Experimental subjects and design
Experimental subjects, obtained from IFC bioterium were: B6; 12P2-Pvalbtm1(cre)Arbr/J (PV-Cre; Silvia Arber, Friederich Miescher Institute; Jackson Labs, stock# 008069), called PV+ mice from now on. Experimental subjects were housed in acrilic cages (4–5 mice per cage; 19 × 29 × 12 cm) with wood-based bedding and cardboard cylinders, kept on a 12:12 light/dark (light beginning at 8 am) period with a temperature maintained at 20–21 °C in IFC vivarium after surgery (see below) until used for experiments. All animals had standard rodent chow and water ad libitum. In order to identify isolated PV+ interneurons, PV-Cre transgenic mice at PD 21 (21 days, mean ± 4 days, 30 g mean ± 4, at 14–18 h), were anesthetized i.p. with ketamine (Bayer 75 mg/kg) and xilazine (Bayer 10 mg/kg) and injected stereotaxically in a laminar flow hood (Telsar technologies. Model PV-30/60) in a dedicated, sterile room, with the following viral constructs (University of Pennsylvania Vector Core): AAV2/1.CAG.Flex.tdTomato.WPRE.bGH (Honguki Zeng) for whole cell recordings in isolated cells, AAV1.Syn.Flex.GCaMP6f.WPRE.SV40 [29], for calcium imaging recordings and AAV1.CAG.Flex.eGFP.WPRE.bGH (Allen institute) for some current clamp experiments in slices at the following coordinates relative to bregma (in mm): AP = 0.9, ML = ± 1.2, DV = − 3.2. The total virus volume injected was 0.8 µl over a period of 10 min (Fig. 1a). Animals were monitored for two weeks to ensure full recovery and fluorescent protein expression (Fig. 1b). A total of 45 infected PV-Cre mice were randomly assigned to 6 independent groups: for voltage clamp recordings of calcium currents (see next sections for details of the techniques) to observe contribution of Ca2+ channels classes (Fig. 2; n = 19 recordings from 18 different mice, below); effects of DA on Ca2+ currents (Fig. 3a, b; n = 8 recording from 8 different mice); SCH + SKF control group (Fig. 3c, d; n = 6 recordings from 4 different mice); nicardipine on DAergic actions (Fig. 4; n = 8 recordings from 6 different mice); current clamp recordings in slices (Fig. 5; n = 6 recordings from 6 different mice for SKF-nicardipine experiments and n = 4 for SKF-SCH experiments) and calcium imaging experiments (Fig. 6; n = 33; for imaging PV-cre identified FSI were extracted from 6 different experiments/slices from 3 different mice). The experimental units were single neuron recordings or changes in fluorescence (∆F/F where ∆F = changes in fluorescence and F = basal fluorescence). Subject numbers were minimized to obtain statistical significance.
Preparation of dissociated neurons and slices
Brain slices and acutely dissociated neurons were obtained and described in previous work [30,31,32,33,34]. Briefly, infected PV-Cre mice were anesthetized (see above). The mice were decapitated, their brains were removed and submerged in iced saline solution containing (in mM): 126 NaCl, 3 KCl, 26 NaHCO3, 2 CaCl2, 1 MgCl2, 11 glucose, 0.2 thiourea and 0.2 of ascorbic acid (25 °C; pH: 7.4 with HCl, 300 ± 5 mOsm/l with glucose; saturated with 95% O2 and 5% CO2). Using a vibratome (1000 Classic, Warner Instruments, Hamden, USA), sagittal brain slices of 300 µm thick were cut and placed in the same saline solution for 1 h at 34 °C. When recordings were done in slices, they were transferred to a submerged chamber and superfused at 5 ml/min with saline solution. When recordings were done in dissociated cells, the dorsal striatum was dissected from the slices and returned into the saline solution containing 10 mM HEPES plus 0.5 mg/ml of papain (Carica papaya; Calbiochem, Cat# 5125. San Diego CA) at 34 °C. After 20–25 min of digestion, the striatum slices were transferred to a low Ca2+ (0.4 mM CaCl2) saline solution. To obtain individual cells, the striatal slices were mechanical dissociated with a graded series of fire-polished Pasteur pipettes. The cell suspension (1 ml) was plated into a Petri dish mounted on the stage of an inverted microscope (Nikon Instruments, Melville, NY, 20 ×/0.4 NA). Cells were left for 10–15 min for neurons to adhere to the bottom of the dish. The dish contained 1 ml of the whole-cell recording saline solution (in mM): 0.001 tetrodotoxin (TTX), 140 NaCl, 3 KCl, 5 BaCl2, 2 MgCl2, 10 HEPES, and 10 glucose (pH: 7.4 with NaOH; 300 ± 5-mOsm/l with glucose). Thereafter, the cells were superfused at 1 ml/min with saline of the same composition at room temperature (approximate 25 °C). Tomato-positive neurons were visualized using a UV lamp (X-Cite; EXFO, Ontario, Canada; Fig. 1b). Dissociated neurons lack their distal dendrites and axon, so currents reported are somatic.
