PMC

New synaptic clusters depend on Lis1 levels.
A–F. Hippocampal cultures from Lis1^+/+ (A,C,E) & Lis1^+/− (B,D,F), analysed at DIV7 (A,B), DIV14 (C,D) and DIV21 (E,F) after immunostaining of pre- (vGlut1) & post- (PSD95) synaptic components. Only puncta were counted with clear overlap of the two markers (circles). Scale bar 10 µm.
G. Mean synaptic cluster density of Lis1^+/+ and Lis1^+/− neurons at DIV7 (p = 0.01), DIV14 (p = 0.05) & DIV21. (n = 3–4 animals/genotype; 5–10 neurons per animal).

Rho kinase inhibitors can rescue Actin polymerization defects in Lis1^+/− neurons.
A–C. Incorporation of fluorescently labelled actin monomers into filamentous actin in spines of mouse hippocampal neurons was visualized using FRAP at DIV12–14. GFP-actin was photobleached from the spines and time-lapse imaging followed the fluorescence recovery of F-actin. Lis1^+/− neurons were treated with 10 µM HA1077 + 10 µM Y27632 (HY) for 1 h before imaging (Lis1^+/− HY). Scale bar 4 µm.
D. The averaged recovery curves from Lis1^+/+ (blue), Lis1^+/− (pink) and Lis1^+/− HY (yellow) revealed a diminished recovery in Lis1^+/− neurons, but significant rescue in Lis1^+/− HY. (n_+/+ = 35 total protrusions, n_+/− = 25 total protrusions, n_+/−HY = 18 total protrusions).

Model: Increased RhoA activity in Lis1^+/− neurons impacts dendritic protrusion motility with consequences for proper circuit formation.
A. In Lis1 wild type neurons, appropriate RhoA levels permit protrusive motility of filopodia, open to making early synaptic connections. Filopodia with weak and non-functional connections get eliminated as part of circuit maturation and refinement. Some of these filopodia with early synaptic connections morph into spines.
B. Developing Lis1^+/− neurons exhibit fewer filopodia and elevated GTP-RhoA levels. High RhoA activity impairs filopodial protrusive motility, causing them to be rigid and overly stable, and synaptic connections accumulate more slowly. Overstability prevents elimination and appropriate pruning. Absence of pruning ends in accumulation of dendritic protrusions with synapses, which results in spine densities that are similar between adult mutants and wild types.
C. Lis1^+/− neurons treated with Rho kinase inhibitors display restored protrusive motility and length in a subset of filopodia. This allows formation of proper connections and pruning.

Lis1^+/− mutant mice exhibit asocial behaviour in a 3-chamber social interaction test. Abbreviation: S1, stranger 1; M, medial; E, empty; S2, stranger 2.
A–D. In the first phase (A,B; p = 0.01), the mouse is presented in one chamber with an empty cup and in another chamber with a strange mouse in the cup (stranger 1). The second phase of the test (C,D; p = 0.01) examines social novelty where the test mouse is presented with a novel strange mouse (stranger 2) in the previously empty cup. Total time spent in each compartment is measured (A,C), as well as time spent in close proximity/interaction with a cup/stranger mouse (B,D). Lis1^+/− mice spent significantly more time interacting with an empty cup (B) and failed to show interest in social novelty (D).
E. As a control, Lis1^+/− mice performed equally well as Lis1^+/+ mice upon testing olfaction (n_+/+ = 12 animals, n_+/− = 10 animals).
F. Quantification of buried marbles shows no difference between Lis1^+/+ and Lis1^+/− animals.

Reduced spine density in young pyramidal neurons of Lis1^+/− mice.
A–D. Golgi-stained hippocampal CA1 pyramidal neurons in brain sections of control (A, B) and Lis1^+/− (C, D) P21 litter-mates. Scale bar 50 µm (A, C), 10 µm (B, D).
E. Quantification of spine density along dendrites of hippocampal CA1 pyramidal neurons in Lis1^+/+ and Lis1^+/− mice (n = 110 dendritic segments per genotype; p = 0.0001).
F–I. Two-dimensional projection of a three-dimensional stack of dendritic branches in Lis1^+/+ (F,G) and Lis1^+/− (H,I) barrel cortex. Scale bar 15 µm (F,G).
J. Spine density quantification in barrel cortex (n = 3/genotype; p = 0.009).
K. Neurolucida drawings of pyramidal neurons representative for both genotypes.
L,M. Quantification of the number of intersections per circle. (L) Significant differences were found in the circles ranging from 50 to 120 µm of apical dendrites in Lis1^+/+ compared to Lis1^+/− (n = 10 neurons/genotype; p = 0.001). (M) Significant differences were found in the circles ranging from 60 to 160 µm of basal dendrites in Lis1^+/+ compared to Lis1^+/− (p = 0.001).
N. Golgi-stained hippocampal CA1 pyramidal neurons in sections from Lis1^+/+ and Lis1^+/− mice collected 24 h after i.p. injection of saline or Y-27632 (n = 3/genotype/treatment; 60 dendritic segments/animal). Scale bar 10 µm.
O. Quantification of spine density along apical CA1 pyramidal dendrites in Lis1^+/+ and Lis1^+/− mice treated with saline or Y-27632 for 24 h (p = 0.01).

