These studies provide the first evidence that calcineurin selectively enhances L-type Ca²⁺ channel activity in neurons. Moreover, this action appears to be increased concomitantly with the well-characterized increase in L-type Ca²⁺ channel availability in hippocampal neurons with age-in-culture.

In turn, calcineurin may exert negative feedback over its own activation by inactivating specific Ca²⁺ sources, including voltage-sensitive Ca²⁺ channels (VSCCs) and N-methyl-d-aspartate (NMDA) receptors (NMDARs). However, this regulation is somewhat controversial. Although a number of studies suggest that calcineurin is important in Ca²⁺-dependent inactivation of VSCCs and NMDARs (Armstrong, 1989; Tong et al., 1995; Schuhmann et al., 1997; Lukyanetz et al., 1998) other reports have not reached a similar conclusion (Legendre et al., 1993; Krupp et al., 1996; Branchaw et al., 1997; Victor et al., 1997; Zeilhofer et al., 2000). One possible basis for conflicting results, at least for VSCCs, could be the differential expression of multiple VSCC types, or their association with disparate regulatory proteins in different cell types. There are at least five functional types of high-voltage-activated (HVA) VSCCs (L-, N-, P-, Q-, and R-type) in neurons (Fox et al., 1987; Bean, 1989; Fisher et al., 1990; Llinas et al., 1992; Mintz and Bean, 1993; Tsien et al., 1995) and it is well-documented that the mechanisms that regulate different VSCC types are highly selective, and moreover, can vary across cell types (Holz et al., 1986; Hescheler et al., 1987; Bean, 1989; Cox and Dunlap, 1992; Hille, 1994; Dolphin, 1995; Surmeier et al., 1995; Currie and Fox, 1997; Zamponi and Snutch, 1998; Ikeda and Dunlap, 1999). Even for a particular VSCC type, regulatory mechanisms may differ depending on state or environment. Regulation of the L-type VSCC by protein phosphorylation, for example, can depend on whether the full or truncated form of the α_1C (Ca_v1.2) subunit is present (Hell et al., 1993a,b, 1996), or on whether L-type VSCCs are examined in native vs. non-native cell types (Zong et al., 1995).

It is less well-recognized, however, that calcineurin may also positively regulate some ion channels, including the GABA_A receptor (Jones and Westbrook, 1997) and a renal Cl⁻ channel (Marunaka et al., 1998). Moreover, the possibility that calcineurin positively regulates VSCCs in some cell types has been raised recently. A series of preliminary studies in our laboratory revealed that the calcineurin inhibitor, FK-506, reduced rather than enhanced VSCC currents in cultured hippocampal neurons (Norris et al., 1999). A similar action of FK-506 on VSCCs also has been reported for arterial smooth muscle cells (Yasutsune et al., 1999). In addition, overexpression of calcineurin in transgenic mice enhances Ca²⁺ current density in cardiac myocytes (Yatani et al., 2001). Positive regulation of VSCCs by calcineurin in the hippocampus may be of particular interest, as calcineurin and multiple VSCC types are highly expressed in this brain structure (Fisher et al., 1990; Polli et al., 1991; Kuno et al., 1992; Llinas et al., 1992; Westenbroek et al., 1992; Hell et al., 1993a,b) and appear to be coupled functionally (Graef et al., 1999).

Here, we tested the possible positive modulation of VSCCs by calcineurin in hippocampal neurons using three distinct calcineurin inhibitors [the immunosuppressants FK-506 and cyclosporin A (CsA), and a calcineurin autoinhibitory peptide (CN-AIP)] and two negative controls (rapamycin and okadaic acid). Moreover, the effects of calcineurin inhibition on isolated VSCC subtype currents were examined to determine if hippocampal VSCC regulation by calcineurin is VSCC-type selective. Because L-type VSCC density and N-type VSCC regulation change with age in-culture (Porter et al., 1997; Blalock et al., 1999), somewhat analogously to changes with aging in vivo (Thibault and Landfield, 1996), we further examined the effect of age in culture on VSCC regulation by calcineurin.

