Translational research approach to social orienting deficits in autism: the role of superior colliculus-ventral tegmental pathway

Abstract

Autism Spectrum Disorder (ASD) is characterized by impairments in social interaction and repetitive behaviors. A key characteristic of ASD is a decreased interest in social interactions, which affects individuals' ability to engage with their social environment. This study explores the neurobiological basis of these social deficits, focusing on the pathway between the Superior Colliculus (SC) and the Ventral Tegmental Area (VTA). Adopting a translational approach, our research used Shank3 knockout mice (Shank3^-/-), which parallel a clinical cohort of young children with ASD, to investigate these mechanisms. We observed consistent deficits in social orienting across species. In children with ASD, fMRI analyses revealed a significant decrease in connectivity between the SC and VTA. Additionally, using miniscopes in mice, we identified a reduction in the frequency of calcium transients in SC neurons projecting to the VTA, accompanied by changes in neuronal correlation and intrinsic cellular properties. Notably, the interneuronal correlation in Shank3^-/- mice and the functional connectivity of the SC to VTA pathway in children with ASD correlated with the severity of social deficits. Our findings underscore the potential of the SC-VTA pathway as a biomarker for ASD and open new avenues for therapeutic interventions, highlighting the importance of early detection and targeted treatment strategies.

Fig. 1. Shank3 KO mice and autistic children show deficits in social orientation.

a Schematic representation of the breeding and the social orientation test. b Upper panel: heatmaps reporting the relative position of the social stimulus during orientation test. Lower panel: violin plots reporting the time passed with the juvenile stimulus in the frontal field displayed for Shank3^+/+ (n = 20) and Shank3^−/− (n = 9) mice (unpaired t-test: p-value = 0.034). The mean and s.e.m. are indicated per group. Human sample (panels c–j) demonstrate group comparisons obtained by using t-test comparing the ASD sample (green) to 100 bootstrapped typically developing (TD) samples (blue). c Left panel: example frames of the video used for the experiment. Right panel: violin plots reporting the eye-tracking derived visual preference for social information (TD, n = 46) and ASD (n = 46) children (p-value < 0.001, median cohen D = 0.95). d Left panel: example frames of the video used for the experiment. Right panel: violin plots reporting proximity index derived from viewing a complex social scene TD (n = 38) and ASD (n = 38) children (p-value < 0.001, median cohen D = 1.04). e Symptom severity in social affect domain of Autism Diagnosis Observation Schedule (ADOS) for TD (n = 46) and ASD (n = 46) children (p-value < 0.001, median cohen D = −3.45). f Symptom severity in restricted and Repetitive Behaviors (RRB) domain of ADOS for TD (n = 46) and ASD (n = 46) children (p-value < 0.001, median cohen D = −4.25). Mullen Early Learning scales (MSEL) Developmental quotient (DQ) for TD (n = 46) and ASD (n = 46) children in g Visual reception (p-value < 0.001, median cohen D = 2.04); h Fine Motor (p-value < 0.001, median cohen D = 1.32); i Receptive Language (p-value < 0.001, median cohen D = 2.39); and (j) Expressive Language (p-value < 0.001, median cohen D = 2.21).

Fig. 2. Partial least squares correlation (PLS-C) analysis of superior colliculus (SC) whole-brain functional connectivity with age and symptom severity in 46 children with ASD.

a Selection of the superior colliculus (SC) as region of interest (ROI). b Scatter plot depicting correlation (Pearson) between brain (x-axis) and design (y axis) scores. c The design saliences from the significant latent variable (LV) demonstrate a pronounced positive effect of the level of autistic symptoms and a negligible negative effect of age. d The brain salience pattern reveals regions of decreased connectivity with the SC (blue, negative bootstrap ratio (BSR)) in children exhibiting higher symptom severity. Or more severe ASD symptoms correlate with reduced connectivity between the SC and regions integral to attention and reward processing, including the ventral tegmental area (VTA), pulvinar (PUL), ventral anterior cingulate cortex (vACC), precuneus (PREC), primary visual cortex (VIS), left amygdala (AMY), and vermis (VERM).

Fig. 3. Shank3 KO mice showed reduced activation in neurons controlling orienting response.

