Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2012 Mar 7;31(5):1217-30.
doi: 10.1038/emboj.2011.488. Epub 2012 Jan 17.

Raising Cytosolic Cl- In Cerebellar Granule Cells Affects Their Excitability and Vestibulo-Ocular Learning

Affiliations
Free PMC article

Raising Cytosolic Cl- In Cerebellar Granule Cells Affects Their Excitability and Vestibulo-Ocular Learning

Patricia Seja et al. EMBO J. .
Free PMC article

Abstract

Cerebellar cortical throughput involved in motor control comprises granule cells (GCs) and Purkinje cells (PCs), both of which receive inhibitory GABAergic input from interneurons. The GABAergic input to PCs is essential for learning and consolidation of the vestibulo-ocular reflex, but the role of GC excitability remains unclear. We now disrupted the Kcc2 K-Cl cotransporter specifically in either cell type to manipulate their excitability and inhibition by GABA(A)-receptor Cl(-) channels. Although Kcc2 may have a morphogenic role in synapse development, Kcc2 disruption neither changed synapse density nor spine morphology. In both GCs and PCs, disruption of Kcc2, but not Kcc3, increased [Cl(-)](i) roughly two-fold. The reduced Cl(-) gradient nearly abolished GABA-induced hyperpolarization in PCs, but in GCs it merely affected excitability by membrane depolarization. Ablation of Kcc2 from GCs impaired consolidation of long-term phase learning of the vestibulo-ocular reflex, whereas baseline performance, short-term gain-decrease learning and gain consolidation remained intact. These functions, however, were affected by disruption of Kcc2 in PCs. GC excitability plays a previously unknown, but specific role in consolidation of phase learning.

Conflict of interest statement

The authors declare that they have no conflict of interest.

