Marco Pagani1, Alice Bertero1,2, Alessia De Felice1, Andrea Locarno3, Ieva Miseviciute3, Stavros Trakoshis4,5, Carola Canella1,6, Elizabeth de Guzman1, Kaushtub Supekar7, Vinod Menon7, Alberto Galbusera1, Raffaella Tonini3, Michael V. Lombardo5, Massimo Pasqualetti2, and Alessandro Gozzi1
1Functional Neuroimaging Laboratory, Istituto Italiano di Tecnologia, Rovereto, Italy, 2Biology Department, University of Pisa, Pisa, Italy, 3Neuromodulation of Cortical and Subcortical Circuits Laboratory, Istituto Italiano di Tecnologia, Genova, Italy, 4Department of Psychology, University of Cyprus, Nicosia, Cyprus, 5Laboratory for Autism and Neurodevelopmental Disorders, Istituto Italiano di Tecnologia, Rovereto, Italy, 6Center for Mind and Brain Sciences, University of Trento, Rovereto, Italy, 7Stanford University, Stanford, CA, United States
Synopsis
Altered brain
functional connectivity is a hallmark finding in autism but the neural basis of
this phenomenon remains unclear. We show that a mouse line reconstituting synaptic
pruning deficits observed in postmortem autistic brains exhibits widespread functional hyper-connectivity, and that
pharmacological normalization of synaptic aberrancies completely rescues behavioral
and functional connectivity deficits. We also show that a similar connectivity
fingerprint can be isolated in human rsfMRI scans of people with autism, and
linked to overexpression of genes related to this dysfunctional pathway. Our
results reveal a possible mechanistic link between deficient synaptic pruning
and functional hyper-connectivity in autism.
Introduction
Post-mortem
examinations have revealed an excess of excitatory synapses in the brain of
children with autism spectrum disorders (ASD) [1, 2]. Recent
investigations have linked this trait to hyper-activity of the mTOR pathway,
resulting in synaptic pruning deficits [1]. ASD is also
characterized by alterations in brain functional connectivity as measured with
resting state functional MRI (rsfMRI) [3, 4]. These observations
raise the question of whether and how mTOR-related deficient pruning affects
rsfMRI dysconnectivity observed in ASD.
Here, we
mapped dendritic synaptic density, rsfMRI connectivity [5, 6] and social behavior
in tuberous sclerosis 2 deficient (Tsc2+/-) mice, a mouse line that
mechanistically reconstitutes the mTOR-dependent dendritic spine surplus
observed in post mortem ASD brain samples [1]. A separate cohort of animals was subjected to
a daily treatment with the mTOR inhibitor rapamycin with the aim to rescue the
synaptic surplus, and corroborate a mechanistic link between synaptic traits
and brain-wide connectivity. We finally sought to
translate these findings across species, by isolating an analogous connectivity
signature in publicly available rsfMRI scans of children with ASD (ABIDE-I) [7].Methods
Mouse studies. All
experiments were carried out in accordance with Italian regulations governing
animal welfare and protection (DL 26/2014). We used a cross-sectional treatment
protocol with four cohorts of n=20 animals each: control Tsc2+/+ mice
treated with rapamycin; control Tsc2+/+ mice treated with vehicle;
mutant Tsc2+/- mice treated with rapamycin; mutant Tsc2+/-
mice treated with vehicle. After daily treatment during pre-pubertal phase [1], Tsc2+/- and
control mice underwent rsfMRI mapping at 7 tesla under halothane sedation [8] using a single-shot EPI
sequence (TR/TE = 1200/15 ms, flip angle 30°, FOV 2.3 × 2.3 cm, matrix 100 ×
100, 18 coronal slices, slice thickness 0.60 mm, 1920 volumes). Data were
preprocessed as previously described [9] and inter-group differences in rsfMRI connectivity were mapped using weighted degree
centrality and seed-based analysis [5, 10]. Post-mortem dendritic
spine density was measured as in [1]. Human rsfMRI studies. We used weighted
degree centrality and seed based analysis to map rsfMRI connectivity in
pre-pubertal (age range 7-13) children with ASD (n=163) and typically
developing controls (n=168) from eight sites of the ABIDE-I collection. Gene
enrichment analyses were carried out as described in [11].Results and discussion
In keeping with previous findings, we observed
increased spine density in juvenile Tsc2+/-
mice (p < 0.05, Figure 1A), an
effect associated with largely altered connectivity, involving long-range rsfMRI
over-synchronization of integrative neocortical regions and the basal ganglia (Figure 1B). Importantly, developmental
treatment with
the mTOR-inhibitor rapamycin completely rescued synaptic density (p < 0.05, Figure 2A) and
normalized rsfMRI hyper-connectivity (t > 2, p < 0.01
FWE cluster corrected, Figure 2B) in Tsc2+/- mice, hence
corroborating a mechanistic link between aberrant mTOR activity and functional over-connectivity.
