Axonal injury detected by in vivo diffusion tensor imaging correlates with neurological disability in a mouse model of multiple sclerosis

Abstract

Recent studies have suggested that axonal damage, and not demyelination, is the primary cause of long-term neurological impairment in multiple sclerosis and its animal model, experimental autoimmune encephalomyelitis (EAE). The axial and radial diffusivities derived from diffusion tensor imaging have shown promise as non-invasive surrogate markers of axonal damage and demyelination, respectively. In this study, in vivo diffusion tensor imaging of the spinal cords from mice with chronic EAE was performed to determine if axial diffusivity correlated with neurological disability in EAE assessed by the commonly used clinical scoring system. Axial diffusivity in the ventrolateral white matter showed a significant negative correlation with EAE clinical score and was significantly lower in mice with severe EAE than in mice with moderate EAE. Furthermore, the greater decreases in axial diffusivity were associated with greater amounts of axonal damage, as confirmed by quantitative staining for non-phosphorylated neurofilaments (SMI32). Radial diffusivity and relative anisotropy could not distinguish between the groups of mice with moderate EAE and those with severe EAE. The results further the notion that axial diffusivity is a non-invasive marker of axonal damage in white matter and could provide the necessary link between pathology and neurological disability.

Figure 1

Clinical course for EAE induced in mice using adoptive transfer (a) and active immunization (b). Adoptive transfer EAE (n = 6) produces a more reproducible clinical course, whereas EAE induced through active immunization (n = 10) produces a greater heterogeneity in clinical course. Values are expressed as mean daily score ± s.d.

Figure 2

The DTI parameter maps at vertebral segment L1 are shown from a single mouse imaged six times in the month following induction of EAE through adoptive transfer. A decrease in white matter axial diffusivity and RA is seen beginning at the onset of clinical signs.

Figure 3

The group-averaged (n = 6) DTI parameters show a decrease in axial diffusivity (a) occurring at the onset of clinical signs that persists into the chronic phase of the disease. Radial diffusivity (b) increases early in the time course but is not significant until after the acute phase of the disease. RA (c) decreases throughout the disease course. T2-weighted SNR (d) was not significantly different at any timepoint. Values are expressed as mean ± s.d. * = post-hoc test significantly different from baseline, p < 0.05.

Figure 4

The DTI parameter maps from a control mouse (left), a mouse with clinical score of 2 (center), and a mouse with a clinical score of 4 (right) one month after the induction of EAE through active immunization. Increasing disease severity leads to an obvious loss of white matter integrity depicted as a decrease in both axial diffusivity and RA and an increase in radial diffusivity.

Figure 5

A Spearman’s rank correlation analysis demonstrates the correlation between EAE scores and DTI parameters from ventrolateral white matter. See text for additional details. * - p < 0.05.

Figure 6

Spinal cord sections were stained for injured axons (SMI-32; top) and myelin (MBP, bottom). No SMI32 positive axons are seen in the ventral white matter of the control cord (a), but the number of positively stained axons increases in moderate EAE (b; clinical score = 2) and further increases in severe EAE (c; clinical score = 4). Well-defined myelin rings are seen in the MBP stain from a control cord (d). A loss of MBP staining is evidence of demyelination in mice with moderate (e) and severe (f) EAE, and the myelin rings have lost their integrity in the EAE cord (e & f, insets). Scale bar indicates 50 μm.

Figure 7

Correlations between quantitative histological parameters and EAE scores demonstrate a strong relationship between axonal damage and EAE score (a), with the number of damaged axons differentiating mice with moderate or severe EAE. Demyelination is moderately correlated with EAE score (b), but does not differentiate mice with different EAE scores. Axial diffusivity highly correlated with the number of injured axons (c). Radial diffusivity did not correlate with the percent of demyelinated area (d). * - p < 0.05.