Voltage clamp recordings of calcium currents
Voltage-clamp recordings were performed on identified striatal PV+ interneurons with 12–15 µM soma diameter and whole-cell capacitance of 6–7 pF with short or absent dendritic trunks [32, 34]. Patch pipettes of borosilicate glass (WPI, Sarasota, FL, USA) were pulled in a Flaming-Brown puller (Sutter Instrument Corporation, Novato, CA, USA) and fire polished prior to use. The internal saline solution contained (in mM): 180 N-methyl-d glucamine (NMDG), 40 HEPES, 10 EGTA, 4 MgCl2, 2 ATP, 0.4 GTP and 0.1 leupeptin (pH = 7.2 with H2SO4; 280 ± 5 mOsm/l; room temperature around 25 °C). Whole-cell recordings used electrodes with D.C. resistance of 3–6 MΩ in the bath. Liquid junction potentials (5-10 mV) were corrected. Recordings of Ca2+ currents were obtained with an Axopatch 200B patch-clamp amplifier (Axon Instruments, Foster City, CA, USA) and controlled and monitored with pClamp (version 8.2, RRID: rid_000085) and a 125 kHz DMA interface (Axon Instruments, Foster City, CA, USA). We recorded currents passing through Ca2+ channels using Ba2+ as a charge carrier as shown in previous articles [31, 34, 35]. Ba2+ is a potent K+ blocker. In addition, intracellular K+ was replaced by 180 mM NMDG. Na+ channels were blocked with 1 µM TTX. Currents isolated in this way were completely blocked by 200–400 µM Cd2+ (Fig. 1f) in this way, and for simplicity, we will refer to these currents as Ca2+ currents. Current–voltage relationships (I–V plots) were generated before and after drug application. Figure 1c shows representative Ca2+ currents evoked with 20 ms rectangular voltage commands from − 80 to 50 mV in 10 mV steps. Figure 1d shows a representative Ca2+ current in response to a voltage ramp command (0.7 mV/ms) from − 80 to 50 mV. When I–V plot from both methods coincide, space-clamp was considered acceptable (Fig. 1e). For clarity, most figures only show representative responses to voltage ramps.
Current clamp recordings in slices
Current clamp recordings were performed with the patch clamp technique in the whole cell configuration of PV+ neurons of infected mice ranging in age 28–60 days. Sagittal slices (250–300 μm thick) were cut using a vibratome (1000 Classic, Warner Instruments, Hamden, USA), transferred to a recording chamber and superfused continuously with oxygenated saline solution (5 ml/min) at room temperature (~ 25 °C). Neurons within the striatum were visualized with infrared differential interference contrast videomicroscopy and PV+ neurons were identified using epifluorescent illumination with a 40 × immersion objective (0.8 NA; Nikon Instruments, Melville, NY). Micropipettes were pulled (Sutter Instrument, Novato, CA) from borosilicate glass tubes (WPI, Sarasota, FL) to an outer diameter of 1.5-mm for a final D.C. resistance of 4–6 MΩ when filled with internal saline. The internal solution contained (in mM): 120 KSO3CH4, 10 NaCl, 10 EGTA, 10 HEPES, 0.5 CaCl2, 2 MgCl2, 2 ATP-Mg, and 0.3 GTP-Na (pH = 7.3, 290 mOsM/l). Recordings were made with an Axopatch 200A amplifier (Axon Instruments, Foster City, CA) and data were acquired with the Im-Patch© software designed in the Lab View environment (freely available for download at im-patch.com). Evoked firing responses at different depolarizing membrane potentials were obtained before and after a selective dopamine receptor agonist was administered. Current–voltage relationships made in current-clamp mode superimposed tightly with those performed in voltage-clamp mode at steady state, suggesting that neither bridge balance, nor series resistance, represented a problem in our recordings.
