|
|
||||||||
Departamento de Biofísica, Instituto de Fisiología Celular, Universidad Nacional Autónoma de México, Mexico City, Mexico
Submitted 3 May 2005; accepted in final form 17 August 2005
| ABSTRACT |
|---|
|
|
|---|
-Conotoxin GVIA occluded most of the Ca2+ current modulation in PD14 neurons. However, this occlusion was greatly decreased in PD40 neurons.
-Agatoxin TK occluded a great part of the modulation in PD40 neurons but had a negligible effect in PD14 neurons. The data indicate that dopaminergic D2-mediated modulation undergoes a change in target during development: from CaV2.2 to CaV2.1 Ca2+ channels. This change occurred while CaV2.2 channels were being down-regulated and CaV2.1 channels were being up-regulated. Presynaptic modulation mediated by D2 receptors reflected these changes; CaV2.2 type channels were used for release in young animals but very little in mature animals, suggesting that changes took place simultaneously at the somatodendritic and the synaptic membranes. | INTRODUCTION |
|---|
|
|
|---|
However, not only CaV1 Ca2+ channels but also CaV2 Ca2+ channels may be controlled by dopamine. Consequently, this study asked if D2 receptor activation also regulates CaV2 Ca2+ channels in striatal neurons. The dopaminergic modulation of CaV2 Ca2+ channels may be important given the roles of these channels in spiny cells (Perez-Garci et al. 2003
; Tecuapetla et al. 2005
; Vilchis et al. 2000
). Moreover, because the contribution of diverse CaV2 Ca2+ channels may change during development (Chameau et al. 1999
), and because medium spiny neurons have a protracted development (Tepper et al. 1998
) with respect to other neurons, it is important to see what happens with dopaminergic modulation during possible channel reconfigurations.
The relevance of this question becomes evident when recalling that CaV2 channels supply the Ca2+ necessary to activate Ca2+-dependent K+ channels (Vilchis et al. 2000
) in mature neurons. In turn, these K+ channels generate the postspike afterhyperpolarization (AHP) that makes up the interspike interval. The AHP controls the firing of spiny cells by setting the gain for a given stimulus (Bargas et al. 1999
; Perez-Garci et al. 2003
; Pineda et al. 1992
; Vilchis et al. 2000
). In addition, CaV2 channels are in charge of GABA release at the synaptic terminals of spiny neurons (Tecuapetla et al. 2005
). In this place too, there is evidence that dopamine exerts a presynaptic modulation of GABA release (Guzman et al. 2003
). Thus theoretically, dopamine could control neostriatal output at two different levels through the regulation of CaV2 Ca2+ channels: first, at the somatodendritic membrane through the regulation of the firing mechanism, and second, at the synaptic terminals (Cooper and Stanford 2001
; Guzman et al. 2003
) through the regulation of transmitter release. Accordingly, the main objective of this study was to find evidence of this dual regulation. In relation to this, it was recently shown that activation of muscarinic M1 receptors controls both the interspike interval and the transmitter release at the synaptic terminals of spiny neurons (Perez-Rosello et al. 2005
). Because muscarinic M1 receptors and dopaminergic D2 receptors seem to be linked to the same signaling cascade (Hernandez-Lopez et al. 2000
; Rakhilin et al. 2004
), it seems interesting to ask if these receptors share some physiological actions in neostriatal cells, and if not, why. This work has been published in abstract form (Salgado et al. 2004
).
| METHODS |
|---|
|
|
|---|
The protocols followed the National University of Mexico and National Institutes of Health guidelines for the use of animals in biomedical experiments. Briefly, and as described elsewhere (Bargas et al. 1999
), male Wistar rats of postnatal days 14 (PD14) and 40 (PD40), from our animal house, were deeply anesthetized, and their brains were quickly removed into ice-cold saline (4°C) containing (in mM) 126 NaCl, 3 KCl, 1 MgCl2, 2 CaCl2, 25 NaHCO3, and 11 glucose (pH 7.4 with NaOH, 298 mOsm/l with glucose; aerated with 95% O2-5% CO2). Parasagittal neostriatal slices (300 µm thick) were cut in 4°C saline using a vibratome (Ted Pella, Reading, CA). Slices were transferred to room temperature saline (2325°C) and left to recover for 1 h.
