Quantitative magnetization transfer imaging in postmortem multiple sclerosis brain

Abstract

Purpose: To investigate the relationship of myelin content, axonal density, and gliosis with the fraction of macromolecular protons (fB) and T2 relaxation of the macromolecular pool (T2B) acquired using quantitative magnetization transfer (qMT) MRI in postmortem brains of subjects with multiple sclerosis (MS).

Materials and methods: fB and T2B were acquired in unfixed postmortem brain slices of 20 subjects with MS. The myelin content, axonal count, and severity of gliosis were all quantified histologically. t-Tests and multiple regression were used for analysis.

Results: MR indices obtained in unfixed postmortem MS brains were consistent with in vivo values reported in the literature. A significant correlation was detected between Tr(myelin) (inversely proportional to myelin content) and 1) fB (r = -0.80, P < 0.001) and 2) axonal count (r = -0.79, P < 0.001). fB differed between 1) normal-appearing white matter (NAWM) and remyelinated WM lesions (rWMLs) (mean: fB 6.9 [SD 2] vs. 4.0 [1.8], P = 0.01), and 2) rWMLs and demyelinated WMLs (mean: 4.2 [2.2] vs. 2.5 [1.3], P = 0.016). No association was detected between T2B and any of the histological measures.

Conclusion: fB in MS WM is dependent on myelin content and may be a tool to monitor patients with this condition.

Figure 1

Correlation of MRI and histopathology in a postmortem brain slice (A) of a subject with MS. Four exemplary regions of high signal (RHS) on T₂-weighted MRI (B) are pathologically confirmed MS WM lesions (WMLs, a-h). The RHS are coregistered (circles) to the respective MTR (C), f_B (D), and T₁-RT (E) maps. On sections stained for LFB all WMLs appear demyelinated (a-d and h). For one chronic inactive lesion sections stained for LFB (d), CD68 (c), BIEL (e and g), and GFAP (f) are shown. The myelin content and extent of gliosis were assessed by measuring transmittance (see text) on sections stained for LFB (a, b, d, and h) and GFAP (f), respectively, at a final magnification of × 125. Axonal counts were determined on sections stained for BIEL at a final magnification of × 1250 (g).

Figure 2

Correlation of MRI (A-D) and histopathology (a-m) in postmortem MS brain. Five exemplary RHS on T₂-weighted MRI (A) are pathologically confirmed MS WMLs (a-m). The respective MTR (B), f_B (C), and T₁-RT (D) maps are shown. For one chronic inactive (a-e) and one chronic active lesion (i-m) sections stained for H&E (a and i), LFB (b and j), CD68 (c and k), GFAP (d and l), and BIEL (e and m) are shown. Note the inflammatory infiltrate on H&E- and CD68-stained sections of the active WML (i and k) and compare with the respective sections of the chronic hypocellular WML (a and c).

Figure 3

Scatterplots of myelin content (Tr_myelin) on slides stained for LFB vs. (a) the fraction of macromolecular protons (f_B), and (b) axonal count in postmortem brains of 20 and 37 patients with MS, respectively (r-values are within-patient correlation coefficients). A high value of Tr_myelin reflects low myelin content.