Digitalized electrophysiological data were imported and analyzed into Origin v8, Microcal (Northampton, MA), and MatLab (The Mathworks Inc. Natick, MA). Data are presented as the mean ± standard error (SEM). Firing rate plots were made by taking firing rate at rheobase in the different pharmacological conditions (Fig. 5c). Free-distribution statistical tests Wilcoxon’s T test and Friedman, one-way ANOVA with post hoc Dunn’s tests were used to assess statistical significance between paired or unpaired samples comparisons. Statistical significance was defined by P-values below 0.05.
Calcium imaging recordings
Calcium imaging recordings were obtained from PV+ neurons of mice infected with a Cre-dependent GCamp6f expression. Recordings were performed in saline solution containing (in mM): 126 NaCl, 2.5 KCl, 26 NaHCO3, 1.2 NaHPO4, 1 CaCl2, 1.3 MgCl2, 10 glucose, 0.2 thiourea and 0.2 of ascorbic acid (25 °C; pH: 7.4 with HCl, 300 ± 5 mOsm/l with glucose; saturated with 95% O2 and 5% CO2). For recordings, a microscope equipped with a 20 × 0.95 NA water-immersion objective (XLUMPlanFI, Olympus, Center Valley, PA) which has an image field of 750 × 750 μm, was used. To observe spontaneous changes in GCamp6f fluorescence intensity, light pulses at 488 nm (15–50 ms exposure) were delivered to the preparation with a Lambda LS illuminator (Sutter instruments, Petaluma, CA) connected to the microscope via optic fiber. Brief image sequences or movies (~ 180 s per epoch) were acquired with open access Im-Patch© software [6] at time intervals of 5–10 min during ≥ 60 min with a cooled digital camera (CoolSnap K4, Photometrics, Tucson, AZ) and 100–250 ms/image frame. Ca2+ entry was seen as spontaneous neuronal intrasomatic Ca2+ transients in PV+ neurons whose first time derivative reflects the time of electrical activity [36]. Activity of each cell was illustrated as dots in a raster plot.
Inmunocytochemical procedures
PV-Cre mice were infected as described earlier. Mice were deeply anesthetized (see above) and perfused transcardially with a solution of 4% paraformaldehyde in PBS. Thereafter, animals were decapitated and their brains removed from the skull and fixed overnight with 4% paraformaldehyde in PBS. The brains were then cut on a vibratome into 40 μm slices that were incubated 30 min with 1% bovine albumina to block nonspecific binding sites and for 36 h with a rabbit polyclonal antibody against parvalbumin (anti PV 1:2000 Abcam dissolved in PBS containing 0.25% Triton-X). The slices were then rinsed thrice with PBS and incubated with a goat versus rabbit secondary antibody (1:200 Vector Laboratories, Burlingame, CA, dissolved in PBS containing 0.25% Triton-X) during 1 h. This antibody was conjugated with FITC (Vector Laboratories, Burlingame, CA). Samples were mounted with vectashield (Vector Laboratories, Burlingame, CA) and observed in a confocal microscope ZEISS LSM 700 (10 ×/1.0 NA) (n = 10).
Drugs
For dissociated cell recordings, drugs were applied with a gravity-fed system that positioned a glass capillary tube 100 μm from the recording cell in the direction of superfusion flow. Solution changes were performed with a D.C. controlled microvalve system (Lee; Essex, CT, USA). This method allowed reversible drug applications [26, 33]. For current clamp recordings drugs were administered into the bath saline. Substances used were the DA receptor D1-like selective agonist SKF 81297 (Cat# S143), DA receptor D1-like antagonist SCH 23390 (Cat# 125941-87-9), Ca2+ CaV1 antagonist nicardipine (Cat# N7510) all from Sigma-Aldrich-RBI (St Louis, MO, USA); Ca2+ CaV2.2 blocker ω-conotoxin GVIA (Cat# C-300), Ca2+ CaV3 blocker TTA-P2 (Cat# T-155), Ca2+ CaV2.3 blocker SNX-482 (Cat# RTS-500), Na+ blocker tetrodotoxin (TTX) (Cat# T-550) from Alomone Laboratories (Israel) and Ca2+ CaV2.1 blocker ω-agatoxin TK (Cat# 4294-s) from Peptides International (Louisville, KY).
