Fig. 1

Corticopontine but not commissural layer V pyramidal neurons exhibit shorter thin dendritic spines in Itgb3 KO mice (A) Major outputs of LV mPFC pyramidal neurons include the pons for pyramidal tract neurons (corticopontine [CP] neurons; green) and the contralateral cortex for intratelencephalic neurons (commissural [COM] neurons; red). (B) The retrograde rAAV AAVrg-hSyn-EGFP was injected in the pons or contralateral mPFC to label CP or COM neurons, respectively. One CP (red) and one COM neuron (green) are shown in sagittal and horizontal views (images generated using the MouseLight interface; ID AA0261 and AA0656, https://www.janelia.org/project-team/mouselight/neuronbrowser). (C) Representative images and Imaris 3D rendering of basal dendrites from CP WT, CP Itgb3 KO, COM WT and COM Itgb3 KO LV pyramidal neurons. Arrows point to representative thin (T), mushroom (M) and stubby (S) spines. Loss of Itgb3 reduces spine length in CP neurons. Intensity profiles of the same dendrites after straightening reveal a ‘spine neck kink’ only in the COM WT and COM KO conditions, suggesting that COM neurons (WT and KO alike) exhibit spines with thinner spine necks than CP neurons (WT and KO alike). (D) Left: sample Imaris 3D rendering of thin, stubby and mushroom spines. Right: pie chart of spine type distribution (*p = 0.03; ***p = 0.0003; Chi-square test; n = 482, 532, 476 and 317 spines for CP WT, CP Itgb3 KO, COM WT and COM Itgb3 KO, respectively). (E) Density of stubby spines (two-way ANOVA; genotype effect: F (1, 90) = 3.548, p = 0.0628; neuron type effect: F (1, 90) = 13.45, ***p = 0.0004; genotype x neuron type interaction: F (1, 90) = 2.670, p = 0.1057; n = 27, 17, 31 and 19 dendritic stretches for CP WT, CP KO, COM WT and COM KO, respectively). (F) Violin plot for the length of thin spines (*p = 0.01, non-parametric Kruskal-Wallis ANOVA followed by Benjamini, Krieger and Yekutieli post-test, which corrects for multiple comparisons by controlling the false discovery rate; n = 245, 276, 242 and 157 thin spines for CP WT, CP Itgb3 KO, COM WT and COM Itgb3 KO, respectively). In each violin plot, the thick dotted line and the two thin dotted lines indicate the median and the quartiles, respectively. (G) Thin spines of CP Itgb3 KO neurons were ranked according to their length, resampled to match the number of thin spines of CP WT neurons and plotted against the ranked thin spines of CP WT neurons (straight line: CP WT vs. CP WT; open circles: CP Itgb3 KO vs. CP WT; line through open circles is a sigmoid fit; p = 0.01; Kolmogorov-Smirnov test; n = 245 and 276 thin spines for CP WT and CP Itgb3 KO, respectively; full data set in supplemental Fig. 2). (H) Histogram of the ratio between length and head width for thin spines of CP WT and CP Itgb3 KO neurons on a logarithmic scale (Log (length / width)) reveals a log-normal distribution for this morphological parameter (**p = 0.004 between CP WT and CP Itgb3 KO, parametric Brown-Forsythe and Welch ANOVA followed by Benjamini, Krieger and Yekutieli post-test; full data set in supplemental Fig. 2). Continuous lines are Gaussian fits with the indicated means (µ) and standard deviations (σ)