Glass electrodes Recording pipets consisted of glass capillary tubes (Drummond Scientific, Broomall, PA, USA) pulled on a horizontal micropipet puller (model P-87; Sutter Instruments, Novato, CA, USA). Whole-cell pipets were coated with polystyrene Q-dope and had a mean tip resistance of 2.16 ± 0.02 MΩ. Cell-attached patch pipets were coated with Sylgard (Dow Corning, Midland, MI, USA) and had a mean tip resistance of 2.7 ± 0.07 MΩ. All recording pipets were fire-polished immediately before recording (Corey and Stevens, 1983).

Recording solutions For whole-cell recordings of isolated HVA VSCC currents, external solution contained (in mM): 111 NaCl, 5 BaCl₂, 5 CsCl, 2 MgCl₂, 10 glucose, 10 HEPES, 20 tetraethylammonium (TEA) Cl, 0.01 6-cyano-7-nitroquinoxaline-2,3-dione disodium (CNQX) and 0.001 tetrodotoxin (TTX). pH was adjusted to 7.35 using NaOH and osmolarity adjusted to 330 mOsm using sucrose. Pipette solution for whole-cell recordings contained (in mM): 145 methane sulfonic acid, 10 HEPES, 3 MgCl₂, 11 EGTA, 1 CaCl₂, 5 MgATP, 13 TEA Cl, 0.1 leupeptin. pH was adjusted to 7.35 using CsOH and osmolarity adjusted to 320 mOsm using high-performance liquid chromatography (HPLC) grade H₂O. This ratio of EGTA to Ca²⁺ buffers the intracellular Ca²⁺ concentration ([Ca²⁺]_i) at levels below resting values (e.g., at <100 nM, Bers et al., 1994). To examine whole-cell Ca²⁺, rather than Ba²⁺, currents some experiments exchanged external BaCl₂ for an equimolar amount of CaCl₂. In one subset of these experiments, the internal solution was unchanged, while in other experiments, EGTA and CaCl were omitted and MgATP was replaced with an equimolar amount of 2NaATP.

For cell-attached patch recordings of multichannel activity, the external bath solution contained (in mM): 140 K⁺ gluconate, 3 MgCl₂, 10 D-glucose, 10 EGTA, 10 HEPES. This solution, commonly used in single-Ca²⁺ channel studies, zeros the membrane and thus provides a convenient reference for setting the patch membrane potential (Fox et al., 1987; Fisher et al., 1990). pH was adjusted to 7.35 using KOH and osmolarity adjusted to 300 mOsm with HPLC grade H₂O. The pipet solution consisted of (in mM): 20 BaCl₂, 90 choline Cl, 10 TEA Cl, 10 HEPES. pH was adjusted to 7.35 using TEA-OH and osmolarity adjusted to 290 mOsm using sucrose. Prior to recording, the culture medium in each 35-mm dish was exchanged for 1.5 ml of external recording solution.

Data acquisition Recordings were obtained according to standard patch-clamp methods (Hamill et al., 1981) using an Axopatch 200A integrating patch-clamp amplifier (Axon Instruments, Foster City, CA, USA). Data were filtered at 2 kHz and digitized at 5 kHz. Voltage commands and data acquisition were controlled by pCLAMP software (Clampex, versions 6 and 7; Axon Instruments). All experiments were carried out at room temperature.

