Abstract
Ca2+ currents, especially those activated at low voltages (LVA), influence burst generation in thalamocortical circuitry and enhance the abnormal rhythmicity associated with absence epilepsy. Mutations in several genes for high-voltage-activated (HVA) Ca2+ channel subunits are linked to spike-wave seizure phenotypes in mice; however, none of these mutations are predicted to increase intrinsic membrane excitability or directly enhance LVA currents. We examined biophysical properties of both LVA and HVA Ca2+ currents in thalamic cells oftottering (tg; Cav2.1/α1A subunit), lethargic (lh; β4 subunit), and stargazer (stg; γ2 subunit) brain slices. We observed 46, 51, and 45% increases in peak current densities of LVA Ca2+ currents evoked at −50 mV from −110 mV in tg, lh, andstg mice, respectively, compared with wild type. The half-maximal voltages for steady-state inactivation of LVA currents were shifted in a depolarized direction by 7.5–13.5 mV in all three mutants, although no alterations in the time-constant for recovery from inactivation of LVA currents were found. HVA peak current densities intg and stg were increased by 22 and 45%, respectively, and a 5 mV depolarizing shift of the activation curve was observed in lh. Despite elevated LVA amplitudes, no alterations in mRNA expression of the genes mediating T-type subunits, Cav3.1/α1G, Cav3.2/α1H, or Cav3.3/α1I, were detected in the three mutants. Our data demonstrate that mutation of Cav2.1 or regulatory subunit genes increases intrinsic membrane excitability in thalamic neurons by potentiating LVA Ca2+ currents. These alterations increase the probability for abnormal thalamocortical synchronization and absence epilepsy in tg, lh, andstg mice.
- Ca2+ currents
- thalamocortical relay cells
- absence seizures
- stargazer
- tottering
- lethargic
- calcium ion channel mutation
Rhythmic burst firing characterizes cellular signaling behavior in the thalamus and depends on intrinsic membrane properties of thalamocortical relay (TC) neurons, as well as the synaptically linked neurons in the adjacent thalamic reticular nucleus (nRT) (McCormick and Bal, 1997; Steriade, 2000). In both cell types, membrane hyperpolarization produces a Ca2+-dependent low-threshold depolarization that serves as the generator potential for bursts of Na+ spikes (Llinás and Jahnsen, 1982). Voltage-clamp studies have identified the ion channel responsible for thalamic postinhibitory rebound burst firing as a T-type Ca2+ channel (Jahnsen and Llinás, 1984; Coulter et al., 1989a; Crunelli et al., 1989; Huguenard, 1996). Three genes, Cacna1g,Cacna1h, and Cacna1i, encoding T-type Cav3.1/α1G, Cav3.2/α1H, and Cav3.3/α1I α subunits have been localized in rat brain; Cav3.1 is predominantly expressed in thalamic relay nuclei and Cav3.3 in the nRT (Talley et al., 1999). The evidence directly linking this pathway to human absence epilepsy is based on thalamic involvement in spike-wave electrogenesis (Williams, 1953; Niedermeyer et al., 1969) and the ability of anti-absence drugs to reduce low-voltage-activated (LVA) currents (Coulter et al., 1989b; Gomora et al., 2001). Indirect evidence from experimental models provides additional support; LVA currents in nRT are increased in a rat model of absence epilepsy (Tsakiridou et al., 1995), and the threshold for spike-wave generation is significantly elevated in Cav3.1 −/− mice (Kim et al., 2001).
The identification of mutations in Ca2+ channel subunit genes intottering (tg; Cav2.1/α1A) (Fletcher et al., 1996),lethargic (lh; β4) (Burgess et al., 1997), and stargazer (stg; γ2) (Letts et al., 1998) mice and in humans (Escayg et al., 2000; Jouvenceau et al., 2001) that display an absence epilepsy phenotype provide an excellent opportunity to determine how defective Ca2+ signaling leads to thalamocortical epilepsy. In tg and its alleletgla , dramatic reductions in single-channel open probability and peak current density of high-voltage-activated (HVA) P/Q-type (Cav2.1/2.2) Ca2+currents in cerebellar Purkinje have been reported; however, it is difficult to explain how this defect could enhance intrinsic membrane excitability if present in the thalamocortical circuit (Dove et al., 1998; Lorenzon et al., 1998; Wakamori et al., 1998). In lhPurkinje cells, loss of functional β4 subunits did not alter Cav2.1/2.2 Ca2+ currents, probably attributable to compensatory interactions with alternative β1–3 subunits (Burgess et al., 1999; McEnery et al., 1998). Similarly, loss of the γ2subunit in stg, shown recently to directly interact with HVA channels (Kang et al., 2001), had no significant effect on Ca2+ currents in cerebellar granule cells, presumably attributable to rearrangements with alternative members of the γ subunit family (Chen et al., 2000; Burgess et al., 2001).
