Elevated Thalamic Low-Voltage-Activated Currents Precede the Onset of Absence Epilepsy in the SNAP25-Deficient Mouse Mutant Coloboma

Cortical spike-wave discharges in the Cm/+ mutant are associated with elevated LVA Ca²⁺ currents in thalamic neurons

Similar to mice with recessive mutations in calcium channel subunit genes, we observed frequent spontaneous bursts of 5–6 Hz bilateral cortical SWDs in all Cm/+ mutants. These stereotyped discharges appeared in the awake animal and were always accompanied by simultaneous behavioral arrest. Representative bilateral EEG traces recorded in adult Cm/+ mice are shown in Figure 1A (top). Intraperitoneal injection of 5 mmol/kg ethosuximide (ETX) completely blocked SWDs within 2–4 min (Fig. 1A, middle). The mean number of SWDs in adult Cm/+ mice was reduced from 29.3 ± 4.2 per hour before injection to 0 per hour for 2–4 hr after injection (Fig. 1A, bottom), followed by full recovery of the SWDs and absence-type seizures. The sensitivity of intraperitoneal injection of ETX in Cm/+ mutants was reduced significantly compared with the effective doses used in mice absence models: tg, lh, and stg models (1–1.4 mmol/kg) (Heller et al., 1983; Aizawa et al., 1997). Spike wave discharges of this general pattern and behavioral phenotype are seen in human absence epilepsy and Ca²⁺ channel mouse mutants (Noebels and Sidman, 1979; Hosford et al., 1992; Qiao and Noebels, 1993), as well as in other genetic models of absence in rat (Marescaux et al., 1992; Coenen and Van Luijtelaar, 2003).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Spike-wave discharges coexist with elevated low-voltage-activated Ca²⁺ peak current in thalamic neurons of Coloboma mutant mice. A, Representative bilateral EEG traces showing 5∼6 Hz spike-wave discharges (top) and pharmacological sensitivity to 5 mmol/kg ethosuximide (ETX, i.p.) in Cm/+ mutants (n = 5; at P30) (middle). L and R designate left and right frontal cortex recording sites. The bottom panel shows the mean incidence of spike-wave discharges in Cm/+ mutant mice before and after ETX injection. B, Increased thalamic LVACa²⁺ peak current in Coloboma. Top panel, Representative LVA current traces from thalamocortical relay cells (TCs) of the LDN in control (C3H) and Cm/+ mice at age P15. The cell capacitance values of the two neurons were 100 and 101 pF in C3H and Cm/+, respectively; holding potential =–70 mV. The membrane potential was prepulsed to –110 mV for 3 sec before stepping to –50 mV for 200 msec. Decay of the current was fitted by a single-exponential function (superimposed dotted line). A significant alteration in the macroscopic current decay was found in Cm/+ neurons. The time constants (τ) for decay of the representative currents are 24.6 msec in C3H and 29.8 msec in Cm/+, respectively. Bottom panel, Elevated LVACa²⁺ current amplitude and peak current density fromCm/+ TCs. LVA currents were evoked at the same membrane potential as described in top panel. **p<0.001 versus control at corresponding age.

We then analyzed low-voltage-activated Ca²⁺ current amplitudes and kinetics in TC neurons to determine whether any alterations coexist with the expression of absence epilepsy in the Cm/+ mutant. In Figure 1B, the top panel shows representative traces of LVA Ca²⁺ currents in response to a test pulse to –50 mV arising from a 3 sec prepulse to –110 mV in TC neurons from wild-type and Cm/+ 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 (see Figs. 4B, 5B), whereas HVA Ca²⁺ currents in these cells have not yet started to activate (see Fig. 6B). The current traces of the LVA calcium channels show fast activation and inactivation, similar to that found in vitro by expression of Ca_v3.1(α_1G) and Ca_v3.2(α_1H) 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) and TC neurons of tg, lh, and stg brain slices (Zhang et al., 2002). The peak current densities (normalized by cell capacitance) of LVA currents at a membrane potential of –50 mV were increased by 60% compared with the corresponding value in control neurons at this age (P14–16) (Fig. 1B, right bottom panel). The mean peak current amplitudes and peak current densities were –1142.3 ± 131.9 (pA) and 12.4 ± 1.1 (pA/pF) in control mice and –1916.2 ± 54.3** (pA) and 19.8 ± 0.4** (pA/pF) in Cm/+ (**p < 0.001 vs control).