Lis1cko mutant mice exhibit deficits in spine density, synaptic cluster formation and social novelty recognition.
A,B. Images of Golgi-stained hippocampal CA1 pyramidal neurons in Lis1^flox/+ (A) and Lis1cko (B) P28 litter-mates.
C. Quantification of spine density (n = 70 dendritic segments per genotype; p = 0.0001; scale bar 50 µm).
D. Neurolucida drawings of pyramidal neurons representative for Lis1^fl/+ and Lis1cko.
E,F. Quantification of the number of intersections per circle revealed no differences in apical or basal dendrites in Lis1^fl/+ compared to Lis1cko animals (P28; n = 10 neurons/genotype).
G–J. Hippocampal cultures from Lis1^flox/+ (infected with inactive, ▵Cre lentivirus) (G,I) and Lis1^cko/+ (Lis1^flox/+ neurons infected with Cre lentivirus) (H,J), analysed at DIV7 (G,H), and DIV21 (I,J) after immunostaining of pre- (vGlut1) & post- (PSD95) synaptic components. Lis1^flox/+ cultures are infected on DIV1 with either EGFP-Cre or EGFP-inactiveCre lentivirus. GFP is used to detect Cre in infected neurons.
K. Mean synaptic cluster density of Lis1^flox/+ and Lis1^cko/+neurons at DIV7 (p = 0.001), DIV14 (p = 0.001) & DIV21. (n = 3–4 animals/genotype; 5–10 neurons per animal; scale bar 10 µm).
L–O. Lis1cko mutant mice exhibit deficits in social behaviour testing. Both control and Lis1cko mice show preference for a stranger mouse in the cup (S1) as opposed to an empty cup (E). In contrast to controls, Lis1cko mice show no preference for social novelty (stranger 2) in the previously empty cup (n_fl/+ = 11 animals, n_cko = 10 animals). Abbreviation: S1, stranger 1; M, medial; E, empty; S2, stranger 2.

Downregulation of RhoA in Lis1^+/− neurons rescues motility in a subset of filopodia.
A,B. Lis1^+/+ (A) and Lis1^+/− (B) DIV2 hippocampal neurons stained with TuJ1 (red) and GFP for myr-Venus. Scale bar 10 µm.
C. Filopodial length quantification shows shorter protrusions in Lis1^+/− neurons at DIV2 (p = 0.01).
D. Filopodial density is reduced 60% in Lis1^+/− neurons at DIV2 (p = 0.001). (n_+/+ = 10 neurons, n_+/− = 7 neurons; 10–15 filopodia/neuron).
E. Number of motile filopodia at DIV2 presented as a percentage of total filopodia per cell (p = 0.001).
F,G. Length Motility Index (F) and Motility (G) were significantly less in Lis1^+/− neurons (p = 0.001). (n_+/+ = 10 neurons, n_+/− = 7 neurons; 10–15 filopodia/neuron).
H–H''. Time lapse video of a DIV2 myrVen;Lis1^+/− hippocampal neuron. Note absence of dynamic filopodia in this part of dendrite.
I–I''. Time lapse video of the same dendrite after 1 h incubation with ROCK inhibitor cocktail. Filopodia marked with arrowheads extend and retract. Scale bar 5 µm (H–I″).
J–L. Filopodia length increases in Lis1^+/−;myrVen neurons after 4 h incubation with 10 µM Y27632. H–L images are color inverted. Scale bar 2 µm (J–L).
M. Quantification of M (motility) and LMI (length motility index) shows increased rates in filopodia of Lis1^+/− neurons treated with ROCK inhibitors, 10 µM HA1077 + 10 µM Y27632 (HY) (n_{total filopodia} = 26; p = 0.0005). 25–30% of non-motile filopodia in Lis1^+/−;myrVen neurons exhibited rescued motility post drug treatment.
N. Filopodia length is significantly increased in Lis1^+/−;myrVen neurons after 4 h incubation with 10 µM Y27632 (n = 250 filopodia/genotype; p = 0.01).

In vivo two photon time lapse imaging in adolescent and young adult Lis1^+/+ and Lis1^+/− mice.
A. Lis1^+/+ neurons have multiple filopodia that are very dynamic (encircled). Arrows indicate other stable filopodia. Images are color inverted. (n_+/+ = 4 animals, n_+/− = 3 animals; n_+/+ = 269 total spines, n_+/− = 205 total spines). Scale bar: 2 µm.
B. Lis1^+/− neurons have fewer dynamic filopodia (e.g. filopodia that neither retract nor extend during imaging period are encircled in red, blue circles depict spines for comparison).
C–F. Quantification of filopodia and spine dynamics in adolescent (P21) and young adult (P30) Lis1^+/+ and Lis1^+/− mice.
C. Filopodia elimination (p = 0.003) and turnover (p = 0.05) rates are lower in Lis1^+/− mice over 1 h (1st view at 0′ followed by 2nd view at 60′), whereas formation rates were similar (p = 0.9).
D. Filopodia stability (p = 0.003) increased over 2 days (1st view at P21, 2nd view of the same region at P23) in adolescent Lis1^+/− mice compared to control mice. In contrast, the elimination (p = 0.003) rate in Lis1^+/− mice decreased over 2 days relative to controls. Formation rates trended lower but did not reach significance (p = 0.2). (n_+/+ = 4 animals, n_+/− = 3 animals; n_+/+ = 182 total filopodia, n_+/− = 88 total filopodia in C; n_+/+ = 173 total filopodia, n_+/− = 135 total filopodia in D).
E. Quantification of spine dynamics in young adult mice over 48 h (1st view at P21, 2nd view of the same region at P23) reveals increased stability (p = 0.007), decreased elimination (p = 0.007) and no difference in formation rates. (n_+/+ = 4 animals, n_+/− = 3 animals; n_+/+ = 246 total spines, n_+/− = 220 total spines).
F. Quantification of spine dynamics in adult mice over a 2-week period (1st view at P30, 2nd view of the same region at P44) show increased stability (p = 0.009) and decreased elimination (p = 0.009) rates in Lis1^+/− mutants, while again, formation showed a reduced trend (p = 0.43).