Disrupted-in-Schizophrenia-1 (DISC1) was first identified as a candidate gene for mental illness. Mutant DISC1 gene may predispose individuals to the development of schizophrenia (1). In spite of a few studies that investigated DSIC1-related macrostructural abnormalities such as volume of the hippocampus and ventricle as well as thickness of cerebral cortex (2-4), mutant DISC1 induced microstructural alterations remain largely unexplored, especially in gray matter regions. Changes in molecular compositions have also not yet been investigated using non-invasive imaging methods. Kurtosis metrics derived from non-Gaussian diffusion kurtosis imaging (DKI) potentially have higher specificity to probe interactions of water protons with cell and tissue components and presumably reflect heterogeneity and irregularity of cellular microstructure (5-8). Another recently developed imaging tool for characterizing tissue properties in both normal and disease states is quantitative susceptibility mapping (QSM) (9-11). This technique directly reflects the molecular composition and cellular architecture of the tissue. The purpose of this study is utilizing DKI and QSM to detect and quantify cerebral microstructural changes and to relate them to alterations in molecular compositions.The Tet-off double transgenic system was used to generate a mouse model of inducible expression of mutant truncated human DISC1 (hDISC1) (12). In the present study, double transgenic mutant hDISC1 mice were included in the mutant group (n = 8), whereas their single tetracycline-transactivator transgenic littermates were included in the control group (n = 7). Pilot experiments showed no expression of mutant hDISC1 in single DISC1 transgenic mice and no significant differences between wild-type mice and single DISC1 transgenic mice in the behavioral tests or neurite outgrowth assays in primary neurons. Brain specimens of these two groups were scanned at a 9.4 T Oxford magnet. For reconstruction of QSM, magnitude and phase data were acquired using a GRE sequence with 8 echoes For calculation of DKI metrics, Diffusion images were acquired with 30 gradient directions and two b-values (2.5 and 5.0 ms/μm2). Significant increases in mean kurtosis (MK) were observed in all gray matter regions including the caudate putamen (control: 0.82 ± 0.03, hDISC1: 0.87 ± 0.06, P = 0.014), cerebral cortex (control: 0.75 ± 0.03, hDISC1: 0.79 ± 0.04, P = 0.021) and hippocampus (control: 0.82 ± 0.03, hDISC1: 0.86 ± 0.02, P = 0.009). Significant decreases of FA were observed in the anterior commissure (control: 0.42 ± 0.04, hDISC1: 0.35 ± 0.05, P = 0.002) and hippocampus (control: 0.19 ± 0.01, hDICS1: 0.8 ± 0.02, P = 0.04). A significant decrease of MD was observed in hippocampus (control: 0.49 ± 0.02 ×10-3mm2/s, hDISC: 0.47 ± 0.01 ×10-3mm2/s, P = 0.29). Significant increases of susceptibility were found in both white matter tracts and gray matter regions: anterior commissure (control: 21.31 ± 2.76 ppb, hDISC1: -18.00 ± 2.22 ppb, P = 0.025); caudate putamen (control: 1.26 ± 0.50 ppb, hDISC1: 2.33 ± 0.81 ppb, P = 0.008); hippocampus (control: 1.88 ± 0.71 ppb, hDISC1: 2.73 ± 0.29 ppb, P = 0.012). These data are shown in Figure 2. Significant correlation between DKI metrics and magnetic susceptibility were observed only in gray matter regions. Linear regression lines are shown in Figure 3.

This is the first study that utilized both DKI and QSM methods to probe microstructural and molecular abnormalities in hDISC1 transgenic mice. In white matter tracts of the anterior commissure, a decrease in FA may result from increased extra-axonal space free-diffusion water, misalignment of axonal fiber orientations, demyelination or loss of myelinated axonal content. Although not completely ruling out other possible mechanisms, increased susceptibility values turned our focus to myelin related mechanisms that may play a dominant role in the observed FA decrease. In gray matter region of the hippocampus, a decrease in FA may be caused by misalignment of axonal fiber orientation and a decrease in MD can be resulted from more densely packed neuronal cell bodies. Being more sensitive to microstructural changes in relatively isotropic diffusional environment, MK captured increased microstructural heterogeneity and irregularity the in caudate putamen and cerebral cortex in addition to the hippocampus. Increased susceptibility in the caudate putamen and hippocampus further confirmed that increased neuronal cell density is one of the leading factors.

In conclusion, DKI method provided sensitive metrics for detecting microstructural changes in not only the anterior commissure but also relatively the isotropic gray matter regions. Relating DKI findings to QSM-derived measurements of molecular compositions enabled clearer interpretations of myelin content and cellular density related mechanisms. Further validations that establish the relationship between imaging metrics and histological measurements such as neuronal cell body density and g-ratio are needed.