Figure 1.
Increased formation and stabilization of dendritic spines in L5 pyramidal neuron apical dendrites of MECP2-duplication mice. A, Example brightfield image of chronic cranial windows over M1 at two months postoperative, showing the well-defined vasculature and pale gray matter characteristic of high-quality preparations. B, left, Image demonstrating the microstimulation experiment, performed post hoc in experimental mice at the end of imaging. A bipolar stimulating electrode was lowered ∼750 μm into the window at nine sites in a 1000-μm grid. Right, Example map of motor cortex generated by microstimulation. Forepaw and hindpaw twitches were generated at low currents in all stimulated cortices (n = 10 mice), confirming localization to M1. C, The dendritic tree of a GFP-labeled L5 complex-tufted pyramidal neuron in area M1 imaged by in vivo two-photon microscopy. Apical tuft terminal dendrites are targeted for time-lapse imaging (yellow box). D, Experiment paradigm: sample images of dendritic segments imaged at baseline (left), 4 d later (middle), and 8 d later (right). Some mice were trained on the rotarod daily while others were left to rest in the cage between D1 and D5. Top, WT controls. Bottom, MECP2-duplication animals. White arrows point to spines present at baseline, green to newly-formed spines, light-green to stabilized spines, and pink to non-stabilized spines. E, Spines formed per 100 μm with and without motor training in terminal dendritic branches of MECP2-duplication animals (orange, training: n = 10 animals, 300 spines formed, 82 dendritic branches; no training: n = 4 animals, 56 spines formed, 27 dendritic branches) and WT littermate controls (black, training: n = 10 animals, 276 spines formed, 99 dendritic branches; no training: n = 6 animals, 42 spines formed, 35 dendritic branches). Squares and circles depict individual data points for trained and untrained animals, respectively. Effect of genotype: ** p = 0.003, F(1,26) = 10.3; effect of training: p = 0.77; interaction: p = 0.7; two-way ANOVA across animals); p values in plot show Tukey-corrected pairwise ANOVA comparisons. F, Newly-formed spines stabilized per 100 μm in each genotype. Effect of genotype: ***p = 0.0003, F(1,26) = 17; training: p = 0.7; interaction: p = 0.88; p values in plot show Tukey-corrected pairwise ANOVA comparisons. G–K, Dendritic spines formed, stabilized, and eliminated, and overall spine turnover rate in L5 apical dendritic arbors with data pooled between trained and untrained animals, in MECP2-duplication mice (orange: n = 14 animals, 356 spines formed, 109 dendritic branches) and WT littermate controls (black: n = 16 animals, 318 spines formed, 134 dendritic branches). Squares and circles depict individual data points for trained and untrained animals, respectively. G, Spines formed per 100 μm, **p = 0.003, Mann–Whitney U test. H, Newly-formed spines stabilized per 100 μm, ***p = 0.0007. I, Baseline spines eliminated between experiment days 1 and 5, p = 0.17. J, baseline spines eliminated between experiment days 5 and 8, p = 0.12. K, Overall spine turnover rate (spines formed + spines eliminated/2 × total dendritic length) in 4 d (from day 1 to day 5), *p = 0.01. L, Mean per-trial rotarod performance (time spent on the accelerating rotarod before falling) across animals in MECP2-duplication animals (orange, n = 17) and WT controls (black, n = 22). This panel contains a larger sample size than other panels, because mice excluded from imaging analysis because of poor imaging quality but who still underwent rotarod training were included; **p = 0.009, effect of genotype, F(1,37) = 7.6, repeated-measures ANOVA. Effect of time: p < 0.0001, F(15,37) = 29.8; interaction: p = 0.47, F(15,37) = 0.99. M, Scatter plot of learning associated spines that stabilized versus rotarod performance per animal, in imaged MECP2-duplication mice (orange squares) and WT controls (black squares); *p < 0.05, r2 = 0.26, n = 20 mice pooled across genotypes, Pearson correlation, Student’s t test. Error bars indicate mean ± SEM. See Extended Data Figure 1-1 for visualization of data as estimation plots with bootstrap-estimated differences between groups (see Materials and Methods).