Voltage-clamp recordings of calcium currents
Neostriatal neurons from PD14 or PD40 rats were acutely dissociated using procedures similar to those previously described (Bargas et al. 1999
). Briefly, after a 1- to 6-h incubation, slices were removed into HEPES-buffered saline, and the dorsal striatum was dissected. Neostriatal slices were placed in the same HEPES solution, now containing 1 mg/ml of pronase E type XIV (Sigma, St. Louis, MO) at 32°C. After about 20 min, the slices were removed into a low-Ca2+ HEPES-buffered saline. Slices were rinsed and dissociated mechanically with a graded series of fire-polished Pasteur pipettes. The cell suspension (2 ml) was plated into a 35-mm Lux petri dish mounted on the stage of an inverted microscope containing 1 ml of the following recording saline (in mM): 0.001 TTX, 140 NaCl, 3 KCl, 5 BaCl2, 2 MgCl2, 10 HEPES, and 10 glucose (pH 7.4 with NaOH; 298 mOsm/l with glucose). After allowing the cells to settle, superfusion began at about 1 ml/min with saline of the same composition. Recordings were made only from medium-sized neurons (1012 µm soma diam). Whole cell currents recordings used standard techniques. Electrodes were pulled from borosilicate glass (WPI, Sarasota, FL) in a Flaming-Brown puller (Sutter Instrument Corp., Novato, CA) and fire polished before use. The internal saline contained (in mM) 180 N-methyl-D-glucamine (NMDG), 40 HEPES, 4 MgCl2, 10 EGTA, 2 Na2ATP, 0.2 Na3GTP, and 0.1 leupeptin (pH = 7.2 with H2SO4, 265270 mOsm/l). Electrode DC resistances were 48 M
in the bath. Recordings were obtained with an Axopatch-1D patch-clamp amplifier (Axon Instruments, Foster City, CA), and controlled and monitored with a PC clone running pClamp (version 5) with a 125-kHz DMA interface (Axon Instruments). After seal rupture, the series resistance (<15 M
) was compensated (7080%) and monitored before and after drug application. Current-voltage relationships (I-V plots) before and during drug applications were built from currents evoked with both 20-ms step voltage commands, from 80 to 50 mV in 10-mV steps, and with responses to voltage ramp commands (0.7 mV/ms) from 80 to 50 mV. Because the results from both methods coincided (see Fig. 1), most figures only show representative I-V plots built from the responses to voltage ramps.
|
|
|
Neostriatal slices from juvenile (PD14) or mature young rats (PD40) were transferred to a custom plexiglass recording chamber and superfused with oxygenated saline (36 ml/min) as above. Individual neurons were visualized (x40 water immersion objective) under differential interference contrast (DIC) enhanced visual guidance using infrared videomicroscopy in an upright microscope (Diaphot, Nikon, Melville, NY) adapted with a CCD camera (CCD-100, Dage-MTI, Michigan City, IN). Synaptic events were evoked with a bipolar concentric tungsten electrode (12 µm at the "pencil shaped" tip; FHC, Bowdoinham, ME) located at the globus pallidus to stimulate antidromically the axons of spiny cells (Guzman et al. 2003
; Tecuapetla et al. 2005
). Paired shock stimulation (4550 ms of interstimulus interval; 0.2- to 0.4-ms duration; 1- to 4-V strength; frequency of 0.1 Hz) was delivered with a computer interface. Isolation units (Digitimer, Hertfordshire, UK) between the computer and the stimulating electrodes were used to adjust stimulus parameters during the experiment. Distance between recording and stimulating electrode was about 1 mm. Synaptic responses in these conditions were of moderate amplitude, but some experiments were performed with either strong or weak stimulation strength to check if D2-class receptor actions were present. Thus evoked inhibitory postsynaptic currents (eIPSCs) had amplitude variation but without exhibiting failures except at the weaker stimulation strengths. Traces shown are the average of
4-min recordings (24 traces) taken once the amplitude had been stabilized in a given condition. A hyperpolarizing voltage command (10 mV) continuously monitored input conductance. Internal solution was (in mM) 72 KH2PO4, 36 KCl, 2 Mg Cl2, 10 HEPES, 1.1 EGTA, 2 Na2ATP, 0.2 Na3GTP, 5 QX 314 (to prevent neuronal firing and enhance input resistance), and 0.5% biocytin (pH = 7.2, 275 mOsM/l). Bath solution contained 10 µM 6-cyano-2,3-dihydroxy-7-nitro-quinoxaline disodium salt (CNQX) and 50 µM D-()-2-amino-5 phosphonovaleric acid (AP5) to block glutamatergic currents so that synaptic responses were eIPSC sensitive to bicuculline. Cells with resting potential more negative than 70 mV (at 0 current), input resistance about 100200 M
, and holding current (in voltage clamp mode)
0.02 nA to maintain a holding potential near the resting potential of the cell were chosen. Whole cell recordings were made using an Axoclamp 2B amplifier (Axon Instruments). Whole cell access resistances were in the range of 520 M
. Access resistance was continuously monitored, and experiments were abandoned if changes >20% were encountered. No cell capacitance, series resistance, or liquid junction potential (2 mV) compensations were made. All recordings were filtered at 13 KHz and digitized with an AT-MIO-6040E interface and a DAQ (NI-DAQ) board (National Instruments, Austin, TX) in a PC clone. On-line data acquisition used custom programs made by L. Carrillo in the LabVIEW environment (National Instruments). The NI-DAQ board was used to save the data on binary files in the computer hard disk for further off-line analysis.
eIPSCs obtained with different simulation strengths (weak to strong; see intensity-amplitude plots in Tecuapetla et al. 2005
) and different ages were chosen to have an array of different amplitudes for eIPSCs. These were used to perform a mean-variance analysis as described in Clements and Silver (2000)
.