Because the behavior of Ca2+ currents observed in cerebellar cells cannot be generalized to thalamic neurons, which express their own pattern of interacting subunits and modulatory signals, the cellular mechanisms of epileptogenesis in these three mutants remain unclear. We therefore evaluated the functional properties of both LVA and HVA Ca2+currents in mutant thalamic neurons and examined in vivogene expression patterns of the three LVA channel genes. Our results demonstrate that elevated thalamic LVA currents are a common feature shared by three distinct absence epilepsy gene mutations.
MATERIALS AND METHODS
Preparation of brain slices
Coronal brain slices (350-μm-thick) were prepared from 14- to 19-d-old homozygous wild-type (C57BL/6J),tottering(C57BL/6J-Cacna1atg/tg),lethargic(B6EiC3H-a/A-Cacnβ4lh/lh ), andstargazer(C57BL/6J-Cacnγ2stg/stg ) mice. Genotypes of the three mutants were confirmed by PCR of tail DNA (Burgess et al., 1997). Slices were obtained at the level of the lateral dorsal nucleus (LDN) of the thalamus, selected because of its projection to frontal cortical regions in which cortical spike-wave discharges in these mutants predominate (J. L. Noebels, unpublished observations). Each slice was perfused with a solution containing (in mm) (Kapur et al., 1998): 125 choline-Cl, 3.0 KCl, 1.25 NaH2PO4, 25 NaHCO3, 1.0 Ca Cl2, 7.0 MgCl2, 10 dextrose, 1.3 ascorbate acid, and 3.0 pyruvate (bubbled with 95% O2–5% CO2). Slices were then incubated in an artificial CSF solution for 40 min at 37°C and then maintained at room temperature (22–25°C). The artificial CSF was gassed with 95% O2–5% CO2 and contained (in mm): 130 NaCl, 3.0 KCl, 2.0 MgCl2, 2.0 CaCl2, 1.25 NaH2PO4, 26 NaHCO3, and 10 glucose.
Electrophysiological recording
Macroscopic Ca2+ currents from thalamocortical cells in the LDN were recorded using the whole-cell configuration of the patch-clamp technique. A Zeiss (Oberkochen, Germany) Axioskop fitted with a 40× water-immersion objective and differential interference contrast optics was used to view the slices and to identify neurons for analysis. Voltage command pulses were generated by a computer using pClamp 8.02 software. Currents were recorded with an Axopatch-1D amplifier, filtered at 10 kHz (−3 dB), and compensated for series resistance (∼70%). Patch electrodes were drawn from borosilicate glass and coated with Sylgard. Ca2+ currents were corrected for leak and capacitive currents by subtracting a scaled current elicited by a 10 mV hyperpolarization from the standard holding potential of −70 mV. All recordings were performed at room temperature (22–25°C). The standard holding potential was −70 mV. To identify and stain the neurons in whole-cell recordings from the slice, 1% biocytin was included in the intracellular patch pipette solution. After recordings, the slices were cut into 50-μm-thick sections and then immediately fixed. A standard histochemical procedure was used to process the sections and stain the injected neurons (Huguenard and Prince, 1992).
Solutions. The recording bath solution consisted of (in mm): 115 NaCl, 3.0 KCl, 10 sucrose, 10 glucose, 26 NaHCO3, 2 MgCl2, 2.5 CaCl2, 0.5 4-aminopyridine, 5 CsCl, 10 tetraethylammonium-Cl (TEA-Cl), and 0.001 TTX, pH 7.4 (gassed with 95% O2–5% CO2). The intracellular pipette solution contained (in mm): 78 Cs-gluconate, 20 HEPES, 10 BAPTA-Cs4 (cell-impermeant), 0.5 CaCl2, 1.0 MgCl2, 4 Mg-ATP, 0.3 GTP-Tris, 6 phosphocreatine (Di-Tris salt), 4.0 NaCl, and 20 TEA-Cl, pH 7.3 (titrated with CsOH).
