|Home | About | Journals | Submit | Contact Us | Français|
Hypoxia regulates neuronal ion channels, sometimes resulting in seizures. We evaluated the effects of brief sustained hypoxia (1% O2, 4 h) on voltage-gated calcium channels (VGCCs) in cultured rat primary cortical neurons. High-voltage activated (HVA) Ca2+ currents were acquired immediately after hypoxic exposure or after 48 h recovery in 95% air/5% CO2. Maximal Ca2+ current density increased 1.5-fold immediately after hypoxia, but reverted to baseline after 48 h normoxia. This enhancement was primarily due to an increase in L-type VGCC activity, since nimodipine-insensitive residual Ca2+ currents were unchanged. The half-maximal potentials of activation and steady-state inactivation were unchanged. The calcineurin inhibitors FK-506 (in the recording pipette) or cyclosporine A (during hypoxia) prevented the post-hypoxic increase in HVA Ca2+ currents, while rapamycin and okadaic acid did not. L-type VGCCs were the source of Ca2+ for calcineurin activation, as nimodipine during hypoxia prevented post-hypoxic enhancement. Hypoxia transiently potentiated L-type VGCC currents via calcineurin, suggesting a positive feedback loop to amplify neuronal calcium signaling that may contribute to seizure generation.
Brain neurons can tolerate moderate to severe hypoxia for several hours (Erecinska and Silver, 2001; Scheufler et al., 2002) as long as glucose is present to meet minimal metabolic demands via glycolysis (Kahlert and Reiser, 2000; Vannucci and Vannucci, 2000; Vannucci and Vannucci, 2001). However, hypoxia can predispose to seizures, which occur in 9% of patients after stroke (Bladin et al., 2000) and up to 36% of patients after cardiopulmonary arrest (Krumholz et al., 1988). The risk of epilepsy after perinatal hypoxia is 5-fold greater than in normal neonates (Bergamasco et al., 1984; Jensen, 2002). Thus, understanding the cellular mechanisms mediating hypoxia-induced neuronal dysfunction is clinically relevant and could lead to new treatments for hypoxia-induced seizures and epilepsy.
Transmembrane ion channels play important roles in oxygen (O2)-regulated cellular processes (Hammarstrom and Gage, 1998; Jiang et al., 1994; Tai and Truong, 2007). We previously reported that transient hypoxia reduces inhibitory GABAA receptor (GABAAR) function in NT2-N neurons (Gao et al., 2004) and primary cortical neurons (Wang and Greenfield, Jr., 2009) and that the reduction in cortical neurons was blocked by the L-type voltage-gated calcium channel (L-VGCC) antagonist, nitrendipine (Wang and Greenfield, Jr., 2009). L-VGCCs play a critical role in mediating intracellular Ca2+ signaling, activity-dependent synaptic plasticity and neuronal survival (Berridge, 1998; Catterall, 2000), and are likely involved in post-hypoxic signaling mechanisms. However, the sensitivity of neurons to hypoxia varies by neuron type and brain region, and cortical neurons are more resistant to O2 deprivation than hippocampal or cerebellar neurons (Krnjevic, 2008). For example, Li et al. (Li et al., 2007) reported a delayed reduction in Ca2+ currents in hippocampal CA1 but not CA3 neurons, associated with increased cell death in CA1 neurons. Moreover, regulation of L-VGCCs after deprivation of both O2 and glucose may differ markedly from hypoxia alone. The effects of short term hypoxia on high voltage activated (HVA) Ca2+ currents in primary rat cortical neurons have not been thoroughly investigated. Regulation of cortical neurons HVA Ca2+ currents by hypoxia may contribute to cortical neuron hyperexcitability and thus predispose to seizures.
At least four types of HVA VGCCs (L-, N-, P/Q- and R-type) have been identified in neurons, with diverse regulatory mechanisms that are specific to VGCC subtypes (Catterall, 2000). One important regulatory factor is the Ca2+/calmodulin-dependent phosphatase, calcineurin (also known as protein phosphatase 2B). Both positive and negative regulation of VGCC activity by calcineurin have been reported, varying across cell types (Groth et al., 2003) and model systems (Zeilhofer et al., 2000; Norris et al., 2002; Oliveria et al., 2007; Norris et al., 2008), suggesting that the physiological regulation of VGCCs by calcineurin is complex and variable depending on the neurons involved. Since seizure generation is a predominantly cortical activity, it is critical to understand how seizure-inducing stimuli like hypoxia affect cortical neuron VGCCs and the possible role of calcineurin in this response.
Here, we evaluated HVA Ca2+ currents in rat primary cortical neurons in culture using whole-cell voltage clamp recordings after exposure to 1% O2 for 4 h. HVA Ca2+ currents were increased immediately (0–2 h) after 4 h hypoxia but returned to baseline when recorded after 48 h normoxic recovery. The increase in HVA current was blocked by nimodipine (NIM), and hence L-VGCC-dependent. Inhibition of calcineurin activity with FK-506 or cyclosporine A (CsA) blocked the post-hypoxic increase in L-VGCC current. Our results suggest that O2 deprivation transiently increases L-VGCC activity in cortical neurons via a calcium dependent process requiring L-VGCC activation and calcineurin, suggesting a positive feedback loop to amplify neuronal calcium signaling after hypoxia. These findings may have clinical significance, since hypoxia-induced increases in intracellular Ca2+ after stroke or cardiopulmonary arrest may contribute to post-hypoxic neuronal hyperexcitability, cell death or epileptogenesis.
