Impact of the leaner P/Q-type Ca²⁺ channel mutation on excitatory synaptic transmission in cerebellar Purkinje cells

Abstract

Loss-of-function mutations in the gene encoding P/Q-type Ca²⁺ channels cause cerebellar ataxia in mice and humans, but the underlying mechanism(s) are unknown. These Ca²⁺ channels play important roles in regulating both synaptic transmission and intrinsic membrane properties, and defects in either could contribute to ataxia. Our previous work described changes in intrinsic properties and excitability of cerebellar Purkinje cells (PCs) resulting from the leaner mutation, which is known to reduce whole-cell Ca²⁺ currents in PCs and cause severe ataxia. Here we describe the impact of this mutation on excitatory synaptic transmission from parallel and climbing fibres (PFs, CFs) to PCs in acute cerebellar slices. We found that in leaner PCs, PF-evoked excitatory postsynaptic currents (PF-EPSCs) are ˜50% smaller, and CF-evoked EPSCs are ˜80% larger, than in wild-type (WT) mice. To investigate whether reduced presynaptic Ca²⁺ entry plays a role in attenuating PF-EPSCs in leaner mice, we examined paired-pulse facilitation (PPF). We found that PPF is enhanced in leaner, suggesting that reduced presynaptic Ca²⁺ entry reduces neurotransmitter release at these synapses. Short-term plasticity was unchanged at CF–PC synapses, suggesting that CF-EPSCs are larger in leaner PCs because of increased synapse number or postsynaptic sensitivity, rather than enhanced presynaptic Ca²⁺ entry. To investigate the functional impact of the observed EPSC changes, we also compared excitatory postsynaptic potentials (EPSPs) elicited by PF and CF stimulation in WT and leaner PCs. Importantly, we found that despite pronounced changes in PF- and CF-EPSCs, evoked EPSPs in leaner mice are very similar to those observed in WT animals. These results suggest that changes in synaptic currents and intrinsic properties of PCs produced by the leaner mutation together maintain PC responsiveness to excitatory synaptic inputs. They also implicate other consequences of the leaner mutation as causes of abnormal cerebellar motor control in mutant mice.

Mutations in the gene (Cacna1a) encoding the pore-forming subunit of P/Q-type Ca²⁺ channels (Ca_v2.1) cause inherited defects in motor control in mice and humans (Ophoff et al. 1996; Fletcher et al. 1996; Pietrobon, 2002). While these channels are expressed throughout the central nervous system (Hillman et al. 1991; Stea et al. 1994; Volsen et al. 1995; Westenbroek et al. 1995), ataxias resulting from Cacna1a mutations suggest a primary defect in cerebellar motor control (Sidman et al. 1965; Victor & Ropper, 2001; Zwingman et al. 2001; Stahl, 2004; Pietrobon, 2005; Katoh et al. 2007). Despite this selective vulnerability, the mechanistic link between P/Q-type Ca²⁺ channel mutations and cerebellar dysfunction is unknown.

Mice expressing mutations in the Cacna1a gene provide useful models for investigating how P/Q-type Ca²⁺ channels participate in normal cerebellar motor control, and how defects in channel function lead to ataxia. There are several mouse mutants that display different degrees of ataxia, including (from least to most severe): rocker (rkr,Zwingman et al. 2001), tottering (tg, Green & Sidman, 1962), rolling Nagoya (tg^rol, Oda, 1973) and leaner (tg^la, Meier & MacPike, 1971) (for reviews see Pietrobon, 2002; Liu et al. 2003). Voltage-clamp studies of dissociated cerebellar Purkinje cells (PCs) have shown that these mutations lead to reduced whole-cell Ca²⁺ current to an extent that roughly parallels the severity of ataxia, with little or no change in channel expression (in leaner, Lau et al. 1998) or in the voltage dependence of channel gating (Dove et al. 1998; Wakamori et al. 1998; Lorenzon et al. 1998; Mori et al. 2000; Kodama et al. 2006). This raises the possibility that reduced voltage-sensitive Ca²⁺ entry through P/Q-type Ca²⁺ channels is an important factor linking channel mutations to disruptions of cerebellar motor control.

Ca²⁺ entry through P/Q-type Ca²⁺ channels is important in regulating multiple neuronal functions, each representing a potential site at which channel mutations could interfere with cerebellar function. Perhaps the most well-known function is regulation of presynaptic Ca²⁺ entry and neurotransmitter release; hence it has been proposed that P/Q channelopathies are primarily diseases of synaptic transmission (Pietrobon, 2005). Consistent with this idea, excitatory postsynaptic currents in PCs from tg^rol mutant mice elicited by parallel fibre stimulation are smaller than those measured in WT cells and show enhanced paired-pulse facilitation (Matsushita et al. 2002). However, P/Q-type Ca²⁺ channels also play important postsynaptic roles, including regulation of intrinsic membrane properties that control excitability. A good example is provided by PCs, which rely on these channels for regulating > 90% of their voltage-sensitive Ca²⁺ entry (Mintz et al. 1992). In these cells, Ca²⁺ entry through P/Q channels is critical for generating dendritic Ca²⁺ spikes and for controlling the activity of Ca²⁺-dependent K⁺ channels that regulate repetitive spike firing (Llinás & Sugimori, 1980a, b; Tank et al. 1988; Womack et al. 2004; Walter et al. 2006). Previous work has shown that the leaner mutation leads to a ˜60% reduction in whole-cell Ca²⁺ current in PCs (Dove et al. 1998; Wakamori et al. 1998; Lorenzon et al. 1998); correspondingly, leaner PCs fail to generate dendritic Ca²⁺ spikes and are unable to sustain repetitive Na⁺ spike firing during steady depolarizing current injection (Ovsepian & Friel, 2008). In addition, loss-of-function P/Q Ca²⁺ channel mutations (leaner, ducky) reduce the regularity of spontaneous spike firing in PCs in a manner that is consistent with diminished Ca²⁺ entry and SK channel activity (Walter et al. 2006).

The observations above suggest that P/Q-type Ca²⁺ channel mutations may have multiple effects that are relevant to cerebellar function, including changes in synaptic currents as well as changes in intrinsic electrical properties that determine how these currents influence somatic membrane potential and spike firing. How might such effects jointly influence synaptic integration by PCs? In principle, changes in synaptic currents and intrinsic membrane properties that result from a channel mutation could produce complementary or antagonistic effects on evoked postsynaptic potentials. For example, a mutation that decreases the amplitude of a synaptic current could also decrease postsynaptic sensitivity to this current (e.g. by lowering input resistance or increasing dendritic filtering in the postsynaptic cell), which would exaggerate the effect of the mutation on synaptic responses. On the other hand, an increase in sensitivity (e.g. caused by increased input resistance or reduced dendritic filtering) would tend to be compensatory. Similarly, if a mutation increases the amplitude of a synaptic current, changes in intrinsic properties of the postsynaptic cell could either increase or decrease its impact on evoked postsynaptic potentials. Thus, to evaluate how a Ca²⁺ channel mutation affects synaptic information transfer, it is necessary to consider both the effects of the mutation on synaptic currents and on the postsynaptic voltage responses they elicit.

