine whether inhibition of L-type Ca²⁺ current requires activation of G_q proteins, PPNs were dialyzed with a peptide designed to mimic its carboxyl-terminal docking site (Akhter et al. 1998). An excess of this peptide should prevent the agonist-bound 5-HT₂ receptors from binding and activating G_q proteins. In these experiments, control measurements were taken from PPNs dialyzed with a nonsense peptide (100 µM) of equal size (Fig. 6A). The traces in Fig. 6B show that introduction of the G_q peptide (100 µM) results in a blockade of the DOI-mediated inhibition of tail current. Figure 6C summarizes these experiments, showing that the G_q peptide significantly reduces the DOI-mediated inhibition of L-type tail current (median = 0%) when compared with control neurons dialyzed with the nonsense peptide (median =17%; Kruskal-Wallis ANOVA, P < 0.01, n = 10). In fact, in 50% of the neurons dialyzed with the G_q peptide, the DOI-mediated inhibition of L-type tail current was completely abolished.

View larger version (30K): [in this window] [in a new window]   Fig. 6. Blocking 5-HT2 receptor interaction with Gq proteins attenuates 5-HT2-mediated inhibition of L-type current. A: DOI (10 µM) inhibits L-type tail current in PPNs dialyzed with a nonsense peptide control (100 µM). B: the DOI-mediated inhibition is reduced in PPNs dialyzed with a with a Gq peptide (100 µM) designed to block docking of Gq to the 5-HT2 receptor. C: the box plot summarizes the data, showing that dialyzing PPNs with a Gq peptide (100 µM) reduces or completely blocks the DOI-mediated inhibition of the L-type tail current when compared with the nonsense peptide controls (P < 0.01, n = 10).

PPNs coexpress PLC1 and PLC4 mRNAs

There are four known isoforms of PLC that are expressed in the brain (PLC1-4) (Exton 1996). PCR profiling revealed that pooled PFC cDNA contains mRNA for all PLC isoforms (Fig. 7A, top left). All of the PPNs examined with scRT-PCR techniques had detectable levels of PLC1 mRNA (n = 8; Fig. 7A, top right). Half of this sample also had detectable levels of PLC4 mRNA. None of the individual neurons profiled had detectable levels of PLC2 or PLC3 mRNA.

View larger version (37K): [in this window] [in a new window]   Fig. 7. Blocking phospholipase C (PLC) activation diminishes 5-HT2-mediated inhibition of L-type current. A: PCR analysis of PFC tissue and PPN expression (scRT-PCR) of PLC mRNAs. The first set of amplicons (top left), labeled "pooled PFC cDNA," shows that 5-HT2A, PLC1, PLC2, PLC3, and PLC4 mRNAs are all expressed in tissue from rat PFC. Top right: amplicons, from an individual PPN (scRT-PCR), show coexpression of 5-HT2A, PLC1, and PLC4 mRNA. The frequency distribution results indicate that from this sample of PPNs, 63% of the individual neurons had detectable levels of 5-HT2A mRNA, 100% express PLC1 mRNA, and 50% had detectable levels of PLC4 mRNA (n = 8). No individual PPNs were positive for PLC2 or PLC3 mRNAs. B: the box plot shows that inhibition of PLC activity with U-73122 (200 nM), in conditions where constitutive protein kinase C (PKC) activity is blocked with chelerythrine (1 µM), reduces DOI-mediated inhibition of L-type tail current when compared with U-73343 (200 nM), a weak PLC inhibitor (P < 0.01, n = 6).

PLC activation is required for 5-HT₂-mediated inhibition of L-type Ca²⁺ current

To determine if DOI inhibits L-type tail currents by stimulating PLC1 or PLC4, L-type tail currents were isolated and PLC activity was blocked. In these experiments, constitutive protein kinase C (PKC) activity was inhibited with chelerythrine chloride (1 µM). Although inhibition of PKC activity reduced Ca²⁺ channel currents (data not shown), it did not block the modulation of L-type currents by DOI. Co-perfusion of the PLC blocker U-73122 (200 nM) significantly reduced the ability of DOI to inhibit L-type tail currents (Fig. 7B). The median percent inhibition of L-type current by DOI was 17% in the presence of chelerythrine chloride alone and 0% when U-73122 was co-perfused (Kruskal-Wallis ANOVA, P < 0.01, n = 6). The DOI modulation in the presence of the weak PLC inhibitor U-73343 (200 nM) was not significantly different from that seen in control recordings (Kruskal-Wallis ANOVA, P > 0.05, n = 5, Fig. 7B).

