Article Figures & Data
Figures
- Figure 1.
In vivo imaging of AMPAR with motor skill training. a, Schematic of the experimental design. b, Images of transfected region of cortex showing overlap of tdTom (magenta) and sGluA2 (white). Scale bar = 15 μm. c, Behavioral performance (normalized success rate) of mice trained on forelimb reaching task. Thin lines represent individual mice and the bold line is the average [Fmr1 KO, blue, n = 6 mice and litter mate control (WT) black, n = 7 mice] mean ± SEM. Two-way repeated measures ANOVA followed by one-way ANOVA with Bonferroni correction. Genotype × Time, F(4,44) = 7.03, p = 0.0002, *p < 0.05, ***p < 0.001, and ****p < 0.0001. Red stars indicate WT versus KO comparisons, and black stars indicate D1 versus Di. See Extended Data Figure 1-1 and Extended Data Table 1-1 for more information.
- Figure 2.
Motor skill training results in transient increase in spines in WT and Fmr1 KO mice. a, Images of layer 2/3 neuron dendrites in the forelimb M1 of WT and Fmr1 KO mice imaged before (T0) and at different times after the first training session. The hemispheres ipsilateral and contralateral to the trained forelimb are referred to IH and CH, respectively. Red, blue, and yellow arrows mark newly formed spines, spines eliminated in the following imaging session, and new spines that were eliminated in the following imaging session, respectively. Scale bar = 10 μm. b, Total spine number. WT-IH (gray) 36 dendrites, 881 spines at T0. WT-CH (black) 37 dendrites, 963 spines at T0. KO-IH (light blue) 35 dendrites, 911 spines at T0. KO-CH (dark blue) 36 dendrites, 965 spines at T0. n = 5 mice per group. Nested random effects mixed model analysis. Genotype, F(1,9.5) = 0.72, p = 0.42; Training, F(1,61.4) = 21.25, p < 0.0001; Time, F(4,564. 9) = 12.99, p < 0.0001. Red stars indicate CH versus IH comparisons and black stars indicate T0hr versus Ti. *p < 0.05, **p < 0.01, and ***p < 0.001. c, Spine formation and elimination between imaging days. Nested random effects mixed model analysis. Formation: Genotype × Hemisphere, F(1,86.84) = 6.98, p < 0.01; Genotype × Time, F(4,560) = 1.45, p = 0.22; Hemisphere × Time, F(4,560) = 5.77, p = 0.0001. Elimination: Genotype × Hemisphere, F(1,100.3) = 6.06, p = 0.016; Time, F(1,563.2) = 44.43, p < 0.0001; Genotype × Time, F(1,563.2) = 0.58, p = 0.68; Hemisphere × Time, F(1,563.2) = 0.46, p = 0.76; Genotype × Hemisphere × Time, F(1,563.2) = 2.19, p = 0.07. Stars indicate IH versus CH. mean ± SEM *p < 0.05 and ***p < 0.001. d, Proportion of new spines (formed at 2 or 18 h after training) that were stable until the last imaging day. WT-IH (135) spines, WT-CH (264 spines), KO-IH (178 spines), KO-CH (226 spines), mean ± SEM, n = 5 mice per group. Two-way ANOVA with Sidak correction. Hemisphere, F(1,16) = 9.39, p = 0.007; Hemisphere × Genotype, F(1,16) = 0.15, p = 0.7. See Extended Data Table 2-1 for more information.
- Figure 3.
Motor skill training results in accumulation of sGluA2 in the contralateral hemisphere of WT but not Fmr1 KO mice. a, Images of sGluA2 changes. Arrows and arrowheads point to spines with persistent and transient increases in sGluA2, respectively. Scale bar = 5 um. b, sGluA2 changes in persistent spines. WT-IH (gray, 614 spines, 36 dendrites), WT-CH (black, 630 spines, 37 dendrites), KO-IH (light blue, 616 spines, 35 dendrites). KO-CH (dark blue, 564 spines, 36 dendrites). n = 5 mice per group. Geometric mean ± SEM. Nested random effects ANOVA model. Genotype × Hemisphere × Time: F(4,11941) = 5.38, p = 0.003. Red stars indicate CH versus IH comparisons and black stars indicate T0hr versus Ti. *p < 0.05, **p < 0.01, and ***p < 0.001. c, Proportion of persistent WT and KO spines with increased (Up), decreased (Down) and unchanged (Same) levels of sGluA2. WT-IH n = 36 dendrites, WT-CH n = 37 dendrites, KO-CH n = 36 dendrites and KO-IH n = 35 dendrites. Three-way ANOVA with post hoc one-way ANOVA and Sidak correction, Same: Time, F(4,700) = 20.71, p < 0.0001; Down: Genotype × Hemisphere, F(1,699) = 5.45, p = 0.02; Up: Time, F(4,700 = 22.55, p < 0.0001; Genotype × Hemisphere, F(1,700) = 11.42, p < 0.001. *p < 0.05 and **p < 0.01. d, Distribution histograms and cumulative distributions (insets) of average D6 and D10 sGluA2 change. WT: p < 0.0001, KO p = 0.44 (Kolmogorov–Smirnov test). e, Proportion of spines with increased sGluA2 at 18 h that maintain the increase until day 10. N = dendrite. Two-way ANOVA with Sidak correction, Genotype, F(1,139) = 5.28, p = 0.2; Hemisphere, F(1,139) = 2.03, p = 0.16; Genotype × Hemisphere, F(1,139) = 5.28 p = 0.02, *p < 0.05 See Extended Data Figures 3-1, 3-2, 3-3, 3-4, and 3-5 and Extended Data Table 3-1 for additional information.
