Downregulation of KCNMB4 expression and changes in BK channel subtype in hippocampal granule neurons following seizure activity

Abstract

A major challenge is to understand maladaptive changes in ion channels that sets neurons on a course towards epilepsy development. Voltage- and calcium-activated K+ (BK) channels contribute to early spike timing in neurons, and studies indicate that the BK channel plays a pathological role in increasing excitability early after a seizure. Here, we have investigated changes in BK channels and their accessory β4 subunit (KCNMB4) in dentate gyrus (DG) granule neurons of the hippocampus, key neurons that regulate excitability of the hippocampus circuit. Two days after pilocarpine-induced seizures, we found that the predominant effect is a downregulation of the β4 accessory subunit mRNA. Consistent with reduced expression, single channel recording and pharmacology indicate a switch in the subtype of channels expressed; from iberiotoxin-resistant, type II BK channels (BK α/β4) that have higher channel open probability and slow gating, to iberiotoxin-sensitive type I channels (BK α alone) with low open probability and faster gating. The switch to a majority of type I channel expression following seizure activity is correlated with a loss of BK channel function on spike threshold while maintaining the channel's contribution to increased early spike frequency. Using heterozygous β4 knockout mice, we find reduced expression is sufficient to increase seizure sensitivity. We conclude that seizure-induced downregulation of KCNMB4 is an activity dependent mechanism that increases the excitability of DG neurons. These novel findings indicate that BK channel subtypes are not only defined by cell-specific expression, but can also be plastic depending on the recent history of neuronal excitability.

Fig 1. Seizure effect on BK channel subunit expression.

(A) RT-PCR of dissected hippocampi shows that seizure activity decreases β4 mRNA expression by 41% ± 0.3. (B) Pilocarpine results in 20% ± 1 less Kcnma1 and kainic acid decrease α-subunit mRNA expression by 49% ± 1. (C) STREX specific primers show a drop in STREX containing mRNA. (D) The relative values of STREX containing and total BK mRNA are unaffected by either seizure model, suggesting that seizure activity does not lead to early regulation of STREX retention in BK channel mRNA. (n = 5 for each group.) (E) Immunoflourescent staining of β-actin (488 nm) and BK α-subunit (555 nm) in dentate gyrus granule cell layer and hilus of hippocampal slices. The wild type sections were taken at a 40X magnification. The knockout sections (bottom row) were taken at a 20X magnification to ensure that there is no staining throughout the section. (F) Average BK α-subunit protein immunofluorescence (normalized to β-actin immunofluorescence) shows no changes 48 hours following pilocarpine. (Control n = 6, Seizure n = 5, and BK α KO n = 2).

Fig 2. Pilocarpine-induced seizures reduce KCNMB4-EGFP reporter expression.

(A) Anti-EGFP (top panels) and anti-MAP2 (middle panels) immunofluorescence staining from heterozygous mice containing the EGFP cDNA that replaces the first coding exon of KCNMB4 gene [10]. The pictures are oriented from left, medial (the dentate gyrus) to right, temporal (hilar mossy fibers extending to CA3 region). Bottom panels are merged of above images. (B) Average fluorescence intensity of EGFP normalized to MAP2 measured at the granule cell area. Control is 4.1 ± 0.20, N = 7, pilocarpine 3.4 ± 0.18, N = 6, P = 0.034.

Fig 3. cfos immunostaining marks neurons with reduced gene expression of KCNMB4.

(A) Anti-EGFP (top panels) and anti-c-Fos (middle panels) immunofluorescence staining from heterozygous mice containing EGFP knockin into the KCNMB4 locus [10]. Left panels are control mice, right panels are pilocarpine treated mice. The pictures are focused on the granule cell layer of the dentate gyrus. Bottom panels are merged images from panels above. (B) Average anti-EGFP immunofluorescence in c-Fos positive cells normalized to EGFP immunofluorescence in c-Fos negative, neighboring cells. Control cells show a reduced EGFP immunofluorescence in c-Fos positive cells (bar value) normalized to c-Fos negative cells in the section (0.91 ± 0.017, N = 12, P < 0.001). Pilocarpine-treated mice show no reduction in EGFP immunofluorescence in c-Fos positive cells relative to c-Fos negative cells. (0.98 ± 0.019, N = 15, P = 0.38).