Intracellular recordings
Slices obtained as above (see Preparation of slices), from PD40 rats, were also recorded in a submerged chamber and superfused with the same saline at 1 ml/min (3436°C). Intracellular recordings were performed with sharp microelectrodes filled with 3 M K-acetate (DC resistances: 80120 M
) and the help of an active bridge electrometer (Neuro Data, Cygnus Tech, DWG) (Pineda et al. 1992
). Records were digitized and saved on VHS tapes (40 KHz) and analyzed off-line in a PC clone. Stimulation consisted of intracellular injections of constant current steps to evoke the AHP after a single action potential (Pineda et al. 1992
). Stimulus were of suprathreshold intensity and given at a holding potential of 55 to 60 mV by adjusting constant current. Bridge balance as well as recovery periods (without DC current) were monitored between sample records. After recording, some neurons were injected with biocytin as previously described (Galarraga et al. 1999
). All neurons identified in this study were medium-sized spiny projection neurons.
Drugs used
Drugs were dissolved in the bath saline from daily made stock solutions using a gravity-driven superfusion system. Equilibrated concentrations of the drugs were achieved in 45 min. L-
-amino-3-hydroxy-5-methyl-isoxazolepropionate and 2-carboxy-3-carboxymethyl-4-isopropenylpyrrolidine/kainate (AMPA/KA) antagonist, CNQX (10 µM), N-methyl-D-aspartate (NMDA) antagonist, AP5 (50 µM), nitrendipine, bicuculline, QX-314, TTX, sulpiride, quinelorane, and quinpirole were all purchased from Sigma-RBI (St. Louis, MO). Calcium channels antagonists
-conotoxin GVIA (
-CgTx) and
-agatoxin TK (
-AgaTK) were obtained from both Peptides international (Louisville, KY) and Alomone Labs (Jerusalem, Israel). Almost all drugs were dissolved in water to get stock solutions and added to the superfusate to reach the final concentration. Nitrendipine was disolved in dimethylsulfoxide (DMSO; 0.01%).
Data analysis
Digitized data were imported for analysis and graphing into commercial software (origin v.6. Microcal, Northampton, MA). Mean ± SE of all ICa2+ and IPSCs are reported. Free-distribution statistical tests were used to assess statistical significance; Mann-Whitney U test or Wilcoxon's t-test (depending on nonpaired or paired samples) and one-way ANOVA with post hoc Tukey's test were used to assess significance between multiple samples in which quinelorane modulation was tested in the presence or the absence of Ca2+ channel blockers (for details see Vilchis et al. 2000
).
To approximate the contribution for each type of Ca2+ channel, we took the amount of Ca2+ current blocked by a given antagonist (nitrendipine for L-type;
-AgaTK for P/Q-type; and
-CgTx for N-type) as the contribution of a given channel type to the whole cell calcium current. Current through R-type Ca2+ channels was calculated as the amount of current left after all three blockers were added together. The sum was normalized to 100% so that L + N + P/Q + R = 100%. To determine the amount of quinelorane modulation for each type of Ca2+ channel, we compared quinelorane modulation on Ca2+ currents in the absence and presence of the different Ca2+ channels blockers and introduced the data into a system of linear equations (for details, see Vilchis et al. 2002
).