Figures

Figure 1
Figure 1
Cell type-specific deletion of Kcc2 in the cerebellum. (A) Cerebellar neuronal circuitry. Input into the cerebellum is through mossy fibres (MFs) and climbing fibres (CFs). GABAergic Purkinje cells (PCs) provide the only output. Inhibitory interneurons (SCs, stellate cells; BCs, basket cells; GoCs, Golgi cells) generate feedforward inhibition on PCs (1) and granule cells (GCs) (2), as well as feedback inhibition on GCs (3). Cell bodies in which Kcc genes were deleted are not filled and synapses that are expected to be attenuated in our mouse models are highlighted in pink. (B) Targeted allele (Kcc2lox) of the Kcc2 (Slc12a5) locus. Exons 2–5 were flanked by loxP sites (arrowheads). Exons are indicated by black bars. E1, EcoRI restriction site used for genotyping by Southern blotting. (C) Western blots showing that Kcc2lox/lox mice express the Kcc2 protein at WT levels (left) and that the anti-Kcc2 antibody is specific (right; no signal in Kcc2−/− mice). (D) Progression of Kcc2 expression on Purkinje cells from Kcc2lox/lox (control) and PC-ΔKcc2 mice, averaged over all regions in the vermis. PCs showing somatic membrane staining were counted as positive cells (average±standard deviation, n=2, 1, 4, and 2 mice for P17, P21, P25, and P30, respectively (for each genotype); ∼200 cells counted per animal). Also in granule cells of the vermis, Kcc2 deletion was complete at P25. (EL) Immunofluorescent labelling for Kcc2 (green), parvalbumin (red), and nuclei (blue) (IL) on sagittal sections of cerebellar cortex from adult mice (scale bar: 40 μm). (E, I) Kcc2lox/lox mice (control). (F, J) GC-ΔKcc2 mice with specific Kcc2 deletion in granule cells. (G, K) PC-ΔKcc2 mice with specific deletion of Kcc2 in Purkinje cells. (H, L) PC;GC-ΔKcc2 mice with Kcc2 deletion in both PCs and GCs. Kcc2 expression remains in cerebellar interneurons (G, H, arrows). Asterisks in (G, H) indicate PC somata.
Figure 2
Figure 2
Cell type-specific disruption of Kcc3 in Purkinje cells. (A) Targeted allele (Kcc3lox) of the Kcc3 (Slc12a6) locus. Exons 5 and 6 were flanked by loxP sites (arrowheads). Exons are indicated by black bars. S, SpeI restriction sites used for genotyping by Southern blotting. (B) Western blots showing that Kcc3lox/lox mice express the Kcc3 protein at WT levels and the specificity of the Kcc3 antibody. (C) In-situ hybridization for Kcc3 mRNA of sagittal sections of cerebellar cortex from adult control, PC-ΔKcc3, and GC-ΔKcc3 mice. The probe is directed against sequence of exons 5 and 6, which are flanked by loxP sites. Strong hybridization of the antisense probe is observed in Purkinje cells of control and GC-ΔKcc3, but not of PC-Kcc3 mice. The signal in GC-ΔKcc3 mice did not differ from Kcc3lox/lox control. In the granule cell layer, Golgi cells were positive for Kcc3, but no staining could be assigned to granule cells. No staining was observed with the control sense probe on sections from Kcc3lox/lox (control) cerebellum (scale bar: 100 μm).
Figure 3
Figure 3
Inhibitory and excitatory synapses appear unchanged upon cell-specific Kcc2 deletion. (A) Confocal sections of cerebellar cortex from adult mice stained against VGAT (vesicular GABA transporter, green), VGLUT1 or VGLUT2 (vesicular glutamate transporter 1 and 2, green) as marker for inhibitory or excitatory synapses reveal no differences between the genotypes. Co-staining against parvalbumin is shown in red (ML, PCL, and GL, molecular layer, Purkinje cell layer, and granule cell layer, respectively). Scale bars: 40 μm. (B) Biocytin filling of PCs from adult Kcc2lox/lox (control) (above) and PC;GC-ΔKcc2 mice (below) reveals no obvious difference in spine number and morphology. (C) Cumulative distribution of PC dendritic spine lengths from control and PC;GC-ΔKcc2 mice. (D) Electron micrographs of the granular layer (top panels) and molecular layer (bottom panels) of control mice, GC-ΔKcc2, PC-ΔKcc2, and PC;GC-ΔKcc2 mice (from left to right). Arrowheads indicate symmetric synapses. Asterisks indicate synapses in glomeruli (granular layer) and asymmetric synapses in the molecular layer. Scale bar: 500 