Fig 4. Control and seizure animals express different ratios of type II and type I BK channels.

(A-B) Type II channels have longer open dwell times and were resistant to IBTX. Type I channels had observably shorter open dwell times and were consistently blocked by IBTX. All channels were blocked by paxilline. (C) β4 KO mice only express Type I channels that are sensitive to IBTX and paxilline. (D) Type I BK channels were more predominant from seizure experienced mice (Control n = 15, Seizure n = 19). (E) Single channel current amplitude was similar in control, pilocarpine treated, and β4 KO mice. Single channel current was an average of 7 ± 1.5 pA at a 0 mV holding potential. (Control n = 8, Seizure n = 7, β4 KO n = 5) (F) The average number of BK channels per membrane patch was similar in control and seizure mice. (Control n = 15, Seizure n = 19). Panels A-E data was acquired at 0 mV potential with 60 μM buffered calcium internal solution.

Fig 5. Seizure activity switches predominant BK channel gating properties from type II to type I.

(A) The average open dwell time of BK channels recorded from pilocarpine treated mice was more similar to β4 KO channels than they were to control recordings. (B) The seizure driven loss of β4 reduces the NP_O of channels at a -40 mV holding potential. (Control n = 8, Seizure n = 7, β4 KO n = 5) (C) Independent of seizure history, the dwell times of IBTX sensitive (and IBTX resistant) BK channels were similar. (Control IBTX sensitive n = 4, Control IBTX resistant n = 6, Seizure IBTX sensitive n = 5, Seizure IBTX resistant n = 4). Recordings were made with 60 μM buffered calcium internal solution.

Fig 6. BK channels shape the action potential afterhyperpolarization.

(Top) Representative traces of 1^st action potential comparing BK channel block (paxilline, grey trace) in control (A) and pilocarpine treated (B) granule cells. (C) Average fAHP amplitude before and after paxilline from control and seizure experienced animals. (D) The relative effect of paxilline on fAHP amplitude. (E) Average 10% width of first action potential before and after paxilline from control and seizure experienced animals. (F) The relative effect of paxilline on the 10% width of the first action potential. (Control n = 9, seizure n = 8).

Fig 7. Seizures do not change BK channel contribution to early spike frequency.

(Top) Representative traces of first 20 ms of voltage response to a 200 pA constant current injection. The voltages traces are aligned at the first spike (pre-spike tracings are not shown) to display effects on instantaneous frequency. Before (black trace) and after BK channel block (paxilline, grey traces) in control (A) and pilocarpine treated (B) granule cells. (C) Average instantaneous firing frequency of 1^st action potential before and after paxilline from control and seizure experienced animals. (D) Average change in instantaneous firing frequency resulting from blocking BK channels in control and pilocarpine treated mice. (Control n = 9, seizure n = 8).

Fig 8. BK channels affect the action potential threshold and delay to first spike.

(A) The first action potential threshold voltage is hyperpolarized following paxilline in control mice (A, left), while there is no effect of blocking BK channels in seizure-experienced neurons (A right). (B) Average of absolute change in threshold potential of first spike in response to paxilline block in control and seizure treated mice. (C) The average time to spike is decreased by paxilline in control and seizure treated animals. (D) The average absolute change in time to spike resulting from BK channel block from control and seizure treated animals. (E, F) Representative voltage traces showing the change in threshold and time to first action potential for control recordings (E) and recordings after seizure (F). (Control n = 9, seizure n = 8).

Fig 9. Reduced β4 expression has consequences on seizure threshold.

(A) EEG recordings of mice monitored for 24 hours for spontaneous EEG activity (top panels) and in response to cumulative increased doses of kainic acid (KA, 10 mg/Kg) at 30 minute intervals (bottom panels). Wild type β4^+/+ do not have spontaneous seizures. One heterozygous β4^-/+ had a spontaneous focal seizure (above) and had generalized seizures (bottom) within 10 minutes of a single dose (10 mg/Kg) of KA. β4^+/+ control mice required a second dose of kainic acid (20 mg/Kg total) to demonstrate seizures (bottom left). Time scale is 1 second per division. The channels recorded were from left temporal (top channel), right frontal (middle channel), and right temporal (bottom channel). B. Average time to 1st seizure and C. total seizure time after KA in β4^+/+ and β4^+/- mice (N = 7 each β4^+/+ and β4^+/-).