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

Depolarized shift of the voltage dependence of low-voltage-activated calcium channel availability (steady-state inactivation) in Coloboma mutants. A, Representative current traces for SSI of LVA Ca²⁺ currents. A standard double-pulse protocol for steady-state inactivation was given from the holding potential of –70 mV. A 4 sec prepulse at potentials ranging from –120 to –40 mV preceded each depolarization, followed by a subsequent voltage step to –50 mV for 200 msec. The interpulse interval was 10 sec. B, Normalized current–voltage curves for SSI of LVA Ca²⁺ currents. Current amplitude from the inactivation protocol, normalized to maximum, was plotted as a function of prepulse membrane potentials and best fitted with a Boltzmann function: I/I_max = {1 + exp(V – V_1/2)/k}–1.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5.

Recovery from inactivation of LVA Ca²⁺ currents. A, Representative current traces for recovery from inactivation of LVA currents in control and Cm/+. The holding potential was set to –50 mV, and 50 mV hyperpolarizations of incremental duration were applied. LVA peak amplitude was measured after returning to –50 mV. B, Recovery from inactivation curves. Recovery curves were established by plotting the normalized peak amplitude versus duration. The recovery curves followed a two-exponential time course, and the fast time constant (τ₁) and slow time constant (τ₂) were derived from curve fitting.

• Download figure
• Open in new tab
• Download powerpoint

Figure 6.

HVA Ca²⁺ currents in Cm/+ and wild-type mice. A, Representative superimposed HVA Ca²⁺ current traces from thalamic cells in control and Cm/+ mice. The I–V protocol consisted of a 3 sec prepulse potential at –60 mV followed by voltage steps (200 msec) ranging from –80 to +60 mV in 5 mV increments, with the holding potential maintained at –70 mV. B, Current density–voltage curves for HVA Ca²⁺ currents, constructed by plotting the normalized current amplitude at various membrane potentials. The voltage protocol used was identical to that described in A. C, Peak current density and current amplitude from B. D, SSA of HVA Ca²⁺ currents in control and Cm/+. The steady-state conductance (G) and voltage (V) data were transformed from I–V data shown in B. The solid and dotted curves are fits of the data to the Boltzmann equation of the form: G/G_max = 1/(1 + exp(V_{1/2 –} V)/k), where G_max is maximum conductance, V_1/2 is half-maximal voltage, and k is the slope.

Age of onset and mean incidence of spontaneous cortical SWDs in Cm/+ mutants

The onset of cortical SWDs in Cm/+ mice, like those in mice bearing Ca²⁺ channelopathies, displayed a highly reproducible developmental profile. Representative bilateral EEG traces of both Cm/+ mutants and C3H control mice recorded at various postnatal ages from P13 through adulthood are shown in Figure 2A and are compared with EEG traces at the same ages from tg and stg mutants as well as their control genotypes. In our study, no synchronous discharges were ever recorded from Cm/+ mutants at ages P11–13. SWDs emerged at P14–15 and were always present by P20 in Cm/+ mice. As shown in Figure 2A, most of the spike-wave discharges in the immature (P14–18) Cm/+ mouse brain display a pattern of prolonged and frequent 3–4 Hz synchronous discharges, intermixed with occasional shorter 5–6Hz bursts. From P20 onward, a 5–6 Hz spike-wave pattern becomes the predominant pattern in Cm/+ mutants. In comparison, cortical SWDs were also first visible at P13 and P14, respectively, in stg and tg mutants. Similar to Cm/+, SWDs exhibited a variable spike frequency pattern of 3∼6 Hz during the period from P13 to P18 and then stabilized at 5–6 Hz from P20 onward in both tg and stg mice. No SWDs were ever observed in C3H (wild-type littermates of Cm/+) and C57/BL6 (+/+ control for tg and stg) mice within the age range of P11 to P25. Group data showing the mean incidence of cortical discharges in Cm/+ mutants and C3H +/+ control mice during development are demonstrated in Figure 2B. As shown, the onset of synchronous cortical discharges began with rare bursts (1–30 sec duration; some as long as 150–300 sec) in Cm/+ mutant mice at the age of P14 and P15 (two to five bursts per hour) and gradually increased to a stable rate exceeding 20 per hour. The mean incidence of spike-wave discharges in Cm/+ mice at P20 was ∼15 bursts per hour and then reached 25–45 bursts per hour between P21 and P25. Behavioral video image data obtained simultaneously with EEG recordings showed that except for very brief duration discharges, the SWDs were seen when Cm/+ mutants were motionless in a state of quiet wakefulness, similar to the correlation with behavioral arrest in other absence seizure models. The patterns of onset and incidence of spike-wave discharges in Cm/+ mutants are essentially identical to those exhibited in Ca²⁺ channel mutants (Noebels and Sidman, 1979; Noebels, 1984; Noebels et al., 1990). No synchronous discharges were ever recorded in adolescent C3H control mice; however, rare isolated SWDs were noted in recordings from some +/+ C3H aged ≥5 weeks, suggesting that this genetic background may express additional permissive susceptibility genes for this trait.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Onset and mean incidence of spike-wave discharges in Cm/+ mutant mice during development. A, Representative bilateral EEG traces recorded in Cm/+, compared with those in tg and stg mutants at the same age. B, Group data showing the mean incidence (i.e., number of SWDs per hour) of cortical discharges in Cm/+ and C3H control mice at ages ranging from P11 to P25. In C3H (n = 8), no SWDs were ever observed throughout the recording period. For Cm/+ mice, each age point includes mean data from two to seven mice ± SE; ^#p < 0.01; *p < 0.001 versus corresponding values of Cm/+ at P14–17, respectively.