nm. (E) The density of inhibitory synapses in granular layer glomeruli of GC-ΔKcc2 mice (n=4) was indistinguishable from controls (n=4), as is that in PC-ΔKcc2 (n=4) and PC;GC-ΔKcc2 mice (n=4) (P=0.89, 1.0, and 0.99, respectively; one-way ANOVA). The density of inhibitory and excitatory synapses onto Purkinje cells in GC-ΔKcc2, PC-ΔKcc2, and PC;GC-ΔKcc2 mice neither differed from controls (P=0.97, 0.75, and 0.72, respectively; one-way ANOVA). (F) The density of parallel fiber (PF) synapses onto Purkinje cell spines in control mice was indistinguishable from that in GC-ΔKcc2, PC-ΔKcc2, and PC;GC-ΔKcc2 mice (P=0.26, 0.99, and 0.18, respectively; one-way ANOVA). Similarly, the width of Purkinje cell spine necks did not differ between controls, GC-ΔKcc2, PC-ΔKcc2, and PC;GC-ΔKcc2 mice (P=0.82, 0.61, and 0.09, respectively; one-way ANOVA). Colour codes as indicated on top of panels in (D). Error bars, s.e.m.
Figure 4
Figure 4
Frequency of miniature and spontaneous excitatory and inhibitory postsynaptic currents in Purkinje cells. (A) The frequency of mIPSCs recorded from PCs in the presence of tetrodotoxin is not changed by PC-specific Kcc2 disruption, nor (B) is there a change in the frequency of mEPSCs. The amplitudes of mEPSCs (control: 15.2±1.1 pA, n=11 cells and PC-ΔKcc2: 14.7±0.8 pA, n=14 cells) and mIPSCs (control: 50±5.8 pA, n=19 cells and PC-ΔKcc2: 62±5.3 pA, n=15 cells) were not changed. (C) Spontaneous IPSCs and (D) EPSCs of PCs measured without tetrodotoxin. Each panel shows representative traces (left) and diagrams (right) that display individual data points, with mean values±s.e.m. indicated by solid and dashed lines. Control, Kcc2lox/lox mice (P25–34).
Figure 5
Figure 5
Role of Kcc2 and Kcc3 in setting [Cl]i of Purkinje cells. (A) Effect of the GABAAR agonist muscimol on PC membrane voltage V from control and PC-ΔKcc2 mice. (B) Currents elicited by puff application of muscimol to PCs of Kcc2lox/lox (control) and PC-ΔKcc2 mice held at different potentials (indicated left from traces; series resistance corrected) in the perforated patch configuration. (C) Determination of EGABA. Maximum current after muscimol application plotted against voltage. The intersection of the fitted line with I=0 gives EGABA. (D) Summary of V and EGABA. ‘Floxed' animals served as paired controls (number of cells indicated on bars). (E) Protocol to determine Cl extrusion following intraneuronal Cl loading. A series of voltage ramps (500 ms, from −110 to −40 mV) combined with a 50-ms application of 50 μM muscimol (arrows) was used to determine changes in EGABA during the loading phase (for 60 s, puff application and voltage ramps at 0.2 Hz, each ramp followed by a 400-ms step to 0 mV) and the recovery phase (one muscimol puff/minute). The first voltage ramp (t=−5 s) was not combined with muscimol application. Analysis of the recovery phase started 60 s after the loading phase (t=120 s). (F) Time course of EGABA during and after Cl-loading phase for PC-ΔKcc2 mice (n=7) and control littermates (n=5), and PC-Δ(Kcc2+Kcc3) mice (n=6) and control littermates (n=5). Coloured background displays initial EGABA of each genotype and asterisks significant differences between PC-ΔKcc2 and control mice. Plots display averages±s.e.m. Age of mice: P25–P65.
Figure 6
Figure 6
Kcc2 deletion in granule cells alters resting V, but not the Cl driving force (DFGABA). (A) Cl driving force determined by currents through GABAA receptors in cell-attached recordings. Top: single-channel recordings at different pipette voltages from Kcc2lox/lox (control) and GC-ΔKcc2 GCs. ‘c', closed state. Below: mean single-channel currents as function of V. Linear regression (line) shows that Kcc2 deletion does not change the Cl driving force (arrow). Single-channel conductance and mean open time were similar for both genotypes (18.4±0.3 and 18.3±0.4 pS, and 0.40±0.02 and 0.41±0.02 ms for control and GC-ΔKcc2, respectively). (B) V determined from KV channel currents measured in cell-attached mode measured from other GCs of the same cerebellar region. Typical recordings for a GC-ΔKcc2 (green) and control (black) GC. The