| RESULTS |
|---|
|
|
|---|
Figure 1A shows a family of Ca2+currents obtained in response to depolarizing step voltage commands (see METHODS) from a PD14 neuron. Ba2+ was the charge carrier. Figure 1B shows a similar experiment done in a PD40 neuron. Figure 1C shows the Ca2+ current evoked in the same PD14 neuron (Fig. 1A) after a depolarizing ramp command from 80 to 50 mV. A similar experiment is shown in Fig. 1D for the PD40 neuron. Figure 1, E and F, shows the current-voltage relationships (I-V plots) obtained from the experiments depicted in Fig. 1, AD. Filled circles are measurements from the currents depicted in Fig. 1, A and B, measured at the end of the voltage step. Continuous traces are taken from currents depicted in Fig. 1, C and D, plotted as a function of ramp voltage. Note that ramp-evoked I-V plots look as the "fits" of the I-V plots obtained with step commands; indicating a virtual agreement between both recording methods. For the sake of clarity, most figures will show I-V plots obtained with ramp commands only. However, most experiments were done with both protocols. Histogram in Fig. 1G compares mean current amplitudes at both developmental stages. In PD14 neurons, peak current amplitude was 234 ± 14 (SE) pA (n = 49), whereas it was 254 ± 12 pA in PD40 neurons [n = 54; not significant (NS); Mann-Whitney U test]. Because current density could change during development, this parameter was compared in Fig. 1H by dividing ionic current over whole cell capacitance (CN). CN in PD14 cells was 7.3 ± 0.2 pF, whereas it was 7.5 ± 0.3 pF in PD40 cells (NS; Mann-Whitney U test), indicating that somatic surface does not change between these developmental stages. Accordingly, normalized currents were 32 ± 2.2 pA/pF in PD14 and 34 ± 1.4 pA/pF in PD40 cells (Fig. 1H; NS; Mann-Whitney U test). These measurements made us confident that the total number of Ca2+ channels does not increase between these two developmental stages. Experiments to document possible changes in kinetics or voltage dependence were not done in this work (cf., Fig. 1, A and B). Instead we focused on the channel types that contribute to the whole cell current, because, even if channel numbers do not increase, the channel types that compose the current may change. To test this expectation, a pharmacological dissection of Ca2+ currents at both stages, PD14 and PD40, was done with the help of known Ca2+ channel antagonists.
Figure 2A shows that, in PD14 neurons, 10 µM nitrendipine blocked 29 ± 3% (n = 15) of the Ca2+ current, whereas in PD40 neurons, the percentage of Ca2+ current blocked by dihydropyridine was 23 ± 2% (n = 24; this difference was not statistically significant; Mann Whitney U test; Fig. 2C). Therefore it was concluded that L-type Ca2+ channels do not change their numbers between these developmental stages. However, further study is needed to see if the subtypes of L-type channels (
1D and
1C) do change (Olson et al. 2005
).
|
-AgaTK (400 nM). This peptide reduced Ca2+ current in PD14 cells by 24 ± 3% (n = 7; Fig. 2D), whereas it reduced Ca2+ current in PD40 cells by 40 ± 1% (n = 15; Fig. 2E). This difference was statistically significant (Fig. 2F; P < 0.0001; Mann-Whitney U test); indicating that P/Q-type Ca2+ channels increase their numbers and their relative contribution to the whole cell Ca2+ current between these developmental stages.
Equivalent experiments were done using the CaV2.2 (N-type) channel blocker,
-CgTx (1 µM). This peptide reduced Ca2+ current in PD14 cells by 43 ± 2% (n = 11; Fig. 2G), whereas it reduced Ca2+ current in PD40 cells by 28 ± 2% (n = 8; Fig. 2H). This difference was statistically significant (Fig. 2I; P < 0.002; Mann-Whitney U test), indicating that N-type channels decrease their numbers and their relative contribution to the whole cell Ca2+ current during the PD14 to PD40 developmental window. Comparable findings have been reported in cultured pyramidal neurons (Chameau et al. 1999
).
Finally, to isolate the CaV2.3 (R-type) current, we added the three blockers together (nitrendipine,
-AgaTK, and
-CgTx) at the above concentrations and measured the current that was left. Less than 15% of the initial current was left in both PD14 and PD40 cells (Fig. 2, JL), indicating that, as L-type channels, R-type channels do not change their numbers during this developmental period (Bargas et al. 1994
; Foehring et al. 2000
).
In PD14 cells, the percentage that each current type contributes to the whole cell current was (see METHODS and Vilchis et al. 2002
) as follows: L-type, 27%; P/Q-type, 22%; N-type, 39%; R-type, 12% (Fig. 3A). In PD40 cells, the percentages were as follows: L-type, 22%; P/Q-type, 38%; N-type, 26%; R-type, 14% (Fig. 3B). Therefore the results show that expression of CaV2.2 (N-type) channels diminishes, whereas that of CaV2.1 (P/Q-type) channels augments during this developmental interval. Because one of the main functional roles of N- and P/Q-type Ca2+ channels in spiny neurons is the triggering of transmitter release (Tecuapetla et al. 2005
), we chose to start searching into the functional significance of these changes by hypothesizing that GABA released from the terminals of mature spiny neurons should be mainly triggered by P/Q-type Ca2+ channels.
