Effects of heterozygous deletion of autism-related gene Cullin-3 in mice

Abstract

Autism Spectrum Disorder (ASD) is a developmental disorder in which children display repetitive behavior, restricted range of interests, and atypical social interaction and communication. CUL3, coding for a Cullin family scaffold protein mediating assembly of ubiquitin ligase complexes through BTB domain substrate-recruiting adaptors, has been identified as a high-risk gene for autism. Although complete knockout of Cul3 results in embryonic lethality, Cul3 heterozygous mice have reduced CUL3 protein, demonstrate comparable body weight, and display minimal behavioral differences including decreased spatial object recognition memory. In measures of reciprocal social interaction, Cul3 heterozygous mice behaved similarly to their wild-type littermates. In area CA1 of hippocampus, reduction of Cul3 significantly increased mEPSC frequency but not amplitude nor baseline evoked synaptic transmission or paired-pulse ratio. Sholl and spine analysis data suggest there is a small yet significant difference in CA1 pyramidal neuron dendritic branching and stubby spine density. Unbiased proteomic analysis of Cul3 heterozygous brain tissue revealed dysregulation of various cytoskeletal organization proteins, among others. Overall, our results suggest that Cul3 heterozygous deletion impairs spatial object recognition memory, alters cytoskeletal organization proteins, but does not cause major hippocampal neuronal morphology, functional, or behavioral abnormalities in adult global Cul3 heterozygous mice.

Fig 1. Cul3 heterozygous deletion in Cul3 global knockout mice.

A) Diagram of Cul3 global knockout mice model and Cre-LoxP knockout strategy. Exons 3–8 of the Cul3 gene are represented along with LoxP, neomycin-resistance, and frt sites. Expression of Cre-recombinase in the germline results in recombination and loss of Cul3 expression. B) Representative immunoblots of cortical and hippocampal homogenates for CUL3, KCTD13, RhoA and α-tubulin (WT N = 6 mice, male; HET N = 6 mice, male). C) Graphs of pooled data for Cul3, KCTD13, RhoA and α-tubulin. “HET” represents Cul3 heterozygous deletion. Cul3 expression decreased to ∼56% in cortical neurons and decreased to ∼49% in hippocampal neurons respectively (F = 3.21, P = 0.0018 for Cul3 in cortex; F = 18.01, P = 0.0002 for Cul3 in hippocampus). Cul3 heterozygous deletion did not significantly alter KCTD13 or RhoA expression in cortex or hippocampus (F = 3.492, P = 0.8296 for KCTD13 in cortex; F = 2.671, P = 0.9163 for KCTD13 in hippocampus; F = 1.001, P = 0.8833 for RhoA in cortex, F = 11.16, P = 0.7174 for RhoA in hippocampus). *P<0.05, **P<0.01, ***P<0.001, unpaired student’s t-test; graphs depict mean ± SEM.

Fig 2. Body weight, motor skills, sensory ability, and anxiety-like behavior in Cul3 global knockout mice.

A) Body weights measured periodically over time do not differ substantially between Cul3 HETs and WT littermates, a main effect of sex or genotype was observed; P < 0.0001 (9–10 wks, WT = 19, HET = 10 (male, WT = 10, HET = 7; female, WT = 9, HET = 3) , P = 0.103 for male, P = 0.9974 for female. 11–12 wks, WT = 15, HET = 11 (male, WT = 10, HET = 6; female, WT = 5, HET = 5) , P = 0.9009 for male, P = 0.1927 for female. 13–15 wks, WT = 18, HET = 13 (male, WT = 9, HET = 7; female, WT = 9, HET = 6), P = 0.8276 for male, P = 0.0058 for female. 16–17 wks, WT = 20, HET = 19 (male, WT = 10, HET = 9; female, WT = 10, HET = 10), P = 0.6733 for male, P = 0.0686 for female. 20–21 wks, WT = 20, HET = 20 (male, WT = 10, HET = 10; female, WT = 10, HET = 10), P = 0.258 for male, P = 0.4851 for female). B) Locomotor activity measured over 2 h in clean mouse cages using beam break counts was unchanged in Cul3 HETs vs. WT littermates, a main effect of bin was observed; P < 0.0001 (P = 0.5895, male, WT = 10, HET = 10; female, WT = 10, HET = 10). C) No difference in total distance travelled in an open field arena (F = 1.499, P = 0.8636, male, WT = 10, HET = 10; female, WT = 10, HET = 10). D) No significant difference in time to fall from the accelerating rotarod was observed between Cul3 HETs and WT littermates. A main effect of trail was observed; P < 0.0001 (P = 0.159, male WT = 10, HET = 9; female, WT = 10, HET = 10). E) No difference in time to paw-lick response on the hotplate sensory task between Cul3 HETs and WT littermates (F = 1.28, P = 0.2635, male, WT = 10, HET = 9; female, WT = 10, HET = 10). F) Time spent in the mouse scent interaction zone was unchanged between Cul3 HETs and WT littermates. A main effect of target was observed; P < 0.0001 (P = 0.9194 for water, P = 0.883 for mouse scent, male, WT = 10, HET = 9; female, WT = 10, HET = 10). G) Time spent in center zone of an open field arena is increased in Cul3 HETs vs. WT littermates (F = 1.533, P = 0.0283, male, WT = 10, HET = 10; female, WT = 10, HET = 10). H) Distance travelled in center zone of an open field arena is increased in Cul3 HETs vs. WT littermates (F = 1.177, P = 0.042, male, WT = 10, HET = 10; female, WT = 10, HET = 10). I) No change in time spent in open arms of elevated plus maze (F = 1.818, P = 0.3832, male, WT = 10, HET = 10; female, WT = 10, HET = 10). J) No change in time spent in dark chamber in dark/light test (P = 0.8892 for time in dark chamber, male, WT = 10, HET = 10; female, WT = 10, HET = 10). K) Startle response in arbitrary units in response to white noise auditory stimulus pulse of various volumes in decibels (dB). A main effect of trail was observed; P < 0.0001 (P>0.9999 for no stimulus group, P>0.9999 for 80 dB group, P = 0.7174 for 90 dB group, P>0.9999 for 100 dB group, P = 0.9423 for 110 dB group, P = 0.6434 for 120 dB group, male, WT = 10, HET = 10; female, WT = 10, HET = 10). L) Prepulse inhibition of startle (%) as a function of the prepulse volume above background noise (4, 8, & 16 dB above background). A main effect of trail was observed; P < 0.0001 (P = 0.0531 for PP4 group, P>0.9999 for PP8 group, P>0.9999 for PP16 group, male, WT = 10, HET = 10; female, WT = 10, HET = 10). *P<0.05, **P<0.01, ***P<0.001, unpaired student’s t-test, 2-way ANOVA or 3-way ANOVA test, compared between genotypes; graphs depict mean ± SEM. (See S1 Table in S1 File for detailed statistics).