This effect was especially prominent in the mouse default mode (Figure
3A-B) and salience networks (Figure 3C-D). Developmental
treatment with rapamycin also rescued autism-like behavioral deficits in Tsc2
mutant mice (Fig. 4). Notably, voxel-wise correlational mapping revealed that
functional hyper-synchronization of insular and striato-prefrontal
regions was predictive of motor stereotypies in these mice (Figure
4). Collectively, these results establish a causal link
between mTOR-related pruning deficits, fronto-insular hyperconnectivity and ASD-relevant
motor stereotypies.
Given the prominent
implication of mTOR pathway dysfunction for human ASD [1], we hypothesized that
a similar hyper-connected phenotype should be identifiable in clinical ASD
populations. In keeping with this notion, rsfMRI connectivity mapping in pre-pubertal children with ASD
from ABIDE revealed foci of increased long-range connectivity in the anterior
insula, basal ganglia and prefrontal cortices of affected participants (Figure 5A-B). Furthermore, insular-prefrontal
over-connectivity was significantly associated with repetitive behaviors (Figure
5C-D), hence reconstituting both the connectional and behavioral phenotype observed in Tsc2+/-
mice. By coupling gene decoding and enrichment analysis, we finally found that
TSC2/mTOR-network genes appear to be expressed in a similar topological pattern
as the identified ASD hyper-connectivity phenotype, suggesting a putative
mechanistic link between the observed hyper-connectivity and mTOR-related
aberrant activity in human ASD.Conclusion
Taken together, our results establish a cross-species
mechanistic link between mTOR-dependent
deficient synaptic pruning and rsfMRI over-connectivity in ASD.Acknowledgements
This
work was supported by Simons Foundation Grants (SFARI 400101) to A. Gozzi. A.
Gozzi was also supported by Brain and Behavior Foundation 2017 (NARSAD -
National Alliance for Research on Schizophrenia and Depression) and the
European Research Council (ERC - DISCONN, GA802371). M. Pagani was supported by
European Union's Horizon 2020 research and innovation programme (Marie
Sklodowska-Curie Global Fellowship – CANSAS, GA845065).References
1. Tang, G., et al., Loss of mTOR-Dependent Macroautophagy Causes
Autistic-like Synaptic Pruning Deficits. Neuron, 2014. 83(5): p. 1131-1143.
2. Hutsler,
J.J. and H. Zhang, Increased dendritic
spine densities on cortical projection neurons in autism spectrum disorders.
Brain research, 2010. 1309: p. 83-94.
3. Nomi,
J.S. and L.Q. Uddin, Developmental
changes in large-scale network connectivity in autism. NeuroImage:
Clinical, 2015. 7: p. 732-741.
4. Uddin,
L.Q., K. Supekar, and C.J. Lynch, Salience
network-based classification and prediction of symptom severity in children
with autism. JAMA Psychiatry, 2013. 70(8):
p. 869-879.
5. Pagani,
M., et al., Deletion of autism risk gene
Shank3 disrupts prefrontal connectivity. The Journal of Neuroscience, 2019:
p. 2529-18.
6. Sforazzini,
F., et al., Distributed BOLD and
CBV-weighted resting-state networks in the mouse brain. Neuroimage, 2014. 87: p. 403-15.
7. Di
Martino, A., et al., The autism brain
imaging data exchange: towards a large-scale evaluation of the intrinsic brain
architecture in autism. Mol Psychiatry, 2014. 19: p. 659-667.
8. Bertero,
A., et al., Autism-associated 16p11.2
microdeletion impairs prefrontal functional connectivity in mouse and human BRAIN,
2018. 141(7): p. 2055–2065.
9. Suetterlin,
P., et al., Altered neocortical gene
expression, brain overgrowth and functional over-connectivity in Chd8
haploinsufficient mice. Cerebral Cortex, 2018. 28(6): p. 2192-2206.
10. Michetti,
C., et al., The knockout of Synapsin II
in mice impairs social behavior and functional connectivity generating an
ASD-like phenotype. Cerebral Cortex, 2017. 27(10): p. 5014-5023.
11. Gorgolewski,
K.J., et al., NeuroVault. org: a
web-based repository for collecting and sharing unthresholded statistical maps
of the human brain. Frontiers in neuroinformatics, 2015. 9: p. 8.