We previously reported that leaner PCs display several differences in intrinsic properties compared to WT cells. For example, they have smaller dendrites than age-matched WT controls, without showing detectable changes in somatic size (Ovsepian & Friel, 2008). Such a modification would be expected to alter somatic responses to a given synaptic input (Roth & Hausser, 2001) by changing the intrinsic electrical properties of PCs and by altering the electrical distance between distal synaptic inputs and the soma. Leaner PCs also have a greatly reduced ability to generate dendritic Ca²⁺ spikes, which would be expected to alter dendritic voltage responses elicited by large synaptic currents. Understanding the overall effect of such changes on synaptic signal processing by PCs, however, requires information about the postsynaptic currents that are elicited by presynaptic inputs, which may also be affected by the mutation. Despite this, there have been no reports describing the effect of the leaner mutation on synaptic currents in PCs, or on the postsynaptic potentials they elicit.

In the present study, we investigated the impact of the leaner mutation on excitatory synaptic currents elicited by PF and CF stimulation under voltage clamp, as well as on the postsynaptic potentials they produce under current clamp. It was reasoned that since the leaner mutation produces the most pronounced reduction in macroscopic Ca²⁺ currents and the most severe ataxia of the mouse Cacna1a mutants, defects in synaptic information transfer should be at their worst in this mutant and therefore most readily detected. We set out to answer the following questions. How does this mutation affect excitatory postsynaptic currents in PCs evoked by PF and CF stimulation? Are changes in EPSCs consistent with changes in presynaptic Ca²⁺ entry? Do changes in EPSC properties in leaner PCs lead to parallel changes in evoked postsynaptic potentials? The main finding was that despite causing pronounced changes in synaptic currents, the leaner mutation has little effect on evoked synaptic potentials. These results suggest that changes in excitatory synaptic currents and PC intrinsic properties caused by the mutation together maintain the ability of fast excitatory inputs to elicit postsynaptic voltage responses in leaner PCs.

Methods

Animals

C57BL/6J wild-type mice (++/++) and mice that were heterozygous for the leaner mutation of the Cacna1a gene (Os+/+tg^la) were obtained from the Jackson Laboratory (Bar Harbor, ME, USA) and maintained in the Animal Resource Center at Case Western Reserve University in accordance with guidelines provided by our Institutional Animal Care and Use Committee. All animals were given food and water ad libitum and maintained at 22 ± 2°C with a 12 h/12 h light/dark cycle. Mice were maintained and propagated by brother–sister mating as previously described (Herrup & Wilczynski, 1982). Os is a tightly linked, semidominant mutation carried in repulsion with tg^la whose expression leads to abnormal paw structure. Since (Os/Os) is embryonic lethal, the only offspring that survive are leaner homozygotes (+tg^la/+tg^la) and heterozygotes (Os+/+tg^la). Leaner homozygous mice (which will be referred to simply as leaner mice) were distinguished from their heterozygous littermates based on their normal paw structure and pronounced ataxia. C57BL/6J wild-type pups (without either the tg^la or Os mutation) were used as wild-type controls.

Preparation of cerebellar slices

Mice from postnatal days 17–21 (P17–21) were killed by decapitation after being deeply anaesthetized (ketamine, 150 mg kg⁻¹ i.p.). The cerebellum was removed and placed in ice-cold sucrose-based high-Mg²⁺, low-Ca²⁺ artificial cerebrospinal fluid (ACSF) containing (in mm): 75 sucrose, 85 NaCl, 2.5 KCl, 1.25 NaH₂PO₄, 25 NaHCO₃, 0.5 CaCl₂, 4 MgCl₂, 25 glucose, bubbled with 95% O₂, 5% CO₂ (Aghajanian & Rasmussen, 1989). Parasagittal slices (300 μm) were cut using a Vibratome, Series 1000 (Vibratome, St Louis, MO, USA) and immediately transferred to a holding chamber containing a Na⁺ -based ACSF with the same composition as the slicing solution except that sucrose was omitted and the NaCl concentration was increased to 125 mm. This solution was maintained at 32°C and bubbled continuously with 95% O₂, 5% CO₂. After incubation for 30 min, slices were transferred to normal ACSF containing 2 mm CaCl₂ and 2 mm MgCl₂ that was maintained at room temperature (23–25°C) and bubbled continuously (95% O₂, 5% CO₂). For recordings, individual slices were transferred to a chamber that was continuously superfused with pre-warmed, bubbled (95% O₂, 5% CO₂) ACSF at a rate of 3–4 ml min⁻¹. Recordings were made from slices for up to 8 h after preparation.

EPSC measurements under voltage clamp

Whole-cell patch-clamp recordings were made from visually identified lobule V Purkinje neurons under IR-DIC at 33–35°C using a ×40 water-immersion objective on an upright microscope (Leica DMLFSA). Currents were recorded under voltage clamp using an Axopatch 200B (Molecular Devices, Sunnyvale, CA, USA). Patch electrodes (2.5–3.5 MΩ) were made from borosilicate glass capillaries (World Precision Instruments (WPI), Sarasota, FL, USA) on a Flaming/Brown puller (P-97, Sutter Instruments, Novato, CA, USA). Electrodes were filled with a solution containing (in mm): 100 caesium gluconate, 30 caesium chloride, 4 NaCl, 5 QX-314, 10 Hepes, 10 EGTA, 4 Mg-ATP, 0.4 Na₃-GTP, with pH adjusted to 7.35 with CsOH. After establishing the whole-cell configuration, the membrane potential was held at −70 mV. Series resistance (R_s) was estimated based on the integral and decay time constant of capacity transients following small (5 mV) voltage steps and was compensated by 65–70%. During the experiments, input resistance was periodically measured based on the steady-state currents evoked by small (5 mV) hyperpolarizing voltage steps. Cells were rejected if input resistance changed by over 13 MΩ over the course of the recording. Currents were filtered at 5–10 kHz. To maintain voltage control during measurement of climbing fibre EPSCs (CF-EPSCs), NBQX (0.7 μm) was included in the external solution to reduce EPSC size, and QX-314 was included in the internal solution to block fast voltage-gated Na⁺ channels.