5-HT₂-mediated inhibition of L-type Ca²⁺ current requires release of Ca²⁺ from intracellular stores

Activation of PLC is known to stimulate PI hydrolysis, IP3 production, and liberation of Ca²⁺ from endoplasmic reticulum (ER) stores. In parallel, PLC increases 1,2-diacylglycerol (DAG) levels leading to activation of PKC. The failure of the PKC inhibitor chelerythrine chloride to block the 5-HT₂-mediated inhibition of L-type current implicates the IP3 limb of the signaling cascade. The next set of experiments tests this hypothesis.

If release of Ca²⁺ from intracellular stores is a critical step in the modulation, then it should be attenuated by depletion of these stores. To accomplish this, neurons were exposed to the Ca²⁺ pump blocker thapsigargin (5 µM). As shown in Fig. 8, thapsigargin markedly attenuated the ability of DOI to modulate Bay-K-8644-enhanced tail currents; records prior to (Fig. 8A) and after thapsigargin treatment (Fig. 8B) are shown. Because thapsigargin elevates cytosolic Ca²⁺ levels by disrupting sequestration (Thastrup et al. 1990), the reduction in the slow tail currents revealed by the time-course plot (Fig. 8C) is expected of a Ca²⁺-dependent modulation. The data are summarized in Fig. 8D. Thapsigargin treatment significantly reduced the median inhibition of L-type current from 23 to 3% (Kruskal-Wallis ANOVA, P < 0.01, n = 5).

View larger version (43K): [in this window] [in a new window]   Fig. 8. Depleting Ca2+ or blocking inositol trisphosphate (IP3)-mediated release of Ca2+ from intracellular stores reduces 5-HT2-mediated inhibition of L-type current. A: L-type tail current is evoked as before with Bay K 8644 (1 µM) in the absence of external Ca2+ and Ca2+ chelators and the DOI-mediated inhibition is recorded. B: after the PPN is perfused with thapsigargin (5 µM) and the current stabilizes, the DOI-mediated inhibition is markedly reduced. C: the time course of this experiment demonstrates 1st, a robust inhibition of L-type current by DOI (10 µM), 2nd, temporal increases in cytosolic Ca2+ caused by thapsigargin inhibit L-type current, and 3rd, subsequent application of DOI fails to further inhibit tail current. D: the box plot shows that application of DOI (10 µM) inhibits the current significantly more prior to depletion of intracellular Ca2+ stores with a 3- to 5-min application of thapsigargin (5 µM; P < 0.01, n = 5). E: this box plot shows that blocking IP3-mediated Ca2+ release with xestospongin C (1 µM) significantly reduces the DOI inhibition of L-type tail current when compared with control records taken prior to inhibitor treatment (P < 0.02, n = 4).

To test whether the DOI-mediated effect on tail currents required IP3-mediated release of Ca²⁺ from intracellular stores, PPNs were perfused with xestospongin C (1 µM), a specific blocker of IP3-mediated Ca²⁺ release (Gafni et al. 1997). Xestospongin significantly reduced the DOI-mediated inhibition of L-type tail current (Kruskal-Wallis ANOVA, P < 0.02, n = 4; Fig. 8E). The median DOI inhibition of L-type tail current was 18% prior to xestospongin exposure and 4% after.