Experimental protocols involving the use of vertebrate animals were approved by the University of Toledo College of Medicine Institutional Animal Care and Use Committee (IACUC) and conformed to United States National Institutes of Health guidelines.
Primary cultures of cortical neurons were prepared from E18 fetal Sprague-Dawley rats according to a protocol slightly modified from established techniques (Porter et al., 1997). Briefly, E18 rat fetuses were removed under sterile conditions after euthanization of the dam. Fetal cortices were dissected in sterile Hank’s Balanced Salt Solution (HBSS) and digested for 5 min at 37°C using 0.25% trypsin-ethylene diaminotetraacetic acid (EDTA) in HBSS, then repeatedly washed in HBSS. Tissues were further triturated in Spiner’s modification of Eagle’s minimum essential medium (SMEM) containing both horse serum (HS, 5%) and fetal bovine serum (FBS, 5%). The cell suspension was plated onto 35 mm plastic culture dishes (Corning Inc., Corning, NY) pre-coated with poly-D-lysine at 2–3×105 cells/ml in SMEM plus 5% FBS and 5% HS. 5-Fluoro-2′-deoxyuridine (FUDR) and uridine were added 48 h after plating to prevent non-neuronal cell proliferation. Half of the medium was exchanged for fresh SMEM/HS three times a week. Cells were maintained in humidified atmosphere with 5% CO2 at 37 °C for 2 weeks and used for experiments between 13–15 days in vitro (DIV).
Culture medium (SMEM/HS) was deoxygenated by bubbling for 10 min with 95% N2/5% CO2 and warmed to 37°C. Regular SMEM/HS in culture dishes were then replaced with deoxygenated SMEM/HS and placed in an O2- and CO2-controlled incubator (Innova CO-48, New Brunswick Scientific Co. Inc.) pre-equilibrated to 1% O2, 5% CO2 at 37 °C for 4 h. After hypoxic exposure, the medium was replaced with fresh aerated SMEM/HS and neurons were either studied immediately (within 2 h of termination of hypoxia) or returned to the normoxic incubator (95% air/5% CO2, 37 °C) for 48 h prior to recording (within ± 2 h after 48 h recovery). Control neurons were similarly handled but maintained in a normoxic environment and solutions.
HVA currents were recorded under whole-cell voltage-clamp conditions at room temperature. The external solution contained (in mM): NaCl 110; HEPES 10; TEA chloride 25; KCl 5.4; CaCl2 5; 4-AP 5; MgCl2 1; D-glucose 25; TTX 1 μM pH 7.4. The patch pipettes (4–6 MΩ) contained (in mM): CsF 110; TEA chloride 25; phosphocreatine 20; phosphocreatine kinase 50 units/mL; EGTA 10; HEPES 10; NaCl 5; MgCl2 2; CaCl2 0.5; BaCl2 0.5; MgATP 2; NaATP 0.1; pH 7.3. Currents were recorded with an Axoclamp 200B amplifier (Molecular Devices, Union City, CA) using a Digidata 1440A AD/DA converter and pClamp 10.2 acquisition software (Molecular Devices, Union City, CA). Current signals were low-pass filtered at 1 kHz and digitized at 10 kHz. After establishing the whole-cell configuration, neurons were allowed to stabilize for 5 min before current recording protocols were initiated. Neurons were voltage clamped at −80 mV. Peak HVA calcium currents (pA) were activated with a protocol modified from (Norris et al., 2002) and (Sochivko et al., 2003) using 200 ms voltage steps to voltage levels from −75 to +40 mV in 5 mV increments. The steady-state inactivation of Ca2+ currents was evaluated with a 200 ms test pulse to +10 mV, preceded by a 1500 ms conditioning pre-pulse ranging from −80 mV to +10 mV in 5 mV increments. Currents were corrected for linear, non-specific leak currents and capacitive transients. Junction potentials were nulled in the bath using the pipette offset control. Series resistance was typically under 10 MΩ, compensated by at least 80%, and did not differ between control and post-hypoxic cells. Neurons in which the access resistance changed >20% during the course of an experiment were excluded from statistical analyses. Cell membrane capacitance was measured shortly after obtaining whole-cell configuration using the “Membrane Test” function of Clampex 10.2. Current density was calculated as a function of cell membrane capacitance (pF) as an estimate of cell size.
Whole-cell calcium currents were analyzed off-line using Clampfit 10.2 software. Current density (pA/pF) was defined as peak current amplitude divided by membrane capacitance and plotted as a function of membrane potential (mV). Activation curves were constructed using individual conductance values for each cell from the current/voltage plot using the equation g = I/(Vt−Vr), where Vt was the test voltage, Vr was the reversal potential extrapolated from individual I/V curves (usually near +60 mV) and I was the measured current (Sochivko et al., 2003). The conductance was then normalized to the maximum conductance (gmax) and plotted against activating voltage. Activation and inactivation curves were constructed from peak currents elicited by the current/voltage or steady state inactivation protocols by fitting normalized peak conductance, g/gmax, with a Boltzmann equation: g/gmax= 1/(1+exp[(V−Vh)/k]), where Vh was the potential where g was half of gmax and k was a factor proportional to the slope at Vh. Curve fitting was performed using Prism 4.0 software (Graph Pad Software Inc., San Diego, CA). Data are reported as mean ± standard error of the mean (SEM). Statistical analyses were performed using Student’s t-test or ANOVA with post hoc Bonferroni correction for multiple comparisons. Asterisks denote significant differences between compared groups at p<0.05.
