Reciprocal Alterations of White Matter Microstructure in Carriers of Deletions Versus Duplications at the 16p11.2 Chromosomal Locus Are Associated with Cognitive and Behavioral Impairments

Saturday, May 16, 2015: 11:30 AM-1:30 PM
Imperial Ballroom (Grand America Hotel)
Y. Chang1, J. Owen1, T. Thieu2, N. Pojman3, P. Bukshpun4, M. Wakahiro4, E. Marco5, J. I. Berman6, J. E. Spiro7, W. Chung8, R. L. Buckner9, T. P. Roberts6, S. Nagarajan10, E. H. Sherr5 and P. Mukherjee4, (1)Radiology, UCSF, San Francisco, CA, (2)Neurology, UCSF, San Francisco, CA, (3)Neurology, Radiology, UCSF, San Francisco, CA, (4)UCSF, San Francisco, CA, (5)University of California, San Francisco, San Francisco, CA, (6)Children's Hospital of Philadelphia, Philadelphia, PA, (7)Simons Foundation, New York, NY, (8)Pediatrics, Columbia University, New York, NY, (9)Harvard University, Cambridge, MA, (10)Univeristy of California San Francisco, San Francisco, CA
Background:  Copy number variants (CNVs) at the BP4-BP5 16p11.2 chromosomal locus are associated with autism, schizophrenia, and epilepsy. Individuals with 16p11.2 deletions tend toward increased body mass index (BMI) and larger head size, while individuals with the reciprocal duplications often demonstrate lower BMI and smaller head size. 

Objectives:  To compare white matter (WM) alterations in pediatric and adult carriers of 16p11.2 duplications and deletions using diffusion tensor imaging (DTI), and to relate these alterations to cognitive and behavioral function.

Methods:  The study included 30 pediatric deletion carriers, 13 pediatric duplication carriers, 7 adult deletion carriers, 23 adult duplication carriers and matched control participants for each cohort. Nonverbal IQ (NVIQ) and social responsiveness scale (SRS) scores served as measures of cognition and social behavior, respectively. 3T structural MRI and DTI using 30 directions at b=1000 s/mm2 were acquired and fractional anisotropy (FA), mean diffusivity (MD), radial diffusivity (RD), and axial diffusivity (AD) maps were constructed using tract-based spatial statistics (TBSS). Voxel-wise group differences of each DTI parameter were assessed for each of the four CNV carrier cohorts and their matched controls. Global and regional group differences were also assessed separately for the children and adults using analysis of variance (ANOVA). Finally, correlations of NVIQ and SRS were performed with the absolute value of z-scored DTI values in global and regional white matter.

Results:  The voxel-wise TBSS results reveal extensive WM increases of FA and AD in the pediatric deletion carriers relative to their controls, whereas the pediatric duplication carriers show extensive decreases of FA and elevations of MD and RD. The adult duplication carriers similarly exhibit decreased FA and increased RD relative to their controls, but with additional extensive decreases of AD. While the adult deletion carriers do not show significant voxel-wise group differences due to lack of statistical power with only 7 carriers, they do exhibit significant elevations of AD in the association and limbic tracts from the ANOVA group analysis. Significant correlations between NVIQ and the absolute values of z-scored DTI metrics in the callosal, association, and projection tracts were found in children, with both increases and decreases of DTI metrics from control values associated with lower NVIQ. The adults showed significant correlations of NVIQ and SRS with the absolute value of z-scored DTI metrics in the callosal and limbic tracts. 

Conclusions:  We demonstrate widespread and opposing WM alterations in carriers of the 16p11.2 deletion versus the reciprocal duplication, consistent with the theory of gene dosage dependence. We also show associations of cognitive and behavioral impairment with deviation in either direction from the WM microstructural values of the control cohorts. The common cognitive/behavioral effects of these opposite changes in WM microstructure may help elucidate the heterogeneity in prior DTI studies of neuropsychiatric disorders that are associated with this CNV, such as autism and schizophrenia. These significant findings with modest sample sizes support the view that specific genetic variations may be more strongly associated with changes in WM microstructure than are the more mechanistically complex shared neuropsychiatric diagnoses.