a Schema reporting the viral injections (AAVrg-Ef1α-mCherry-IRES-Cre in the VTA and AAV-hSyn1-dlox-jGCaMP7f(rev)-dlox-WPRE in the SC) and the GRIN lens implantation above the SC. b Right panel: schema of the free social interaction test. Left panel: example traces of calcium signals recordings (z-score) from SC-VTA-projecting neurons during the free social interaction test. c Intrinsic parameters of the calcium transients recorded in Shank3^+/+ (in blue, 119 neurons from 9 mice) and Shank3^−/− (in orange, 118 neurons from 9 mice): frequency (Mann-Whitney test: p-value < 0.001), decay time (Mann-Whitney test: p-value = 0.04), and amplitude (Mann-Whitney test: p-value = 0.85). d interneuronal correlation. Left panel: example of estimation of partial correlation networks using EBIC selection (hyperparameter = 0.5). Right panel: violin plots reporting the percentage of significant correlations for a neuron in Shank3^+/+ (in blue, 119 neurons from 9 mice) and Shank3^−/− (in orange, 118 neurons from 9 mice) brains during habituation phase (Mann-Whitney test: p-value < 0.001) and interaction with a juvenile (Mann-Whitney test: p-value < 0.001). e Neural responses to social orientation. Heatmaps reporting the peri-event time histogram (PETH) of normalized calcium signals (mean z-score) for SC-VTA-projecting neurons in Shank3^+/+ (in blue, 119 neurons from 9 mice) and Shank3^−/− (in orange, 118 neurons from 9 mice) centered on ipsi- and contra-recorded orientations. f AUC (periods [−2:0] and [0:2]) calculated from the mean z-score reported in e per Shank3^+/+ and Shank3^−/− neurons (two-way ANOVA with Bonferroni post-hoc comparison. Ipsi: genotype main effect F_{(1, 235)} = 2.580, p  =  0.109, period main effect F_{(1, 235)} = 21.34, p  < 0.001, genotype x period interaction F_{(1, 235)} = 23.19, p  < 0.001. Contra: genotype main effect F_{(1, 235)} = 0.7460, p  =  0.389, period main effect F_{(1, 235)} = 8.489, p  = 0.004, genotype x period interaction F_{(1, 235)} = 0.09, p  = 0.764). g Barplot reporting the percentage of responding neurons according to the increase of AUC (AUC^{[0; 2]} - AUC^{[−2; 0]}) per Shank3^+/+ and Shank3^−/− group. Violin plots are reporting the mean +/− s.e.m. as error bars.

Fig. 4. Differences in intrinsic electrophysiological properties in Shank3 deficient neurons.

a Schema reporting the viral injections (AAVrg-Ef1α-mCherry-IRES-Cre in the VTA and AAVrg-Ef1α-mCherry- IRES-Cre in the SC). b Example traces at 300 pA depolarizing current injection for Shank3^+/+ (in blue) and Shank3^−/− (in orange) neurons. c Number of action potentials (N.APs) across increasing depolarizing current steps (0–500 pA) for Shank3^+/+ (in blue, n = 15, mice = 3) and Shank3^−/− (in orange, n = 11, mice = 3, two-way ANOVA with Bonferroni post-hoc comparison, genotype main effect F_[1,14] = 9.49, p  =  0.0081, current steps main effect F_{(10, 140)} = 45.58, p  < 0.001, genotype x current step interaction F_[10,96] = 11.57, p  < 0.001). Intrinsic properties of recorded neurons: d Capacitance (Cp, unpaired t-test: p = 0.070, Shank3^+/+ = 15, Shank3^−/− = 12); e Input resistance (unpaired t-test: p < 0.0001, Shank3^+/+ = 14, Shank3^−/− = 12); f Resting membrane potential (unpaired t-test: p = 0.0037, Shank3^+/+ = 15, Shank3^−/− = 12); g After-hyperpolarization current (AHP, unpaired t-test: p = 0.012, Shank3^+/+ = 15, Shank3^−/− = 12); h Rehobase (unpaired t-test: p = 0.058, Shank3^+/+ = 6, Shank3^−/− = 8). Each graph reports the mean +/− s.e.m. as error bars.

Fig. 5. Relation between deficits severity and the SC to VTA pathway in rodents and humans.

Relation between percentage of significant correlations for a neuron and interaction time during (a) social interaction (during social interaction phase) or (b) future interaction (during habituation phase) for Shank3^+/+ (in blue, 119 neurons from 9 mice) and Shank3^−/− (in orange, 118 neurons from 9 mice) mice. The mean per each mouse is reported. c Selection of the superior colliculus (SC) and ventral tegmental area (VTA) as region of interest (ROI). Panels (d–k) display scatterplots of SC-VTA functional connectivity (x-axis) against various measures (y-axes) encompassing behavioral symptoms, eye-tracking metrics of visual attention to social stimuli, and developmental outcomes. All panels depict the results of the Pearson correlation, the panel d included Spearman correlation. A median absolute deviation (MAD) threshold of 3 was used to identify and exclude outliers in the correlation analyses. Of note, none of the results changed significance status due to the outlier removal. Panel (d) shows the total level of symptoms evaluated by ADOS (n = 45); panel (e) percentage of time spent on fixating social stimuli (n = 45); panel (f) total number of saccades, n = 40; panel (g) number of ambient fixations, n = 41; panel (h) number of focal fixations, n = 42. Mullen Early Learning Composite Score rate of change is depicted in (i) for the total score, n = 28, (j) for the receptive language subdomain, n = 28, and (k) for the fine motor subdomain, n = 28. All panels depict the results of the Pearson correlation.