FK-506 (tacrolimus), cyclosporine A (CsA), rapamycin (RAP), okadaic acid (OKA) and nimodipine (NIM) were dissolved in DMSO and stored in light-tight vials at −20 °C. Drugs were diluted ≥1000-fold in extracellular buffer to the specified concentration, the DMSO concentration in the recording solution not exceeding 0.1%. VGCC antagonists were dissolved in external recording solution and delivered using a pressure-driven, modified U-tube “multipuffer” system (Greenfield, Jr. and Macdonald, 1996) with continuous, rapid perfusion of external solution when drug was not being applied. For incubation studies, drugs were added to the deoxygenated medium and were present during 4 h hypoxia treatment. Culture medium was then exchanged with several washes of drug-free medium and recording solution immediately prior to the experiment. For intracellular application studies, drugs were dissolved in the pipette solution on the day of experiment. All other drugs were dissolved in dH2O for stock solutions and diluted in external buffer to final concentration. Tetraethylammonium chloride (TEA), 4-aminopyridine (4-AP)), NIM, cadmium chloride, CsA, and OKA were all from Sigma-Aldrich Chemical Co. (St Louis, MO). FK-506 and RAP were from Alexis Biochemicals Inc. (Lausen, Switzerland). Tetrodoxin (TTX) was obtained from Alamone Laboratories (Jerusalem, Israel).
AKAP: protein kinase A anchoring protein; CsA: cyclosporine A; DIV: days in vitro; dH2O: distilled water; FUDR: 5-Fluoro-2′-Deoxyuridine; FK506: tacrolimus; GABAAR: GABA type A receptor; HIF-1α: hypoxia-inducible factor 1 alpha; HS: horse serum; HVA: high voltage activated; OKA: okadaic acid; PI3K phosphoinositol 3-kinase; RAP: rapamycin; SMEM: Spiner’s modification of Eagle’s minimum essential medium; SEM: standard error of the mean; TEA chloride: tetraethylammonium chloride; TRP channel: transient receptor potential channel; VGCC: voltage-gated calcium channel; 4-AP: 4-aminopyridine
Cultured rat primary cortical neurons were inspected by phase-contrast microscopy prior to recordings to confirm normal morphology and viability after 4 h exposure to 1% O2. Consistent with our previous reports (Gao et al., 2004; Wang and Greenfield, Jr., 2009), this hypoxic exposure did not result in obvious somatic swelling or dendrite retraction, features commonly associated with neuronal injury, and showed no increase in cell death or injury as measured by trypan blue staining or LDH release (Gao et al., 2004; Wang and Greenfield, Jr., 2009).
VGCC blockers were used for pharmacological identification of the elicited currents. At maximal activation (+10 mV), the selective L-VGCC antagonist nimodipine (NIM, 10 μM) reduced HVA Ca2+ currents to 55.9 ± 5.2% of baseline (n=5) indicating that L-type currents represented a major component of HVA Ca2+ current in investigated neurons (Fig. 1). Sample traces from a typical experiment are shown in Fig. 1A. The nimodipine-resistant current was likely attributable to other HVA Ca2+ currents, including N, P/Q and R-type (Catterall, 2000), as well as TTX-insensitive sodium currents and non-specific cation currents. T-type channel currents were unlikely to contribute significantly as no rapidly desensitizing, low voltage activated component was observed. Subsequent addition of the non-selective Ca2+ channel blocker CdCl2, (10 μM), further suppressed all but 16.5 ± 5.9% of the whole currents (Fig. 1A, B). Blockade by NIM and Cd2+ did not alter the current/voltage relationship or activation kinetics of the residual currents (Fig. 1C). The residual current component remaining after Cd2+ may be attributed to TTX-resistant sodium currents or non-selective cation currents (Fadulova et al., 1991; Ma et al., 1997; Shkryl et al., 1999). Washout of Cd2+ partially reversed the blockade by this cation (to 60.9 ± 8.8% of baseline, n=5). Gradual rundown of HVA currents over the time of recording was sometimes observed, but most cells had stable maximal currents, and those with > 20% rundown over the course of the experiment were excluded from the analysis.
Representative Ca2+ current traces elicited in cortical neurons within 2 h after 4 h exposure to 1% O2 (Fig. 2A, Hypox.4/Rec.0) were approximately 1.5-fold larger than those from control cells maintained at atmospheric 02 (20.9%, Normox). The atmospheric O2 concentration at sea level is relatively hyperoxic compared to the 2.5 – 5.3% (19–40 mm Hg) found in cerebral cortex in vivo (Erecinska and Silver, 2001) but is routinely used for neuronal culture and thus considered “normoxic” for these studies. Peak current amplitudes at each activation voltage were normalized to cell capacitance to yield Ca2+ current densities and plotted across the activating voltage range studied (Fig. 2B). Current density was significantly increased immediately after 4 h hypoxia exposure. Maximal Ca2+ current density for both groups was observed at voltage steps to +5 mV, at which post-hypoxic neuron current density was significantly larger than that from normoxic controls (Normox: 14.1 ± 2.3 pA/pF, n=10; Hypox.4/Rec.0: 20.4 ± 1.8 pA/pF, n=10, p<0.05 by unpaired t-test). The changes in maximal current were not due to altered cell size, as membrane capacitance was not significantly different between normoxic and post-hypoxic groups (Normox: 21.4 ± 0.5 pF, n=10; Hypox.4/Rec.0: 21.2 ± 1.0 pF, n=10, p>0.05).