If an elevation in intracellular Ca²⁺ concentration is necessary for the DOI-mediated reduction of L-type tail current, chelation of intracellular Ca²⁺ should reduce the modulation. To test this prediction, neurons were dialyzed with a high concentration of the fast Ca²⁺ chelator BAPTA (15 mM). The first set of traces in Fig. 9A show the typical DOI-mediated inhibition of L-type current in the absence of chelator. The next records (Fig. 9B) show that this inhibition is reduced by dialysis with BAPTA. These experiments are summarized in Fig. 9C. The median percent DOI inhibition of L-type current was 20% in controls (n = 10) and 10% in neurons dialyzed with high BAPTA (n = 7). Kruskal-Wallis ANOVA confirmed that the BAPTA reduction of the DOI-mediated inhibition of L-type tail current was significant (P < 0.001). The -m-5-HT-mediated reduction of L-type tail current also was examined for sensitivity to Ca²⁺ chelation. Like DOI, neurons dialyzed with BAPTA show significantly less modulation by -m-5-HT. In these experiments, the median percent of -m-5-HT inhibition was 27% in the controls and 10% in neurons dialyzed with high BAPTA (Kruskal-Wallis ANOVA, P < 0.05, n = 5). The fact that high concentrations of BAPTA were less effective than thapsigargin or zestospongin in reducing the DOI/-m-5-HT modulation is not surprising given its hydrophilicity and potentially limited subcellular distribution.

View larger version (36K): [in this window] [in a new window]   Fig. 9. Chelating intracellular Ca2+ reduces 5-HT2-mediated inhibition of L-type tail current. A: traces recorded in chelator-free conditions show the DOI (10 µM) inhibition of L-type tail current. B: the 2nd set of traces shows that chelating intracellular Ca2+ by adding bis-(o-aminophenoxy)-N,N,N',N'-tetraacetic acid (BAPTA, 15 mM) to the internal pipette solution reduces the DOI (10 µM) inhibition of L-type tail current. C: the box plot summarizes the effect that chelating intracellular Ca2+ has on the 5-HT2-mediated inhibition of L-type tail current. The 1st panel shows that chelating intracellular Ca2+ by dialyzing the cells with BAPTA (15 mM) reduces the DOI (10 µM) inhibition of L-type tail (n = 7) when compared with controls without BAPTA (P < 0.001, n = 10). The 2nd panel shows similar results were obtained when -m-5-HT (10 µM) was used (P < 0.04, n = 5).

5-HT₂-mediated inhibition of L-type Ca²⁺ current requires activation of calcineurin

Elevations in intracellular Ca²⁺ have been shown to reduce L-type Ca²⁺ currents by activation of the Ca²⁺-dependent phosphatase calcineurin (Chad and Eckert 1986; Lukyanetz et al. 1998a; Schuhmann et al. 1997). To determine whether a similar mechanism was mediating the 5-HT₂ receptor modulation, calcineurin activity was blocked by dialysis with the calcineurin autoinhibitory peptide (10 µM) (Hashimoto et al. 1990). Recordings taken from control PPNs (Fig. 10A) were compared with those from neurons dialyzed with the calcineurin autoinhibitory peptide (Fig. 10B). This peptide significantly reduced the median DOI inhibition of L-type tail current from 20 to 8% (Kruskal-Wallis ANOVA, P < 0.001, n = 7; Fig. 10D). As an additional test, neurons were treated with the membrane permeable calcineurin inhibitor FK-506 (1 µM) (McCall et al. 1996; Raufman et al. 1996). FK-506 also decreased the ability of DOI to inhibit tail current (Fig. 10C). The median inhibition of L-type Ca²⁺ current was dramatically reduced to 0% by FK-506 preincubation (Fig. 10D; Kruskal-Wallis ANOVA, P < 0.001, n = 7).

View larger version (51K): [in this window] [in a new window]   Fig. 10. Blocking calcineurin activation attenuates 5-HT2-mediated inhibition of L-type current. A: a representative trace from a control set of recordings where L-type tail current was evoked as before with Bay K 8644 (1 µM) and the DOI-mediated inhibition measured. B: the inhibition of tail current is reduced in a PPN dialyzed with calcineurin autoinhibitory peptide (50 µM). C: preincubation with the calcineurin inhibitor FK-506 (1 µM) blocks the inhibition of tail current by DOI (10 µM) (P < 0.001, n = 7). D: this box plot summarizes the effect of blocking calcineurin on the DOI-mediated inhibition of the tail current. First, dialyzing the neurons with the calcineurin autoinhibitory peptide (50 µM) significantly reduces the inhibition of tail current by DOI (P < 0.001, n = 7). Second, perfusion with the calcineurin inhibitor FK-506 (1 µM) also reduces inhibition of tail current by DOI (P < 0.001, n = 7).