Excitatory synaptic currents were evoked by 200 μs current pulses generated with a stimulus isolator (A-360, WPI) under computer control and delivered with a bipolar electrode (tip diameter ˜10 μm) made from theta-tubing (TST150-6, 1.5 mm o.d., WPI) filled with normal ACSF. Electrical stimuli were delivered in the molecular layer above and ˜100 μm away from the recorded Purkinje neuron to activate parallel fibre (PF) inputs, or in the granule cell layer below and ˜50 μm away from the recorded Purkinje neuron for stimulating climbing fibres (CFs). To determine the position of the stimulating electrode, the electrode was initially placed within the visual field and was repositioned until synaptic responses were evoked with minimum stimulus intensity. Stimuli were delivered at low frequency (0.1–0.5 Hz for stimulation of PFs, 0.1 Hz for stimulating CFs) except to examine short-term plasticity. Bicuculline (20 μm) or SR 95531 (10 mm) was included in the perfusion solution to block fast inhibitory synaptic transmission (Konnerth et al. 1990; Perkel et al. 1990). Voltages that are presented have not been corrected for liquid junction potentials.

EPSP measurements under current clamp

To evaluate the impact of altered EPSCs in leaner mice on excitatory postsynaptic potentials (EPSPs), postsynaptic responses were elicited by stimulation of the respective pathways under current clamp, using the same methods for stimulation that were employed for measuring EPSCs under voltage clamp. These measurements were made using an Axoclamp 2B in bridge mode, with pipettes (3.0–4.5 MΩ) filled with (in mm) 128 potassium gluconate, 8 NaCl, 2 MgCl₂, 10 Hepes, 4 sodium ATP, 0.4 sodium GTP, 14 Tris-creatine phosphate, pH adjusted to 7.35 with KOH. PCs in acute cerebellar slices are spontaneously active (Llinás & Sugimori, 1980a; Tank et al. 1988; Häusser & Clark, 1997; Womack & Khodakhah, 2002; Edgerton & Reinhart, 2003; McKay & Turner, 2005) so we compared evoked synaptic potentials at hyperpolarized potentials in the absence of spontaneous spike generation. This was achieved by stimulating synaptic inputs during injection of steady hyperpolarizing current, or while cells occupied the hyperpolarized ‘silent’ state during trimodal activity (Womack & Khodakhah, 2002). Bridge balance was maintained throughout the experiments using responses to small current pulses to provide estimates of R_s.

Data acquisition and analysis

Signals were filtered at 5–10 kHz, sampled at 20–50 kHz using an ITC-16 (Instrutech, Port Washington, NY, USA) and saved on a laboratory computer using software written in IgorPro (Wavemetrics, Lake Oswego, OR, USA) and data acquisition XOPs provided by Instrutech. All population values are expressed as means ±s.e.m., with the number of animals and cells from which measurements were made indicated by N and n, respectively. Statistical comparisons between normal and mutant mice were made using Student's unpaired t test. In illustrations of evoked responses, stimulus artifacts have been removed for clarity.

To characterize PF-EPSPs, we measured EPSP amplitude, maximal slope during the rising phase, and the exponential time constant of the recovery. Measurements of these parameters were made from multiple responses in each cell, which were then averaged to describe responses from the respective cells. These measurements were then averaged over WT and leaner cells to give values describing the corresponding cell populations. To describe single action potentials evoked by the second of two PF-EPSPs, we measured spike amplitude, maximal slope and threshold. Threshold was defined as the voltage where dV/dt exceeded 5 V s⁻¹, and spike amplitude was determined as the difference between the peak and threshold voltages. To analyse complex spikes (CSs) elicited by CF stimulation, we measured the amplitude of the first spike, its latency, the number of spikelets, and the plateau voltage level after the last spikelet. The latency to first spike was taken as the time difference between the end of the stimulus and the time of the peak response in cases where the CS was not preceded by an antidromic spike. The number of spikelets was determined as the number of zero crossings of the first derivative of the voltage where the second derivative was negative. The plateau level during the repolarizing phase of the CS was defined as the voltage where the magnitude of the first derivative falls below 0.5 V s⁻¹ after the last spikelet. In cases where antidromic spikes were observed, antidromic spike amplitude and latency were measured as with the first spike within the CS.

Chemicals

2,3-Dihydroxy-6-nitro-7-sulphonyl-benzo[f]quinoxaline disodium salt (NBQX) and (+/–)-2-amino-5-phosphonopentanoic acid (APV) were from Tocris. All other chemicals were purchased from Sigma.

Results

In the first part of this study, we compare excitatory postsynaptic currents elicited in PCs from WT and leaner mice under voltage-clamp conditions by stimulation of parallel or climbing fibre inputs, the two excitatory synaptic inputs received by these cells (see Fig. 1). To investigate whether observed differences in synaptic currents in leaner mice are due to changes in presynaptic Ca²⁺ entry, we next evaluate changes in short-term plasticity at PF–PC and CF–PC synapses. Finally, we ask if the observed changes in synaptic currents lead to parallel changes in synaptic potentials by measuring both evoked EPSPs and suprathreshold responses under current-clamp conditions.

Figure 1. Circuitry of the cerebellar cortex illustrating excitatory inputs to Purkinje cells.

Two pathways provide excitatory input to Purkinje cells, parallel fibres and climbing fibres. These inputs arise from granule cells and cells in the inferior olive (IO), respectively. Purkinje cells integrate these excitatory inputs and send projections to target cells in the deep cerebellar nuclei (DCN). Inputs from intrinsic inhibitory neurons (which were pharmacologically inhibited in this study) are omitted for clarity. Parallel fibres were stimulated with electrodes positioned near the surface of the molecular layer, and climbing fibres were stimulated with electrodes placed in the granule cell layer (see diagrams in Figs 2 and 3).

Impact of the leaner mutation on synaptic transmission from parallel fibres to PCs

When WT Purkinje neurons were held at −70 mV under voltage clamp, brief (200 μs) field stimuli delivered in the molecular layer elicited transient inward currents that could be blocked by NBQX (10 μm) + APV (100 μm) (Fig. 2A, top left), indicating that they are carried by ionotropic glutamate receptors. Synaptic currents were identified as PF-induced EPSCs (PF-EPSCs) based on two criteria (Konnerth et al. 1990; Perkel et al. 1990; Matsushita et al. 2002): (a) EPSC amplitudes increase in a graded manner with stimulus intensity (Fig. 2A, top left) and (b) the second of two stimuli elicits a synaptic current that is larger than the first, indicating paired-pulse facilitation (PPF, see below). PF-EPSCs measured in WT PCs had a time to peak (T_peak) of 4.4 ± 0.1 ms and decayed exponentially with time constant τ_decay = 10.7 ± 0.4 ms (n = 30, N = 6), showing little or no change in time course as the stimulus intensity was increased. This can be seen by superimposing EPSCs elicited from a given cell by stimuli of different intensities (15–40 μA) after normalizing to the same amplitude (Fig. 2A, top right).

Figure 2. Leaner Purkinje cells (PCs) show smaller parallel fibre (PF-) EPSCs than WT cells.