|
To test the above hypothesis, we observed eIPSCs elicited by the activation of axon collaterals that interconnect spiny neurons. These were obtained by electrically stimulating the globus pallidus to antidromically activate striofugal axons and, therefore axon collaterals that stay inside the neostriatum and interconnect spiny neurons (Guzman et al. 2003
). At the PD14 stage, both N- and P/Q-types Ca2+ channels trigger GABA release from these terminals (Tecuapetla et al. 2005
) (Fig. 4, A and D):
-CgTx and
-AgaTK reduced the amplitude of antidromically eIPSCs by 65 ± 7 (n = 10) and 90 ± 5% (n = 8), respectively (Tecuapetla et al. 2005
)both peptides were used at saturating concentrations (1 µM and 400 nM, respectively).
|
-CgTx on antidromically eIPSCs was significantly less important in PD40 neurons compared with PD14 neurons: in PD40 neurons,
-CgTx blocked 18 ± 7% of the current (n = 5; Fig. 4B; P < 0.009; Mann Whitney U test; Fig. 4C). On the other hand, the P/Q-type Ca2+ channel blocker,
-AgaTK, virtually blocked all eIPSCs in PD40 cells: 98 ± 1% (n = 5; Fig. 4E). Thus the effect of
-AgaTK was not significantly different when comparing both developmental stages (Fig. 4F; Mann Whitney U test); it is only the participation of N-type Ca2+ channels in synaptic transmission that is decreased during maturation. The participation of P/Q-type Ca2+ channels becomes the most important. These experiments show that the reconfiguration of Ca2+ current types at the terminals parallels that occurring at the soma. A similar developmental change has been reported in other synaptic terminals, some of them, inhibitory (Iwasaki and Takahashi 1998Dopaminergic D2 receptor modulation of Ca2+ currents in young and mature neurons
An important characteristic of medium spiny neurons is their having D2 dopaminergic receptors whose activation modulates Ca2+ currents. However, only the modulation of CaV1 Ca2+ channels has been extensively studied (Hernandez-Lopez et al. 2000
; Olson et al. 2005
; Rakhilin et al. 2004
). Here we report that modulation of CaV2 Ca2+ channels is also very important. Figure 5 confirms the D2 modulation of Ca2+ current in neostriatal cells, showing that this modulation is reversible during the time course of a typical experiment. This modulation reduces the slower tail currents recorded after a step command (Fig. 5B, inset), finding that has been taken as evidence of L-type current modulation (Olson et al. 2005
; Rakhilin et al. 2004
). D2 modulation of Ca2+ current was observed in 80% of the recorded dissociated neurons at each developmental stage.
Figure 6, AC, shows that the D2 modulation of Ca2+ current (10 µM quinelorane) is present in both juvenile (PD14) and mature (PD40) neurons. This modulation comprises about the same amount of current reduction at both developmental stages (Fig. 6C; NS). Quinelorane acted in a concentration-dependent fashion (data not shown) at both stages. Thus 10 µM quinelorane reduced the current in PD14 by 35 ± 2% (n = 12) of neurons and by 36 ± 3% (n = 12) in PD40 neurons (NS; Mann-Whitney U and ANOVA tests). All the modulation produced by quinelorane and other D2 receptor agonists (e.g., quinpirole) was virtually blocked by 1 µM of the unselective D2 receptor class antagonist, sulpiride (1 µM), at both stages (n = 12; data not shown).
|
PD14) (Cooper and Stanford 2001
In juvenile animals (PD14), the mean paired pulse ratio (PPR) of eIPSCs after a pair of stimulus (Fig. 6D, inset) in the control condition was 1.3 ± 0.16 and in the presence of D2 receptor agonists was 1.9 ± 0.26 (n = 12; P < 0.002; Wilcoxon's t-test). In mature animals (PD40; Fig. 6E, inset), control PPR was 0.82 ± 0.1 and in the presence of D2 receptor agonists it was 0.97 ± 0.08 (n = 6; P < 0.05; Wilcoxon's t-test). Because the PPR value is linearly proportional to the probability of release, the paired pulse stimulation has been an accepted protocol to search for presynaptic actions (e.g., Baldelli et al. 2005
). Thus the modulation is suggested to be presynaptic in both juvenile and mature animals (Guzman et al. 2003
). However, to further support this inference, a mean-variance analysis was performed (Clements and Silver 2000
) and accompanied with an independent analysis of amplitude histograms.