Ca_v1.2 channels are targeted by the 5-HT₂ signaling cascade

Neocortical pyramidal neurons express two types of L-type Ca²⁺ channels: Ca_v1.2 (class C) and Ca_v1.3 (class D) (Ertel et al. 2000). Ca_v1.2 channels are found primarily in the soma and proximal dendrites, whereas Ca_v1.3 channels have a more diffuse distribution (Hell et al. 1993). Although both channel 1 subunits are phosphorylated by serine/threonine kinases (Hell et al. 1994), the functional consequences of cardiac Ca_v1.2 channel phosphorylation and dephosphorylation are best understood (e.g., McHugh et al. 2000). To determine if Ca_v1.2 channels are targeted by the 5-HT₂ receptor signaling pathway, the ability of DOI to modulate L-type currents was examined in neurons from Ca_v1.3 knockout mice (Platzer et al. 2000). In these neurons, all of the L-type current is attributable to Ca_v1.2 channels. As shown in Fig. 11A, Ca_v1.2 channel currents activated at relatively hyperpolarized membrane potentials, having a threshold near 60 mV, a median half-activation voltage of 18 mV, and median slope factor of 7.5 (n = 3). Bay K 8644-enhanced Ca_v1.2 tail currents were robustly modulated by application of DOI. A typical plot of peak tail current amplitude as a function of time is shown if Fig. 11B. Representative current traces showing the DOI induced reduction in tail currents is shown in Fig. 11C. The median modulation by DOI was 28% in Ca_v1.3 knockout PPNs and 29% in heterozygotes (Fig. 11D). Although a role for Ca_v1.3 channels cannot be excluded, these results demonstrate that Ca_v1.2 Ca²⁺ channels are prominent targets of 5-HT₂ receptor modulation in PPNs.

View larger version (38K): [in this window] [in a new window]   Fig. 11. Cav1.2 channels are targeted by the 5-HT2 receptor signaling pathway. A: individual record of a nifedipine-subtracted (5 µM) trace, elicited by a slow voltage ramp, shows that Cav1.2 channel currents activate at relatively hyperpolarized membrane potentials (arrow). B: the time course demonstrates that DOI (10 µM) reversibly inhibits Bay K 8644-induced L-type tail current in a PPN from a Cav1.2 KO mouse. C: the DOI (10 µM) trace shows the 5-HT2-mediated inhibition of the Bay K 8644-induced L-type tail current. This inhibition is clearly seen in the magnification box to the right. D: the data summarized in the box plot show no difference between the median percent inhibition of L-type tail current by DOI (10 µM) in control heterozygote mice and Cav1.3 KO mice.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Deep-layer PPNs express 5-HT₂ receptor mRNAs

The 5-HT innervation of the rodent cortex arising from the raphe nuclei is diffuse and is found in all cortical layers. There are several classes of 5-HT receptor that transduce signals arising from this cortical innervation. The most prominent of these in the frontal cortex appear to be the 5-HT_1A, 5-HT_2A, and 5-HT_2C classes (Boess and Martin 1994). Recently, 5-HT_2A receptors have been localized on the apical dendrite of layer III and V pyramidal neurons in the primate PFC (Goldman-Rakic 1999; Willins et al. 1997). These receptors are also found in some interneurons (Vysokanov et al. 1998). The expression of 5-HT_2A receptors by a substantial population of rat layer V-VI PPNs is consistent with our scRT-PCR profiling and suggests that the physiological effects described here are mediated primarily by 5-HT_2A receptors.

5-HT₂ receptor activation reduces Ca_v1.2 L-type Ca²⁺ channel currents in PPNs through G_q/PLC/IP3/ calcineurin signaling pathway

In agreement with studies of pyramidal neurons from other cortical regions, HVA currents in PPNs arose from N-, P/Q-, R-, and L-type Ca²⁺ channels (Lorenzon and Foehring 1995). 5-HT₂ agonists DOI and -m-5-HT inhibited aggregate HVA Ca²⁺ currents. A large fraction of the DOI/-m-5-HT modulation resulted from a targeted suppression in L-type Ca²⁺ channel currents. Blockade of L-type currents with channel-specific concentrations of nifedipine reduced the effects of these receptor agonists. Furthermore, isolation of L-type tail currents with the dihydropyridine agonist Bay K 8644 clearly showed an agonist-mediated modulation of slowly deactivating L-type tail currents. The reduction in tail currents was not accompanied by an alteration in deactivation kinetics, arguing that fluctuation in the quality of the voltage clamp did not contribute the observed effects.

