Syngap1 haploinsufficiency damages a postnatal critical period of pyramidal cell structural maturation linked to cortical circuit assembly

Abstract

Background: Genetic haploinsufficiency of SYNGAP1/Syngap1 commonly occurs in developmental brain disorders, such as intellectual disability, epilepsy, schizophrenia, and autism spectrum disorder. Thus, studying mouse models of Syngap1 haploinsufficiency may uncover pathologic developmental processes common among distinct brain disorders.

Methods: A Syngap1 haploinsufficiency model was used to explore the relationship between critical period dendritic spine abnormalities, cortical circuit assembly, and the window for genetic rescue to understand how damaging mutations disrupt key substrates of mouse brain development.

Results: Syngap1 mutations broadly disrupted a developmentally sensitive period that corresponded to the period of heightened postnatal cortical synaptogenesis. Pathogenic Syngap1 mutations caused a coordinated acceleration of dendrite elongation and spine morphogenesis and pruning of these structures in neonatal cortical pyramidal neurons. These mutations also prevented a form of developmental structural plasticity associated with experience-dependent reorganization of brain circuits. Consistent with these findings, Syngap1 mutant mice displayed an altered pattern of long-distance synaptic inputs into a cortical area important for cognition. Interestingly, the ability to genetically improve the behavioral endophenotype of Syngap1 mice decreased slowly over postnatal development and mapped onto the developmental period of coordinated dendritic insults.

Conclusions: Pathogenic Syngap1 mutations have a profound impact on the dynamics and structural integrity of pyramidal cell postsynaptic structures known to guide the de novo wiring of nascent cortical circuits. These findings support the idea that disrupted critical periods of dendritic growth and spine plasticity may be a common pathologic process in developmental brain disorders.

Keywords: Autism spectrum disorder; Development; Epilepsy; Intellectual disability; Mouse model; Synapse; Syngap1.

Figure 1. Pathogenic Syngap1 mutations accelerate the rate of normal neocortical pyramidal cell growth and maturation

(A) 3D reconstructed confocal imaging sections of somatosensory cortex obtained in post-fixed brain slices in Syngap1+/− and wt (PND 21 and PND 60). (B) Representative example of 3D reconstruction layer V pyramidal neurons (PND 21 and 60), scale bar = 5μm. (C–E) Graphs depicting the total length, intersections and volume of traced neurons * p<.05 within genotype group comparison, # p<.05 within age group comparison. (F) Scatter plot graphs showing the Correlation Index between cell body volume and the total length of the neuronal dendrites measured by Sholl analysis quantification. (G) Immunostaining for p-S6 (red) and NeuN (green) in L5 neurons in prefrontal cortex of PND 4 SynGAP1+/− (n=5 animals, 150 cells) mice and wild type (n=3 animals, 90 cells) littermates. * p < 0.05, Student t-Test. Values represent means; error bars indicate SEMs. Complete statistical reporting can be found in Table S1.

Figure 2. Accelerated spine formation and premature spine pruning in developing neocortical neurons in Syngap1 hets

(A) Representative dendritic branches at different time points. (B–C) Line graphs show spine density and filopodia density trajectories over time; PND14 (wt-n=5, Syngap1+/− n=4), PND21 (wt-n=6, Syngap1+/− n=6), PND30 (wt-n=5, Syngap1+/− n=5), PND60 (wt-n=5, Syngap1+/− n=4). Values represent means; error bars indicate SEMs. * p<.05 within genotype group comparison, # p<.05 within age group comparison. Complete statistical reporting can be found in Table S2.

Figure 3. The impact of whisker deprivation on dendritic protrusion dynamics in developing Syngap1 hets

(A) Whisker trimming experimental design. (B) Representative dendritic samples of layer 5 pyramidal cell tuft for Syngap1+/− and wt and conditions (un-trimmed and trimmed) in PND 21 mice, (scale bar=5μm). (C) Spine density for trimmed (wt, n=4; Syngap1+/−, n=4) and un-trimmed (wt, n=3; Syngap1+/− n=3) animals. Values represent means, error bars indicate SEMs. Multiple comparisons were evaluated within each independent factor. * p<.05. Complete statistical reporting can be found in Table S3.

Figure 4. Premature cellular maturation causes altered spine dynamics in Syngap1 hets

(A) Schematic demonstrating the sequential time-lapse recording experiment. (B) Representative dendritic segments. In vivo time-lapse imaging of the same dendritic segments over 2 and 4 h in the primary somatosensory cortex (scale bar = 5μm). (C–E) Graphs depicting filopodia density (C), proportion of stable spines over 2 and 4 h recording (D), or proportion of filopodia that converted into spines in Syngap1+/− (n=5) and wt (n=5); # p<.05. (F–I) Diagram showing time lapse recording experimental design at PND21-23 and PND30-32. (G–M) 2PLMS imaging session of the same dendritic segments over 2 days (scale bar = 5μm). (H–I) PND 21-23 dynamic properties of dendritic spines. Graphs showing the proportion of new formed spines and spine eliminated detected over 2 days each time point (H) and Turnover index (I), Student t-Test [t(5) =3.223, p=.023] Syngap1+/− (n=4) and wt (n=3) animals. (J–L) Spine formation and elimination from PND30 to PND32. Graphs showing the dynamic plasticity of spine imaged over 2 days (K) and Turnover index (L), Student t-Test [t(6) =2.46, p=.0492] Syngap1+/− (n=4) and wt (n=3) animals. Arrowheads: green = new formed; yellow = lost; red = stable spines; white = filopodia; Purple filopodia converted into spines. Values represent mean, error bars indicate SEM. LSD post-hoc test was used for multiple comparisons. Complete statistical reporting can be found in Table S4.