A, top, left: EPSCs elicited in a representative WT PC by PF stimuli of increasing intensity (15, 20, 25, 30, 35, 40 μA), and by a 40 μA stimulus after treatment with NBQX (10 μm) + APV (100 μm), which elicited no detectable synaptic current (top trace). Top, right. EPSCs from left panel after normalizing to the same amplitude, indicating similar EPSC kinetics. Bottom: diagram illustrating sites of stimulation (Stim) and recording (I_m). B, input–output relations for WT (◯, n = 30, N = 6) and leaner PF-EPSCs (•, n = 8, N = 3). Inset compares representative PF-EPSCs elicited by 40 μA stimuli in WT (left) and leaner (right) PCs. Holding potential: −70 mV. * P < 0.05, ** P < 0.01. Arrowheads indicate time of stimulation. Stimulus artifacts have been removed for clarity.

We found that PF-EPSCs in leaner mice were smaller than those observed in WT cells, having amplitudes that were approximately one-half as large as those elicited in WT PCs with the same stimulus intensity (Fig. 2B). However, despite the difference in amplitude, the kinetic properties of leaner PF-EPSCs were not detectably different from WT responses (T_peak = 4.9 ± 0.2 ms, τ_decay = 9.8 ± 0.4 ms, n = 8, N = 3, (not significant (n.s.) compared to WT).

Impact of the leaner mutation on synaptic transmission from climbing fibres to PCs

We also tested for differences in transmission via the climbing fibre pathway, the other pathway that provides excitatory input to cerebellar Purkinje cells. This pathway is composed of axons arising from cells in the inferior olive (Fig. 1); although these cells are not present in parasagittal cerebellar slices, their axons (climbing fibres, CFs) remain and can be stimulated in the granule cell layer (Llinás & Sugimori, 1980a). To investigate how the leaner mutation affects transmission between CF inputs and PCs, CF-EPSCs were recorded after stimulation of ascending CF axons in the granule cell layer (Fig. 3A). As with PF-EPSCs, CF-EPSCs could be blocked by exposure to NBQX (10 μm) + APV (100 μm), indicating they are carried by ionotropic glutamate receptors (data not shown). CF-EPSCs were distinguished from PF-EPSCs based on (a) their all-or-none response characteristics as the stimulus intensity is increased (Fig. 3B), and (b) the form of short-term plasticity they display, paired-pulse depression (PPD, see below) (Eccles et al. 1966; Konnerth et al. 1990; Perkel et al. 1990; Matsushita et al. 2002). For CF-EPSC measurements, extracellular solutions were supplemented with NBQX (0.7 μm) to reduce the size of the evoked currents and facilitate voltage control (see Methods). Under these conditions, suprathreshold CF stimulation elicited transient inward currents in WT PCs with mean amplitude 1.5 ± 0.1 nA, time to peak (T_peak) 3.4 ± 0.2 ms and decay time constant (τ_decay) 8.5 ± 0.4 ms (n = 17, N = 8; Fig. 3A, top left).

Figure 3. Leaner PCs show larger climbing fibre (CF-) EPSCs with faster decay kinetics than WT cells.

A, top: comparison between representative EPSCs elicited by suprathreshold CF stimulation in WT (left) and leaner (right) PCs, illustrating larger amplitude and faster decay kinetics in leaner. EPSC decays could be described by a single exponential function in each cell type (see smooth curves). Bottom: diagram illustrating sites of stimulation (Stim) and recording (I_m). B, left: input–output relations from four WT and four leaner PCs, illustrating all-or-none properties of CF-EPSCs and differences in EPSC amplitudes. Right: comparison between CF-EPSC amplitude distributions in WT (open bars) and leaner (grey bars) PCs. Symbols at right show mean CF-EPSC amplitudes ± s.e.m. (** P < 0.01). Holding potential: −70 mV.

In contrast to the smaller PF-EPSCs observed in leaner mice, CF-EPSCs were typically larger than those observed in WT PCs, with mean amplitude 2.7 ± 0.3 nA (n = 11, N = 4), nearly twice as large as that observed in WT PCs (P < 0.01, Fig. 3A and B). CF-EPSC amplitudes were also more broadly distributed in leaner PCs than in WT cells. This is illustrated in Fig. 3B by representative input–output curves from leaner and WT PCs (left) and the distribution of CF-EPSC amplitudes in these two cell populations (right). While there was overlap between the two distributions, most leaner PCs showed larger responses than WT cells (Fig. 3B, right). The larger CF-EPSCs observed in leaner PCs cannot be readily explained by an increase in the number of climbing fibres that innervate a given PC, since leaner PCs that displayed large EPSCs also exhibited all-or-none activation as in WT cells (Fig. 3B, left). The enhancement of CF-EPCSs in leaner PCs is similar to, but more pronounced than, that observed in tg^rol mutant mice (Matsushita et al. 2002). However, in contrast to the effect of the tg^rol mutation, which causes a slowing of CF-EPSC decays, we found that the leaner mutation accelerates the decay (τ_decay = 6.8 ± 0.4 ms, n = 11, N = 4, P < 0.01 compared to WT), without producing a detectable change in time to peak (T_peak = 3.5 ± 0.2 ms, n.s. compared to WT).

Altered short-term plasticity at PF–PC synapses

Given that P/Q-type Ca²⁺ channels represent a major pathway for presynaptic Ca²⁺ entry at PF–PC synapses in WT mice (Mintz et al. 1995; Matsushita et al. 2002), one possible explanation for the observed reduction in synaptic strength at these synapses in leaner mice is reduced presynaptic Ca²⁺ entry. However, it has been shown that granule cells undergo degeneration in leaner cerebellum starting at ˜P10 (Herrup & Wilczynski, 1982; Lau et al. 2004), raising the possibility that a reduction in the number of parallel fibres is responsible for the reduction in evoked EPSCs.

To investigate whether reduced presynaptic Ca²⁺ plays a role in attenuating PF-EPSCs in leaner mice, we examined short-term plasticity. In WT mice, transmission at PF–PC synapses displays paired-pulse facilitation (PPF). For example, if a second PF stimulus is delivered within ˜300 ms of an initial stimulus, it elicits an EPSC that is larger than the first (Fig. 4A). PPF is thought to result from incomplete Ca²⁺ clearance from presynaptic terminals after the first stimulus, leading to larger [Ca²⁺] elevations produced by the second stimulus and increased neurotransmitter release (Zucker & Regehr, 2002). If PF-EPSCs are smaller in leaner cerebellar slices because of reduced presynaptic Ca²⁺ entry, then a change in facilitation would be expected; if the only effect of the mutation was a reduction in the number of parallel fibre inputs, no change would be predicted. Facilitation can be quantified using the paired-pulse ratio, here defined as the ratio (A₂/A₁) of the amplitudes of the second to the first EPSC elicited by two successive stimuli. We found that PPF is significantly stronger in leaner PCs than in WT cells. For example, in leaner PCs tested with an interpulse interval (IPI) of 30 ms (Fig. 4A), the second of two evoked EPSCs was on average 3.5-fold larger than the first, as compared to a 2.7-fold enhancement in WT cells (Fig. 4B, see vertical dashed line). PPF did not depend on stimulus intensity (not shown) and decayed biexponentially as the interpulse interval was increased (Fig. 4B) without a detectable difference in recovery kinetics between leaner and WT PCs (Fig. 4B, inset).