Figure 7A shows an amplitude histogram of spontaneous IPSCs (sIPSCs) recorded in a neostriatal neuron (Fig. 7B). These events may come from interneurons or spiny cell terminals and fitted to Gaussian functions with a first modal amplitude of 8 pA (range, 512 pA), similar to that reported for other GABAergic synapses (e.g., Ling and Benardo 1999
). The second mode was 16 pA (range, 1220 pA) (cf., Paulsen and Heggelund 1994
). Thereafter, some eIPSCs were recorded after applying quasi-minimal stimulation strength. The amplitudes histogram of these eIPSCs was multimodal (Fig. 7, C and D), obtaining values of 9 and 17 pA for the first modes; the interval between peaks was 89 pA (Ling and Benardo 1999
). Similarity between sIPSCs and eIPSCs histograms supports the notion that terminals that interconnect spiny neurons are fairly homogeneous (Tecuapetla et al. 2005
), and their modal values can now be used to independently ascertain the results of the mean-variance analysis (Clements and Silver 2000
). eIPSCs were obtained from a range of eIPSCs amplitudes from different experiments (see intensity-amplitude plots in Tecuapetla et al. 2005
). Mean eIPSCs amplitudes were graphed against their peak variance (Fig. 7E, black circles) and the resultant plot could be fitted wit a parabola (see Clements and Silver 2000
) that yielded a weighted quantal amplitude, Qw = 8 ± 2 pA, notably similar to the first modal amplitude in the histograms. In the presence of the D2-class receptor agonist, quinelorane (1 µM), the cases appeared intermingled, with control cases of lower mean amplitudes at the initial portion of the parabola (Fig. 7E, gray circles), and linear fits of the two samples were not significantly different. Qw was 9 ± 2 pA for the D2 cases (NS; Mann-Whitney U test; also see amplitude histograms). Therefore mean-variance analysis supported the view that D2 receptor actions are presynaptic (i.e., initial parabola slopes do not diverge; Clements and Silver 2000
).
|
D2-class receptormediated actions on other ion channels have been linked to intracellular Ca2+ mobilization and calcineurin (CaN) (Hernandez-Lopez et al. 2000
; Hu et al. 2005
; Rakhilin et al. 2004
). However, it was shown (Fig. 8, AD) that the percentage of D2-mediated modulation on Ca2+ currents was not significantly different when the neurons were internally perfused with either 1 or 10 mM EGTA: 34 ± 2 and 36 ± 3%, respectively (NS; n = 6; Mann-Whitney U test). Nevertheless, D2-class receptormediated actions were virtually abolished when 15 mM BAPTA were perfused instead (Fig. 8, C and D), suggesting that D2 receptor actions were mediated by a Ca2+-dependent diffusible pathway (Dargan and Parker 2003
; Stefani et al. 2002
). Figure 8F shows Ca2+ current evoked with the double pulse protocol (depicted in Fig. 8E; see METHODS) in a PD14 neuron. It can be seen that the amount of current modulated by the D2 receptor agonists did not change significantly (Fig. 8H; Wilcoxon's t-test). The same result was found on PD40 neurons (Fig. 8, G and I) (Hernandez-Lopez et al. 2000
). The results suggest that the signaling pathway used by D2-class receptors to modulate CaV2 Ca2+ channels is a diffusible Ca2+-sensitive cascade, as is the case of CaV1 Ca2+ channels (Hernandez-Lopez et al. 2000
; Hu et al. 2005
; Olson et al. 2005
; Rakhilin et al. 2004
). In view of these results, the next question was whether this modulation was being mediated by the same Ca2+ channels at both developmental stages.
Ca2+ channels mediating dopaminergic D2 receptor modulation change during development
Because the amount of D2 modulation on the Ca2+ current was the same in juvenile (PD14) and mature (PD40) neurons, it was decided to use different Ca2+ channel blockers in conjunction to the D2-class receptor agonist, quinelorane (10 µM), to see if the channels being modulated were the same in juvenile and mature neurons. The calcium channel blockers, nitrendipine (10 µM),
-AgaTK (400 nM), and
-CgTx (1 µM) were used, the rationale being that, if one channel type mediates the D2-induced modulation, its blockage would occlude a part of this modulation (Perez-Rosello et al. 2005
; Vilchis et al. 2002
). However, if a channel type does not mediate most of the D2-induced modulation, its blockage would not occlude the modulation significantly because percentage of modulation is calculated from the current left after the action of a given blocker. Nonetheless, the percentages of current components vary somewhat from cell to cell, and multiple samples were compared with the same control samples (quinelorane's effect without channel blockers). Therefore both parametric and nonparametric (Kruskal-Wallis) ANOVA tests were used to compare all samples altogether taken into account intrasampling variance. Both tests yielded significant differences when the D2-class agonist was administered (F = 15.2; K-W: 53.8; P < 0.0001 in both cases). Therefore post hoc (Tukey's test) pair to pair comparisons were used to contrast different groups. At the PD14 stage, pretreatment with nitrendipine did not completely block quinelorane action but reduced it from 35 ± 2 to 26 ± 2%; (n = 11; Fig. 9A;
9% occlusion; P < 0.05; ANOVA with post hoc Tukey's test), as expected after a block of a modulated channel. This shows that D2 receptor activation modulates CaV1 Ca2+ channels at this stage, but also shows that there is a remaining modulation on CaV2 Ca2+ channels. Still, a more important occlusion of the D2 agonist effect by the L-type channel blocker was observed in PD40 neurons: quinelorane action was reduced from 36 ± 3 to 17 ± 1% (n = 11; Fig. 9B;
19% occlusion; P < 0.0001; ANOVA with post hoc Tukey's test). When the degree of occlusion by the L-type channel blocker was compared between both developmental stages, the difference was statistical significant (P < 0.04; ANOVA and post hoc Tukey's test). These experiments show that the dopaminergic D2 modulation of L-type Ca2+ channels is not only preserved but enhanced significantly with age. This suggests that the modulation of the dynamic firing range (Perez-Garci et al. 2003
) enhances its importance with network development.
