Motor-Skill Learning Is Dependent on Astrocytic Activity

Abstract

Motor-skill learning induces changes in synaptic structure and function in the primary motor cortex through the involvement of a long-term potentiation- (LTP-) like mechanism. Although there is evidence that calcium-dependent release of gliotransmitters by astrocytes plays an important role in synaptic transmission and plasticity, the role of astrocytes in motor-skill learning is not known. To test the hypothesis that astrocytic activity is necessary for motor-skill learning, we perturbed astrocytic function using pharmacological and genetic approaches. We find that perturbation of astrocytes either by selectively attenuating IP3R2 mediated astrocyte Ca(2+) signaling or using an astrocyte specific metabolic inhibitor fluorocitrate (FC) results in impaired motor-skill learning of a forelimb reaching-task in mice. Moreover, the learning impairment caused by blocking astrocytic activity using FC was rescued by administration of the gliotransmitter D-serine. The learning impairments are likely caused by impaired LTP as FC blocked LTP in slices and prevented motor-skill training-induced increases in synaptic AMPA-type glutamate receptor in vivo. These results support the conclusion that normal astrocytic Ca(2+) signaling during a reaching task is necessary for motor-skill learning.

Figure 1

IP₃R2 mediated Ca²⁺ signaling is necessary for motor-skill learning. a(i) Dual loading of acute M1 slices with SR101 (left) and fluo-4 AM (right) in control (top panel) and mIP₃R2 (bottom panel) mice. ROIs of astrocyte soma are outlined and numbered. Time-lapse images of the fluo-4 channel for control and mIP₃R2 KO mice showing spontaneous Ca²⁺ transients in soma. Scale bar: 10 μm. (a′) Spontaneous Ca²⁺ transients in soma in control and mIP₃R2 mice. Scale bar: 50 s, 100% ΔF/F. a(ii) Quantification of the proportion of active cells with spontaneous Ca²⁺ transients in soma in control and mIP₃R2 slices. a(iii) Quantification of ATP-induced Ca²⁺ transients in control and mIP₃R2 slices. (b) Motor-skill learning curves for control and mIP₃R2 mice. mIP₃R2 mice show a deficit in motor-skill learning. Repeated measures two-way ANOVA revealed a significant interaction between genotype and training factors (P = 0.0055). (c) Number of reaching attempts made on day 1 and day 5 in control and mIP₃R2 mice. (d) Relationship between ATP-induced Ca²⁺ transients and learning (success rate on day 4/success rate on day 1). ^* P < 0.05, ^** P < 0.01, and ^*** P < 0.0001.

Figure 2

Astrocytic activity is necessary for motor skill learning. (a) Motor skill learning curves for saline and FC injected mice. Two-way ANOVA revealed a significant effect with training (P = 0.0001) and treatment (P < 0.0001). FC injected mice show a deficit in motor skill learning that is rescued by D-serine on days 6 and 7. (b) No difference was detected between number of reaching attempts made on day 1 and day 5 with FC injection. (c) Representative images of dendritic spines in saline and FC injected conditions. Scale bar: 3 μm. FC injection does not alter the dendrite width and spine density. ^* P < 0.05 and ^** P < 0.01.

Figure 3

Fluorocitrate blocks astrocytic Ca²⁺ transients and LTP in slices. (a) Dual loading of acute M1 slices with SR101 (left) and fluo-4 AM (right) in control and after perfusion of FC. ROIs of astrocyte soma are outlined and numbered. Time-lapse images of the fluo-4 channel for control and after FC conditions. Scale bar: 10 μm. (a′) Spontaneous Ca²⁺ transients in soma in control and FC conditions. Scale bar: 75 s, 100% ΔF/F. (b) The percent of active soma, the number of Ca²⁺ transients, and the peak value of the Ca²⁺ transients significantly decreases after treatment with FC. (c) Average time course of the change in FPs in M1 slices after TBS (indicated by arrow) shown for control and FC conditions. Incubation of slices with FC blocks LTP. Representative FPs (average of 5 traces) during baseline condition and 20–30 min after TBS are shown as black and gray traces. Scale bar: 0.4 mV, 4 ms. (d) Average time course of the change in FPs shown for FC coapplied with D-serine, FC coapplied with D-serine and D-AP5. D-Serine partially rescues LTP. (e) Summary of LTP experiments with FPs measured 20–30 min after TBS relative to baseline. ^** P < 0.01 and ^*** P < 0.0001.

Figure 4

Astrocyte function is necessary for motor-skill learning induced synaptic GluA1 insertion. (a) An example electron micrograph demonstrating the presence of synaptosomes. Scale bar: 1 micrometer. (b) Western blots of synaptic GluA1 from forelimb M1 regions of untrained (utr) and trained (tr) hemispheres of 1 day trained mice following ICV injections with saline or FC. (c) Interhemispheric ratios (trained/untrained) of synaptic GluA1. ^* P < 0.05.