Voltage protocols. To generate Ca2+ channel current–voltage (I–V) curves, currents were elicited by applying voltage step commands (200 msec) to varying potentials from a 3 sec prepulse potential at −60 or −110 mV. The I–V protocol for HVA Ca2+ currents consisted of voltage steps from −80 to +60 mV in 5 mV increments triggered from a 3 sec prepulse potential at −60 mV. To define LVA Ca2+ currents, difference currents obtained by digital subtraction of the currents elicited during depolarizing voltage steps from −60 and −110 mV were used. Standard voltage protocols for steady-state activation (SSA) of HVA Ca2+ currents, as well as the steady-state inactivation (SSI) and recovery from inactivation of LVA currents, respectively, were applied, and are explained in further detail below (see figure legends). In our study, we did not find any significant time-dependent ICa2+ run down within 30–40 min after membrane breaking. Statistical data analysis was tested by one-way ANOVA with the post hoc test. Differences with p < 0.05 were scored as statistically significant. The data shown represent means ± SE
In situ hybridization
In situ hybridization of mRNAs encoding α subunits for three calcium channel subtypes mediating LVA T-type Ca2+ currents, Cav3.1/α1G, Cav3.2/α1H, and Cav3.3/α1I, was performed in 2- to 3-week-old homozygous tg, lh, andstg mutants and C57BL/6 +/+ mice using standard techniques described previously in detail (Burgess et al., 1999). Briefly, horizontal brain sections (12-μm-thick) from 14- to 19-d-old mice were fixed in 4% paraformaldehyde in PBS and dehydrated through an ascending ethanol series. Antisense oligonucleotide probes were end labeled using terminal deoxynucleotidyl transferase (Promega, Madison, WI) and [α-35S]dATP (1250 Ci/mmol; NEN, Boston, MA) to a specific activity of ∼109 dpm/μg. The hybridization solution contained 50% (v/v) formamide, 4× SSC, 25 mm sodium phosphate, 1 mmsodium pyrophosphate, 10% dextran sulfate (w/v), 5× Denhardt's solution, 200 μg/ml sonicated herring sperm DNA (Promega), 100 μg/ml polyadenylic acid [5′] (Sigma-Aldrich, Milwaukee, WI), and 5 × 102 dpm of [α-35S]dATP-labeled probe. Control sections were hybridized with an additional 100-fold excess of unlabeled oligonucleotide. The sequences of the 45-mer probes were as follows: Cacna1g, 5′-GATGCAGCTGGTGTCTGCTGGTTGGGAGTGAACAGACAAGATGG-3′; Cacna1h, 5′-CAAGAAGGTCAGGTTGTTGTTCCTGACGAAGGCGCTGTCCA-GGAA-3′; andCacna1i, 5′ GCGGATGGCTGACAGGTTGATGTTC-TGTAGGTCCAGAGAGTACTC-3′. The probes were hybridized to the sections overnight at 42°C, washed in 1× SSC (22°C, 20 min), 0.3× SSC (55°C, 40 min), and 2× SSC (22°C, 5 min), and then dehydrated and exposed to Kodak BioMax MR film (Eastman Kodak, Rochester, NY) for 1 week. Developed autoradiographs were digitized (Sprintscan 35; Polaroid, Cambridge, MA) and arranged using Photoshop 5.0 (Adobe Systems, San Jose, CA), and all images were processed simultaneously. Optical density reflecting relative abundance of mRNA was determined by Scion (Frederick, MD) Image–NIH Image software. The in situ results were collected from wild-type and mutant mice (three animals per group). Eight brain sections were obtained from each group. Multiple small square areas within several anatomically distinct brain areas were selected, including the lateral dorsal nucleus and the adjacent white matter as an internal standard. These values were used to determine the lateral dorsal/white matter density ratio between affected mice and the respective homozygous wild-type mice. Differences in the optical densities were analyzed for statistical significance using Student's t test.