|
-AgaTK in PD14 neurons was not enough to significantly occlude the D2 agonist modulation in PD14 neurons:
-AgaTK reduced quinelorane effect from 35 ± 2 to 27 ± 2%; (n = 6; Fig. 9C; P < 0.3; ANOVA with post hoc Tukey's test). However, in PD40 neurons, the P/Q-type channel blocker produced a significant occlusion of the D2 agonist effect: quinelorane action was reduced from 36 ± 3 to 20 ± 1% (n = 6; Fig. 9D;
44%; P < 0.0001; ANOVA with post hoc Tukey's test). When the degree of occlusion by P/Q-type channel blockers was compared between both developmental stages, the difference was statistical significant (P < 0.05; ANOVA with post hoc Tukey's test). This means that modulation of P/Q-type channels by D2 receptor activation is small in juvenile (PD14) neurons but becomes important in mature (PD40) neurons.
A completely different outcome was obtained with the N-type channel blocker
-CgTx. In the presence of this blocker, the D2 agonist action on Ca2+ current of PD14 neurons was reduced from 35 ± 2 to 12 ± 2% (n = 8; Fig. 9E;
66% occlusion; P < 0.0001; ANOVA with post hoc Tukey's test). Therefore
-CgTx resulted in the most potent agent capable to occlude the dopaminergic modulation in juvenile neurons, suggesting that modulation of CaV2 Ca2+ channels by dopamine is more important than modulation of CaV1 Ca2+ channels at this stage. In contrast,
-CgTx was unable to significantly decrease the action of D2 agonists on the Ca2+ current of PD40 neurons. Ca2+ current reduction by quinelorane in the presence of
-CgTx in PD40 neurons was from 36 ± 3 to 31 ± 3% (n = 8; Fig. 9F; P < 0.9; ANOVA and post hoc Tukey's test). When the degree of occlusion by the N-type channel blocker was compared between both developmental stages, the difference was statistical significant (P < 0.0001; ANOVA and post hoc Tukey's test). This means that modulation of N-type channels by D2 receptor activation is large or very important in juvenile (PD14) neurons, but it is much smaller in mature (PD40) neurons.
A summary of all samples compared can be seen in Fig. 10. It was observed that the most potent blocker of the D2 agonist action in young (PD14) neurons was
-CgTx, followed by nitrendipine. However, in mature neurons (PD40), both nitrendipine and
-AgaTK were potent blockers of the D2-agonist action on Ca2+ currents, but
-CgTx did not longer occlude dopaminergic actions. In summary, a considerable switch in target for D2 receptor signaling occurs in medium spiny neurons during development.
|
L (0.25) + P/Q (0.17) + N (0.63) + R (0), where the L, P/Q, N, and R coefficients denote the percentage that each channel type contributes to the whole cell Ca2+ current in PD14 neurons (see Fig. 3A), and numbers in parentheses are the solved unknowns. The same system of equations applied for PD40 neurons explained the 36% modulation at this stage: 36
L (0.81) + P/Q (0.45) + N (0.04) + R (0) (see Fig. 3B). This analysis helped to approximate a quantitative assessment of the results described above: 1) L-type channel D2 modulation increased during development from 25 to 81% becoming the most important Ca2+ channel modulation with age, 2) P/Q-type channel modulation increased during development to 45%, becoming the second most important channel being modulated by D2-class receptors, and 3) N-type channels are the most modulated in juvenile neurons (63%); however, their modulation tends to disappear in mature neurons.