- Figure 4.
sGluA2 changes in contralateral hemisphere of WT mice correlate with Increased behavioral performance. a, Correlation of reaching performance and sGluA2 levels at individual training sessions for which imaging was performed 2–24 h thereafter. Small symbols represent individual mice, and dashed lines are linear regressions for each mouse. Bold blue symbols represent average sGluA2 levels and behavioral performance of n = 5 mice at each imaging session for WT and Fmr1 KO (b) mice. Bold line is the linear regression for the average values (WT, R2 = 0.75, p = 0.13; KO, R2 = 0.46, p = 0.32). Error bars, SEM. Correlations (Spearman coefficient) between sGluA2 changes at different time points and behavioral performance on different days in (c) WT and (d) KO mice. *p < 0.05. See Extended Data Figure 4-1 for additional information.
Extended Data
Extended Data Figure 1-1
Motor skill learning and number of attempts. a, Behavioral performance (success rate) of mice trained on forelimb reaching task. Thin lines represent individual mice and the bold line is the average [Fmr1 KO, blue, n = 6 mice and litter mate control (WT) black, n = 7 mice] mean ± SEM. Two-way repeated measures ANOVA followed by one-way ANOVA with Bonferroni correction. Genotype × Time, F(4,44) = 6.07, p = 0.0006. b, Number of attempts performed. Genotype × Time, F(4,44) =1.916, p = 0.12 Download Figure 1-1, TIF file.
Extended Data Table 1-1
Full statistical information for Figure 1. Download Table 1-1, DOCX file.
Extended Data Figure 2-1
Increased basal level spine density in the Fmr1 KO mouse. Download Figure 2-1, TIF file.
Extended Data Table 2-1
Full statistical information for Figure 2. Download Table 2-1, DOCX file.
Extended Data Figure 3-1
Basal levels of sGluA2 in spines is lower in the Fmr1 KO mice. Cumulative distributions of sGluA2 at T0 in contralateral and ipsilateral hemispheres of WT and Fmr1 KO mice. Download Figure 3-1, TIF file.
Extended Data Figure 3-2
Normalized geometric mean of dendritic shaft tdTomato intensity in IH and CH hemispheres of WT and KO mice. Download Figure 3-2, TIF file.
Extended Data Figure 3-3
Raster plots of percent sGluA2 change relative to T0 at different times following training in IH and CH of WT and KO mice. Spines are sorted by D10 changes. Download Figure 3-3, TIF file.
Extended Data Figure 3-4
Spines intensity changes WT-IH (grey, 614 spines, 36 dendrites), WT-CH (black, 631 spines, 37 dendrites), KO-IH (light blue, 616 spines, 35 dendrites). KO-CH (dark blue, 564 spines, 36 dendrites). n = 5 mice per group. Geometric mean ± SEM. Nested random effects ANOVA model. Genotype × Hemisphere × Time, F(4,12055) = 1.54, p = 0.5085. Red stars indicate CH versus IH comparisons and black stars indicate T0h versus Ti. *p < 0.05, **p < 0.01, and ***p < 0.001. b, Correlation between average (D6 and D10) spine size and spine sGluA2 change with linear fit. Pearson R values are indicated. Download Figure 3-4, TIF file.
Extended Data Figure 3-5
Proportion of spines with increased sGluA2 at 18 h that maintain the increase until day 10 in untrained WT mice (WT). N = dendrite. Two-way ANOVA with Sidak correction, Genotype, F(1,139) = 5.28, p = 0.2; Hemisphere, F(1,139) = 2.03, p = 0.16; Genotype × Hemisphere, F(1,139) = 5.28 p = 0.02, *p < 0.05. Download Figure 3-5, TIF file.
Extended Data Figure 3-6
sGluA2 levels in the WT-IH (grey), WT-CH (black), KO-IH (light blue), and KO-CH (dark blue) hemispheres in four groups based on increasing percentile rank of sGluA2 intensity in spines at T0. Number of spines in each group (IH/CH) are indicated in the legend. Geometric means ± SEM. Nested random effects ANOVA model. Genotype × Hemisphere × Time, F(4,11948) = 5.67, p = 0.0001; Genotype × Hemisphere × Group, F(3,11948) = 5.81, p = 0.0006; Hemisphere × Time, F(4,11948) = 8.58, p < 0.0001; Hemisphere × Group, F(3,11948) = 5, p = 0.002. Stars indicate T0 h versus Ti for Group 1. **p < 0.01. Download Figure 3-6, TIF file.
Extended Data Table 3-1
Full statistical information for Figure 3. Download Table 3-1, DOCX file.
Extended Data Figure 4-1
Spine formation in contralateral hemisphere of WT mice does not correlate with increased behavioral performance. a, Correlation of reaching performance and spine formation at individual training sessions for which imaging was performed 2–24 h thereafter. Small symbols represent individual mice, and dashed lines are linear regressions for each mouse. Bold blue symbols represent average spine formation and behavioral performance of n = 5 mice at each imaging session for WT and Fmr1 KO (b) mice. Bold line is the linear regression for the average values WT (R2 = 0.07, p = 0.74; KO, R2 = 0.01, p = 0.88). Error bars, SEM. Correlations (Spearman coefficient) between spine formation at different time points and behavioral performance on different days in (c) WT and (d) KO mice. Download Figure 4-1, TIF file.
In this issue