Data analysis
Collected digitalized data were analyzed and plotted using commercial software (Origin v8, Microcal, Northampton, MA, USA; RIDD: rid_000069). We report mean ± SEM of peak Ca2+ currents changes for dissociated FSI recordings without assuming normal distributions. We also used the 5, 25, 50 (median), 75 and 95 percentile ranges of absolute current values represented as Tukey box plots. Friedman, Kruskal–Wallis or Wilcoxon test with post hoc Dunn for multiple comparisons tests were used (signaled in each Result). Friedman and Wilcoxon test were used when we compared the same samples in two or three different conditions (before, during and after application of a drug). P< 0.05 was used as significance threshold. Analysis was conducted by GraphPad Prism 6.01 (La Joya, CA). Here, Ba2+ currents are reported as Ca2+ currents and graphs summarizing sampling results are illustrated. For current clamp recordings, we report mean ± SEM of firing rate. For calcium imaging experiments, activity of each FSI was determined as the total number of active frames/total number of frames. Finally, to quantify the amount of activity on each experiment, a cumulative activity plot was built on each condition.
Contribution of each class of Ca2+ channel to the whole-cell Ca2+ current
The method to obtain the average contribution of a given class of Ca2+ channel to the whole cell Ca2+ current was described in previous work [37]. Briefly, to approximate the contribution of each class of Ca2+ channel, the amount of Ca2+ current blocked by a given antagonist: nicardipine, ω-conotoxin GVIA (ω-CgTx), ω-agatoxin TK (ω-AgTx), TTA-P2 (TTA) and SNX-482 (SNX) was obtained by subtraction in the same or different experiments. Hardly all antagonists could be tested in a single experiment, but the amount blocked by each antagonist was taken no matter the number or order of the antagonists tested. This amount of blocked current was defined as the contribution of that specific channel class to the whole-cell control Ca2+ current normalized to 100% without any antagonist. Thereafter the data was introduced in the following system of linear equations:
$$\begin{array}{*{20}l} {0_{{{\text{X}}1}} + N_{{{\text{X}}2}} + PQ_{{{\text{X}}3}} + T_{{{\text{X}}4}} + R_{{{\text{X}}5}} = A} \hfill \\ {L_{{{\text{X}}1}} + 0_{{{\text{X}}2}} + PQ_{{{\text{X}}3}} + T_{{{\text{X}}4}} + R_{{{\text{X}}5}} = B} \hfill \\ {L_{{{\text{X}}1}} + N_{{{\text{X}}2}} + 0_{{{\text{X}}3}} + T_{{{\text{X}}4}} + R_{{{\text{X}}5}} = C} \hfill \\ {L_{{{\text{X}}1}} + N_{{{\text{X}}2}} + PQ_{{{\text{X}}3}} + 0_{{{\text{X}}4}} + R_{{{\text{X}}5}} = D} \hfill \\ {L_{{{\text{X}}1}} + N_{{{\text{X}}2}} + PQ_{{{\text{X}}3}} + T_{{{\text{X}}4}} + 0_{{{\text{X}}5}} = E} \hfill \\ \end{array}$$
where L, PQ, N, T and R are the contributions in percentage (± SEM) of each channel class: CaV1, CaV2.2, CaV2.1, CaV3 and CaV2.3 to the whole-cell Ca2+ current. For example, PQ refers to the current blockade by the selective P/Q type Ca2+ channel antagonist (ω-AgTx). Zero in the linear equation system means a blockade of a given Ca2+ channel class, thus, coefficients L, N, PQ, T or R were replaced by zero when the corresponding Ca2+ channel class was blocked. A, B, C, D or E stand for the mean percentage of Ca2+ current in the control (100%) with one channel class blocked (< 100%). Subscripts X1–X5 are the unknown variables, in other words, the values that multiply the coefficients L, N, PQ, T and R in order to determine percentage contribution of each channel to the whole-cell Ca2+ current.