Several lines of evidence implicate 5-HT₂ receptors in the modulation of L-type currents. First, as mentioned in the preceding text, 5-HT_2A receptors are expressed by these neurons. Second, 5-HT₂ receptor-selective agonists DOI and -m-5-HT effectively modulated L-type currents. In contrast, activation of the other major 5-HT receptor population in these neurons5-HT_1A receptorswas ineffective in modulating L-type currents, as in sensorimotor cortex pyramidal neurons (Foehring 1996). The next logical step in verifying 5-HT₂ receptor involvement would have been to show selective antagonism. However, the application of 5-HT₂ receptor-specific antagonists reduced currents in the absence of receptor stimulation. Because these effects were qualitatively similar to those of the agonists, the change was not due to inverse agonism. Rather experiments examining the interaction of the antagonist effects with those of the nonspecific Ca²⁺ channel blocker Cd²⁺ suggested that much of the reduction in currents could be attributed to a direct channel block. Neuroleptics also block voltage-dependent Ca²⁺ channels, albeit through what appeared to be an allosteric mechanism (Sah and Bean 1993). More work will be required to cleanly delineate the mechanism by which the 5-HT antagonists are blocking Ca²⁺ channels. Nevertheless, these direct actions precluded their use in confirming activation of 5-HT₂ receptors in the modulation of L-type Ca²⁺ channel currents.

As an alternative to receptor antagonism, the signaling elements mediating the agonist-induced modulation were characterized. These studies lend strong support to the proposition that the modulation was mediated by 5-HT₂ receptors. Biochemical studies have shown that cortical 5-HT₂ receptor regulation of intracellular events is dependent on coupling to the G protein G_q (Adlersberg et al. 2000). Dialysis with a peptide mime of the G_q docking domain (that should act as a competitive inhibitor of 5-HT₂ receptor coupling) significantly attenuated the DOI-mediated inhibition of L-type current, providing direct evidence for G_q involvement and 5-HT₂ receptor mediation. Even more compelling evidence for this conclusion came from subsequent experiments that delineated the signaling mechanism linking G_q proteins to L-type channels.

G_q proteins stimulate PLC isoforms, leading to the generation of IP3 and DAG (Exton 1997; Katan 1996; Taylor et al. 1991; Umemori et al. 1997). scRT-PCR profiling revealed that PPNs co-express two PLC isoform mRNAs (PLC1 and PLC4). Inhibition of these isoforms with U-73122 blocked the DOI-mediated reduction of L-type Ca²⁺ current. DAG generated by PLC1/4 results in the activation of PKC isoforms (Tanaka and Nishizuka 1994). However, this limb of the 5-HT₂ receptor signaling cascade did not appear to participate in the modulation of L-type channel currents as perfusion with the PKC inhibitor chelerythrine chloride did not alter the DOI effect. This was true in spite of the fact that PKC can suppress Ca_v1.2 channel currents in certain circumstances (McHugh et al. 2000) and that 5-HT₂ receptors activate PKC targeting Na⁺ channels in PPNs (D. B. Carr and D. J. Surmeier, unpublished observations). Rather, the other limb of the PLC pathwaythe IP3 limbwas necessary. IP3 binds to ER receptors and promotes Ca²⁺ release (Berridge 1993). Blockade of IP3 receptors with xestospongin C disrupted the agonist-mediated modulation as did depletion of intracellular Ca²⁺ stores with the pump blocker thapsigargin. Last, chelation of intracellular Ca²⁺ with BAPTA reduced the L-type channel modulation. This line of evidence implicates a PLC/IP3/Ca²⁺ signaling cascade in the 5-HT₂ agonist-mediated modulation.