Figure 4.

Paired-pulse facilitation is enhanced at PF–PC synapses in leaner mice A, comparison between PF-EPSCs elicited by two 35 μA stimuli separated by 30 ms in representative WT (left) and leaner (right) PCs. This pathway displayed enhanced paired-pulse facilitation (PPF) in leaner, quantified as the ratio of the amplitudes of the second to the first EPSCs (A₂/A₁). B, time course of recovery from PPF. Recoveries could be described by biexponential functions with limiting values of unity (see smooth curves). Dashed vertical line indicates the interpulse interval (30 ms) separating stimuli in A. Inset compares recovery kinetics after normalizing A₂/A₁ measurements to the values at the shortest interpulse interval (20 ms) and subtracting unity. Results in B are based on analyses of 26 cells in 6 WT mice and 8 cells in 3 leaner mice. Holding potential: −70 mV. * P < 0.025.

These results suggest that reduced presynaptic Ca²⁺ entry causes diminished neurotransmitter release at PF–PC synapses in leaner PCs, although additional factors, such as a reduction in the number of parallel fibres that contribute to PF-EPSCs, or reduced postsynaptic sensitivity to glutamate, may also play a role in lowering the strength of synaptic transmission at these synapses.

Normal short-term plasticity at CF–PC synapses

Results presented in the last section suggest that one factor underlying attenuation of transmission strength at PF–PC synapses is a reduction in presynaptic Ca²⁺ entry. This would be consistent with previously described effects of the leaner mutation on whole-cell Ca²⁺ currents in PCs (see Introduction). However, reduced Ca²⁺ entry does not readily explain the increase in CF-EPSC amplitude observed in leaner PCs. This points to other mechanistic effects of the leaner mutation on CF–PC synapses. Previous work has shown that non-P/Q-type Ca²⁺ channels account for a larger fraction of presynaptic Ca²⁺ entry at CF–PC synapses than at PF–PC synapses (Regehr & Mintz, 1994; Matsushita et al. 2002), which could render transmission at CF–PC synapses less sensitive to reduced Ca²⁺ entry via P/Q-type Ca²⁺ channels. However, this would not, by itself, lead to an increase in CF-EPSC amplitudes. One possibility is that overcompensation by other Ca²⁺ channels increases overall presynaptic Ca²⁺ entry, leading to enhanced vesicular release and larger CF-EPSCs.

Measurement of paired-pulse ratios provides a means to test this possibility. Previous work has shown that CF–PC synapses display paired-pulse depression (PPD), in which the second of two CF stimuli elicits an EPSC that is smaller than the first (Eccles et al. 1966; Konnerth et al. 1990; Perkel et al. 1990). It is thought that CF–PC synapses display depression instead of facilitation because the quantal content is large (Wadiche & Jahr, 2001), resulting in appreciable vesicle depletion following the first stimulus that limits release in response to the second stimulus (Hashimoto & Kano, 1998; Dittman & Regehr, 1998; Zucker & Regehr, 2002).

To investigate whether a presynaptic mechanism can account for increased CF-EPSC amplitudes in leaner mice, we again measured the paired-pulse ratio (A₂/A₁) using responses to paired stimuli (Fig. 5). If increases in CF-EPSCs are caused by an increase in the number of synaptic vesicles released from individual release sites, then it would be expected to cause more pronounced depression. However, we found that neither the magnitude nor the kinetics of recovery from PPD were detectably different in leaner compared to WT PCs (Fig. 5A and B). This suggests that the observed increase in CF-EPSC amplitude in leaner PCs is not due to changes in vesicular release at individual CF–PC release sites, but is consistent with an increase in the number of release sites, in postsynaptic sensitivity, or a combination of the two.

Figure 5.

Paired-pulse depression at CF–PC synapses is unchanged in leaner mice A, comparison between EPSCs elicited by two 30 μA stimuli separated by 80 ms in representative wild-type (WT, left) and leaner (right) PCs. Both WT and leaner PCs showed paired-pulse depression (PPD) at CF synapses, quantified as the ratio of the amplitudes of the second to the first EPSCs (A₂/A₁). B, time course of recovery from PPD. Recoveries follow biexponential time courses with limiting values of unity (see smooth curves) that are indistinguishable in the two cell populations. Dashed vertical line indicates the interpulse interval (80 ms) separating stimuli in A. Results in B are based on analysis of 15 PCs in 8 WT mice (except for the 2800 ms time point where n = 8) and 8–11 PCs in 4 leaner mice. Holding potential: −70 mV.

Effect of the leaner mutation on excitatory postsynaptic potentials elicited by PF stimulation

Given the observed changes in PF- and CF-EPSCs in leaner mice, it is important to evaluate their impact on evoked voltage responses in PCs, which are arguably one of the most important immediate consequences of synaptic transmission. These voltage responses are expected to depend on the amplitude and kinetics of synaptic currents, as well as on the intrinsic membrane properties of PCs, which are also modified by the leaner mutation (Ovsepian & Friel, 2008). To address this point, we compared postsynaptic potentials elicited by PF and CF stimulation in WT and leaner PCs under whole-cell current-clamp conditions.

Figure 6A compares representative EPSPs elicited by weak parallel fibre stimulation that did not elicit spikes in WT (left) and leaner (right) PCs. Cells were hyperpolarized by steady injection of hyperpolarizing current to suppress spontaneous PC spike activity (see Methods, Mittmann et al. 2005). As in the voltage-clamp experiments described above, cells were exposed to a GABA_A antagonist to block inhibitory transmission that would confound analysis of the EPSPs (see Methods). For a given stimulus intensity (40 μA, Fig. 6A, top), WT (left) and leaner (right) PCs responded with comparable EPSPs, which were quantified in terms of their amplitude (A), maximal slope during the rising phase (S) and exponential time constant of recovery (τ_D) (Fig. 6A, bottom). For these comparisons, responses elicited by stimuli within the range 40–50 μA were pooled, since for each cell population, differences in EPSCs within this range were small compared to differences between the cell populations (see Fig. 2B). We did not detect a difference in either EPSP amplitude or maximal slope during the rising phase between leaner and WT PCs. However, EPSPs decayed faster in leaner PCs than in WT cells, with a time constant that was approximately one-half the value observed in WT cells (leaner: 35.6 ± 6.2 ms, n = 4, N = 2; WT: 69.1 ± 6.6 ms, n = 4, N = 3, P < 0.01). Thus, despite having considerably smaller PF-EPSC amplitudes under voltage clamp, PF stimulation in leaner mice elicits EPSPs with normal amplitude and maximal slope, but with faster decay kinetics.

Figure 6.

