Sex Chromosome Dosage Effects on White Matter Structure in the Human Brain

Abstract

Sex chromosome aneuploidies, a group of neurogenetic conditions characterized by aberrant sex chromosome dosage (SCD), are associated with increased risks for psychopathology as well as alterations in gray matter structure. However, we still lack a comprehensive understanding of potential SCD-associated changes in white matter structure, or knowledge of how these changes might relate to known alterations in gray matter anatomy. Thus, here, we use voxel-based morphometry on structural neuroimaging data to provide the first comprehensive maps of regional white matter volume (WMV) changes across individuals with varying SCD (n = 306). We show that mounting X- and Y-chromosome dosage are both associated with widespread WMV decreases, including in cortical, subcortical, and cerebellar tracts, as well as WMV increases in the genu of the corpus callosum and posterior thalamic radiation. We also correlate X- and Y-chromosome-linked WMV changes in certain regions to measures of internalizing and externalizing psychopathology. Finally, we demonstrate that SCD-driven WMV changes show a coordinated coupling with SCD-driven gray matter volume changes. These findings represent the most complete maps of X- and Y-chromosome effects on human white matter to date, and show how such changes connect to psychopathological symptoms and gray matter anatomy.

Trial registration: ClinicalTrials.gov NCT00001246.

Keywords: aneuploidy; neuroimaging; sex chromosomes; white matter.

Published by Oxford University Press 2021.

Figures

Figure 1

Local WMV changes as a function of X- and Y-chromosome dosage. (A, B) Thresholded t-maps, where colored regions represent cortical and subcortical tracts showing statistically significant WMV changes with (A) increasing supernumerary X-chromosome dosage, or (B) increasing supernumerary Y-chromosome dosage (P < 0.05 for all tracts after correction for multiple comparisons; blue represents areas where WMV decreases as chromosome dosage increases, and orange represents areas where WMV increases as chromosome dosage increases). (C) Conjunction t-map showing the high degree of overlap between areas where X-chromosome dosage affects WMV and areas where Y-chromosome dosage affects WMV (blue represents regions where WMV decreases with both mounting X- and mounting Y-chromosome dosage, and orange represents regions where WMV increases with both mounting X- and mounting Y-chromosome dosage). (D) Density plot demonstrating that the observed cross-voxel correlation between X- and Y-chromosome effects on local WMV is greater than 1000 correlations generated by randomly permuting SCA labels across participants.

Figure 2

Correlation between SCD-driven changes and internalizing and externalizing psychopathology. (A) Significant interaction effect of externalizing psychopathology on the relationship between WMV in the anterior limbs of the internal capsules and supernumerary X-chromosome dosage. As X-chromosome dosage increased, WMV decreased, but this relationship was attenuated in individuals with more severe externalizing psychopathology. These interactions were significant in both the left (t = 3.7, P < 0.001) and right (t = 3.5, P < 0.001) internal capsules, but only the left-side line graph is shown here. (B) Significant interaction effect of internalizing psychopathology on the relationship between WMV in the genu of the corpus callosum and supernumerary Y-chromosome dosage (t = −3.9, P < 0.001). WMV in this region increased with mounting Y-chromosome dosage, but this relationship was less dramatic in individuals with greater internalizing psychopathological symptoms. In both cases, line graphs were generated based on the mean WMV of all voxels within the respective regions of interest, with error bars indicating standard deviation. Models were run using continuous measures of externalizing and internalizing psychopathology from the CBCL and ABCL, but are placed in highest and lowest 50th percentile categories here for visualization purposes. Supernumerary X- and Y-dosages refer to X and Y dosage counts that are above gonadal baseline (i.e., 0 in graph 2A encompasses XX and XY individuals, 1 encompasses XXY and XXX individuals, 2 encompasses XXYY individuals, and 3 encompasses XXXXY individuals).

Figure 3

Local GMV changes as a function of X- and Y-chromosome dosage. (A, B) Thresholded t-maps, where colored regions represent cortical and subcortical regions showing statistically-significant GMV changes with (A) increasing supernumerary X-chromosome dosage, or (B) increasing supernumerary Y-chromosome dosage (P < 0.05 for all regions after correction for multiple comparisons; blue represents regions where GMV decreases as chromosome dosage increases, and orange represents regions where GMV increases as chromosome dosage increases). (C) Conjunction t-map showing the high degree of overlap between areas where X-chromosome dosage affects GMV and areas where Y-chromosome dosage affects GMV (blue represents regions where GMV decreases with both mounting X- and mounting Y-chromosome dosage, and orange represents regions where GMV increases with both mounting X- and mounting Y-chromosome dosage). (D) Density plot showing that the observed cross-voxel correlation between X- and Y- chromosome effects on regional GMV is greater than 1000 correlations drawn from random permutations of SCA labels across participants.

Figure 4

Coordination of regional WMV and GMV changes with SCD variation. (A) Yellow gray matter regions are those connected by white matter tracts that have at least 50% of their voxels scaling positively with increased X-and Y-chromosome dosage. (B) Yellow gray matter regions are those showing significant volumetric increases with increased X- and Y-chromosome dosage. (C) Blue gray matter regions are those connected by white matter tracts with at least 50% of their voxels scaling negatively with increased X- and Y-chromosome dosage. (D) Blue gray matter regions are those showing significant volumetric decreases with increased X- and Y-chromosome dosage. Fisher’s exact tests demonstrated a significant overlap between maps A and B (odds ratio = 5.3, P = 0.02), as well as between maps C and D (odds ratio = 2.7, P = 0.04), suggesting a strong congruence between SCD effects on white and gray matter in the human brain. Maps are presented in Desikan-Killiany space, and were generated with the help of IIT Atlas resources (see methods for details).