Ca²⁺ released from intracellular stores is known to activate the Ca²⁺-dependent phosphatase calcineurin (PP2B) in a variety of cell types (Hernandez-Lopez et al. 2000; Nishi et al. 1999; Yan et al. 1999). Inhibition of calcineurin with either an autoinhibitory peptide or with FK-506 significantly attenuated the agonist-induced modulation of L-type tail currents. Calcineurin is also known to dephosphorylate skeletal (Ca_v1.1) and cardiac (Ca_v1.2) L-type Ca²⁺ channels at serine residues leading to a reduction in macroscopic currents (Armstrong et al. 1991; Chad and Eckert 1986; Lukyanetz et al. 1998b; Schuhmann et al. 1997). Although the phosphorylation sites critical to this reduction have not been definitively mapped, the PKA site at S1928 is a strong candidate (Gao et al. 1997). PKA phosphorylation of Ca_v1.2 channels enhances open probability and macroscopic currents (Yue et al. 1990). The anchoring proteins necessary to support PKA modulation of Ca_v1.2 channels (AKAPs) also bind calcineurin, bringing both enzymes into close proximity to the channel complex (Coghlan et al. 1995; Klauck et al. 1996). The demonstration that the 5-HT₂ receptor signaling cascade targets Ca_v1.2 channels suggests that a similar AKAP complex is present in PPNs. Our findings do not exclude the possibility that Ca²⁺ released from intracellular stores directly interacts with the calmodulin-like domain of the channel, further reducing currents (Imredy and Yue 1994; Peterson et al. 1999; Schuhmann et al. 1997). However, the impact of calcineurin inhibitors on the modulation argue that this is not an important mechanism in the response to 5-HT.

Functional implications

In a variety of cell types, L-type Ca²⁺ currents are important determinants of neuronal integration, synaptic plasticity, and gene expression. For example, entry of Ca²⁺ specifically through L-type Ca²⁺ channels has been linked to the induction of activity-dependent changes in gene expression in hippocampal pyramidal neurons (Bading et al. 1993; Boukhaddaoui et al. 2000; Mermelstein et al. 2000; Murphy et al. 1991). Related mechanisms may mediate L-type channel involvement in certain forms of long-term synaptic potentiation and depression (Bolshakov and Siegelbaum 1994; Calabresi et al. 1994; Kapur et al. 1998). At this point, the roles of Ca_v1.2 and Ca_v1.3 L-type channels in these processes have not been determined. In hippocampal and cortical pyramidal neurons, both channels are found not only at distal dendritic sites close to synapses but at perisomatic sites that have preferential access to Ca²⁺-dependent transcriptional regulators (Hell et al. 1993; Westenbroek et al. 1990). This differential subcellular localization is particularly prominent for Ca_v1.2 channels. The modulation of these channels by 5-HT₂ receptors establishes a potential mechanism by which serotonin could regulate gene expression and long-term changes in PFC physiology. It also provides a means by which atypical neuroleptics could induce immediate early gene expression and longer term changes in PFC function (Robertson et al. 1994; Robertson and Fibiger 1996; Wiesel et al. 1994). However, the ostensibly direct action of clozapine and risperidone on L-type Ca²⁺ channels makes it unclear whether the atypical neuroleptics are acting as 5-HT antagonists or 5-HT mimes.

In addition to regulating long-term changes in excitability, Ca²⁺ entry through L-type channels influences short-term integration of synaptic input. For example, dendritic L-type Ca²⁺ channels are thought to make a significant contribution to the nonlinear integration of strong excitatory synaptic input to pyramidal neurons (Kim and Connors 1993; Larkum et al. 1999; Magee et al. 1998; Schiller et al. 1997). This contribution may be particularly important in the amplification of NMDA receptor currents and prolongation of synaptically triggered dendritic depolarization (Cepeda et al. 1998; Dudek and Fields 2001; Hernandez-Lopez et al. 1997). Even a modest (10-20%) suppression of these regenerative currents could have profound effects on these nonlinear dendritic events. In fact, activation of 5-HT₂ receptors in PPNs does lead to a suppression of NMDA receptor-mediated currents (Arvanov et al. 1999a), possibly through suppression of boosting L-channel currents. The 5-HT₂ receptor suppression of L-type currents may also narrow the temporal window through which back-propagating action potentials can influence orthodromically conducted events and burst generation (Larkum et al. 1999; Schiller et al. 1995; Stuart et al. 1997; Svoboda et al. 1997).