To develop novel gene-specific diffusion tensor imaging biomarkers in the assessment and characterization of autism spectrum disorder.

Ex vivo imaging methods were used in male rats to examine the structural differences of the Nrxn1 genetic model of ASD (derived from outbred Sprague Dawley rats; n = 4) as compared to age-matched male controls (outbred Sprague-Dawley rats; n = 4). At PND 45, animals were deeply anesthetized with isoflurane and were transcardially perfused with fresh 4% paraformaldehyde (PFA). Fixed brains were removed and stored in 4% PFA until imaging, whereupon they were then rinsed in 0.9% saline for 48 hours prior to imaging to minimize attenuating effects of fixative on the MRI signal. The brains were placed in a custom-built holder and immersed in Fluorinert (FC-3283, 3M, St. Paul, MN, USA) for image acquisition.

For ex-vivo diffusion tensor imaging (DTI) acquisition, brains were simultaneously imaged for ∼7 h using a 4.7-T Agilent MRI system and 3.5-cm diameter quadrature volume RF coil. A series of multi-slice, diffusion-weighted, spin echo images were acquired with three non-weighted (b ∼ 0) and 30 diffusion weighted (b ∼ 1200 s/mm2), using non-colinear weighting directions. Other imaging parameters were TE/TR = 24.17/2000 ms, FOV = 30 × 30 mm2, matrix = 192 × 192 reconstructed to 256 × 256, slice thickness = 0.5 mm, number of slices = 35 and two signal averages. DTI maps were created offline using a combination of FSL software and custom MatLab code. Raw images were corrected for eddy currents by rigid body co-registration and the diffusion tensor was determined at each voxel using a non-linear least squares fitting algorithm. DTI maps for fractional anisotropy (FA), mean diffusion (MD) and axial and radial diffusivity (D|| and D⊥) were generated as well as images containing the eigenvector information for the principal eigenvalue used for orientation analysis. Offline registration of the non-weighted image to a template was performed to correct any misalignment of the measured orientation vectors (1). Subsequent region of interest (ROI) analysis was performed in the native imaging space. DTI values of the corpus callosum (CC) and other principle white matter tracts were performed (e.g., cingulum, fimbria of the hippocampus, fornix, medial longitudinal fasiculus, and internal capsule). MRI and DTI values for each ROI were compared using ANOVA. Two-tailed t-tests and Pearson's correlation were also performed where applicable.

Animals harboring the Nrxn1 biallelic deletion demonstrated widespread changes in gray and white matter structure and organization. These also included areas of significant change in FA as previously reported in the superior and inferior colliculus, cerebral cortex of the frontal lobe, and in several areas within the deep gray nuclei including the globus pallidus and thalamus. Significant differences between our Nrxn1 genetic model of ASD as compared to wild-type animals were also noted in our ROI analysis including the corpus callosum and several white matter tracts including the medial longitudinal fasiculus, internal capsule, forceps minor, and occipitotemporal tracts.

The etiology of ASD remains elusive with no objective measurable clinical markers. The development of advanced structural and functional MR-imaging based methodologies, including fMRI and DTI, has spurred tremendous interest towards their application in ASD research, especially towards the identification of a neuroimaging biomarker (2,3). While much progress has been made, the identification of specific imaging biomarkers in ASD remains opaque. One contributing factor is the marked genetic heterogeneity of ASD (4) contributing to inconsistent and varying neuroimaging findings. New targeted genome editing technologies have yielded the first cogent genetic animal models of ASD including our working model with a biallelic deletion of the Nrxn1 gene, a high-risk susceptibility factor in the neuropathogenesis of ASD. Our results are the first to demonstrate a truly cogent gene-specific neuroimaging marker in the assessment of ASD on DTI. As further work proceeds with the identification of additional gene-specific DTI neuroimaging markers in other gene-specific knockouts, the collective sum of these findings will allow for the generation and identification of cogent gene-specific neuroimaging biomarkers, which can set the stage and contribute to a new neuroimaging diagnostic paradigm in ASD.