Authors :
Presenting Author: Alexander Ngo, BSc – McGill University
Lang Liu, MBI – McGill University; Sara Larivière, PhD – Harvard Medical School; Valeria Kebets, PhD – McGill University; Serena Fett, BS – McGill University; Clara Weber, MS – University of Lübeck; Jessica Royer, PsyD – McGill University; Eric Yu, MS – McGill University; Raúl Rodríguez-Cruces, MD, PhD – McGill University; Luis Concha, MD, PhD – National Autonomous University of Mexico; Zhiqiang Zhang, MD – Nanjing University School of Medicine; Leon Qi Rong Ooi, MS – National University of Singapore; B.T. Thomas Yeo, PhD – National University of Singapore; Birgit Frauscher, MD, PhD – McGill University; Neda Bernasconi, MD, PhD – McGill University; Andrea Bernasconi, MD – McGill University; Ziv Gan-Or, MD, PhD – McGill University; Boris Bernhardt, PhD – McGill University
Rationale:
While temporal lobe epilepsy (TLE) is traditionally conceptualized as a prototypical limbic epilepsy, it is commonly associated with widespread alterations in adult case-control series.
1 With the seizure onset often in youth, genetic effects and atypical brain development may contribute to epilepsy risk in many patients and relate to alterations in brain structure. Here, we mapped associations between polygenic risk scores (PRS) related to TLE and brain morphology in childhood and assessed its concordance with case-control atrophy.
Methods:
Participants. We aggregated high-resolution structural T1w magnetic resonance imaging (MRI) and whole-genome sequencing data of neurotypical children and adolescents from the Pediatric Imaging, Neurocognition and Genetics
2 (n = 313, 171 males, mean ± SD age = 12.0 ± 5.2 years) and Adolescent Brain Cognition Development (n = 3826, 2052 males, mean ± SD age = 10.0 ± 0.6 years) datasets.
3
PRS-enriched morphology analysis. We derived vertex-wise cortical thickness maps, harmonized across sites. Individualized PRS were computed based on the weighted sum of common genetic risk variants for hippocampal sclerosis (HS), obtained from the latest genome-wide association study on epilepsy.4 Surface-based linear models related PRS-HS to the imaging-derived phenotype in pediatric participants.
Relation to disease-specific atrophy. Leveraging the imaging data from four independent case-control datasets comprised of patients with TLE and healthy controls (HC), ENIGMA-Epilepsy,1 Jinling Hospital,5 Epilepsy and Cognition,6 and MICA-MICs,7 we evaluated the correspondence between imaging-genetic and TLE-specific atrophy maps using spatial correlations with autocorrelation preserving null models.
Results:
Widespread cortical thickness decreases were observed in children with high PRS-HS that mainly targeted temporo-limbic cortices (
pFWE < 0.05; F1A). Indeed, stratification of PRS effects according to an atlas of functional zones further identified peak values in paralimbic regions. PRS-HS associations at the study and site level were consistent with the pooled ABCD × PING sample (F1B). However, no significant correlation between PRS-HS mediated shifts in cortical thickness and atrophy in left (F2A) and right (F2B) TLE were found (all
p > 0.15).
Conclusions:
Capitalizing on recent imaging-genetic initiatives, our analyses revealed widespread thinning associated with increased PRS-HS, with peak effects in temporo-limbic regions— pathophysiological epicentres of the condition. As PRS-HS effects did not strongly resemble adult case-control patterns of atrophy, our findings suggest that adult TLE atrophy likely reflects additional factors related to disease progression and aging.
1Whelan CD et al., 2018, Brain, 141:391-408
2Jernigan et al., 2016, NeuroImage, 124:1149-1154
3Casey et al., 2018, Dev Cogn Neurosci, 32:43-54
4International League Against Epilepsy Consortium on Complex Epilepsies, 2018, Nat Commun, 9:5269
5Weng et al., 2020, Communications Biology, 3:244
6Rodríguez-Cruces et al., 2020, Neuroimaging, 213:116706
7Royer et al., 2022, Sci Data, 9(1):569
Funding:
FRQ-S, CIHR