Carrie E. Bearden, Ph.D. Psychiatry and Biobehavioral Sciences and PsychologySemel Institute for Neuroscience and Human BehaviorUniversity of California, Los [email protected]
Brain Development in 22q11.2 Deletion Syndrome
Convergence of genes on neural systems
Geschwind 2011
Schizophrenia
Key Questions
ØLimited data on individuals carrying the same highly penetrant CNV;What are the most robust/reproducible effects of a 22q11.2 deletion on the brain?
-cortical and subcortical structures
ØDo imaging biomarkers identified in idiopathic psychosis also predict psychotic symptoms in 22q11 DS?
ØInvestigation of genetically homogeneous, known etiology- may tell us how genes within and/or outside deleted region contribute to disrupted biological pathways that ultimately contribute to psychiatric phenotypes
ØPremise: Hemizygous deletion may have a more powerful effect on phenotypic variation than common genetic variation.
Gray Matter Thinning in anterior cingulate and medial frontal cortex in 22q11DS
Bearden et al. Cerebral Cortex 2009
Kana et al. Soc Neurosci 2009
Frontal “Theory of Mind” Network
http://enigma.ini.usc.edu/?p=4900
22q-ENIGMA Current Demographics
Brain Imaging Analysis: Cortical Thickness vs. Surface Area
• Driven by distinct genetic mechanisms1,2
• Different underlying neurobiology– Surface area is considered to
represent the number of cortical columns3
– Cortical thickness is a proxy for the number of cells in a column4
Methods
-All raw data processed via Freesurfer5.3.0 using ENIGMA processing pipeline -Covaried for age, sex, and scanner
1Winkler et al., 2009; 2Panizzon et al., 2009;3Rakic et al. 1988;4Pontius et al., 2008
• Volume measures are derived from indices of both cortical thickness and surface area
Most Significant Cortical Measures: 22q11DS Case vs. Control
Jonas et al 2014 (adapted from Tan et al 2009 meta-analysis)
Reduced SA Due to Reduced Gyrification in 22q11DS
Gyrification Index: Measure of cortical folding
Machine Learning Classification of 22q11DS Case-Control Status
• GLMNET algorithm (20 10-fold cross-validations)• Feature selection based on Random Forest algorithm
-Accuracy peaks at 94.44% (125 features)-Cuneus & lingual SA, and caudal anterior cingulate volume
among most important features for classification
Accuracy of 86.36% achieved using bilateral cuneus and lingual SA for classification
Sun et al, in prep
High Consistency of Brain Measures Across Sites
Prodromal/Psychotic (n=45) vs. Non-Psychotic (n=81): Greatest Effects for SA in Somatosensory, Heteromodal Association Regions
Sun et al, in prep
Greatest Effects for SA in Somatosensory, Heteromodal Association Regions: Overlap with Idopathic (non-22q) Psychosis
SCZ data courtesy of van Erp & Turner et al. in submission
Longitudinal Findings: Three-way (brain region x time x group) interactions associated with psychotic symptom change over time
Green et al Nat Rev Neurosci 2015
Brain Regions associated with Social Processes
22q11DS(n=184)vs.Controls(n=137)
• Red/Yellow=22q11DS>surfacearea/volumevs.Controls• Blue/Green=22q11DS<Controls
GlobusPallidus
CaudatePutamen
HippocampusHippocampus
Amygdala
ThalamusThalamus
PutamenNucleusAccumbens
Ching et al, in prep
Summary & Implications²Largest MRI sample to date of this CNV; Neuroanatomic patterns allow highly accurate classification of 22q11DS patients relative to controls.
²Reductions in heteromodal association cortex (particularly in SA) converge with regions affected in idiopathic schizophrenia
²Utility of high-penetrance models: 22q11DS may account for only small proportion of risk for psychotic illness overall, but dysregulated genes/pathways may contribute to broader psychosis susceptibility
²Translational investigation -mechanistic studies in animal and cellular models.
Neurobehavioral Genetics Giovanni CoppolaNelson Freimer
Roel OphoffDan Geschwind
USC Imaging Genetics CenterJulio VillalonJustin GalvisConor CorbinPaul Thompson
22q-ENIGMA Working GroupMaria Gudbrandsen, Eileen Daly,
Christine Ecker, Clodagh Murphy, Declan Murphy, Michael Craig, Geor Bakker,
Therese van Amelsvoort, Linda Campbell, Wendy R. Kates, Liz Gras, Ania Faskinski, Jacob Vorstman, Tony Simon, Eva Chow,
Nancy Butcher, Anne Bassett
22q11.2 International Brain Behavior Consortium (Raquel Gur, Thomas Lehner)
Lab MembersMaria Jalbrzikowski Matthew Schreiner
Rachel JonasLeila KushanCarolyn Chow
Amira IbrahimCaroline MontojoAriel SchvarczSteffie TomsonTherese VesagasChristopher Ching
Daqiang (Frank) SunJamie Zinberg
Katie YoungAmy Lin
Acknowledgements
This work supported by: the National Institute of Mental Health (NIMH), ENIGMA Center for Worldwide Medicine, Imaging &
Genomics ( NIH/NIBIB), the Staglin/IMHRO Music Festival, the Brain-Behavior Research Foundation (NARSAD), and all the
patients and their families who participated.