The increased peak current density and prolonged current decay of thalamic LVA calcium currents precede seizure onset

We examined thalamic LVA calcium currents in brain slices obtained at age P13, at least 24 hr and typically several days before the earliest evidence of cortical SWDs in Cm/+ mutants. The data collected at P13 are compared with those obtained at P14–16 when seizures began to emerge. Figure 3A 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 and Cm/+ mice. The current traces of the LVA calcium channels at both P13 (before the onset of SWDs) and P15 (after the emergence of SWDs) display fast activation and inactivation kinetics as well as activation at low voltages, which together define the characteristic properties of T-type Ca²⁺ channels (Destexhe et al., 1998; Lee et al., 1999; Zhang et al., 2000, 2002). The peak current densities of LVA currents at a membrane potential of –50 mV increased by 54% in Cm/+ at P13 and 60% in Cm/+ at P14–16 compared with corresponding values in control +/+ neurons at same age (Fig. 3B). The mean peak current amplitude and peak current density were –1418.0 ± 132.9 (pA) and 12.7 ± 0.5 (pA/pF) in control and –2015.8 ± 117.9* (pA) and 19.5 ± 1.2** (pA/pF) in Cm/+ at age P13, –1142.3 ± 131.9 (pA) and 12.4 ± 1.1 (pA/pF) in control, and –1916.2 ± 54.3** (pA) and 19.8 ± 0.4** (pA/pF) in Cm/+ aged P14–16 (**p < 0.001; *p < 0.01 vs control). The amplitude and peak current density of control and Cm/+ at P13 were not significantly different from corresponding data at P14–16, respectively. Because synchronous discharges in Cm/+ mutant mice are never seen until P14–16 (Fig. 2B), this result demonstrates that thalamic LVA Ca²⁺ currents are elevated before the onset of discharges and therefore are not induced by the seizure itself.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Increased low-voltage-activated Ca²⁺ peak current in Coloboma. A, Representative LVA current traces from thalamocortical relay cells (TCs) in control (C3H) and Cm/+ mice at ages P13 and P15. The cell capacitance values of these neurons were 104 and 105 pF in C3H and Cm/+ at P13, and 100 and 101 pF in C3H and Cm/+ at P15, respectively. The holding potential and voltage protocol used were the same as described in Figure 1 B. Decay of the current was fitted by a single-exponential function (superimposed dotted line). A significant alteration in macroscopic current decay was found in Cm/+. The time constants τ for decay of representative current traces at P13 and P15 are 22.53 and 24.6 msec in C3H and 29.4 and 29.8 msec in Cm/+, respectively. B, Elevated LVA Ca²⁺ current amplitude and peak current density from Cm/+ TC neurons. LVA currents were evoked at the same membrane potential as described in A. **p < 0.001; *p < 0.01 versus control at corresponding age.