Leaner PCs respond to PF stimulation with subthreshold membrane potential responses resembling those found in WT cells A, top: comparison between EPSPs elicited by PF stimulation (40 μA) in WT (left) and leaner (tg^la) PCs at hyperpolarized voltages. Upper traces: membrane potential (V); lower traces: –dV/dt. Bottom (left to right): mean EPSC amplitude (A), maximal slope (S) and decay time constant (τ_D) from 4 WT and 4 leaner PCs (N = 3 and 2, respectively); responses elicited by stimuli within the range 40–50 μA were pooled. B, top: EPSPs elicited by paired stimuli in WT (left) and leaner (right) PCs with interpulse intervals (IPI) 150 ms and 80 ms, respectively. Upper traces: membrane potential (V); lower traces: –dV/dt. Stimulus intensity: WT: 30 μA, leaner: 40 μA. Bottom: collected results describing paired-pulse ratios of EPSP amplitude (A₂/A₁) and maximal slope (S₂/S₁) from 5 WT (N = 2) (left, open bars) and 4 leaner cells (N = 1) (right, grey bars). For both cell populations, A₂/A₁ and S₂/S₁ were significantly greater than unity (** P < 0.01). Tick mark to left of voltage traces: −65 mV.

To determine if paired-pulse facilitation of synaptic currents leads to facilitation of EPSPs in leaner as well as WT PCs, we compared synaptic responses to paired stimuli in these two cell populations (Fig. 6B). For this comparison, interpulse intervals were long enough to permit recovery of the membrane potential after the first stimulus, ensuring that the voltage at the time of the second stimulus was close to its initial value. Given that recoveries following single stimuli are slower in WT than in leaner PCs (see above), we used an interpulse interval of 150 ms for WT cells and 80 ms for leaner. In both cases, EPSP facilitation was clearly evident, based on either the ratio of amplitudes (A₂/A₁) or maximal slopes (S₂/S₁) (Fig. 6B, bottom). While smaller than the paired-pulse ratios of EPSC amplitude at corresponding interpulse intervals (cf. Fig. 4), the paired-pulse EPSP ratios were significantly greater than unity in both cell populations. We did not compare EPSP facilitation in leaner and WT PCs at the same IPI; the quantitative difference between facilitation in WT and leaner in Fig. 6B is probably due, at least in part, to differences in IPIs.

Evoked EPSPs were also able to trigger regenerative responses in both leaner and WT PCs. Figure 7A shows responses like those in Fig. 6 except that the second PF-EPSP elicited a spike. Comparison between spikes elicited by such stimuli in four cells from each cell population did not reveal a significant difference in spike amplitude (WT: 53 ± 1.5 mV, leaner: 51.3 ± 2.9 mV), maximal slope (WT: 344.5 ± 8.6 V s⁻¹, leaner: 306.4 ± 15.4 V s⁻¹) or threshold (WT: −45.3 ± 2.4 mV, leaner: −48.0 ± 0.9 mV, see Fig. 7B). This indicates that while the leaner mutation leads to a reduction of PF-EPSC amplitude, it does not interfere with the ability of PF stimulation to elicit PC action potentials.

Figure 7.

Leaner PCs respond to PF stimulation with suprathreshold membrane potential responses resembling those found in WT cells A, illustration of PF-EPSPs that elicit spikes in WT (left) and leaner (right) PCs, using the same interpulse intervals as in Fig. 6B. Stimulus intensities: 40 μA (WT), 43 μA (tg^la). Tick mark to left of voltage traces: −65 mV. B, collected results from 4 WT cells (N = 2) and 4 leaner cells (N = 2) describing mean spike amplitude (A), maximal dV/dt (S) and spike threshold, none of which differed significantly between leaner and WT cells.

Effect of the leaner mutation on postsynaptic potentials elicited by CF stimulation

A comparison between complex spikes (CSs) elicited by CF stimulation in WT and leaner PCs is presented in Fig. 8. We found no differences between CSs in leaner and WT PCs, either in terms of the amplitude of the first peak (A₁), its latency (T_L), or the number of spikelets (WT: 3.2 ± 0.3, n = 8, N = 3; leaner: 3.5 ± 0.7, n = 5, N = 3, n.s.). However, we did find a more depolarized plateau level (V_P) following the last spikelet (leaner: −41.3 ± 1.2 mV; WT: −48.3 ± 1.6 mV, P < 0.05), which may result from reduced Ca²⁺ entry and attenuation of Ca²⁺-activated K⁺ currents that normally help lower the membrane potential after stimulation in WT PCs (Womack et al. 2004; Ovsepian & Friel, 2008). Overall, these results indicate that despite an average increase in CF-EPSC amplitude of ˜80%, CSs in leaner PCs largely resemble those found in WT.

Figure 8.

Leaner PCs respond to CF stimulation with complex spikes closely resembling those found in WT cells A, top: representative CF-stimulation-evoked complex spikes in PCs in WT (left) and leaner (right) PCs at hyperpolarized membrane potentials. Plateau level (V_P) is indicated by horizontal dashed line. Bottom: complex spikes from top panels (see regions enclosed by dotted rectangles) on expanded time scale, along with definitions of amplitude (A₁) and latency (T_L) of the first spike. Tick mark to left of voltage traces: −60 mV. B, collected results from 8 WT cells (N = 3) and 5 leaner cells (N = 3) describing mean amplitude (A₁), spike latency (T_L) and plateau level (V_P) (* P < 0.05).

Leaner PCs generate antidromic spikes like those observed in WT cells

In some experiments, in addition to triggering a complex spike, field stimulation in the granule cell layer also elicited an antidromic spike (AS), providing an opportunity to compare axosomatic excitability in leaner and WT PCs separately from transmission at CF–PC synapses. Figure 9A illustrates representative recordings from WT (left) and leaner PCs (right), in each case showing three CSs accompanied by ASs along with three ASs without CSs from the same cell elicited with the same stimulus intensity. In these cells, stimuli were consistently suprathreshold for eliciting an AS, but only near threshold for eliciting a CS, leading to occasional CS failures. Figure 9A lower panels show these evoked responses on an expanded time scale, indicating their reproducibility from trial to trial. Antidromic spikes could be distinguished from CSs based on their short latency (< 0.5 ms), unitary waveform, and insensitivity to NBQX (not shown); note that all results in Fig. 8 were obtained under conditions where CSs could be elicited without also triggering ASs. We measured the spike latency (T_L = 0.28 ± 0.01 ms) and amplitude (A = 70.4 ± 3.13 mV) from a total of 91 antidromic spikes in four cells from two WT mice (Fig. 9B). Based on measurements from 65 ASs in two PCs from one leaner mouse, neither parameter was significantly different between the two cell populations (Fig. 9B, compare open and filled symbols), although the range of AS amplitudes was somewhat lower in leaner compared to WT PCs (Fig. 9B, right). Since leaner PCs can generate ASs with amplitudes and latencies similar to those seen in WT PCs, it would appear that the leaner mutation does not grossly impair the ability of these cells to generate axosomatic action potentials.