Current amplitudes elicited from individual neurons by voltage steps were transformed into normalized conductances, plotted against activation voltage and then averaged for each voltage step to yield activation curves (Fig. 2C). There was no significant change in activation between hypoxia-exposed and normoxic control neurons. The voltages of half-maximal activation (Vh) were not significantly different (Normox: 0.8 ± 2.0 mV, n=10; Hypox.4/Rec.0: −1.9 ± 3.7 mV, n=10; p>0.05), nor was there a significant difference in the Boltzmann slope (k) of the activation curves for post-hypoxic cells and matched normoxic controls (Fig 2C. Normox: 12.3 ± 1.6, n=10; Hypox.4/Rec.0: 13.2 ± 1.8, n=10, p>0.05).
Steady-state inactivation was assessed using test pulses of 200 ms duration at +10 mV evoked immediately following 1500 ms conditioning pre-pulses at voltages ranging from −80 mV to +10 mV, in 5 mV increments. Boltzmann inactivation curves derived from individual fits of Ca2+ currents elicited from control and post-hypoxic neurons are shown in Fig. 2C. There was no significant shift in the inactivation curve after 4 h exposure to 1% O2 compared to matched normoxic control neurons.. The voltage of half-maximal inactivation (Vh) was not significantly different (Normox: −40.3 ± 1.6 mV, n=11; Hypox.4/Rec.0: −37.6 ± 1.3 mV, n=11, p>0.05), nor was there a significant difference in the slope (k) of the inactivation curves (Normox: 12.7 ± 0.6, n=11; Hypox.4/Rec.0: 13.8 ± 1.6, n=11, p> 0.05).
We previously reported that maximal GABA current in NT2-N neuronal cells was increased immediately after 8 h exposure to 1% O2 but reduced after 48 h normoxic recovery (Gao et al., 2004), while in cortical neurons, 4 h hypoxic treatment reduced GABAA receptor currents both immediately and 48 h after exposure (Wang and Greenfield, Jr., 2009). For cortical neurons, both phases were blocked by L-type VGCC antagonists. Since the 48 h time point was critical for post-hypoxic GABAA receptor changes in two neuronal cell types, we hypothesized that post-hypoxic HVA Ca2+ current changes might also remain affected after 48 h normoxic recovery. In contrast to the time course for GABAA receptor regulation, HVA Ca2+ currents recorded after 4 h hypoxia exposure followed by 48 h normoxic recovery (95% air/5% CO2, 37°C) showed no significant difference in current–voltage relationship compared to normoxic controls (Fig. 3A). Maximal Ca2+ current density (Fig. 3B) was also unchanged in post-hypoxic neurons after 48 h recovery (Hypox4/Rec.48: 14.6 ± 3.3 pA/pF, n=7) compared to matched controls (Normox: 14.7 ± 0.7 pA/pF, n=7, p>0.05), indicating that the augmentation of maximal Ca2+ current density immediately after hypoxic exposure was transient and reverted to baseline within 48 h of normoxic recovery. There were no changes in the voltage dependence of Ca2+ current activation (Normox: −3.4 ± 0.5 mV, n=7; Hypox.4/Rec.48: −2.5 ± 0.6 mV, n=7, p>0.05, data not shown) or steady-state inactivation (Normox: −42.3 ± 0.6 mV, n=6; Hypox.4/Rec.48: −39.9 ± 0.2 mV, n=6, p> 0.05, data not shown).
To evaluate whether hypoxia selectively regulates L-VGCC function, we used nimodipine (NIM, 10 μM) to identify the L-VGCC Ca2+ current component. The nimodipine-resistant residual Ca2+ current, presumed to be a mixture of N, P/Q and R-types as well as TTX-resistant sodium currents and nonspecific cation currents, was considered the non-L-type current component. The NIM-dependent Ca2+ current was calculated as the difference between total HVA current and residual current. Total HVA Ca2+ current, L-type Ca2+ current and non-L-type current were measured in post-hypoxic and normoxic control neurons within 2 hours of treatment. As previously noted, Ca2+ current density was significantly greater immediately after 4 h hypoxia compared with normoxic controls (at +5 mV, Normox: 15.9 ± 1.4 pA/pF, n=8; Hypox.4/Rec.0: 21.8 ± 2.1 pA/pF, n=8, p<0.05 by unpaired t-test). There was also a significant difference in the L-type (NIM-sensitive) Ca2+ current component (Fig. 4A; Normox-L: 6.3 ± 0.7 pA/pF, n=8; Hypox-L: 11.0 ± 2.0 pA/pF, n=8, p<0.05). There was no significant difference in non-L-type (nimodipine-insensitive) current after hypoxia (Normox-NonL: 9.6 ± 1.0 pA/pF, n=8; Hypox-NonL: 10.8 ± 1.0 pA/pF, n=8, p<0.05). Thus, the enhancement of total HVA Ca2+ current after hypoxia appears to have been due primarily to increased L-type current. In normoxic control neurons, the percentage of L-type component was significantly lower than the non-L-type component (Fig 4B. Normox-L: 39.4 ± 3.5 %, n=8 vs Normox-NonL: 60.6 ± 3.4 %, n=8, *p<0.05). However, after acute hypoxia exposure, there was no difference in the percentage of L-type and non-L-type current (Hypox-L: 48.7 ± 4.5 %, n=8; Hypox-NonL: 51.3 ± 4.5 %, n=8. p > 0.05), consistent with a change in the proportion of functional L-type channels. Moreover, the magnitude of the post-hypoxic increase in L-type current density (about 4.7 pA/pF) is nearly sufficient to account for the increase in total HVA Ca2+ current density (about 5.9 pA/pF). Collectively, these data suggested that acute hypoxia selectively enhanced L-VGCC current in rat primary cortical neurons.
