There is a clinical need to understand mechanisms of disease progression in MS^[1,2]. This can be achieved using experimental autoimmune encephalomyelitis (EAE), an animal model of MS, in which experimental lesions are induced and longitudinally investigated with MR. However, MR studies investigating disease progression in EAE are under-explored due to technical challenges and the need for very high magnetic field scanners^[3,4].To investigate longitudinal changes in metabolite concentrations and lesion development, in relation to neurological deficits in EAE, using 9.4T MRI and ¹H-MRS.

Study Design: Experiments were performed on 18 Dark Agouti (DA) adult rats. EAE was induced in 12 animals and compared with 6 controls. EAE disease progression was assessed longitudinally over 5 time-points, where animals were scored for neurological deficits and received an MR scan. Time-points included were: (1) Baseline: 0 days (no immunization, no deficits), (2) Pre-symptomatic: 2-4 days post-immunization (dpi) (no deficits), (3) Onset: 9-11 dpi (visible deficits), and (4) Recovery: 16-18 dpi (deficits may improve), (5) Relapse: 20-25 dpi (highest deficits).

EAE Induction: Animals were anesthetized with 2% isoflurane. EAE animals were immunized with a subcutaneous injection (0.75mg) of recombinant myelin oligodendrocyte glycoprotein (rMOG) in incomplete Freund’s adjuvant (IFA), at the tail base. Controls were immunized with IFA alone.

Neurological Deficit Scoring (deficit-score): Animals were scored daily using the following 10-point system, receiving one point for each deficit: (1) tip weakness, (2) whole-tail paralysis, (3) toe spread reflex, (4) unsteady gait, (5) hind-limb weakness, (6) hind-limb paralysis, (7) righting reflex, (8) fore-limb weakness, (9) fore-limb paralysis, and (10) moribund.

MR Experiments: A 9.4 T Agilent scanner with a transmit volume coil (Ø =72 mm) and two-element array coils were used to obtain ¹H-MRS and T2-weighted (w) images.

Single-voxel ¹H-MRS: Two volumes of interest (VOIs) were studied: the thalamus (Thal), composed of gray matter, and corpus callosum-hippocampus (CC-Hipp), including gray and white matter (Figure 1). We used sequence combining VAPOR^[5] and PRESS^[6]. Acquisition parameters were: averages=128, TE/TR=7.5/5000ms, sample-points=512, BW=3613Hz, CC-Hipp voxel=6.9x2.0x3.0mm3(41μl), Thal voxel=6.5x1.9x3.0mm3(37μl). Spectra were phase-corrected and normalized with a non-water-suppressed reference, then quantified using Matlab and LCModel. Quantified metabolites included: total N-acetyl aspartate (NAA+NAAG) ([NAA]), total choline (Cho+PCho+CPC), total creatine (Cr+PCr), glutamate, glutamine, myo-inositol, GABA and lipids (Figure 1).

T2-w imaging: TE/TR=20/3000ms, FOV=27.7x27.7x0.5mm3, slices=40, scan-time=13min, in-plane resolution=0.1mm. T2-w images were registered to their corresponding baseline using NiftyReg to assess longitudinal lesion development. T2-w image contrasts were normalised to the baseline using a LUT-histogram-matching algorithm in MIPAVv5.4.3 ensuring the image contrasts were accurately compared longitudinally. VOIs which corresponded to the spectroscopic voxels were manually defined on the T2-w images using MIPAVv5.4.3.

Statistics: EAE animals were divided into two subgroups: ‘Mild-EAE’ if their deficit-scores never exceeded 6 and ‘Severe-EAE’ if their deficit-scores were greater than 6 after Onset. ANOVAs were used to investigate differences in metabolite concentrations between groups; linear regressions were used for associations between metabolite concentration and neurological deficit-score. Only results with p-values <0.05 were reported. T2-w lesions in VOIs were visually detected and only hyper-intense lesions appearing after Baseline, and in the VOI were manually counted.

Six EAE animals developed mild disease (mean peak deficit-score=3), six developed severe disease (mean peak deficit-score=8), and six were control animals (mean deficit-score=0). When investigating longitudinal metabolite changes, Severe-EAE animals had higher [NAA] in the Thal and CC-Hipp, than both Mild-EAE and Controls at the Pre-symptomatic time-point (Thal: p<0.015,p<0.003 respectively; CC-Hipp: p<0.037,p<0.047 respectively (Figure 2a)). When investigating T2-w lesions, hyper-intense lesions were only visible in the CC-Hipp VOI. Severe-EAE animals had more lesions compared to Mild-EAE and Controls (Severe-EAE: mean 3[range 1-4], Mild-EAE: 1[0-2], Control: 0[0])(Figure 2b). When investigating the association between metabolites and deficit-scores, higher [NAA] in both the Thal and CC-Hipp were associated with greater EAE disability. More specifically, higher [NAA] at the Pre-symptomatic time-point correlated with higher deficit-score at Onset (regression coefficient 2.19 [[95% confidence interval]: 1.08; 3.30],p<0.003), Recovery (2.81[0.82; 4.81],p<0.019), and Relapse (4.08[2.73; 5.43],p<0.001), after adjusting for Thal [NAA] at Baseline (Figure 3). Similarly, there was an association between higher [NAA] in the CC-Hipp at the Pre-symptomatic time-point, and higher deficit-score at Recovery (1.23[0.16; 2.39],p<0.041) and Relapse (2.65[0.98; 4.32],p<0.018), after adjusting for CC-Hipp [NAA] at Baseline.Before visible signs of neurological deficits, higher [NAA] predicts the severity of neurological deficits in EAE. Considering NAA is predominantly associated with neuronal mitochondria^[7,8], the increase in Pre-symptomatic [NAA] and its relationship with disability may indicate an early-stage mitochondrial response in rMOG-induced EAE. Therefore, the Pre-symptomatic association between [NAA] and disability-score in EAE suggests that [NAA] may reflect pathological processes relevant to MS disease course and treatment targets.