RESULTS
Increased peak current density of LVA calcium channels in TC neurons of tottering, lethargic, andstargazer mutants
We investigated the effects of mutation of Cav2.1/α1A (tg), β4 (lh), and γ2 (stg) Ca2+ channel subunits on Ca2+ currents in mouse LDN thalamic neurons defined by intracellular staining of biocytin. In mouse, these cells show a round-shaped cell body with multiple dendrites, similar in morphology to TC neurons from rat ventrobasal nucleus (Destexhe et al., 1998) (Fig.1A). Figure1B shows representative traces of LVA Ca2+ currents in response to a test pulse to −50 mV from a 3 sec prepulse to −110 mV in TC neurons from wild-type, tg, lh, and stg mice. At a membrane potential of −50 mV, all LVA Ca2+ currents have recovered from inactivation and are thus available for opening in both wild-type and mutant neurons (Figs.2B, 3), whereas HVA Ca2+ currents in these cells have not yet started to activate (see Fig. 5A). The current traces of the LVA calcium channels show fast activation and inactivation, similar to those studied in vitro by expression of Cav3.1/α1G and Cav3.2/α1H T-type calcium channels (Lee et al., 1999; Delisle and Satin, 2000; Zhang et al., 2000), as well as native LVA currents from dissociated rat TC neurons (Destexhe et al., 1998). The peak current densities (i.e., normalized by cell capacitance) of LVA currents at a membrane potential of −50 mV increased by 46% in tg, 51% in lh, and 45% instg mutants compared with control (Fig. 1C). The mean peak current amplitude and peak current density were −926.3 ± 182.2 pA and 9.5 ± 1.3 pA/pF in control, −1571.9 ± 106.8* pA and 17.63 ± 1.6** pA/pF in tg, −1852.1 ± 118.9** pA and 19.6 ± 1.0** pA/pF in lh, and 1514.9 ± 142* pA and 17.4 ± 1.2** pA/pF in stgmice (*p < 0.05; **p < 0.01 vs control).
The duration of macroscopic inactivation for LVA Ca2+ currents in both control and mutant mice is closer to the time scale observed in Cav3.1 but not Cav3.2 T-type calcium channels expressed in mammalian cells (Lee et al., 1999; Zhang et al., 2000). The decay of macroscopic LVA currents evoked at −50 mV was fitted by a single-exponential function (Fig. 1B). No significant alterations in macroscopic LVA current decay were found in any of the three mutants compared with the wild type. The time constants for decay were 29.2, 29.7, 30.1, and 31.8, respectively, in control, tg, lh, andstg mice.
Depolarizing shifts in voltage dependence of steady-state inactivation and kinetics for recovery from inactivation of LVA currents
We next examined the kinetics of LVA Ca2+ currents in both wild-type and mutant neurons. The current traces of SSI of LVA are shown in Figure2A. For the SSI protocol, we used a 4 sec prepulse to various membrane potentials before delivering a second test stimulus to −50 mV. The 4 sec prepulse was long enough to bring channels to a steady-state condition, because all LVA Ca2+ channels in TC neurons recover from inactivation within 3 sec (Fig. 3). As demonstrated in Figure2A, LVA currents elicited at −50 mV from different premembrane potentials in both control and mutants show fast inactivation and decay completely within 150 msec. We did not observe any sustained component for current decay in tg,lh, or stg mice. Nevertheless, we found a significant depolarizing shift of the steady-state inactivation curves of LVA currents in the mutants in contrast to wild-type neurons (Fig.2B). The mean half-maximal voltage (V1/2) and slope (k) for SSI curves were −92.3 ± 0.16 and 6.8 ± 0.16 in control, −84.8 ± 0.17* and 6.51 ± 0.15 mV in tg, −78.72 ± 0.3** and 6.0 ± 0.27 mV in lh, and −78.6 ± 0.3** and 6.0 ± 0.27 mV in stg, respectively (*p < 0.05; **p < 0.01). The 7.5–13.5 mV depolarizing shifts of the voltage dependence for SSI of LVA currents in TC neurons of the mutants suggest that, at physiological membrane potentials varying from −70 to −75 mV, a higher fraction of all LVA calcium channels are available for opening in the mutant relative to control mice. Recovery from inactivation of the LVA Ca2+ currents in all mice was complete within 3 sec (Fig. 3). The recovery from inactivation curve was best fitted with a two-exponential function, and the fast and slow time constant derived from curve fitting did not significantly differ in any of the mutants compared with control mice. Thus, the depolarizing shift of SSI curves for LVA currents provides an additional biophysical mechanism that may contribute to increased neuronal burst synchronization observed in tottering,lethargic, and stargazer mice.
