Autonomous Purkinje cell activation instructs bidirectional motor learning through evoked dendritic calcium signaling

Abstract

The signals in cerebellar Purkinje cells sufficient to instruct motor learning have not been systematically determined. Therefore, we applied optogenetics in mice to autonomously excite Purkinje cells and measured the effect of this activity on plasticity induction and adaptive behavior. Ex vivo, excitation of channelrhodopsin-2-expressing Purkinje cells elicits dendritic Ca²⁺ transients with high-intensity stimuli initiating dendritic spiking that additionally contributes to the Ca²⁺ response. Channelrhodopsin-2-evoked Ca²⁺ transients potentiate co-active parallel fiber synapses; depression occurs when Ca²⁺ responses were enhanced by dendritic spiking. In vivo, optogenetic Purkinje cell activation drives an adaptive decrease in vestibulo-ocular reflex gain when vestibular stimuli are paired with relatively small-magnitude Purkinje cell Ca²⁺ responses. In contrast, pairing with large-magnitude Ca²⁺ responses increases vestibulo-ocular reflex gain. Optogenetically induced plasticity and motor adaptation are dependent on endocannabinoid signaling, indicating engagement of this pathway downstream of Purkinje cell Ca²⁺ elevation. Our results establish a causal relationship among Purkinje cell Ca²⁺ signal size, opposite-polarity plasticity induction, and bidirectional motor learning.

Fig. 1. Optogenetically evoked Ca²⁺ signaling in PC dendrites.

a Representative images from a single Pcp2::Cre;Ai27 mouse showing tdTomato-tagged ChR2 expression in the flocculus. In the magnified view, PCs are marked by calbindin immunostaining (molecular layer, ML; Purkinje cell layer, PCL; and granule cell layer, GCL). b The average fluorescence intensity profile of ChR2-tdTomato in the cerebellum of an example mouse. c ChR2-expressing PCs were filled with Fluo-5F during whole-cell recording. Two-photon imaging was used to measure Ca²⁺ transients evoked by optogenetic excitation or climbing-fiber (CF) stimulation. d Average Ca²⁺ transients from the same PC dendrite evoked by single pulses of light (λ 461 nm; 5 ms; artifacts blanked for clarity), shown relative to the climbing-fiber-evoked response. e Summary plot of peak Ca²⁺ transient amplitude for different stimulus conditions. In addition to single stimuli, trials also included bursts of closely spaced light pulses to elicit activity (n = 21 dendritic sites obtained from seven PCs; six mice). Descriptive statistics: 1 pulse: 0.13 vs. 1.68 mW P = 0.0345, 0.13 mW vs. CF, P < 0.0001, 0.73 mW vs. CF, P = 0.0404; 3 pulses: 0.13 vs. 0.73 mW, P < 0.0001, 0.13 vs. 1.68 mW, P < 0.0001, 0.13 mW vs. CF, P < 0.0001; 12 pulses: 0.13 vs. 0.73 mW, P < 0.0001, 0.13 vs. 1.68 mW, P < 0.0001, 0.13 mW vs. CF, P < 0.0001, 0.73 vs. CF, P = 0.0002). f Simultaneous patch-clamp recordings from the soma and dendrite of ChR2-expressing PCs allowed for high-resolution measurement of optogenetically induced electrogenic activity. g Electrophysiological responses from the same PC to single pulses of light at increasing powers (λ 461 nm; 5 ms). The evoked dendritic response to climbing-fiber stimulation is shown for a separate cell. h Plot shows the average number of optogenetically evoked dendritic spikes as a function of light power; the threshold for evoking reliable dendritic spiking is indicated by the dashed line. All data are mean ± SEM; asterisk indicates P < 0.05, two-way repeated-measures ANOVA with Tukey’s post test.

Fig. 2. Optogenetic PC activation induces synaptic plasticity.

a During PC recordings, a parallel fiber (PF) tetanus (100 Hz, 70 ms) preceded an optogenetically induced stimulus (Δ 25 ms); light power was subthreshold for evoking dendritic spike firing (λ 461 nm; 5 ms; 0.15 mW/mm²). This pairing was delivered for 300 repetitions (1 Hz). b Top: average test EPSPs to parallel fiber stimulation before and after the conjunctive pairing procedure. Bottom: plot of change in EPSP amplitude across PCs shows optogenetically induced LTP (red). Test EPSP amplitude was not affected by the light-pairing procedure when recording in non-ChR2-expressing control PCs (gray). Data were obtained from 12 mice. c In separate recordings, the parallel fiber tetanus was paired with an optogenetic stimulus suprathreshold for evoking dendritic spike initiation (λ 461 nm; 5 ms; 1.7 mW/mm²). d Top: average test EPSPs before and after the conjunctive pairing procedure with high-power light. Bottom: summary plot of EPSP amplitude across PCs shows optogenetically induced LTD (red), the same polarity of plasticity induced by pairing the parallel fiber tetanus with a climbing-fiber stimulus in place of the light pulse (black). Data were obtained from 14 mice. In the plots, data are presented as mean ± SEM.