Whole-cell studies Prior to recording, junction potentials were nulled in the bath using the pipet offset control on the Axopatch 200A. Pipette capacitance was compensated. To estimate whole-cell membrane capacitance and pipet access resistance, a membrane current (filtered at 10 kHz, digitized at 91 kHz) was elicited at the beginning of each experiment with a 15-ms, 5-mV hyperpolarizing step from the holding potential (−70 mV). Current elicited by a 5-mV depolarizing pulse was equal in magnitude, but opposite in polarity. Due to their highly elaborate dendritic arbors, hippocampal neurons are not isopotential and exhibit capacitive current decay kinetics that are probably best fit by multiple exponential functions (Brown and Johnston, 1983; Johnston and Brown, 1983; Spruston et al., 1994). Therefore, we calculated whole-cell capacitance for each cell by integrating the area under the curve of the capacitive transient. The instantaneous peak current measured during the onset of the capacity transient was used to derive the pipet access resistance, which averaged 8.6 ± 0.14 MΩ, and was not significantly affected by the age of the cells or by the pharmacological agents used. Neurons in which the access resistance and/or the holding current changed dramatically during the course of an experiment were excluded from statistical analyses. Series resistance compensation, using the amplifier’s whole-cell correction parameters, was not performed because we, and others, have consistently found no significant differences in the shape or amplitude of activated currents before and after correction (Randall and Tsien, 1995; Porter et al., 1997; Blalock et al., 1999). For most experiments, the uncompensated error in the membrane potential during peak activation (i.e. voltage step to +10 mV) was estimated at 5–7 mV. Holding potential in all experiments was −70 mV. To generate current–voltage (I–V) relationships, cells were stepped in 10-mV increments (10-s inter-step interval) from their holding potential at −70 mV, to +60 mV. Drug effects were assessed using a voltage protocol that stepped the membrane potential (for 150 ms) from its holding potential to the voltage that elicited essentially maximal/HVA VSCC current (either +10 mV when EGTA was included, or +20 mV when EGTA was excluded from the pipet solution). Unless otherwise stated, five successive current records (30-s interstep interval) were averaged to estimate maximal current for each cell. Online leak subtraction was carried out using a fractional (P/-5) method (as described in Porter et al., 1997 and Blalock et al., 1999), which consisted of five fractionally scaled hyperpolarizing subpulses. Current density (pA/pF) was calculated by dividing the mean whole-cell current (during a voltage step) by the membrane capacitance.

Under these recording conditions, whole-cell Ca²⁺ channel currents are typically followed by a long, post-repolarization Ca²⁺ current (e.g. Campbell et al., 1996; Porter et al., 1997). The source of this current remains controversial (cf. discussion in Thibault et al., 1993; Blalock et al., 1999), but the current appears too long to be due to space clamp artifact. Because it does not appear to alter the activated current and its source is not clear, it was not analyzed in the present studies.

Data obtained from I–V curves were normalized and fit to a Boltzmann equation of the form:

where I is the maximal current from the I–V function, V is the voltage, V_½ is the voltage of half-maximal activation, and k represents the steepness of the sigmoid curve. To measure inactivation of HVA Ca²⁺ currents, current amplitude sampled at 1 kHz during a 500-ms pulse was normalized to the peak current amplitude and fit with a double exponential function of the form:

where 1/b = τ_fast, 1/d = τ_slow, A is the amplitude of each exponential component, and C is the steady-state asymptote. Fitted parameters were then compared across treatment groups using a z test.

Cell-attached multichannel patch studies Seal resistance in cell-attached patch studies averaged 21.9 ± 0.72 GΩ and was not altered by drug treatments. Patches with a seal resistance less than 15 GΩ were excluded from statistical analysis. For I–V relationships, patches were stepped in 10-mV increments (10-s interstep interval) from their holding potential to +40 mV. Maximal current in each patch was achieved by stepping the patch membrane from its holding potential of −70 mV to +10 mV. A series of 15 steps to +10 mV was used to generate an average ensemble current for each patch. In prior studies with similar methods (e.g., Thibault et al., 1993; Thibault and Landfield, 1996; Porter et al., 1997; Brewer et al., 2001), we have observed that multichannel responses from the same cell are relatively consistent, and consequently, 15 step-ensembles generate results that are highly similar to those from 30- or 45-step-based ensembles. The relatively large patches employed here, which can contain numerous channels (Thibault and Landfield, 1996; Porter et al., 1997; Brewer et al., 2001), appear to confer stochastic stability. Each depolarization step was separated by 15 s to prevent channel inactivation from confounding the ensemble current amplitude. In each patch, currents generated from a series of hyperpolarizing pulses (equal in amplitude to the depolarizing steps) were averaged and added offline to ensemble currents to generate leak-subtracted waveforms (Fetchan version 6, Axon Instruments). Current density (pA/μm²) for each patch was derived by dividing the average ensemble current by the patch area. The patch area, which is inversely proportional to the pipet resistance, was estimated for each patch using the equation a = 12.6(1/R+0.018), where a is the patch area and R is the pipet resistance (Sakmann and Neher, 1983).