Based on the growing evidence of the involvement of the Ca2+/calmodulin-dependent protein phosphatase, calcineurin, in VGCC regulation (Norris et al., 2002; Norris et al., 2008; Oliveria et al., 2007) and post-hypoxic signaling (Liu et al., 2007; Marunaka et al., 1998), we hypothesized that the increase in HVA Ca2+ current was related to activation of calcineurin. To determine whether calcineurin was required for upregulation of Ca2+ currents after hypoxia, we included the calcineurin inhibitor, FK-506, in the recording pipette at a concentration (5 μM) previously shown to suppress HVA Ba2+ currents in cultured hippocampal neurons (Norris et al., 2002). Representative HVA Ca2+ current traces (Fig. 5A) from post-hypoxic and normoxic control neurons demonstrate that FK-506 prevented the increase of HVA Ca2+ current following 4 h hypoxic exposure without affecting the current density in normoxic control neurons (maximal current density: Normox-FK506: 12.5 ± 3.1 pA/pF, n=8; Hypox-FK506: 13.0 ± 1.0 pA/pF, n=7, p>0.05). FK-506 also had no significant effect on the current-voltage relationship for HVA Ca2+ currents (Fig. 5B) or the voltage dependence of VGCC activation (Fig. 5C, Normox-FK506 Vhact: −7.3 ± 0.2 mV, n=8; Hypox-FK506 Vhact: −4.8 ± 0.2 mV, n=7, p>0.05.). Steady-state inactivation was likewise unchanged by FK-506 (Fig. 5C, Normox-FK506 Vhinact: −43.0 ± 0.2 mV, n=8; Hypox-FK506 Vhinact: −39.1 ± 0.2 mV, n=7, p>0.05.). Thus, FK-506 prevented the enhancement of HVA Ca2+ current density in cortical neurons after hypoxia exposure without altering voltage-dependent channel kinetics.
To evaluate whether this effect of FK-506 was specific to the post-hypoxia augmentation of L-type VGCC currents, we included 5 μM FK-506 in the whole-cell recording pipette and measured total and nimodipine-resistant HVA Ca2+ currents in neurons acutely treated with normoxia or hypoxia. As previously observed, FK-506 reversed the post-hypoxic increase in HVA maximal Ca2+ current density without significantly affecting the current density in nomoxic controls (Normox-FK506: 13.4 ± 1.3 pA/pF, n=8; Hypox-FK506: 14.3 ± 0.7 pA/pF, n=8, p>0.05). FK-506 also prevented the post-hypoxic increase in NIM-sensitive L-type Ca2+ current density (Normox-FK506-L: 4.5 ± 0.9 pA/pF, n=8; HypoxFK506-L: 5.5 ± 0.9 pA/pF, n=8, p>0.05, Fig. 5D). The non-L-type Ca2+ current density was unchanged by inclusion of FK506 (Normox-FK506-NonL: 9.0 ± 1.0 pA/pF, n=8; Hypox-FK506-NonL: 8.7 ± 0.6 pA/pF, n=8, p>0.05).
FK-506 also prevented the post-hypoxic change in relative percentage of L-type and non-L-type components in total HVA Ca2+ current. In the absence of FK-506, hypoxia increased the percentage of L-type component such that the significant difference between L-type and non-L-type currents was abolished (Fig. 4B). With FK-506 in the recording pipette, the significantly higher proportion of non-L-type current in control cells (Normox-FK506-L: 32.6 ± 5.0%, n=8; Normox-FK506-NonL: 67.4 ± 5.0%, n=8, p<0.05) was still present in post-hypoxic neurons (Hypox-FK506-L: 37.9 ± 5.2%, n=8 vs. Hypox-FK506-NonL: 62.1 ± 5.2%, n=8, p<0.05, Fig. 5E). Prevention of the post-hypoxic increase in L-type Ca2+ current by FK-506 thus suggests that calcineurin-dependent enhancement of HVA Ca2+ current may be specific for L-type VGCCs. Indeed, even in normoxic control neurons, there was a slight (but not significant) decrease in the L-type VGCC component in neurons recorded in the presence of FK-506 compared to non-FK506 normoxic control neurons (Normox-L: 39.4 ± 3.5 %, n=8; Normox-FK506-L: 32.6 ± 5.0%, n=8, p>0.05). This is consistent with studies demonstrating that FK-506 selectively inhibited L-type VGCC current in hippocampal neurons (Norris et al., 2002) and again supports a role for calcineurin-dependent regulation of L-type VGCCs.
