Abstract
The association between copy number variations (CNVs) and schizophrenia is well-documented, yet the specific impact of these genetic alterations on brain structure remains poorly understood. This pilot study investigates the collective effects of CNVs within specific cytogenetic bands on schizophrenia risk and gray matter concentration (GMC). Analyzing 324 participants (151 patients, 173 controls), we identified a significant genetic "hotspot" at 22q13.31. Deletions in this region were significantly more prevalent in schizophrenia patients and were strongly associated with reduced GMC in a brain network comprising the cingulate gyrus and insula ($p = 1.44 \times 10^{-4}$). These findings suggest that regional CNVs, regardless of size, can influence schizophrenia risk through measurable downstream effects on the peri-limbic cortex.
Introduction
Schizophrenia is a complex, highly heritable disorder characterized by significant heterogeneity in both symptoms and genetic underpinnings. While rare, large CNVs (typically >500kb) have been linked to the disorder, they only account for a small fraction of cases. This study shifts the focus toward the "collective effect" of all CNVs—both large and small—within specific chromosomal regions known as cytogenetic bands. By examining these hotspots, we aim to bridge the gap between genetic variation and the structural brain changes, such as gray matter reduction, frequently observed in neuroimaging studies of schizophrenia.
Methods
The study utilized a multi-modal approach combining high-resolution genomics and structural neuroimaging from the MCIC and COBRE datasets.
- CNV Profiling: DNA was analyzed using Illumina 1 million assays. Deletions and insertions were detected using PennCNV and Circular Binary Segmentation algorithms, with findings validated via real-time PCR.
- Neuroimaging: Structural MRI scans were processed using Voxel-Based Morphometry (VBM) to measure GMC.
- Network Analysis: Spatial Independent Component Analysis (ICA) was employed to identify 18 independent brain networks, allowing for a more robust analysis of structural covariance than traditional voxel-wise methods.
- Statistical Modeling: Linear regression models tested associations between CNV features in 811 cytobands and brain network GMC, strictly controlling for age, gender, scan site, and population structure (ethnicity).
Results
The Impact of Population Structure
A critical finding was the discovery of population structure in CNV features. Rare CNV burden was significantly higher in African American participants regardless of diagnosis. Controlling for this ethnic group difference was essential to avoid false-positive associations between CNV burden and schizophrenia.
22q13.31: A Genetic Hotspot
After controlling for ethnicity, deletions in the 22q13.31 cytoband were significantly more frequent in schizophrenia patients ($p < 0.05$). This region contains genes such as SAMM50 and PARVB, which are involved in essential cellular functions like mitochondrial maintenance and cell adhesion.
Structural Brain Correlations
The deletion load at 22q13.31 was strongly associated with a specific brain network primarily involving the cingulate gyrus and insula (the peri-limbic cortex). Carriers of these deletions showed a 9% to 19% reduction in GMC loadings within this network compared to non-carriers. This association explained approximately 5% of the total variance in the network's concentration.
Discussion and Conclusion
This pilot study provides proof-of-concept that analyzing the collective burden of CNVs within chromosomal hotspots can reveal biological downstream effects on brain structure. The 22q13.31 region emerges as a significant site where deletions correlate with structural deficits in the peri-limbic cortex, a region vital for emotional processing and executive function.
While the sample size is modest for a genetic study, these results highlight the potential for using regional genetic markers to sub-categorize the heterogeneous landscape of schizophrenia. Future research with larger cohorts is necessary to replicate these findings and further explore the functional mechanisms by which these deletions alter neurodevelopment.
Summary updated