The duration of macroscopic inactivation of LVA Ca²⁺ currents in both control and mutant mice is closer to the time scale observed in Ca_v3.1 but not Ca_v3.2 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. 3A); however, unlike what we observed in tg, lh, and stg neurons, we found significant prolongation of the time constant for decay of macroscopic LVA currents in Cm/+ (28.1 ± 0.69** msec; P13–16) compared with that in wild-type neurons at the same age range (23.6 ± 0.86 msec; **p < 0.001). We here combined data derived from mice at P13 and at ages ranging from P14 to P16 in both the control group and the Cm/+ group, because theτ for decay at P13 is not significantly different from that at P14–16 in either control or mutant mice.

Depolarizing shifts in voltage dependence of steady-state inactivation and unaltered kinetics for recovery from inactivation of LVA currents in Cm/+ mice

We next examined the voltage-dependent inactivation and recovery from inactivation of LVA Ca²⁺ currents in both wild-type and Cm/+ neurons. The current traces of SSI of LVA are shown in Figure 4A. 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. 5B). As demonstrated in Figure 4A, LVA currents elicited at –50 mV from different premembrane potentials in both control and Cm/+ mutant show fast inactivation and decay completely within 200 msec. We found a significant depolarizing shift of the steady-state inactivation curves of LVA currents in the mutant both before the onset of seizure (P13) and after seizure generation (P14–16) in contrast to wild-type neurons (Fig. 4B). The mean half-maximal voltages (V_1/2) for SSI curves were –84.0 ± 0.7 mV in control (P13–16), –76.0 ± 2.1* mV in Cm/+ at P13, –75.5 ± 3.1^# mV in Cm/+ at P14–16, respectively (*p < 0.01; ^#p < 0.05 vs control). The 8–8.5 mV depolarizing shifts of the voltage dependence for SSI of LVA currents in TC neurons of Cm/+ mice suggest that at physiological membrane potentials varying from –70 to –75 mV, a higher fraction of all LVA calcium channels is available for opening in the mutant relative to control mice. As seen for the LVA peak current density in Cm/+ (Fig. 4), the V_1/2 for SSI in mutant mice also shifted significantly in a depolarizing direction at P13, a time preceding the onset of SWDs (Fig. 2B). This depolarized shift of V_1/2 for SSI of LVA currents remained after seizure onset (Fig. 4B). Accordingly, in parallel with an increased current density (Fig. 3), the depolarizing shift of V_1/2 for SSI of LVA currents in Cm/+ also preceded the emergence of SWDs. The depolarizing shift of the SSI curve for LVA currents in Cm/+ is consistent with our previous data showing 7.5–13.5 mV depolarizing shifts of SSI for thalamic LVA currents in tg, lh, and stg mutants (Zhang et al., 2002). This shift in the voltage dependence of SSI of LVA currents to a more depolarized level in Cm/+ mice will increase membrane excitability and therefore provide an additional biophysical mechanism that may contribute to neuronal burst synchronization in Cm/+ mutants.

Figure 5A displays raw current traces of the RFI of thalamic LVA Ca²⁺ currents in control and Cm/+ mice. As shown, RFI in both control and mutant cells was complete within 3 sec (Fig. 5B). The recovery from inactivation curve was best fitted with a two-exponential function, and the fast and slow time constants derived from curve fitting did not significantly differ in Cm/+ when compared with control mice, either before or after the onset of SWDs. The values of the fast time constant (τ₁) were 260 and 240 msec for control and Cm/+ at P13 and 250 and 210 msec for control and Cm/+ at P14–16, respectively. The values of slow time constant (τ₂) were 900 and 860 msec for control and Cm/+ at P13 and 820 and 920 msec for control and Cm/+ at P14–16, respectively.

High-voltage-activated Ca²⁺ peak currents and voltage dependence of steady-state activation are unaffected in Cm/+ mutants

Finally, we investigated whether thalamic HVA Ca²⁺ currents are altered by SNAP25 deficiency as seen in Ca²⁺ channel mutants. 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; Meir et al., 2000; Kang et al., 2001). Our previous study demonstrated increased HVA Ca²⁺ currents as well as altered channel kinetics in the tg, stg, and lh absence seizure models (Zhang et al., 2002), and there is evidence that SNAP25 directly interacts with HVA Ca²⁺ channels (Rettig et al., 1996; Sheng et al., 1998).