To determine whether the dramatic depolarizing shifts of SSI curves in all three mutants involves phosphorylation of thalamic T-type channels, we examined the effects of a specific protein kinase A inhibitor (PKA-I), as well as a protein kinase C inhibitor (PKC-I) on the voltage dependence of SSI for LVA currents. PKA-I at 50 μm(fragment 6–22, amide) or 100 μm PKC-I (fragment 19–36) was loaded into the pipette solution as described previously (Zhang et al., 2000), and the SSI protocol for LVA channels was then initiated 15–20 min after membrane rupture. We observed that neither PKA-I nor PKC-I antagonized the shift of V1/2for steady-state inactivation of the LVA currents found in untreatedtg, lh, or stg mutants (data not shown). In addition, neither PKA-I nor PKC-I altered the baseline value of V1/2 for SSI in untreated wild-type neurons, suggesting that the resting level of PKA or PKC could be low in both control and mutant mice. These observations suggest that neither PKA nor PKC significantly modulate the voltage dependence of SSI for LVA channels in these mutants, although consensus sites for both PKA and PKC exist on Cav3.1/α1G and Cav3.2/α1H channels (Cribbs et al., 1998;Perez-Reyes et al., 1998).
High-voltage-activated Ca2+ peak currents are increased in tottering andstargazer but not lethargic mice
We then determined whether mutation of Cav2.1/α1A (tg), β4 (lh), or γ2 (stg) Ca2+ channel subunits affect HVA Ca2+ currents in LDN cells. HVA Ca2+ currents are mediated by pore-forming α1 subunits, with current amplitude and gating regulated by cytoplasmic β subunits and transmembrane α2δ and γ subunits (Ahlijanian et al., 1990; Chien et al., 1995; Witcher et al., 1995;Gurnett et al., 1996; Walker and De Waard, 1998; Yamaguchi et al., 1998; Meir et al., 2000; Kang et al., 2001).
The I–V relationships of Ca2+ currents for control and three mutants are shown in Figure 5A. Figure4 shows representative HVA Ca2+ current traces from wild-type,tg, lh, and stg mutant neurons. Both control and mutant HVA Ca2+ currents start to activate at approximately −45 mV. The currents reach a peak at −10 to −15 mV in control and mutant mice (Fig.5A). Pooled peak currents and peak current densities are shown in Figure 5B. The mean peak current density was 10.12 ± 1.2, 13.03 ± 0.6**, 9.42 ± 0.4, and 17.99 ± 2.1** pA/pF in control, tg,lh, and stg (**p < 0.01 vs control). The mean peak amplitude was −885.51 ± 112.8, 1204.21 ± 86.4**, −901.3 ± 48.7, and −1567.6 ± 188** pA in control, tg, lh, and stgmutants, respectively (**p < 0.01 vs control).
Voltage dependence of steady-state activation for HVA currents
The steady-state activation curves of HVA Ca2+currents in TC neurons are demonstrated in Fig 5C. TheV1/2 for steady-state activation of HVA currents in tottering and stargazer mutants did not change dramatically but shifted by 5 mV in a depolarized direction in lethargic mice relative to that in control. The mean values of V1/2 and slope for SSA of HVA channels were −22.0 ± 0.17 and 5.5 ± 0.15 mV in control, −21.7 ± 0.16 and 4.5 ± 0.14 mV intg, −17.0 ± 0.11* and 4.7 ± 0.1 mV inlh, and −23.9 ± 0.18 and 4.6 ± 0.16 mV instg, respectively (*p < 0.05 vs control). The depolarized shift of the SSA curve in lh neurons is consistent with the in vitro finding that heterologous coexpression of β4 subunits elicited a hyperpolarizing shift for the SSA of Cav2.1 HVA calcium channels (De Waard and Campbell, 1995).