Drugs and drug delivery FK-506, CsA, rapamycin, and okadaic acid were dissolved in 100% ethanol. These drugs were then diluted ≥1000×, such that the ethanol concentration in the recording solutions did not exceed 0.1%. Higher concentrations of FK-506 and CsA (>20 μM) typically required sonication to go into solution. CN-AIP was dissolved in dH₂O (at 1000×). All drugs were purchased from Calbiochem (La Jolla, CA, USA), except the FK-506-binding protein (FKBP-12), and okadaic acid, which were obtained from Sigma (St. Louis, MO, USA). For incubation studies, drugs were added to the cell culture medium 2–4 h prior to recording. Culture medium then was exchanged for drug-free recording medium immediately prior to the experiment. For intracellular application studies, drugs were dissolved in the pipet solution. In channel antagonist and aging studies, FK-506, dissolved in external recording medium, was delivered to the recording site by free diffusion from a low-resistance glass (weeper) pipet (Blalock et al., 1999). In these latter experiments, CNQX and TTX were included in the weeper pipet, but excluded from the external solution. Consequently, full diffusion from the weeper pipet could be verified by the absence of a fast Na⁺ current transient and the paucity of synaptic events in the whole-cell current record.

Neuron collection For mRNA studies, pyramidally shaped neurons from 2-week and 4-week-old sister cultures were aspirated into glass pipets (1–1.5 MΩ tip resistance), filled with 5 μl of standard whole-cell internal recording solution. For each culture dish, a single pipet was used to aspirate one neuron at a time until approximately 15 neurons were collected. With this method, the entire neuronal cell body and a portion of the neuron’s proximal neurites were cleanly extracted with minimal collection of extraneous surrounding glial or neuronal processes. Pyramidally shaped neurons of the type recorded were selected for collection. For statistical analyses of mRNA data, n equaled the number of culture dishes analyzed in each age group.

Following collection, neurons were immediately pipetted into a pre-chilled microcentrifuge tube containing RNase inhibitor (1 U/μl) and random hexamers (0.25 mM). This cellular lysate was heated for 10 min at 70°C, cooled on ice and followed by reverse transcription (RT) for cDNA synthesis. The RT was carried out in a 20-μl reaction containing 1× polymerase chain reaction (PCR)II buffer (Perkin Elmer, CA, USA), 3.75 mM MgCl₂, 0.5 mM dNTPs and 50 U MMLV reverse transcriptase (Gibco BRL). The RT reaction mixture was incubated at 42°C for 1 h followed by heating at 95°C for 10 min and then stored at −20°C until further PCR analysis.

Quantitative real-time PCR Quantitation of relative mRNA content was performed using an ABI prism 7700 sequence detection system (Applied Biosystems, CA, USA) and the TaqMan methodology, which utilizes the 5′ nuclease activity of Taq DNA polymerase to generate a real-time quantitative DNA assay. In brief, mRNA-specific oligonucleotide probes (TaqMan probes) containing 5′-fluorescent reporter and 3′-quencher dyes were designed and used for the extension phase of the PCR. The degradation and release of the reporter dye (i.e. FAM) results in fluorescence at 518 nm, which is monitored during the complete amplification process.

All the cDNA samples from the 2-week-old (n = 13) and 4-week-old (n = 14) groups were analyzed simultaneously by real-time PCR, with each reaction run in duplicate. The PCR solution was based on the Single Reporter real-time PCR protocol, as described in the ABI 770 system’s manual. For each mRNA species, the amount of reverse transcriptase used for the PCR was optimized to ensure that the results fell within the standard curve. To normalize the amount of calcineurin and calmodulin cDNA in each RT sample, the ribosomal RNA 18S level was measured using a 100-fold dilution of the RT as the template for real-time PCR. The primers for 18S were: forward, 5′-AGTCCCTGCCCTTTGTACACA-3′; reverse, 5′-GATCCGAGGGCCTCACTAAAC-3′; and the TaqMan probe was: 6FAM-CGCCCGTCGCTACTACCGATTGG-TAMRA (nt 1684–1752, GenBank X01117). For the calcineurin message, 1 μl of RT was used as the template in a 50-μl PCR reaction mixture. The primers for calcineurin were: forward, 5′-TCCCGGTGACTGGAGATGTC-3′; reverse, 5′-TTTCAC-CACCCTGTCCGTAGT-3′; the TaqMan probe was: 6FAM-CCCAAGGCGATTGATCCCAAGTTGT-TAMRA (nt 193–264, GenBank D90036). For calmodulin, 0.5 μl of RT was used as the template in each PCR reaction and the primers were: forward, 5′-GTCTGTTCTGGTCTCGGAAACC-3′; reverse, 5′-TGAATTCTGCGATCTGCTCTTC-3′; the Taq-Man probe was: 6FAM-CCTTGCAGCATGGCTGACCAACTGA-TAMRA (nt 34–110, GenBank M17069).