Figure 6A shows representative HVA Ca²⁺ current traces from wild-type and Cm/+ mutant neurons. The I–V relationships of Ca²⁺ currents for control and Cm/+ mutant are shown in Figure 6B. Both control and mutant HVA Ca²⁺ currents start to activate at around –40 mV and reach a peak between –10 and –15 mV (Fig. 6B). Pooled peak currents and peak current densities are shown in Figure 6C. The mean peak current density was 10.3 ± 1.2 and 11.1 ± 1.4 (pA/pF) in control and Cm/+ mutant neurons at P13 and 10.9 ± 1.2 and 10.9 ± 0.9 (pA/pF) in control and Cm/+ at P14–16, respectively. The mean peak amplitude was –1198.4 ± 219.5 and –1228.6 ± 135.6 (pA) in control and Cm/+ mice at P13 and –1149.9 ± 217.8 and –1122.4 ± 129.6 (pA) in control and Cm/+ mice at P14–16, respectively. These results show that there is no significant alteration in either peak current density or current amplitude in Cm/+ mutants in comparison with those of wild-type mice both before the onset of SWDs and after seizure generation.

Similar to the unaltered peak current in Cm/+, the voltage dependence of SSA for HVA Ca²⁺ currents in mutant TC neurons also did not shift significantly relative to that of control. The SSA curves of HVA currents in both control and mutant are demonstrated in Figure 6D. The mean values of V_1/2 and slope for SSA of HVA channels were –20.1 ± 1.3 mV and 4.3 ± 0.3 in control and –20.4 ± 1.1 mV and 3.8 ± 0.2 in Cm/+ at P13, and –19.1 ± 1.2 mV and 3.7 ± 0.2 in control and –18.6 ± 1.3 mV and 4.0 ± 0.3 in Cm/+ at P14–16. The unchanged peak current density and voltage dependence of SSA differ from what we observed in tg, lh, and stg mutants (Zhang et al., 2002), which demonstrated increased peak current densities in tg and stg as well as a depolarizing shift of SSA curve in lh mice. Thus, the insignificant changes in HVA peak current and voltage dependence of SSA in Cm/+ mice suggest that the Cm/+ mutation does not directly or indirectly affect HVA Ca²⁺ channel gating.

Thalamic Ca²⁺ currents in Cm/+ compared with calcium channel mutations

Potentiation of thalamic LVA Ca²⁺ current precedes the onset of SWDs in Cm/+ mutants

Because potentiation of LVA currents could be activity driven, we examined thalamic LVA currents in Cm/+ and control mice before and after the onset of seizures. At P13, a full day before even the first SWD ever appears, we found a 54% increase in peak current density of LVA Ca²⁺ currents, as well as an 8 mV depolarizing shift of the inactivation curve, demonstrating that potentiation of thalamic LVA currents precedes the emergence of synchronous discharges and therefore was not seizure induced. In fact, even after several days of seizures occurring with an incidence of two to five bursts per hour, we still observed a comparable (60%) elevation in peak current density of LVA currents together with an 8.5 mV depolarizing shift of the voltage dependence of steady-state inactivation that was changed little from pre-seizure levels. Thus, the onset of SWDs did not add an appreciable supplemental enhancement in either LVA peak current or the shift of inactivation curve in Cm/+. These data correspond well with other evidence that the spike-wave seizure pattern, unlike that induced by convulsants (Nahm and Noebels, 1998), does not significantly induce gene expression patterns.

The increased thalamic LVA currents and channel availability near resting membrane potentials in the Cm/+ mutant favor augmented burst firing and membrane hyperexcitability, because T-type Ca²⁺ channels begin to activate at relatively hyperpolarized membrane potentials (Huguenard and Prince, 1992; Zhang et al., 2000) attributable in part to rhythmic input from GABAergic nucleus reticularis thalami neurons. Because SNAP25 deficiency reduces transmitter release in Cm/+ synaptosomal preparations (Raber et al., 1997), other complex changes may contribute to the altered synchronization pattern. Although we have not established a unique causal relationship between the potentiated LVA currents and seizure initiation in Cm/+, the evidence for involvement of thalamic T-type channels in spike-wave generation, as well as the exclusion of activity-dependent causes for LVA current increases in Cm/+, suggests an important primary contribution. Interestingly, a comparable depolarizing shift of SSI with prolongation of LVA current decay has just been described in human mutations identified in human childhood absence epilepsy patients (Khosravani et al., 2004).