Fig 3. Repetitive behaviors and social interaction and approach.

A) No difference in time spent grooming between Cul3 HETs and WT littermates (F = 1.778, P = 0.8323, male, WT = 10, HET = 10; female, WT = 10, HET = 10). B) No difference in number of marbles buried in a marble burying task (F = 2.041, P = 0.2847, male, WT = 10, HET = 10; female, WT = 10, HET = 10). C) No difference in time spent in direct physical contact during a reciprocal social interaction task examining sex-matched WT/WT pairs and HET/HET pairs of mice in an open arena (F = 1.633, P = 0.2316, male, WT = 5 pairs, HET = 5 pairs; female, WT = 5 pairs, HET = 5 pairs). D) No difference in time spent in close proximity during a reciprocal social interaction task of genotype and sex-matched WT/WT and HET/HET pairs (F = 1.438, P = 0.1863, male, WT = 5 pairs, HET = 5 pairs; female, WT = 5 pairs, HET = 5 pairs). E) No difference in time spent in proximity (approaching) a caged social target in an open arena (F = 1.035, P = 0.8911, male, WT = 10, HET = 10; female, WT = 10, HET = 10). *P<0.05, **P<0.01, ***P<0.001, unpaired student’s t-test; graphs depict mean ± SEM.

Fig 4. Spatial object recognition and novel object recognition tasks.

A) Experimental schematic of the sequential Spatial Object and Novel Object recognition tasks. B) Initial time spent interacting with objects during baseline habituation is equivalent for Cul3 HETs and WT littermates (P = 0.1533). C) WT mice exhibit spatial object recognition learning by spending significantly more time with the object (Obj C) in a new spatial location compared to objects in same location as before. A main effect of object was observed; P = 0.0092 (P = 0.0048 for Obj A vs. Obj C, P = 0.05 for Obj B vs. Obj C). Cul3 HETs show no spatial object recognition learning by showing no preference for the newly localized object (P = 0.6016 for Obj A vs. Obj C, P = 0.9245 for Obj B vs. Obj C). D) Cul3 HETs and WT littermates both demonstrate significant novel object recognition with significantly more time spent with the novel object compared to the two familiar objects (P<0.0001). A main effect of object was observed; P<0.0001. Male, WT = 9, HET = 9; female, WT = 10, HET = 10. *P<0.05, **P<0.01, ***P<0.001, 2-way ANOVA with Tukey’s post-hoc analyses; graphs depict mean ± SEM.

Fig 5. Sholl analysis and spine density analysis in mouse hippocampal neurons.

