While cuprizone induces a well-established model of demyelination in the mouse, much less is known about the effects of cuprizone on rat brain. It would be advantageous to have a rat model of cuprizone induced demyelination for MRI studies as the white matter structures are more substantial in that species. Additionally, while several different quantitative MRI methods exist for assessing myelin, the very small size of the mouse brain can pose a problem for techniques like quantitative T2 which are affected by the process of diffusion at ultra-high resolution¹. The rat model may also be desirable in comparing behavioural data with MRI. Initial histological studies in rats reported spongiform encephalopathy or intrymyelinic edema with no loss of myelin^2,3,4, while more recent histological work has demonstrated demyelination in rat brain⁵. The goal of this study is to characterize a rat model of cuprizone induced demyelination using MRI for the first time, and validate this using luxol fast blue (LFB) to assess demyelination, and X-Ray fluorescence to quantify changes in transition metals which also affect MRI contrast.

16-week old male Wistar rats were fed either a 1% cuprizone-enriched diet (treated) or normal rat chow (control) ad libitum. 5 treated and 5 control animals were sacrificed after 4 weeks, brains removed and frozen for metal quantification. 6 treated and 2 control animals were imaged after 2, 4, or 6 weeks with a 7T Bruker BioSpec MRI. Following MRI, animals were perfused with 10% formalin, brains removed and embedded in paraffin. 5 and 10 µm sections were stained with H&E and LFB respectively.

T1-weighted images optimized for myelin contrast⁶ consisted of an MP-RAGE sequence with TE 2.3 ms, TR 5.9 s, 4 segments with duration 384 ms, inversion time 1.3 s, 2 averages, FA 12º, and 167um isotropic resolution. PD-weighted images were acquired at the same resolution using a FLASH sequence with TE 2.3ms, TR 12.3ms, 6 averages, and FA 6º.

Ratio images were calculated by dividing the MP-RAGE by the PD-weighted images and normalizing by the signal in muscle in the head. 2 rats per group (Control Week 2, Treated Week 2, 4, and 6) were spatially registered to a common template using FLIRT and FNIRT with FSL Version 5.0⁷. ROI measurements were made in the corpus callosum (CC), deep cerebellar white matter (DCW), and cerebellar nuclei (CN).

Transition metals (iron, manganese, copper and zinc) were quantified in ROIs representing the DCW and CN in sections from 5 4-week control and 4-week treated animals. Metals were quantified using X-Ray fluorescence (XRF) at the Canadian Light Source (CLS) in Saskatoon using the VESPERS (Very powerful Elemental and Structural Probe Employing Radiation from a Synchrotron) beamline.

Ratio images at each timepoint were combined for display. Figure 1 shows a transverse slice through the corpus callosum (top) and cerebellum (bottom) for the control animals at 2 weeks and in cuprizone treated animals at 2, 4, and 6 weeks. The signal in CC and DCW does not decrease, however a marked decrease in signal is evident in the CN. This signal decrease can be quantified relative to the signal in muscle (Fig. 2). Corresponding H&E and LFB stained sections (Fig. 3) show that the changes in CN are due to a combination of spongiform encephalopathy and loss of myelin.

The XRF measurements revealed significant (p<0.01) reductions in iron and copper in the CN of treated rats. No significant changes in zinc or manganese were noted in the CN and there were no changes in any transition metals in the DCW.

High resolution myelin-contrast optimized MR images allowed the changes in rat brain due to cuprizone diet to be observed in vivo, and were correlated with changes observed on histology. Unlike the mouse model, which shows widespread demyelination throughout the brain, the demyelination we observed in the rat was highly localized to the cerebellar nuclei. The timing of the diet, amount of cuprizone, and strain of rat or mouse are all known to affect the degree of demyelination, and varying these parameters may also alter the pattern of demyelination. Histologic validation of this imaging technique in cuprizone rat model of demyelination will allow us to optimize the diet and timing of experiments using far fewer animals than with histology alone.

The decrease in copper in the cerebellar nuclei is consistent with the idea that cuprizone chelates copper, which leads to demyelination.