Expression of T-type Ca2+ channel Cav3.1/α1G mRNA in tottering,lethargic, and stargazer thalamus
To ascertain whether alterations in T-type calcium channel gene expression could underlie the increased LVA currents observed in mutant TC neurons, we compared regional mRNA expression levels of Cav3.1/α1G, Cav3.2/α1H, and Cav3.3/α1I. Figure6 shows representative autoradiograms of coronal sections taken at the thalamic LDN level from mutant and wild-type brains. The three genes encoding neuronal T-type Ca2+ channels are expressed in distinct and primarily non-overlapping patterns that are consistent with localization patterns in adult rat brain (Talley et al., 1999). In wild-type mouse brain, Cav3.1 mRNA was expressed at high levels in thalamic relay nuclei, Cav3.3 mRNA was detected at high levels in nRT, and Cav3.2 mRNA was detected at only trace levels in either. No extrathalamic alterations were noted in any of the three mutants, and only the expression patterns of each gene in the thalamus are described.
Cav3.1/α1G mRNA
Transcripts of this gene were detected at highest levels in thalamic relay nuclei, with barely detectable levels present in nRT. This pattern was not appreciably altered in homozygous tg,lh, or stg mutant mice.
Cav3.2/α1H mRNA
Only barely detectable levels of Cav3.2 appeared in the wild-type thalamic relay nuclei, and trace levels were found in nRT. This pattern was not visibly altered in the tg,lh, or stg mutant mice.
Cav3.3/α1I mRNA
Cav3.3 was the most broadly distributed of the three T-type genes in the mouse thalamus, with high levels of mRNA observed in nRT. Low and barely detectable levels of Cav3.3 mRNA were detected in principle thalamic relay nuclei. The pattern of this gene was not appreciably altered in homozygous tg,lh, or stg mutant mice. Optical densitometry confirmed that there were no significant differences in optical densities of thalamic expression of Cav3.1–3 mRNA between the three mutants and the wild-type mice.
Although the correlation between mRNA intensity and LVA current density is unknown, the lack of appreciable differences in the in situ hybridization results between mutant and control provides no evidence to support the hypothesis that the Ca2+ channel subunit mutations intg, lh, and stg mutants lead to functionally significant dysregulation of Cav3.1, Cav3.2, or Cav3.3 mRNA levels, and thus the ∼50% increases in thalamic T-type Ca2+ currents cannot be simply accounted for by a proportional upregulation of channel gene expression.
DISCUSSION
Our results demonstrate increased peak current densities and a depolarizing shift of the steady-state inactivation curves of LVA currents, as well as subunit-specific HVA current modifications intottering, lethargic, and stargazerthalamic neurons. The enhanced LVA currents could not be simply explained by major changes in thalamic Cav3.1, Cav3.2, or Cav3.3 gene expression in any of the three mutants. Both the increased LVA current density and shifted voltage dependence of inactivation may favor spike-wave burst firing, the excitability phenotype shared bytg, lh, and stg mice. An additional finding was that HVA peak current densities, which may also facilitate network oscillations, are increased in tg and stgbut not in lh thalamic neurons. This selectivity emphasizes the diversification of cellular compensatory mechanisms modulating Ca2+ currents after mutation of Cav2.1/α1A, β4, or γ2 Ca2+ channel subunits. Our study also provides a key mechanistic link between mutation of HVA subunits and the expression of thalamocortical hypersynchrony.