intersection of a linear fit to the linear (‘leak') current (from ∼+70 to +20 mV) with recorded outward currents gives V. Below: voltage protocol. (C) V of GCs from GC-ΔKcc2 mice and control littermates in the absence or presence of blockers for GABAARs (picrotoxin and gabazine, 100 μM each), GABAARs+GlyRs (additionally 1 μM strychnine), and Nkcc1 (10 μM bumetanide). Data are mean±s.e.m. Number of experiments: ‘no blockers' (n=22, control; n=39, GC-ΔKcc2), picrotoxin+gabazine (n=12, control; n=19, GC-ΔKcc2), picrotoxin+gabazine+strychnine (n=9, control; n=11, GC-ΔKcc2), and bumetanide (n=26, control; n=27, GC-ΔKcc2). Age of mice: P30–P62.
Figure 7
Figure 7
Spontaneous spiking of Purkinje cells in sagittal cerebellar slices of adult (11–21-week-old) mice. (A) Representative traces (left) and ISI histograms (right) of PCs recorded at room temperature in the absence of inhibitors in loose cell-attached patches from PC-ΔKcc2 mice and control littermates (upper traces) and GC-ΔKcc2 mice and control littermates (lower traces). (B) The coefficient of variation of interspike interval (CV) for PC-ΔKcc2 mice (n=36, average frequency: 21.6±2 Hz), GC-ΔKcc2 mice (n=55, 14.4±1 Hz) and their respective control littermates (n=33, 16.6±2 Hz and n=50, 12.7±1 Hz). The plot shows averages (black horizontal lines)±s.e.m. (Boxes) (ISI CV=0.159±0.015 and 0.097±0.012 for control and PC-ΔKcc2 mice, respectively; P<0.05, t-test) (ISI CV=0.096±0.007 and 0.145±0.009 for control and GC-ΔKcc2 mice, respectively; P<0.0001). Individual measurements are shown as circles. Most likely differences in genetic background underlie the difference between control ISI CVs.
Figure 8
Figure 8
Motor performance (AC), learning and consolidation (DG) in Kcc2 mutant mice. (AC) Compensatory eye movements in adult control (n=8), GC-ΔKcc2 (n=8), PC-ΔKcc2 (n=8), and PC;GC-ΔKcc2 (n=10) mice. Mice were subjected to visual (OKR) and/or vestibular stimulation (in dark, VOR; in light, VVOR) and gain (ratio of eye to stimulus velocity) and phase (difference in degrees between eye and stimulus) were calculated. (A) OKR gain was not affected in GC-ΔKcc2 mice (P=1.00), but was significantly lower in PC-ΔKcc2 and PC;GC-ΔKcc2 mice than in controls (P=0.015 and 0.035, respectively). (B) VOR gain values did not differ between Kcc2 conditional KO and control mice (all P>0.9). (C) VVOR gain values did not differ among Kcc2 mutants and control mice (P=0.716, 0.369, and 0.074 for GC-ΔKcc2, PC-ΔKcc2, and PC;GC-ΔKcc2 mice versus controls). PC;GC-ΔKcc2 mice revealed a significant phase lag in OKR (P=0.002), VOR (P=0.006), and VVOR (P=0.001), whereas the other mutants did not (all P>0.14 and >0.24 for GC-ΔKcc2 and PC-ΔKcc2 mice, respectively). (D) Short-term VOR gain-decrease learning was not impaired in GC-ΔKcc2 (n=10) compared with control (n=9) mice (P=0.74; repeated measures ANOVA), whereas PC-ΔKcc2 (n=10) and PC;GC-ΔKcc2 (n=10) mice showed a significant impairment (P<0.005 and 0.001, respectively). (E) Likewise, gain consolidation (calculated as 100%·(ac)/(ab)) overnight was not affected in GC-ΔKcc2 mice (P=0.99 versus controls; one-way ANOVA), whereas PC-ΔKcc2 and PC;GC-ΔKcc2 showed a significant deficit in gain consolidation (P=0.030 and 0.025, respectively). (F, G) VOR phase reversal training induced a significant phase change in GC-ΔKcc2 mice during each consecutive day (P<0.005 for days 2–5, paired Student's t-test). However, from day 4 onwards their VOR phase reversal was impaired compared with controls (P=0.015 and <0.001 on days 4 and 5). This difference is the result of the inability of GC-ΔKcc2 mice to consolidate their phase learning overnight (right panel in (E); P=0.007; Student's t-test). Due to their deficits in VOR gain learning and consolidation, PC-ΔKcc2 and PC;GC-ΔKcc2 mice were also unable to reverse their VOR phase (P<0.001 on day 5 for both mutants versus controls). Error bars denote s.e.m., all P-values are based on repeated measures ANOVA unless stated otherwise.

Similar articles

See all similar articles

Cited by 36 articles

See all "Cited by" articles

Publication types

LinkOut - more resources

Feedback