Figure 9.

Leaner PCs respond to CF stimulation with antidromic spikes like those observed in WT cells A, top: evoked complex spikes (CSs) in WT (left) and leaner (right) PCs accompanied by antidromic spikes (ASs). Bottom: early portion of these recordings (see region enclosed by dotted rectangles in top panel) on an expanded time scale. Each panel shows the superposition of three responses in which the stimulus elicited both a CS and AS, along with three responses without CSs. Tick mark to left of voltage traces: −60 mV. B, plot of antidromic spike latency (T_L, left) and amplitude (A, right) from 91 CSs from four WT PCs (2 mice) and from 65 CSs from two leaner PCs (1 mouse).

Discussion

Understanding the phenotype produced by a mutation that is expressed in multiple cell types requires consideration of the distinct consequences of the mutation in each cell type. In the cerebellar cortex, P/Q-type Ca²⁺ channels are expressed in Purkinje cells, where they regulate intrinsic membrane properties, and in presynaptic terminals that provide synaptic inputs to these cells, where they regulate neurotransmitter release. Therefore, disruptions in channel function are expected to have contex-dependent cellular effects that jointly determine the overall effect of the mutation on circuit function.

We asked how the leaner mutation affects synaptic transmission between parallel and climbing fibre inputs and PCs. Measurements of postsynaptic currents under voltage clamp revealed pronounced changes, including a reduction in PF-EPSC amplitude without a change in kinetics, and an increase in the size of CF-EPSCs accompanied by faster decay kinetics. To investigate how these changes in synaptic current impact evoked membrane potential responses, we also measured EPSPs elicited by the respective stimuli under current-clamp conditions. It was found that despite the large changes in synaptic currents observed in leaner mice, evoked EPSPs were surprisingly normal. A possible explanation comes from our previous work describing altered intrinsic electrical properties in leaner PCs. We propose that the overall effect of the leaner mutation on excitatory synaptic transmission depends on a combination of changes, including changes in synaptic currents, and in the impact of these currents on somatic membrane potential, resulting from changes in intrinsic membrane properties of PCs. Our results provide the first description of evoked EPSCs and EPSPs in leaner PCs, and illustrate how multiple cell-type-specific defects resulting from a Ca²⁺ channel mutation can jointly influence circuit function. The results support emerging evidence that P/Q-type Ca²⁺ channelopathies influence intrinsic membrane properties, as well as synaptic transmission, and that both are important in defining the mutant phenotype.

Effects of the leaner mutation on transmission at PF–PC synapses

We observed a ˜50% reduction in the amplitude of PF-EPSCs in leaner PCs without a detectable change in the time to peak or recovery kinetics. One likely factor contributing to the reduction in EPSC amplitude is attenuation of Ca²⁺ entry in PF presynaptic boutons, leading to reduced neurotransmitter release at individual release sites. This possibility is consistent with previous work showing that (a) the leaner mutation causes a ˜60% reduction in whole-cell Ca²⁺ current density in PC somata (Dove et al. 1998; Wakamori et al. 1998; Lorenzon et al. 1998), (b) P/Q-type Ca²⁺ channels provide the major pathway for presynaptic Ca²⁺ entry at PF–PC synapses (Mintz et al. 1995; Doroshenko et al. 1997; Matsushita et al. 2002), and (c) our demonstration in the present study that reduced PF-EPSC amplitudes are accompanied by enhanced paired-pulse facilitation. Reduced PF-EPSC amplitudes were also reported for the less severe Ca_v1.2 mutants rocker (Kodama et al. 2006), tottering (Matsushita et al. 2002), and rolling Nagoya (Matsushita et al. 2002), although facilitation was altered only in the latter mutant.

Another factor that could attenuate PF-EPSC amplitudes in leaner PCs is a reduction in the total number of parallel fibre synapses that contribute to transmission following PF stimulation. Previous work has described granule cell degeneration in leaner mice starting at ˜P10 (Herrup & Wilczynski, 1982; Lau et al. 2004), which is likely to affect the P17–21 mice used in our study. However, if this were the only change affecting PF–PC transmission, then smaller PF-EPSCs would be expected without a change in paired-pulse facilitation, contrary to what was observed.

It is also possible that other modifications resulting from the leaner mutation contribute to the observed changes in PF-EPSCs. For example, previous work has shown that leaner PCs have weaker somatic Ca²⁺ buffering than WT neurons (Dove et al. 2000), which may compensate for reduced Ca²⁺ entry by increasing the impact of Ca²⁺ on cytoplasmic free Ca²⁺ levels ([Ca²⁺]_i) (Murchison et al. 2002). A similar reduction in Ca²⁺ buffering strength has also been reported in cerebellar granule cells (Nahm et al. 2002). If such a change in Ca²⁺ buffering also occurs in presynaptic boutons of granule cell axons, it could play a role in setting residual [Ca²⁺]_i levels during the interval between paired stimuli, thereby influencing facilitation. Given that the leaner mutation is known to reduce Ca²⁺ current density and promote granule cell degeneration, the simplest explanation of a reduction in PF–PC transmission strength accompanied by increased facilitation is that the mutation attenuates presynaptic Ca²⁺ entry and reduces the number of parallel fibre synapses contributing to transmission. Direct assessments of reduced PF number and its contribution to reduced PF-EPSCs, as well as the potential role of altered presynaptic Ca²⁺ buffering and postsynaptic sensitivity to transmitter (Kodama et al. 2006) are important goals of future experiments.

Despite showing smaller PF-EPSCs, synaptic potentials produced by PF stimulation in leaner mice were indistinguishable from those elicited in WT PCs in terms of their amplitudes and maximal slopes. The observation that a given stimulus can produce similar PF-EPSPs in leaner PCs, despite generating ˜50% smaller postsynaptic currents, raises the possibility that postsynaptic changes increase the sensitivity with which PC membrane potential changes in response to such currents. One possible factor is a reduction in dendritic size in leaner PCs (Ovsepian & Friel, 2008). This could render PF stimulation more effective in promoting somatic depolarization, both by increasing PC input resistance, and by causing a somatopetal shift in the distribution of PF inputs on PC dendrites. This idea is consistent with results of Roth & Hausser (2001), who carried out computer simulations based on detailed reconstructions of rat Purkinje cells, finding that EPSPs resulting from transmission at a given distance from the soma produce a larger somatic depolarization in cells with smaller dendrites, in their study illustrated by comparison between responses in PCs at different developmental stages. However, comparison with that study is complicated by differences between leaner and WT PCs in terms of dendritic conductance as well as size, given that leaner PCs fail to generate dendritic Ca²⁺ spikes in response to somatic current injection (Ovsepian & Friel, 2008), presumably because the Ca²⁺ current density in PC dendrites is insufficient to cause regenerative depolarization. Such a reduction in Ca²⁺ current density could potentially play a role in maintaining normal complex spikes in spite of a nearly 2-fold increase in CF-EPSC amplitude (see next section).