Fig. 3. Optogenetic PC activation induces VOR learning.

a Optical fibers targeting both flocculi were bilaterally implanted into Pcp2::Cre;Ai27 mice to stimulate ChR2-expressing PCs. b Eye movements in a mouse evoked by unilateral optogenetic PC activation using different light powers (λ 471 nm; 12 pulses; 5 ms; 50 Hz, 1 and 2 mW). The direction of lateral eye displacement (T, temporal; N, nasal) depended on whether PCs in the ipsiversive or contraversive flocculus were stimulated (laser pulses indicated by blue tick marks). The absolute amplitude of evoked movements was used to categorize the intensity of optogenetic stimuli (> or <0.75°, high- and low-intensity, respectively). c In darkness, head-fixed mice were trained by pairing sinusoidal vestibular motion (1 Hz) with a low-intensity optogenetic stimulus (λ 471 nm; 12 pulses; 5 ms; 50 Hz) beginning near the completion of head turns (training lasted 30 min). In the example traces below, PC activation was timed to the end of contraversive vestibular motion (average head position, black trace; laser pulses; blue tics). d Average VOR-evoked eye movements measured before and after the mouse were trained with a pairing procedure using low-intensity optogenetic stimuli timed to the end of contraversive vestibular motion. The traces were normalized to that of time-matched, control responses recorded during sessions of training consisting of the vestibular stimulus alone. e The effect of pairing either contraversive or ipsiversive vestibular motion with low-intensity optogenetic stimuli on VOR gain, shown relative to control sessions. Measurements were obtained from the same mice. Changes are shown normalized to a baseline measurement obtained immediately before training (t = 0 min) for each condition. Vestibular stimulation vs. ipsiversive stimulation: P = 0.6903; vestibular stimulation vs. contraversive stimulation: P = 0.0388. f In separate sessions, a cohort of the same mice was trained using pairing that included high-intensity optogenetic stimuli (λ 471 nm; 12 pulses; 5 ms; 50 Hz). In the example head-position traces below, PC activation was timed to the end of ipsiversive vestibular motion. g Average VOR eye movements before and after training with high-intensity optogenetic stimuli timed to the end of ipsiversive vestibular motion. h The effect of pairing vestibular motion with high-intensity optogenetic PC stimulation on VOR gain across mice. Vestibular stimulation vs. ipsiversive stimulation: p = 0.0068, Vestibular stimulation vs. contraversive stimulation: P = 0.1151. Data are shown mean ± SEM with an asterisk indicating P < 0.05, two-way repeated-measures ANOVA with Dunnett’s post tests.

Fig. 4. Optogenetically induced plasticity and learning requires endocannabinoid signaling.

a In the presence of the CB₁ receptor antagonist AM251 (5 μM), a parallel fiber (PF) tetanus (100 Hz, 70 ms) was repeatedly paired (300 repetitions; 1 Hz) with a light pulse (λ 461 nm; 5 ms; either 0.13 or 1.7 mW/mm²) to activate ChR2-expressing PCs. b, c Top: average parallel fiber-evoked EPSPs recorded in PCs before and after conjunctive pairing with an optogenetic stimulus either subthreshold or suprathreshold for dendritic spike firing. Bottom: plots of average EPSP amplitude across PCs for control recordings (gray) or in the presence of AM251 (black). Data obtained from 24 mice total. d During training, ChR2-expressing PCs in the flocculus were activated using optogenetic stimuli (12 pulses, 5 ms, 50 Hz) in conjunction with vestibular motion (1 Hz). Either vehicle (containing saline + DMSO) or AM251 (5 mg/kg in saline + DMSO) was administered 20 min prior to training. e, f Summary data for mice trained with either low- or high-intensity optogenetic stimuli timed to the end of contraversive or ipsiversive vestibular motion, respectively, after being administered vehicle or AM251. Separate sessions also included vestibular-only training in the same pharmacological conditions. Low-intensity stimulation, vehicle: vestibular vs. contraversive stimulation, P < 0.05; AM251: vestibular stimulation vs. contraversive stimulation, P > 0.05. High-intensity stimulation, vehicle: vestibular vs. ipsiversive stimulation, P < 0.05; AM251: vestibular vs. ipsiversive stimulation, P > 0.05. All data are mean ± SEM; asterisk indicates P < 0.05, two-way repeated measures with Sidak’s post test.