Functional properties of LVA Ca2+channels in mutant thalamic neurons
The similarity of the spontaneous six to seven per sec spike-wave phenotype in tg, lh, andstg mice (Noebels and Sidman, 1979; Noebels et al., 1990;Hosford et al., 1992) suggests that a common cellular mechanism, potentiation of LVA currents, may be responsible for enhanced neuronal synchronization attributable to the distinct Ca2+ channel subunit mutations. LVA currents in thalamic neurons play a major and perhaps essential role in the genesis of synchronized oscillations in this system by amplifying postinhibitory high-frequency rebound bursting that enables rhythmic firing patterns. The HVA Ca2+ channel Cav2.1 and β4 subunits have not been shown to physically interact with T-type channel proteins or influence T-type channel biophysical properties in central neurons, although there is emerging evidence that γ2 may interact with both HVA and LVA channel types. Nonetheless, we found dramatic increases (45–51%) in peak current densities of thalamic LVA currents in all three mutants. The increases favor augmented burst firing and membrane hyperexcitability, because T-type channels start to activate at relatively hyperpolarized membrane potentials (Huguenard and Prince, 1992; Delisle and Satin, 2000; Zhang et al., 2000), and thalamic neurons have hyperpolarized resting membrane potentials attributable to the rhythmic inputs from GABAergic nRT neurons (Steriade and Llinás, 1988). In our study, LVA currents started to activate at approximately −65 to −70 mV (data not shown). Interestingly, we also observed a 7.5–13.5 mV depolarizing shift in SSI of LVA channels in all three mutants compared with theV1/2 for SSI in wild-type mice. This very large depolarizing shift in SSI indicates a 7–30% elevation in LVA channel availability in the range of membrane potentials close to the resting membrane potential (−65 to −75 mV) (Fig. 2), and hence more channels will be open once the threshold for activation is reached.
Candidate mechanisms underlying LVA current alterations
We initially considered the possibility that peak LVA current increases might arise from additional de novo T-type channel synthesis, by either upregulation of native Cav3.1/α1G subunit expression or ectopic transcription of Cav3.2/α1H and Cav3.3/α1I subunits not normally expressed in these cells. Expression patterns of all three CaVT genes were unaltered, indicating that neither the mutant subunits nor the thalamocortical seizures they provoke are sufficient to induce CaVT gene transcription. The latter observation is consistent with the absence of c-Fos or c-Jun dysregulation in stg (Nahm and Noebels, 1998), confirming that the rhythmic oscillations during spike-wave synchronization are weak promoters of molecular plasticity compared with the prolonged, continuous depolarizations that stimulate expression of Ca2+ channels and other genes after convulsive seizures (Vigues et al., 1999). Although the increase in LVA peak current density detected in all three mutants is unlikely to result from altered CaVT gene expression, we cannot entirely exclude the possibility of persistent LVA channel elevation attributable to small increases of mRNA below the resolution of the method, increased mRNA translation efficiency, or decreased lability of the membrane protein. A mismatch between elevated LVA T-type Ca2+ currents and gene expression was observed previously in dissociated nRT neurons of the genetically undefined GAERS rat strain with absence epilepsy (Tsakiridou et al., 1995). In this model, one study detected no change in Cav3.1 expression (de Borman et al., 1999), whereas another reported small elevations of Cav3.1 mRNA in ventral posterior lateral thalamic relay nuclei and of Cav3.2 mRNA in nRT neurons (Talley et al., 2000).
Absence of thalamic CaVT gene upregulation suggests the existence of alternative nontranscriptional mechanisms for increased LVA current densities and shifted voltage dependence of SSI in tg, lh, and stg mutants. Because neither Cav2.1/α1A nor β4 proteins physically interact with Cav3.1/α1G subunits, the reduction of P/Q currents in tg and lh mutants must modify LVA current indirectly by altering a downstream Ca2+ signaling pathway. In the case ofstg mice, direct interactions between γ2 and α1G may occur in vivo, and lack of regulatory γ2 subunits could directly alter current density and SSI by allowing novel γ3–8 subunit interactions with LVA channel complexes (Dolphin et al., 1999;Green et al., 2001; Rousset et al., 2001). Other potential mechanisms for elevated LVA current density include the contribution of depolarizing shifts of SSI, which increase channel availability, or increased membrane insertion of channels or spatial redistribution of channels from dendritic to somatic compartments in which they could affect the voltage dependence of inactivation and maximally influence firing properties (Karst et al., 1993). Finally, although we found little evidence for abnormal modulation of T-type channels in mutant neurons by PKA or PKC, modulation by PKG or through other pathways, including pH (Delisle and Satin, 2000; Shan et al., 2001), G-protein (Matsushima et al., 1993; Park and Dunlap, 1998), or molecules such as the endogenous cannabinoid anandamide (Chemin et al., 2001), remain possible candidate mechanisms.