To evaluate the specificity of the FK-506 effect, we also tested the structurally distinct calcineurin inhibitor, cyclosporine A (CsA), which binds to the immunophilin, cyclophilin A (Hamawy, 2003; Snyder et al., 1998). In this case, cortical neurons were co-incubated during the 4 h hypoxic exposure with 20 μM CsA, a concentration shown to maximally inhibit calcineurin in lymphocytes (Kung and Halloran, 2000), which was washed out prior to recording. CsA was previously reported to reduce VGCC current density in cultured hippocampal neurons without altering the I–V relationship (Norris et al., 2002). CsA co-incubation during hypoxia also prevented the hypoxia-induced enhancement of HVA Ca2+ current density in cortical neurons compared to matched normoxic controls (Fig. 6A–C; Normox-CsA: 10.4 ± 1.5 pA/pF, n=8; Hypox-CsA: 9.1 ± 0.8 pA/pF, n=8, p>0.05). Of note, CsA incubation not only occluded the post-hypoxic increase in HVA Ca2+ current, but also reduced the current density in normoxic neurons by 51.4% relative to controls without CsA (Normox-CsA: 10.4 ± 1.5 pA/pF, n=8; Normox: 21.4 ± 0.5 pF, n=10, p<0.05), indicating that constitutive calcineurin activity likely regulates basal VGCC function. CsA did not change the voltage- dependence of VGCC activation in either normoxic or post-hypoxic neurons (Fig. 6C, Normox-CsA Vhact: −2.1 ± 0.1 mV, n=8; Hypox-CsA Vhact: −1.4 ± 0.2 mV, n=8, p>0.05.). The steady-state inactivation was also unaltered by CsA (Fig. 6C, Normox-CsA Vhinact: −40.2 ± 1.4 mV, n=10; Hypox-CsA Vhinact: −36.5 ± 2.7 mV, n=8, p>0.05.). Thus, co-incubation with CsA during hypoxia prevented the post-hypoxic augmentation of HVA Ca2+ current without altering channel activation or inactivation kinetics.
Although FK-506 specifically inhibits calcineurin activity, it also exhibits non-calcineurin-mediated activity through binding to FKBP-12. In addition, FKBP-12 itself can also modulate intracellular Ca2+ homeostasis and calcium-induced calcium release (Loughrey et al., 2004; Xiao et al., 1997) and thus could potentially influence VGCC activity through negative feedback. Similarly, CsA can reduce Ca2+ levels after hypoxia by preventing the opening of the mitochondrial permeability transition pore (MPTP; (Halestrap, 2006; Halestrap, 2009). Therefore, to confirm that reversal of post-hypoxic HVA Ca2+ current enhancement by FK-506 and prevention of enhancement by CsA were specific to inhibition of calcineurin activity, we performed additional control experiments to rule out alternative mechanisms. Rapamycin (RAP) is structurally similar to FK-506 and competes with FK-506 for binding to FKBP-12, but does not inhibit calcineurin activity (Snyder et al., 1998). Inclusion of RAP (5 μM) in the whole-cell recording pipette did not prevent the post-hypoxic increase in HVA Ca2+ currents in hypoxia-treated neurons compared to matched normoxic controls (see Fig. 6D–F). Maximal current density was significantly increased after hypoxia (Normox-RAP: 11.0 ± 1.0 pA/pF, n=9; Hypox-RAP: 17.5 ± 1.2 pA/pF, n=9, p<0.05), but did not alter the voltage dependence of VGCC activation (Fig. 6F, Normox-RAP Vhact: −2.5 ± 0.2 mV, n=9; Hypox-RAP Vhact: −2.4 ± 0.1 mV, n=9, p>0.05.) or steady-state inactivation (Normox-RAP Vhinact: −39.3 ± 5.1 mV, n=7; Hypox-RAP Vhinact: −40.6 ± 3.2 mV, n=10, p>0.05.). Unlike FK-506, RAP did not reverse the post-hypoxic enhancement of HVA Ca2+ current density, supporting the selective effect of FK-506 to inhibit calcineurin-dependent upregulation of VGCC activity after acute hypoxia.
To determine whether other protein phosphatases might be involved in post-hypoxic enhancement of Ca2+ currents, we used okadaic acid to inhibit the serine/threonine phosphatases protein phosphatase 1 (PP1) and protein phosphatase 2A (PP2A). These phosphatases have been shown to regulate cardiac and neuronal L-type Ca2+ channels (Davare et al., 2000; Ono and Fozzard, 1993). Calcineurin disinhibits PP1 in the intracellular Ca2+ signaling cascade (Budde et al., 2002; Guerini, 1997). The effects of calcineurin on HVA Ca2+ current after hypoxia could thus also involve PP1 and PP2A. To test this hypothesis, we co-incubated cortical neurons with 150 nM okadaic acid (OKA) during 4h hypoxia exposure. OKA is a relatively specific inhibitor of PP1 (IC50 of 10–15 nM) and PP2A (IC50 of 100 pM) (Bialojan and Takai, 1988), and only inhibits calcineurin at high concentrations (> 1 μM, IC50 ~ 5 μM) (Bialojan and Takai, 1988; Cohen, 1991). Inclusion of OKA (150 nM) in the medium during hypoxic exposure did not prevent the post-hypoxic increase in HVA maximal Ca2+ current density compared to matched normoxic controls (Fig. 6G–I; Normox-OKA: 12.0 ± 1.1 pA/pF, n=6; Hypoxia-OKA: 18.4 ± 2.2 pA/pF, n=7, p<0.05). OKA did not change the voltage dependence of VGCC activation (Fig. 6I, Normox-OKA Vhact: 0.2 ± 0.2 mV, n=6; Hypoxia-OKA Vhact: −5.2 ± 0.2 mV, n=7, p>0.05.) or steady-state inactivation (Fig. 6I, Normox-OKA Vhinact: −37.4 ± 3.0 mV, n=8; Hypoxia-OKA: -35.6 ± 3.3 mV, n=8, p>0.05). The inability of OKA to prevent the post-hypoxic enhancement of HVA Ca2+ current density again supports the involvement of calcineurin in post-hypoxic regulation of VGCC function.