Effects of the leaner mutation on transmission at CF–PC synapses

EPSCs elicited by CF stimulation in leaner PCs were on average ˜80% larger than those observed in WT PCs. The absence of any detectable change in paired-pulse ratio or in the time course of recovery from paired-pulse depression suggests that changes in presynaptic Ca²⁺ and vesicle dynamics are not the major determinants of enhanced transmission strength at these synapses, pointing instead to changes in the number of CF–PC synapses or the postsynaptic currents elicited at individual synapses. Although it has been reported that some disruptions of P/Q Ca²⁺ channel properties or expression lead to the innervation of PCs by multiple climbing fibres (Miyazaki et al. 2004; Kodama et al. 2006), our results cannot be readily explained in this way, since the large CF-EPSCs observed in leaner PCs were triggered in an all-or-none fashion, although the existence of multiple CF inputs that are activated with the same stimulus threshold remains a possibility. Matsushita et al. (2002) reported qualitatively similar findings in tg^rol mice, although CF-EPSC amplitudes were increased by ˜46%. These investigators reported that CF-EPSC recoveries were slower in PCs from tg^rol mice, while we found that decays in leaner PCs were faster compared to WT cells.

Although CF stimulation in leaner mutants elicited CF-EPSCs with nearly twice the amplitude of those observed in WT PCs, the complex spikes evoked by these stimuli were very similar to those observed in WT cells. Clearly, one possibility is that synaptic currents elicited by CF stimulation are already suprathreshold for eliciting all-or-none responses in PCs, such that an increase in EPSC amplitude does not strongly modify the complex spike. The similarities between CSs in WT and leaner PCs nonetheless provide mechanistic insights into complex spike generation and the contribution of dendritic Ca²⁺ spikes to the somatic complex spike waveform (Schmolesky et al. 2002). Leaner PCs do not generate detectable Ca²⁺ spikes under current-clamp conditions, even in response to strong depolarizing current injection (Ovsepian & Friel, 2008). This may occur simply because the reduction in Ca²⁺ current density in these cells renders them unable to generate Ca²⁺ spikes; similar findings were reported by Mori et al. (2000) in tg^rol PCs. How can the inability to elicit dendritic Ca²⁺ spikes in leaner PCs be reconciled with the observation that complex spikes in these cells are nearly normal? One possible explanation comes from the work of Callaway & Ross (1997), which suggests that complex spikes recorded in Purkinje cell bodies are dominated by regenerative somatic Na⁺ currents. According to this mechanism, synaptic currents elicited by CF stimulation, along with regenerative dendritic Ca²⁺ currents, provide the trigger for evoking somatic complex spikes in WT PCs. This raises the possibility that the observed increase in CF-EPSC amplitude in leaner mice compensates for reduced dendritic Ca²⁺ entry, thereby permitting normal triggering of somatic CSs. The reduction in dendritic size found in leaner PCs (Ovsepian & Friel, 2008) may also contribute by increasing the impact of synaptic currents elicited by CF stimulation on somatic membrane potential and spike generation. The only difference between complex spikes in leaner and WT PCs that we observed was a more depolarized plateau level following the last spikelet in mutant PCs. This may be a consequence of reduced Ca²⁺ entry and a reduction in Ca²⁺-activated K⁺ currents (Womack et al. 2004) that speed repolarization in WT cells after the CS.

Functional considerations

The main finding of the present study is that, despite showing pronounced changes in synaptic currents, excitatory synaptic stimuli elicit postsynaptic membrane potential responses in leaner PCs that are quite similar to those observed in WT PCs. There are, nonetheless, several quantitative differences. PF-EPSP recoveries are faster in leaner PCs and CF-EPSPs display slightly more depolarized plateau levels during repolarization, compared to WT PCs. One possibility is that these subtle differences have profound consequences for cerebellar motor control. Another possibility is that the effect of the leaner mutation on excitatory synaptic transmission is not the major factor disrupting cerebellar motor control. If true, this points to effects on other processes where P/Q-type Ca²⁺ channels regulate cell function.

What might these processes be? Purkinje cells have been a focus of attention because they rely on P/Q-type Ca²⁺ channels for regulating most of their voltage-sensitive Ca²⁺ entry, and provide the only output of the cerebellar cortex, in effect serving as the final common pathway for information processing by the cerebellar cortex. From this perspective, a starting point for considering how P/Q-type Ca²⁺ channel mutations interfere with cerebellar motor control is an understanding of the impact of these mutations on intrinsic membrane properties of PCs. Work from this laboratory (Ovsepian & Friel, 2008) shows that the leaner mutation has several effects on passive and active membrane responses elicited by depolarizing current injection in PCs that have been held at a hyperpolarized voltage by steady current injection.

Changes in intrinsic membrane properties may influence the spontaneous activity that is observed in PCs in vivo (Armstrong & Rawson, 1979) and in acute slices (Llinás & Sugimori, 1980a; Tank et al. 1988; Häusser & Clark, 1997; Womack & Khodakhah, 2002; Edgerton & Reinhart, 2003; McKay & Turner, 2005). This activity appears to reflect intrinsic membrane properties of PCs since it continues after cells are exposed to blockers of fast synaptic transmission (Häusser & Clark, 1997; Womack & Khodakhah, 2002; Cerminara & Rawson, 2004). Evidence that P/Q-type Ca²⁺ channels play a critical role in enabling spontaneous activity comes from experiments in cerebellar slices, where partial blockade of these channels leads to reduced precision of spike timing in PCs (Hoebeek et al. 2005; Walter et al. 2006), while complete blockade abolishes activity (Tank et al. 1988; Womack & Khodakhah, 2002). Ca²⁺ entry through these channels appears to influence spontaneous activity, at least in part, by regulating Ca²⁺-activated K⁺ channels that control outward currents which promote membrane hyperpolarization (Raman & Bean, 1999; Womack et al. 2004). Thus, mutations that impair activity of P/Q-type Ca²⁺ channels or Ca²⁺-activated K⁺ channels in PCs disrupt spontaneous action potential generation in these cells (Sausbier et al. 2004; Hoebeek et al. 2005; Walter et al. 2006).

While the results of the present study suggest that transmission of extrinsic excitatory inputs to PCs are only subtly modified by the leaner mutation, effects on inhibitory transmission have not been investigated. PCs receive inputs from intrinsic interneurons that release GABA in a manner that depends strongly on P/Q-type Ca²⁺ channel activity (Stephens et al. 2001). Moreover, P/Q Ca²⁺ channel mutations could impair recurrent inhibition between PCs, or inhibitory transmission to target cells in the deep cerebellar nuclei. Evaluation of these and other possibilities will require further experiments to delineate the cell-specific effects of the mutations and how they together contribute to cerebellar dysfunction.