Fig. 5. Absence of optogenetically induced dendritic Ca²⁺ signaling and plasticity in PCs with soma-restricted ChR2 expression.

a Image from a Pcp2::Cre mouse expressing mRuby-tagged ChR2-K_v2.1 by Cre-dependent AAV transduction; PCs are marked by calbindin immunostaining. b Electrophysiological recordings from the soma of two different PCs, expressing either ChR2 (n = 11 PCs) or ChR2-K_v2.1 (n = 6 PCs), in response to single pulses of light (5 ms; 0.7 mW/mm²). A complex spike in the ChR2-expressing cell was evoked by climbing-fiber (CF) stimulation for comparison (n = 9 PCs). The summary plot on the right shows the average somatic spiking response to the different stimuli (one-way ANOVA with Tukey’s post test; CF vs. ChR2: P > 0.05, CF vs. ChR2-Kv2.1: P ≤ 0.001, ChR2 vs. ChR2-Kv2.1: P ≤ 0.001). Data from 14 mice. c Dendritic Ca²⁺ activity in a ChR2-K_v2.1-expressing PC to a burst of optogenetic stimuli (12 pulses; 5 ms; 50 Hz; 0.7 mW/mm²) or to repeat climbing-fiber activation. Somatic spiking is shown in the gray boxes below. d Lack of dendritic Ca²⁺ signals in response to optogenetic stimuli in PCs expressing ChR2-K_v2.1 (n = 25 dendrites, four cells, three mice; two-way repeated-measures ANOVA with Dunnett’s post test; 1 pulse: CF vs. ChR2-Kv2.1, P < 0.0001, 3 pulses: CF vs. ChR2-Kv2.1, P < 0.0001, 12 pulses: CF vs. ChR2-Kv2.1, P < 0.0001). e A parallel fiber (PF) tetanus (100 Hz; 70 ms) was stimulated in conjunction with optogenetic activation of ChR2-K_v2.1-expressing PCs (2 pulses; 5 ms; 20 Hz); pairing was repeated 300 times (1 Hz). On the right, average parallel fiber-evoked EPSPs are shown before and after conjunctive pairing. f In recordings from ChR2-K_v2.1-expressing PCs, EPSP amplitude remained unchanged following the conjunctive parallel fiber-light-pairing procedure (red) as opposed to a depression of EPSP amplitude following the parallel fiber -CF pairing procedure (black). Data were obtained from ten mice. All data are mean ± SEM; asterisk indicates P < 0.05.

Fig. 6. Absence of motor learning with optogenetically induced PC simple spike firing.

a Pcp2::Cre mice were implanted with bilateral optical fibers to stimulate floccular PCs expressing soma-targeted ChR2-K_v2.1 by AAV-mediated transduction. b To calibrate the intensity of the optogenetic stimulus, evoked eye movements were monitored in response to activation of ChR2-K_v2.1-expressing PCs (12 pulses, 5–10 ms; 50 Hz). c Average VOR-evoked eye movements measured before and after training with low-intensity optogenetic stimuli timed to the end of contraversive vestibular motion. The traces were normalized to that of time-matched, control responses recorded during training sessions consisting of vestibular motion alone. The summary plot shows the lack of effect of pairing low-intensity optogenetic PC activation with vestibular motion on VOR gain. d Pcp2::Cre;Ai27 mice were implanted with bilateral optical fibers targeting the medial vestibular nuclei (MVN) to activate the ChR2-expressing axon projections of floccular PCs. The image on the right shows ChR2-tdTom expressing PC terminals in the MVN (fourth ventricle, 4 V; genu of CN VII, g7). e VOR-evoked eye movements before and after pairing low-intensity optogenetic activation of PC axons with vestibular motion on normalized VOR performance. Summary plot show a lack of an effect of both contraversive and ipsiversive pairing contexts compared to control training sessions. Data are presented as mean ± SEM and analyzed with two-way repeated-measures ANOVA with Dunnett’s post test. f Comparison of VOR performance before and after pairing with high-intensity optogentic stimulation of ChR2-K_v2.1-expressing PCs in the flocculus, timed to the end of ipsiversive vestibular motion. Summary data from all mice is shown on the right.