Figure 5. Widespread spine enlargement in Syngap1 mutants is caused by a disrupted developmental critical period

(A) Representative dendrites of PND 14 layer 2 neurons. Cumulative frequency curves depict spine head diameter in Layer 2 and Layer 5 apical branches (located in Layer 1) in PND14 Syngap1+/− and wt. (B) Representative electron microscopy (EM) images of pre- (yellow) and post- (green) synaptic area at PND14 in which post-synaptic density (PSD) area is outlined. Cumulative frequency curves depict PSD area observed in Layer 1. (C) Cumulative frequency curves depicting spine head diameter observed in Layer 5 at PND60. (D) Cumulative frequency curves showing spine head diameter of newborn (NB) spines detected during a second in vivo transcranial imaging (i.e. spines present at PND23 but not at PND21). (E) Representative dendrites of PND > 90 layer 2 neurons in Syngap1+/− and Syngap1+/fl adult-induced haploinsufficiency model. Cumulative frequency curves show that wider spine heads persist throughout adulthood in conventional Het mutants, while this effect does not occur in animals with Syngap1 haploinsufficiency induced in adulthood. Complete statistical reporting can be found in Table S5.

Figure 6. Gross anatomical long-distance synaptic connectivity appears normal in Syngap1 mutants

(A) Experimental design. (B) Representative coronal sections in the anterior to posterior axis of long-range connectivity to the mPFC. Scale bar, 1 mm. (C) Quantitative characterization of the injection sites and total number of direct-labeled inputs to the mPFC. A.U., arbitrary unit. (D) Relative hemispheric representations of the contributions of inputs form ipsilateral and contralateral sides. (E) Relative representation of inputs clustered by cortical, subcortical and limbic brain areas. (F) Relative representation of primary and secondary cortical inputs including areas from ipsilateral motor, auditory, visual and soma- to sensory cortex and contralateral motor and somatosensory cortices. Values represent means, error bars indicate SEMs. P values indicate two-tailed unpaired t-test comparisons of wt versus Syngap1+/− mice. Data are presented as a proportion of the total number of direct inputs to the mPFC between wt and Syngap1+/− mice.

Figure 7. Evidence of macroscale disruptions in long-distance intracortical synaptic connectivity in Syngap1 mutants

(A–D) Comparison of the relative strength of monosynatpic inputs to the mPFC in the ipsilateral cortical (A), contralateral cortical (B), subcortical (C) and limbic (D) areas. Values represent means, error bars indicate SEMs. P values indicate two-tailed unpaired t-test comparisons between wt (n=9) and Syngap1+/− (n=9) animals. (E) Monosynaptic inputs to the mPFC labeled with RV-GFP in wt and Syngap1+/− for brain areas located in cortical (Layer 5 pyramidal neurons in the somatosensory cortex), subcortical (thalamus), and limbic (amygdala) regions. Pia is indicated with a dotted line. Scale bar shown in all figures, 200 μm. (F) Cumulative frequency curves of numbers of primary somatosensory neurons in both groups; p<0.0001; two-sample Kolmogorov-Smirnov test. (G) Relative strength in connectivity represented as a measure of the location of specific brain regions. Location of the brain area biased the strength of connectivity of cortical inputs to the prefrontal cortex (n = 19 brain areas, p=0.006 for linear regression of connectivity ratio and anatomical location). The connectivity ratio was established as the differential inputs between wt and Syngap1+/− relative to the number of inputs in wt. Grey indicates 95% confidence interval for the regression line.

Figure 8. Early, but not late, postnatal repair of a pathogenic Syngap1 mutation prevents cognitive and behavioral abnormalities

Four groups of mice (Syngap1+/+; CAG-CreER−/−, Syngap1+/lox-stop; CAG-CreER−/−, Syngap1+/+; CAG-CreER+/−, Syngap1+/loxstop; CAG-CreER+/−) were administered either a single injection (SQ) of tamoxifen (TM) (20 mg/kg) on PND1 or five daily injections (IP) of TM starting on PND21. Behavioral batteries consisting of elevated plus maze (EPM), open field (OF) and remote (30d) contextual fear conditioning (FC) measured by activity suppression ratio (1^st 2 min of the test)/[(1^st 2 min of the test)+ (1^st 2 min of the training)] tests, respectively, were conducted starting at 12 wks. Hippocampi were dissected and processed for immunoblot of Syngap protein levels in the four genotype groups after conclusions of the behavioral batteries. (A) Genotypic reversal of Syngap1 is induced in Syngap1+/lox-stop; CAG-CreER+/− mice administered TM at PND1. Densitometric values were normalized to β-tubulin levels and subsequently transformed. (B–D) Performances of Syngap1+/lox-stop; CAG-CreER+/− mice are rescued relative to that of Syngap1+/+; CAG-CreER+/− mice in EPM, OF, and FC tests. (E) Genotypic reversal of Syngap1 is induced in Syngap1+/lox-stop; CAG-CreER+/− mice administered TM at PND21. Densitometric values were normalized to β-tubulin levels and subsequently transformed. (F–H) Performances of Syngap1+/lox-stop; CAG-CreER+/− mice are not rescued relative to that of Syngap1+/+; CAG-CreER+/− mice in EPM, OF and FC tests. Multiple comparisons within each independent factor were calculated. *p<.05, **p<.01, ***p<.001, difference within Cre, #p<.05, ##p<.01, ###p<.001, difference within genotype. Number of animals are indicated within bar graphs. Complete statistical reporting can be found in Table S6.