Mechanisms for enhanced LVA current

Several possibilities may account for elevated LVA currents in Cm/+ mutants. First, T-type channel synthesis might increase because of impaired synaptic transmission caused by deficient SNAP25 expression or phosphorylation (Boschert et al., 1996; Genoud et al., 1999). Presynaptic release defects may lead to developmental differences in the transcriptional regulation of a heterogeneous population of T-type channel isoforms (Bertolesi et al., 2003; Yunker et al., 2003), which in turn have been shown to influence patterns of neuritogenesis during neuronal differentiation (Chemin et al., 2002). Although a minor increase in expression of Ca_v3.1/α1G and Ca_v3.2/α1H has been reported in adult GAERS (genetic absence epilepsy rats from Strasbourg) rats, a genetically undefined absence model (Talley et al., 2000), no alteration in the thalamic expression pattern of any of the three Ca_v3.1–3 genes that closely resemble the Cm/+ model or evidence for abnormal modulation have been detected (Zhang et al., 2002). Second, the currents may be modulated. Modulation pathways potentially responsible for increased LVA currents in Cm/+ include the modification of T-type channels by protein kinase C (Park et al., 2003), CamKII (Welsby et al., 2003), opioid receptors (Schroeder et al., 1991), pH (Delisle and Satin, 2000; Shan et al., 2001), and anandamide (Chemin et al., 2001). It is worth noting that the Coloboma locus also contains the genes for phospholipase C (PLC) β1 and β4; hence Cm/+ mice are also haploinsufficient for these enzyme isoforms. Mice with homozygous deletion of PLC β1 show lethal tonic–clonic seizures in the second postnatal week, and β4 mice show ataxia without seizures; however, heterozygous mice display no neurological abnormalities (Kim et al., 1997). Whether these heterozygotes show SWDs and altered LVA currents that might be related to PLC-β signal transduction abnormalities remains to be determined.

Overlapping mechanisms for absence epilepsy phenotype in Coloboma and calcium channel mutants

The spike-wave phenotype exhibited in developing Cm/+ mice parallels the early onset (P14–16), increasing incidence (reaching >40 per hour in adulthood), and ETX sensitivity of the SWDs found in tg, lh, and stg (Noebels and Sidman, 1979; Noebels et al., 1990; Hosford et al., 1992). Interestingly, in tg and stg mutants, we also observed a 50–60% increase in peak current densities and an 11–12 mV depolarized shift of inactivation over an age range (P8–11) before SWDs appear (data not shown). Significantly, there is additional functional overlap between epileptogenic calcium channelopathies and the SNAP25-deficient Cm/+ mutants, because SNAP25 interacts with HVA channels involved in exocytosis, and both show transmitter release defects. We did find a difference in pharmacological sensitivity: namely, a 3.6- to 5-fold reduction in ETX sensitivity in Cm/+ mice compared with Ca²⁺ channel mutants. In Cm/+, only a bolus dose of 5 mmol/kg (i.p.) completely blocked the occurrence of SWDs within 2–4 min, and lower doses were ineffective, whereas in Ca²⁺ channel mutants, 1–1.4 mmol/kg is an effective dose (Heller et al., 1983; Aizawa et al., 1997). The blockade of SWDs by ETX could be attributable to the action of the drug on either thalamic T-type Ca²⁺ channels (Coulter et al., 1989) or slow inactivated Na⁺ currents and Ca²⁺-activated K⁺ currents (Leresche et al., 1998). Future experiments will address which currents may be less sensitive to ETX in Cm/+ neurons.

In conclusion, our results provide the first demonstration that potentiated thalamic LVA currents precede abnormal neuronal synchronization and, together with the direct effect of the α1G gene on the threshold for thalamocortical SWD generation (Kim et al., 2001), provide the strongest supportive evidence so far for a major role of thalamic LVA currents in murine models of absence epilepsy. The evidence also points to developmental defects in synaptic transmission as a common cellular mechanism for the dysregulation of T-type currents that lead to this epileptic phenotype.