Statistical comparisons VSCC current records were analyzed quantitatively using Clampfit 6.0 software (Axon Instruments) and all statistical analyses were performed using Statview 5.01. In most studies, the significance of drug effects was determined by analysis of variance (ANOVA). For studies utilizing pre- and post-drug application measures, effects were determined by repeated measures ANOVA. Post hoc analyses were performed using Scheffe’s F-test.

Hippocampal neurons were incubated for 2–4 h in medium containing either FK-506 (at three concentrations: 0.5, 5, or 50 μM), CsA (20 μM) or okadaic acid (1μM) (see Experimental procedures). These initial studies used drug incubation instead of acute application to ensure equilibrium with immunophilins, which may be limiting in brain (Kung and Halloran, 2000) and other binding sites. Incubation across multiple concentrations of FK-506, revealed a concentration-dependent reduction in whole-cell VSCC Ba²⁺ current density relative to the respective ethanol-incubated control cells (n = 11–15 control cells at each of the three FK-506 concentrations) [overall ANOVA: F(3,75) = 17.57, P < 0.001 (Fig. 1A, B)]. Post hoc analyses showed that 0.5 μM (n = 9) induced a small, non-significant reduction (90 ± 5% of control, P = 0.11), 5 μM (n = 16) induced an ~25% decrease (75 ± 5% of control, P < 0.05), and 50 μM (n = 13) resulted in ~50% inhibition (52 ± 4% of control, P < 0.001, n = 13). Higher concentrations of FK-506 appeared to induce no further inhibition, although they were difficult to study because of the drug’s lipophilicity. At none of the employed concentrations did FK-506 significantly alter the voltage dependence of VSCCs (Fig. 1C). Overall, the half-maximal activation voltage for FK-506 studies was −7 ± 0.76 mV.

Fig. 1 FK-506 and CsA, but not okadaic acid, inhibited whole-cell HVA Ba2+ current. (A, D) Representative examples of averaged VSCC traces (from five successive sweeps) recorded during repeated voltage steps from −70 mV to +10 mV in individual vehicle (more ...)

CsA at 20 μM also significantly reduced VSCC current density by approximately 25% (CsA, 10.4 ± 0.7 pA/pF, n = 13; control, 13.8 ± 1.2 pA/pF, n = 14) [F(1,25) = 5.372, P < 0.05], without altering the I–V relationship (Fig. 1D–F). In contrast, cells treated with okadaic acid (1 μM) (n = 12), did not show a significant reduction in maximal current. However, they exhibited a shift to the left in the I–V curve, relative to control cells (n = 12) (control V_½ = −11.92 mV; okadaic acid V_½ = −20.34 mV; z = 7.77, P < 0.05; Fig. 1G), indicating that okadaic acid increased voltage sensitivity. Thus, unlike FK-506 and cyclosporin, okadaic acid tends to enhance, rather than inhibit VSCC current. These results indicate that calcineurin activity, but not protein phosphatases 1 and 2A activity, appears to be required for a significant fraction of VSCC current in hippocampal neurons.