A remaining question was the mechanism of calcineurin activation during hypoxia. Concurrent binding of Ca2+ to the regulatory calcineurin B subunit and Ca2+/calmodulin (CaM) to the catalytic calcineurin A subunit is required for the full activation of calcineurin (Groth et al., 2003). One likely source of Ca2+ is influx through the L-type VGCC itself, which could activate calcineurin co-localized with the channel via protein kinase A anchoring protein (AKAP) 79/150 (Oliveria et al., 2007). To test this hypothesis, cortical neurons were exposed to 4 h normoxia or hypoxia in the presence of NIM (10 μM). Cultures were subsequently washed multiple times and replaced with fresh aerated SMEM/HS culture medium. To further eliminate any residual NIM that might have inhibited L-type Ca2+ current, a modification of the “multipuffer” drug-delivery system (Greenfield, Jr. and Macdonald, 1996) was used to wash the selected neuron by delivering external recording solution for at least 1 min prior to patch-clamp recording. There was no significant difference in the amplitude of HVA Ca2+ currents following NIM incubation compared to control medium (see below), suggesting the successful wash-out of residual NIM.
As previously observed, maximal HVA Ca2+ current density increased after hypoxia (Normox: 12.7 ± 1.3 pA/pF, n=7; Hypox.4/Rec.0: 22.2 ± 1.4 pA/pF, n=9, p<0.05). Inclusion of NIM (10 μM) did not change maximal Ca2+ current density in normoxic control neurons (Normox-NIM: 11.8 ± 1.0 pA/pF, n=7), but completely prevented the post-hypoxic Ca2+ current increase (Hypox-NIM: 10.7 ± 1.3 pA/pF, n=9, p>0.05, Fig. 7). Since blockade of L-VGCCs with NIM during hypoxia prevented the enhancement of HVA Ca2+ current, L-VGCCs likely represent the source of Ca2+ required to activate calcineurin, which suggests a positive feedback mechanism in which hypoxic activation of L-VGCCs mediates Ca2+ entry, activation of calcineurin and subsequent enhancement of L-VGCC current.
In cultured rat cortical neurons, exposure to 1% O2 for 4 hours resulted in Ca2+ entry through L-type VGCCs, activation of calcineurin and a transient increase in L-type VGCC currents that returned to baseline within 48 h after exposure. Blockade of L-type channels during hypoxia prevented the increase in HVA Ca2+ currents, suggesting a critical role for L-VGCCs in their own potentiation. Inhibition of calcineurin either by CsA during hypoxia or by FK-506 in the recording pipette after hypoxia prevented the VGCC current increase. The increase in HVA Ca2+ current was specific to L-type channels, as there was no post-hypoxic increase in the residual current after blockade of L-VGCCs with NIM. These findings suggest a positive feedback loop reinforcing neuronal calcium signaling in the immediate post-hypoxic period.
To our knowledge, this is the first demonstration of L-VGCC enhancement by transient hypoxia in cortical neurons, which could contribute to post-hypoxic cortical hyperexcitability (Yechikhov et al., 2002), seizures or status epilepticus (Bladin et al., 2000; Krumholz et al., 1988). Increased intracellular Ca2+ might activate calcium-dependent processes leading to necrotic or apoptotic cell death (Wasterlain et al., 1993), or alternatively, L-VGCC upregulation and calcium entry might facilitate post-hypoxic dephosphorylation and translocation of the Ca2+-dependent potassium channel Kv2.1, which suppresses neuronal excitability, and thus have a compensatory or protective effect (Misonou et al., 2005).
The mechanism responsible for L-VGCC activation by hypoxia is unclear. Hypoxia may have led to depolarization, either by downregulation of O2-sensitive K+ ion channels (Lopez-Barneo et al., 2004),”anoxic depolarization” due to ATP depletion (Allen et al., 2005), or a direct effect of hypoxia on L-VGCCs. L-type VGCCs, which are expressed at high density in the somata of cerebral cortex and hippocampal neurons (Ricci et al., 2002), have a slow rate of inactivation that allows Ca2+ entry even with gradual depolarization, facilitating Ca2+ signaling in response to relatively mild depolarizing stimuli.
Our studies demonstrated enhancement of L-VGCC currents immediately after 4 h exposure to 1% O2, consistent with prior reports showing that brief hypoxia (15–30 mm Hg, equivalent to 2–4% O2) potentiated high-voltage-activated (HVA) Ca2+ currents in hippocampal CA1 neurons by 26–94% (Lukyanetz et al., 2003; Shkryl et al., 1999; Shkryl et al., 2001). More prolonged (24 h) moderate hypoxia (2.5 – 6% O2) enhanced Ca2+ currents in PC12 cells, cerebellar granule neurons and wild-type HEK 293 cells expressing recombinant L-type α1c (Cav1.2) subunits (Brown et al., 2005; Peers et al., 2005) by a trafficking mechanism involving the formation of amyloid beta (Aβ) peptide by γ-secretase and regulated by reactive oxygen species (Peers et al., 2005). It is unclear whether such a mechanism might be involved in our findings, though the lack of a change in voltage-dependent kinetics is consistent with altered trafficking rather than a direct effect on channel function.