Differential effects of Cav2.1/α1A, β4, and γ2 channel subunit mutations on HVA currents
We found that thalamic Ca2+ currents in tg brain slices showed a 22% increase in HVA peak current density, and a similar result was demonstrated recently in dissociated TCs of Cav2.1-deficient mutants (Song et al., 2001). Because point mutation or deletion of the Cav2.1 gene reduces or eliminates P-type currents (Dove et al., 1998; Wakamori et al., 1998; Jun et al., 1999), the small increase in whole-cell HVA peak current density observed in our experiment likely represents upregulation of other HVA currents, such as L-, N-, or R-type. Indeed, decreased Ca2+ currents through tgP/Q-type channels significantly increases mRNA levels for L-type Ca2+ channels in cerebellar Purkinje cells (Campbell and Hess, 1999). In TC neurons, we found that L-type currents account for 55–70% of whole-cell HVA currents in both control and the three mutants (data not shown). Additional pharmacological studies with selective Ca2+ channel blockers will define which specific HVA channels account for the increases intg thalamic HVA currents.
The increased HVA Ca2+ peak current density in stg is consistent with recent functional analysis of γ2 subunit interactions with neuronal calcium channels, indicating that loss of the γ2 subunit may increase HVA currents (Letts et al., 1998; Klugbauer et al., 2000; Kang et al., 2001). These studies found that γ2/γ3 subunits cosedimented and coimmunoprecipitated with either Cav2.1/α1A or Cav2.2/α1B subunit from rabbit cerebellum and that γ2 coexpression in Xenopus oocytes significantly decreased the current amplitude of both Cav2.1 and Cav2.2 calcium channels. Of the neuronal Ca2+ channel γ subunits investigated so far, the γ2 and γ4 subunits shift the steady-state inactivation curve toward hyperpolarized potentials when coexpressed with Cav2.1 (Klugbauer et al., 2000). Thus, absence of γ2 in stg neurons should increase channel availability at membrane potentials (−40 to −30 mV) that start to activate HVA Ca2+ channels. Our data showed a 45% increase in HVA peak current density in TC neurons that is consistent with the non-neuronal expression studies.
In contrast to tg and stg, mutation of the β4 subunit did not significantly alter peak HVA current density inlh thalamic neurons. This result was unexpected, because β subunits regulate assembly and membrane incorporation of Cav2.1 subunits mediating HVA currents (Nishimura et al., 1993; Chien et al., 1995) and influence the amplitude of Ca2+ currents (De Waard and Campbell, 1995; Roche and Treistman, 1998) in vitro. β subunits have been also found to associate and functionally modify native P/Q-, L-, or N-type Ca2+ channels (Chien et al., 1995; Brice et al., 1997; Meir et al., 2000). Interestingly, the lack of an effect on HVA peak current in TCs oflh mice is consistent with the same negative effect on P/Q-type Ca2+ currents in dissociatedlh Purkinje cells (Burgess et al., 1999) and on P/Q channel function mediating transmitter release at lh hippocampal synapses (Qian and Noebels, 2000). Because alternative β subunits are widely localized in the brain, the negative impact of the β4 mutation on thalamic HVA current in lh may be explained by β subunit reshuffling (Burgess et al., 1999). Finally, the 5 mV depolarizing shift of SSA found in lethargic mice is consistent with in vitro data that β4 subunit interaction elicits a hyperpolarizing shift of inactivation curve for HVA currents (De Waard and Campbell, 1995). Because lh mice show both the LVA current increase and the spike-wave EEG phenotype, these findings suggest that the thalamic HVA alterations in tg andstg favor, but are not essential for, spike-wave generation.
In conclusion, our results are the first Ca2+ current defects to be described intg, lh, and stg thalamic neurons and provide supportive physiological evidence for the functional interaction of β and γ regulatory subunits in modulating LVA currents.
Footnotes
This work was supported by an American Epilepsy Society–Epilepsy Foundation Postdoctoral Fellowship (Y.Z.), National Institutes of Health Grant NS97209 (J.L.N.), and the Blue Bird Circle Foundation.
Correspondence should be addressed to Dr. Jeffrey L. Noebels, Department of Neurology, Baylor College of Medicine, One Baylor Plaza, Houston, TX 77030. E-mail: jnoebels{at}bcm.tmc.edu.