If the above data are correct, the following predictions can be made to further corroborate these findings: first, the conjoined administration of nitrendipine (10 µM) and
-AgaTK (400 nM) will virtually abolish quinelorane action on the Ca2+ current of PD40 neurons, but not in PD14 neurons (cf., Fig. 11, A and B). Conversely, the conjoined administration of nitrendipine (10 µM) and
-CgTx (1 µM) will almost totally block the action of the D2 receptor agonist on the Ca2+ current of PD14 neurons, but not in the case of PD40 neurons (cf., Fig. 11, D and E). Thus after L-type and P/Q-type Ca2+ channels were blocked, quinelorane was still capable to reduce the remaining current by 19 ± 1% (n = 6) in PD14 neurons (Fig. 11C). In contrast, when L-type and P/Q-type Ca2+ channels were blocked in PD40 neurons, quinelorane action was virtually abolished to 3 ± 1% (n = 6: Fig. 11C; P < 0.0001; ANOVA with post hoc Tukey's test). Conversely, when L-type and N-type channels were blocked in PD14 neurons, almost no quinelorane action was left: 6 ± 0.7% (n = 6; Fig. 11, D and F). The opposite findings were observed at the PD40 stage: quinelorane was still capable to modulate the remaining Ca2+ current by 18 ± 2% in the same conditions (n = 6; Fig. 11F; P < 0.0001; ANOVA with post hoc Tukey's test). Globally, both parametric and nonparametric (Kruskal-Wallis) ANOVA tests were significant (F = 50.7; K-W: 18.5; P < 0.0001 for both tests). In summary, besides modulating CaV1 Ca2+ channels, D2 receptor activation also modulates CaV2 Ca2+ channels. However, the modulation is mainly on CaV2.2 Ca2+ channels (N-type) in juvenile striatal neurons (PD14), whereas it is mainly on CaV2.1 Ca2+ channels (P/Q-type) in mature striatal neurons (PD40).
|
To study whether the change of target affects presynaptic inhibition (Guzman et al. 2003
), the effect of the D2 receptor agonist, quinelorane (1 µM), in the presence of
-CgTx (1 µM) was examined on the eIPSCs generated by antidromic activation of the axon collaterals. Figure 12A shows that
-CgTx (1 µM) almost abolished the D2 agonist mediated reduction of IPSCs in PD14 terminals: from a 50 ± 8% reduction in the control (Guzman et al. 2003
) to only 10 ± 2% reduction in the presence of
-CgTx (n = 3; Fig. 12A; P < 0.02; Mann-Whitney U test). A posterior application of
-AgaTK (400 nM) blocked the remaining eIPSC (Fig. 12A), confirming that P/Q-type channels were also present at the terminals and that quinelorane only had its full action if the main target, N-type channels, were present in juvenile terminals.
|
-CgTx did not impede D2 agonist actions in PD40 terminals. Reduction of IPSCs was 42 ± 5% in the controls and 40 ± 8% in the presence of
-CgTx (n = 5; Fig. 12B; NS; Mann-Whitney U test). This shows that presynaptic N-type channels were no longer the target for D2 receptors at this stage. These results are summarized in the histogram of Fig. 12C (P < 0.009; Mann-Whitney U test), which shows the effect of quinelorane in the presence of
-CgTx at this stage. Thus the dopaminergic modulation that distinguishes these terminals is present at both developmental stages but mediated by a different Ca2+ channel type. AHP in medium spiny neurons is negligibly affected by D2 receptor agonists
Because P/Q-type Ca2+ channels are a main source of Ca2+ to activate the Ca2+ dependent K+ channels that generate the AHP in medium spiny neurons (Vilchis et al. 2000
), D2 receptor agonists were tested on the AHP of mature spiny neurons after a single action potential. Two D2 receptor selective agonists were used: 5 µM quinelorane and 5 µM quinpirole. Figure 13 shows that, in mature spiny neurons, no change in AHP was observed in n = 6/18 neurons (33%; Fig. 13, AC), and a mild change was observed in n = 12/18 neurons (66%; Fig. 13, DF) for a mean overall reduction of 1.3 ± 0.26 mV (20%). Note that even when D2-mediated modulation of Ca2+ currents was seen in 80% of dissociated cells, only about 60% of intracellularly recorded (PD40) neurons exhibited AHP modulation. This is a number more in agreement with the expected number of cells presenting D2-type responses, probably suggesting that not all members of the D2 receptor family (D2, D3, D4) participate in the same specific cellular functions.
|
| DISCUSSION |
|---|
|
|
|---|
Dopaminergic modulation of Ca2+ currents undergoes a shift in target during development
Previous studies have shown that medium spiny neurons express functional D2-class receptors (e.g., Aizman et al. 2000
; Surmeier et al. 1993
, 1996
). These receptors increase their number between the third and fourth postnatal weeks (Broaddus and Bennett 1990
), when an important escalation in the numbers of dopaminergic synaptic terminals occurs (Antonopoulos et al. 2002
).
This work reports that N and P/Q-type channels are modulated in neostriatal projection neurons by D2-class receptor activation. If only the whole cell Ca2+ current is observed, no difference could be seen in the amount of current modulated when comparing young and mature neurons. Moreover, inhibitory transmission between spiny neurons was similarly modulated in young (Guzman et al. 2003![]()