In our experiments, the increase in HVA current was dependent on the action of the calcium-dependent phosphatase, calcineurin. Calcineurin-dependent upregulation of VGCC activity has been observed previously in hippocampal CA1 neurons. Inhibition of calcineurin by FK506, CsA or calcineurin autoinhibitory peptide reduced whole-cell HVA Ca2+ and Ba2+ currents through L-type Ca2+ channels both in culture (Norris et al., 2002) and in hippocampal “zipper” slices (Norris et al., 2008). In contrast, activation of calcineurin has also been associated with reduced L-VGCC function in cultured hippocampal pyramidal neurons. Neuronal L-type Cav1.2 activity was enhanced by PKA-mediated phosphorylation through interaction with A-kinase activating protein (AKAP) 79/150, which binds both PKA and calcineurin (Oliveria et al., 2007). Local calcium entry through L-VGCCs activated calcineurin which dominantly suppressed PKA enhancement of calcium currents, possibly via Ca2+-dependent inactivation of L-type channels (Sather et al., 2009), though an earlier study found no calcineurin dependence (Zeilhofer et al., 2000). Calcineurin-dependent downregulation of L-VGCCs was also observed in prefrontal cortical neurons, and was specific to Cav1.2 as it was not observed in Cav1.3 knockout animals (Day et al., 2002). These results differ from the present findings, in which calcineurin activation was essential for enhancement of Ca2+ currents, possibly suggesting that Cav1.3 is the primary L-VGCC alpha subunit involved. The lower activation threshold of Cav1.3 channels (Lipscombe et al., 2004) makes them suitable for upregulation by modest hypoxia-mediated depolarization, but altered kinetics might not be detected in a mixed channel population. Other studies have implicated protein phosphatase 2A (PP2A) in L-VGCC regulation, which binds directly to Cav1.2 and reverses PKA phosphorylation of serine 1928 (Davare et al., 2000). Since okadaic acid did not block hypoxia-induced potentiation of Ca2+ currents in our system, PP2A is not likely involved. The increase in L-VGCC current with no change in activation or inactivation parameters suggests a possible increase in channel number rather than a change in conductance or gating. The site of calcineurin action and the detailed mechanisms underlying increased L-VGCC current are important questions for future study.
Calcineurin activity (but not protein expression) was significantly increased in the rat hippocampus after pilocarpine-induced status epilepticus (Kurz et al., 2001) and was specifically enhanced in synaptic fractions and in the dendrites of hippocampal pyramidal neurons (Kurz et al., 2003). Similar increases were also seen in a fluid percussion model of brain injury (Kurz et al., 2005a; Kurz et al., 2005b). In the pilocarpine model of status epilepticus, calcineurin-dependent dephosphorylation and subsequent activation of the actin depolymerization factor, cofilin, led to loss of dendritic spines on hippocampal pyramidal neurons, possibly associated with epileptogenesis (Kurz et al., 2008). Indeed, calcineurin inhibitors prevented neuronal death and blocked epileptogenesis in a kainic acid model of status epilepticus (Moriwaki et al., 1998), inhibited progression of kindled seizures (Moia et al., 1994) and prevented axonal sprouting in a kindling model of epilepsy (Moriwaki et al., 1996), though recent studies did not confirm the latter finding (Ingram et al., 2009). Enhanced L-VGCC function could provide a mechanism for facilitation of calcineurin activity after status epilepticus and other post-hypoxic conditions associated with altered neuronal excitability.
One of the downstream effects of L-VGCC activation by hypoxia is regulation of neuronal inhibition. Blockade of L-VGCCs during hypoxia prevented the hypoxia-induced reduction of maximal GABAAR current both immediately and 48 h after 4h hypoxia (Wang and Greenfield, Jr., 2009), suggesting that L-VGCCs are activated during hypoxia and that the reductions in GABAAR current are Ca2+-mediated. The early component of this regulation was calcineurin-dependent and prevented by calcineurin inhibitors, while the late component (48 h after hypoxia) was not (Wang and Greenfield, 2008). These findings may have clinical significance, as hypoxia-induced reduction of GABAAR function in hippocampal CA1 pyramidal neurons and subsequent seizures in P10 rat pups were associated with calcineurin activation via Ca2+-permeable AMPA receptors, and both seizures and the reduction of miniature GABAAR currents were prevented by FK-506 (Sanchez et al., 2005). In this important study, hypoxia/seizure-induced Ca2+ entry for activation of calcineurin was through calcium-permeable AMPARs rather than L-VGCCs as observed here. This discrepancy is likely related to the marked differences in model systems between in vivo hypoxia followed by hippocampal slice recording vs. dissociated cell culture in which synaptic activity is minimal. The finding that calcineurin-induced downregulation of GABAAR function (Wang and Greenfield, Jr., 2009) was observed in both settings suggests that this may be a “final common pathway” for GABAAR downregulation via dephosphorylation. In Sanchez et al. (2005), dephosphorylation of the GABAAR β2 and/or β3 subunit was observed. However, calcineurin also interacts directly with the γ2 subunit of the GABAAR to mediate long-term depression of GABAergic neurotransmission (Wang et al., 2003).
Post-hypoxic enhancement of neuronal L-type VGCC currents, increased intracellular Ca2+ and augmented calcineurin activity may trigger downstream modulation of other ion channels as well, contributing to hyperexcitability and seizure generation or epileptogenesis. The downstream mechanisms underlying such long-term changes remain to be explored.
This work was supported in part by R01-NS049389 and a research grant from the Myoclonus Research Foundation, Ft. Lee, N.J. to L.J.G.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.