In this series of studies, agents were applied intracellularly to ensure equal cellular access of the compounds being compared. Hippocampal neurons were dialyzed, via the whole-cell recording pipet, with either 5 μM FKBP-12 alone (control, n = 29), or FKBP-12 in combination with either FK-506 (n = 25), rapamycin (n = 17), or FK-506 together with rapamycin (n = 7). Thus, FKBP-12 was present in all experimental conditions. All immunophilin ligand concentrations were set at 5 μM. Recordings were begun after three stable current records were obtained, at approximately 10 min after establishing the whole-cell configuration. At this point, five successive current traces generated by a 150-ms step from −70 mV to +10 mV (interpulse interval 30 s) were averaged. Cells treated with FK-506/FKBP-12 exhibited a significant reduction in whole-cell Ba²⁺ current density, relative to all other conditions [F(3,73) = 8.679, P < 0.001] (Fig. 2A, B). Rapamycin/FKBP-12 differed only from the FK-506/FKBP-12 group (P < 0.01) and had no inhibitory action on VSCC current. Moreover, when rapamycin was included in the pipet with FK-506/FKBP-12, the inhibitory effects of FK-506 were blocked (post hoc comparison between the FK-506/FKBP-12 and the FK-506/FKBP-12+ rapamycin groups, P < 0.05). These findings suggest that rapamycin at 5 μM was able to bind FKBP-12, and compete with FK-506. They also indicate that ligand binding to FKBP-12 alone is not sufficient for VSCC inhibition. In contrast to FK-506/FKBP-12, rapamycin/FKBP-12 induced a small leftward shift in the I–V function, such that half-maximal activation occurred at −15.79 ± 0.81 mV as opposed to −10.63 ± 0.7 mV for FKBP-12 alone cells (z = 4.82, P < 0.05) and −11.52 ± 1.21 for FK-506/FKBP-12 cells (z = 2.92, P < 0.05) (Fig. 2C). Thus, rapamycin and FK-506, which differ primarily in their abilities to inhibit calcineurin, also have strikingly different effects on VSCC current. Rapamycin tended to enhance rather than inhibit VSCC current. Although it inhibits calcineurin activity through a different mechanism from FK-506, CN-AIP (100 μM) also significantly reduced whole-cell Ba²⁺ current density [F(1,20) = 8.627, P < 0.01] (Fig. 2D, E). As with FK-506 and CsA, CN-AIP did not substantially alter the voltage dependence of VSCCs (control V_½ = −11.73 ± 0.55 mV, CN-AIP V_½ = −13.66 ± 1.5) (Fig. 2F). Together, the results indicate that inhibition of calcineurin is the necessary and sufficient factor in the inhibition of HVA VSCC current by these multiple agents.

Fig. 2 Reduction of HVA Ba2+ current was specific to calcineurin inhibition. Whole-cell currents were recorded in neurons dialyzed with either FKBP-12 alone (5 μM), FKBP-12/FK-506 (Rap; 5 μM), FKBP-12/rapamycin (5 μM), or FKBP-12/FK-506 (more ...)

Fig. 3 FK-506 reduced HVA Ca2+ currents, but did not alter VSCC inactivation. Extracellular Ba2+ was replaced with equimolar Ca2+ and neurons were dialyzed with either FKBP-12 alone (5 μM) or FKBP-12/FK506 (5 μM) as described in Fig. 2. FK-506 (more ...)

Fig. 3B shows that removal of EGTA from the pipet solution also did not prevent FK-506 from inhibiting Ca²⁺ current through VSCCs [FK-506/FKBP-12 = 13.35 ± 1 pA/pF, n = 12; FKBP-12 alone = 16.97 ± 1.27 pA/pF, n = 14; F(1,24) = 4.8, P < 0.05]. The exclusion of EGTA from the pipet solution also gave rise to VSCC currents that exhibited substantial inactivation. That is, with EGTA present in the pipet solution (Fig. 3A), Ca²⁺ current dropped by only 8 ± 2% from its peak amplitude to its minimum amplitude at the end of the 150-ms voltage step. However, cells recorded without EGTA in the pipet showed a much greater reduction (34 ± 3%) in VSCC current during the voltage step [F(1,21) = 63.9, P < 0.001], indicating considerable Ca²⁺-dependent inactivation in EGTA-free cells.