A) Representative images of WT and Cul3 HET global knockout hippocampal neuronal apical dendritic spines. Scale bar, 20 μm. B) Spine density analysis of WT and Cul3 HET demonstrates no change of apical dendritic spine density except the stubby spine density decrease in Cul3 HET mutant neurons (Total spine, F = 1.16, P = 0.0781; thin spine, F = 1.285, P = 0.3683; stubby spine, F = 1.641, P = 0.0267; mushroom spine, F = 1.319, P = 0.2034; filopodia spine, F = 1.015, P = 0.2525. WT N = 50/10 cells/mice; HET N = 50/10). *P<0.05, **P<0.01, ***P<0.001, unpaired student’s t-test; graphs depict mean ± SEM. C) Representative images of WT and Cul3 HET global knockout mice hippocampal neuronal dendritic tree. Scale bar, 20 μm. D) Sholl analysis of WT and Cul3 HET knockouts demonstrates decreased dendritic branching in Cul3 HET mutant neurons, a main effect of genotype was observed (P = 0.002, WT N = 30/10 cells/mice; HET N = 30/10). *P<0.05, **P<0.01, ***P<0.001, 2-way ANOVA test; graphs depict mean ± SEM. E) Number of dendrites (left, F = 1.136, P = 0.0669), total dendritic length (middle, F = 1.62, P = 0.1832), and average dendritic length (right, F = 1.219, P = 0.6299) was unchanged in WT and Cul3 HET global knockout hippocampal neurons (WT N = 30/10 cells/mice; HET N = 30/10). *P<0.05, **P<0.01, ***P<0.001, unpaired student’s t-test; graphs depict mean ± SEM.

Fig 6. Insufficiency of Cul3 does not affect baseline synaptic transmission, the presynaptic transmission, or LTP in hippocampus.

A) Representative traces of field recording fEPSPs in hippocampal Schaffer-collateral pathway from Cul3 HETs and WT littermates. B) Input output curves of evoked fEPSP slope versus stimulus intensity demonstrated no significant difference (P = 0.6141) in baseline evoked synaptic transmission between WT and HETs. C) No significant differences (P = 0.6434) were found in relationship between stimulus amplitude and fiber volley. D) Paired-pulse facilitation across different interstimulus intervals revealed no effect of genotype (P = 0.2816). E) Induction of LTP using high-frequency stimulation (3 x tbs) revealed no significant differences between WT and HETs (P = 0.0913). (WT/HET = 8 mice/ group, 3–6 slices/ mouse). *P<0.05, **P<0.01, ***P<0.001, 2-way ANOVA with repeated measures; graphs depict mean ± SEM.

Fig 7. Cul3 heterozygous deletion leads to increased mEPSC frequency in CA1 pyramidal neurons and no change in NMDA/AMPA ratio in mPFC.

A) Representative traces of mEPSCs in dorsal hippocampal CA1 neurons from Cul3 HETs and WT littermates. (WT N = 50/14 cells/mice; HET N = 30/11 cells/mice). B) Cumulative probability and summarized bar graph (inset) showing that the amplitude of mEPSC is comparable between genotypes (F = 1.169, P = 0.968). C) Cumulative probability and summarized bar graphs (inset) showing that the frequency of mEPSC is significantly increased in Cul3 HET neurons compared with WT neurons (F = 2.978, P = 0.0023). D) Representative traces of eEPSCs recorded in mPFC layer 5 neurons following stimulus in layer 2/3. AMPAR current was measured as the peak response when the cells were held at -70 mV; NMDAR current was measured as the mean current over a 5-ms window, 50 ms after stimulation in L2/3. (WT N = 40/6 cells/mice; HET N = 37/6 cells/mice). E) Summarized bar graphs showing that NMDA/AMPA current ratio is comparable between genotypes (F = 1.624, P = 0.7178). *P<0.05, **P<0.01, ***P<0.001, unpaired student’s t-test; graphs depict mean ± SEM.

Fig 8. Cul3 heterozygous deletion in mouse hippocampus induces differential expression of proteins.

A) Global protein changes in the hippocampus of adult (8–10 weeks old) WT vs Cul3 HET mice as represented by Venn diagram. 133 (63↑70↓) proteins were found to be statistically changed in the hippocampus of the Cul3 HET vs WT out of the total of 2,092 proteins. The Venn diagram depicts the total number of proteins identified across both groups, in addition to those proteins found to be significantly changed as a result of Cul3 HET deletion. B). Volcano plot demonstrating the distribution of the entire data set of proteins with upper limits indicating statistically significant changes depicted in red and green points. C) 2-dimensional heat map of top 37 proteins demonstrates the proteins which are increased (red) or decreased (blue). D) Principle Component Analysis (PCA) demonstrating a tight clustering of the samples across each group. E & F) Gene ontology (GO)-annotated molecular function (E) and biological processes (F) were identified for differentially expressed proteins between the WT and Cul3 HET. System analysis used the top 133 statistically significant proteins, and the resultant pie charts are indicative of the normalized percentage of proteins associated with each category within molecular functions and biological processes. G) Representative western blots (left) and pooled densitometry quantification data (right) for Tropomyosin and Transgelin-2 in hippocampal brain regions (F = 1.209, P<0.0001 for Tropomyosin, F = 4.38, P<0.0001 for Transgelin-2. Male WT = 3, HET = 3; female, WT = 3, HET = 3). *P<0.05, **P<0.01, ***P<0.001, unpaired student’s t-test; graphs depict mean ± SEM.