To examine whether FK-506 altered Ca²⁺-dependent inactivation, a 500-ms depolarizing pulse to +20 mV was applied to cells in the absence of internal EGTA (Fig. 3C). The time course of VSCC inactivation could be well-fit using a double exponential function (see Experimental procedures). The results showed that, in contrast to its effects on maximal VSCC amplitude, FK-506 did not significantly alter the rate of VSCC inactivation (Fig. 3C, right panel). However, both the slow and the fast time constants were slightly reduced in the FK-506/FKBP-12 group. Although inactivation kinetics in dendritically ramified neurons can be confounded by multiple compartments (Brown and Johnston, 1983; Johnston and Brown, 1983; Spruston et al., 1994), the absense of an effect on either exponential function indicates that, in hippocampal neurons, FK-506 can reduce whole-cell Ca²⁺, as well as Ba²⁺ currents, without appreciably altering the inactivation kinetics of HVA VSCCs. Interestingly, cells without EGTA exhibited larger Ca²⁺ currents relative to cells in which Ca²⁺ was buffered [F(1,24) = 18.42, P < 0.001] (Fig. 3B vs. Fig. 3A). Although the fraction of current inhibited by FK-506 in these two conditions was similar, the results are at least consistent with the possibility that elevating intracellular Ca²⁺ can, under some circumstances, enhance Ca²⁺ channel activity, perhaps via a calcineurin mechanism.

To enhance access of FK-506 to channels in the patch, cells were incubated in FK-506 (50 μM) for at least 2 h. Neurons exposed to FK-506 (n = 12) exhibited a significant decrease in cell-attached VSCC activity (~45%) relative to control patches (n = 8) [F(1,18) = 9.64, P < 0.01] (Fig. 4A, B), with no change in the voltage dependence of VSCCs (data not shown). Thus, inhibition of VSCCs by FK-506 was not limited to the whole-cell recording configuration. Moreover, the magnitude of inhibition was relatively similar in both whole-cell (Fig. 1) and cell-attached patch recording modes (Fig. 4).

Fig. 4 FK-506 reduced HVA Ba2+ current activity in multichan-nel cell-attached patches. (A) Representative waveforms of five current traces, and the average ensemble current (lowest trace) generated from 15-step depolarizations in multichannel cell-attached (more ...)

Fig. 5 FK-506 selectively inhibited L-type VSCC current. (A–C) Left panels. Time course plots of mean (± S.E.M.), normalized VSCC currents of neurons exposed to acute external perfusion of 0.1% ethanol alone (A, open circles), 50 μM FK-506 (more ...)

Fig. 6 Effects of age in culture on VSCC inhibition by FK-506 and calcineurin mRNA level. (A) Representative time plots of normalized VSCC currents from a representative 2-week (open circles) and a representative 4-week-(filled circles) old neuron during acute (more ...)

In the present work, a series of disparately acting calcineurin inhibitors, phosphatase inhibitors, and immunophilin ligands were compared to show that the common feature among agents that suppressed VSCC current was calcineurin inhibition. That is, these studies found strong evidence that calcineurin inhibition is both necessary (e.g. the similar FKBP-12 immunophilin ligand, rapamycin, which does not inhibit calcineurin, did not inhibit VSCC current) and sufficient (CN-AIP, which inhibits calcineurin but does not bind to immunophilins, also inhibited VSCC current) for mediating inhibition of VSCC current (Fig. 2). Consistent with these results, the PP1/PP2A inhibitor, okadaic acid, did not suppress VSCC current (Fig. 1). Finally, the effect of calcineurin inhibition did not depend on the conditions of the whole-cell recording configuration, as VSCC inhibition was of relatively similar magnitude in the cell-attached patch configuration (Fig. 4). Together, the results indicate clearly that inhibition of VSCCs by these agents specifically required calcineurin inhibition.

The enhancing effect of calcineurin also was targeted selectively to one type of VSCC, the L-type. The functional availability of L-type VSCCs appeared to be both necessary (i.e., nimodipine, the specific L-type antagonist, blocked all action of FK-506 on VSCCs) and sufficient (in the absence of N-, and P/Q-type VSCC activity, the relative FK-506 effect on L-type current was considerably greater) for the full manifestation of FK-506-mediated inhibition of VSCC currents in hippocampal neurons. Thus, the L-type VSCC was the only VSCC that was dependent on some degree of calcineurin phosphatase activity. Moreover, calcineurin-inhibition did not shift VSCC voltage dependence but instead appeared to block fully the activation of L-VSCCs. These data imply that calcineurin action is essential for enabling L-type VSCC availability in hippocampal neurons. However, it is also possible that, in addition to enabling availability, calcineurin enhances L-type VSCC activity through other processes (e.g. increased open probability).

Maximal catalytic activity of calcineurin requires a large rise in [Ca²⁺]_i (to the μM range) and subsequent binding to Ca²⁺/calmodulin (Stemmer and Klee, 1994). However, many of the present studies were performed in neurons in which Ba²⁺ was the charge carrier and/or [Ca²⁺]_i was buffered by EGTA to near-resting (< 100 nM) concentrations (Bers et al., 1994). Thus, [Ca²⁺]_i presumably did not rise substantially during current influx. Nevertheless, calcineurin also shows some phosphatase activity at resting or low [Ca²⁺]_i levels, stimulated by Ca²⁺ bound to the calcineurin B subunit of the calcineurin holoenzyme. This activity is seen in the absence of calmodulin or in the presence of EGTA (Stewart et al., 1982; Liu et al., 1991; Stemmer and Klee, 1994). Therefore, it can be concluded that relatively low levels of calcineurin activity at near-resting [Ca²⁺]_i can mediate the enabling action of calcineurin on L-type VSCCs. However, EGTA is a relatively slow Ca²⁺ buffer and the possibility that intracellular Ca²⁺ release and/or capacitive Ca²⁺ entry increased calcineurin activity above anticipated levels cannot be ruled out. Finally, the preliminary evidence that Ca²⁺ currents were larger when [Ca²⁺]_i was not buffered (Fig. 3A,B) raises the intriguing possibility that, while low or basal levels of [Ca²⁺]_i may be able to maintain the basic enabling effect of of calcineurin on L-type VSCCs, elevating [Ca²⁺]_i, and in turn, increasing calcineurin activity, may recruit previously unavailable L-type VSCCs or enhance channel open probability. Clearly additional studies will be required to clarify the Ca²⁺ dependence of calcineurin’s actions on L-type VSCCs in hippocampal neurons.

Interestingly, this is the third VSCC-type-specific process found to change with age in long-term hippocampal culture. The other two include an increase in L-type VSCC density (Porter et al., 1997), and a substantial decline in membrane-delimited, G-protein-coupled inhibition of non-L-type VSCCs (Blalock et al., 1999). Thus, with age in culture, multiple changes in VSCC-type-specific regulation occur that seem likely to increase Ca²⁺ influx.

Although these VSCC alterations in culture may well reflect basic developmental processes (e.g., (Turrigiano et al., 1994), there also is some evidence that they may be, in part, analogous to brain aging processes. That is, changes in VSCC activity have been found consistently in models of brain aging and likely contribute to age-related alterations in Ca²⁺-dependent processes such as neuronal excitability, plasticity, and memory (cf., reviews, Khachaturian, 1989; Landfield et al., 1989; Disterhoft et al., 1993; Foster and Norris, 1997; Thibault et al., 1998; Verkhratsky and Toescu, 1998; Griffith et al., 2000). Further, growing evidence indicates that calcineurin function may be increased with brain aging (Norris et al., 1998; Foster et al., 2001) and, in preliminary studies, FK-506 inhibited VSCC current in hippocampal slice neurons of aged rats (Norris et al., 2001).

Additional studies will of course be required to determine whether similar aging changes in calcineurin content or sensitivity occur during aging in vivo. However, regardless of whether calcineurin regulation changes with brain aging, the present studies suggest that calcineurin plays a major enabling/enhancing role in normal and pathological processes mediated by L-type VSCC activity. As a corollary, modulation of calcineurin may provide a basis for developing new therapeutic strategies aimed at regulating L-type VSCC activity and L-type VSCC-dependent disorders.

This work was supported in part by Grants from the National Institute on Aging (AG04542, AG10836) and Training Grants (AG00242 and 1F32 AG05903). We wish to thank Elsie Barr, Janice Staton, and Veronique Thibault for important technical assistance, and Kelley Secrest and Judy Hower for excellent assistance with the manuscript.