|Home | About | Journals | Submit | Contact Us | Français|
White matter (WM) lesions are the classic pathological hallmarks of multiple sclerosis (MS). However, MRI-based WM lesion load shows relatively poor correlation with functional outcome, resulting in the “clinico-radiological paradox” of MS. Unlike lesion based measures, volumetric MRI assessment of brain atrophy shows a strong correlation with functional outcome, and the presence of early atrophy predicts a worse disease course. While extensive literature exists describing MRI characteristics of atrophy in MS, the exact pathogenesis and the substrate of atrophy - gray vs WM loss, axonal/neuronal damage vs demyelination, or a combination of the above – remain unclear. Animal models of atrophy would allow for detailed investigations of the pathomechanism, and would contribute to an enhanced understanding of structural-functional connections in this complex disease. We now report that in the Theiler’s Murine Encephalitis Virus (TMEV) model of MS in SJL/J mice, significant brain atrophy accompanies the development of the progressive MS-like disease. We conducted volumetric MRI studies in 8 cases and 4 age, gender and strain matched control mice. While in controls we did not detect any brain atrophy, significant atrophy developed as early as 3 months into the disease course, and reached its peak by 6 months, resulting in ventricular enlargement by 118% (p= 0.00003). A strong correlation (r=−0.88) between atrophy and disability, as assessed by rotarod assay, was also demonstrated. We earlier reported another neurodegenerative feature in this model, the presence of deep gray matter T2 hypointensity in thalamic nuclei. Future studies utilizing this model will allow us to investigate key components of MRI detectable neurodegenerative feature development, their tissue correlations and associations with functional outcome measures. These studies are expected to pave the way to a better understanding of the substrate of disability in MS models.
Multiple sclerosis (MS) is an inflammatory demyelinating disease of the central nervous system (CNS) and is the leading cause of non-traumatic disability among young adults (Noseworthy et al., 2000). MRI studies sensitively detect white matter (WM) lesions, and MRI-based lesion descriptors are now part of the MS diagnostic criteria (McDonald et al., 2001). However, lesion based measures generally show a poor correlation with functional outcome (Zivadinov and Leist, 2005). The apparent disconnect between lesion load and functional status is often referred to as the “clinico-radiological paradox of MS” (Barkhof and van Walderveen, 1999). However, this paradox is at least partially resolved by advanced MRI studies of normal appearing (NA) brain areas and by volumetric measures of brain atrophy (Barkhof, 2002; Bermel and Bakshi, 2006; Zivadinov and Leist, 2005). Both early cortical atrophy (De Stefano et al., 2003) as well as ongoing central atrophy (Kalkers et al., 2002) have been reported. The pathogenesis of atrophy and its relationship to gray or WM damage remains unclear. Numerous papers established a correlation between cortical and subcortical gray matter (GM) pathology and atrophy, while others demonstrated that the main contribution to atrophy is through WM pathology (Bermel and Bakshi, 2006). From the standpoint of GM involvement (Pirko et al., 2007), it is unclear whether a primarily neuronal process causes cell and tissue loss, or whether axonal transection and related Wallerian degeneration is the culprit. Regarding the potential WM involvement, it is unclear whether lesional pathology or changes occurring in NAWM contribute more to atrophy.
Animal models of brain atrophy would be critically important in determining the potential pathomechanism(s), clarifying the substrate(s) of CNS atrophy, and providing additional clarifications to the structural-functional relationship in this complex disease. Recently, cerebellar cortical atrophy was elegantly demonstrated in experimental allergic encephalomyelitis, a model of MS (MacKenzie-Graham et al., 2006). However, the extensive central atrophy leading to enlargement of CSF spaces as seen in MS has not yet been reported in animal models.
Theiler’s Murine Encephalitis Virus (TMEV) infection of mice is an accepted MS model (Nelson et al., 2004; Tsunoda and Fujinami, 2009). In susceptible mice TMEV infection results in a chronic-progressive demyelinating disease, with clinical features similar to progressive forms of MS, eventually rendering the animal paraplegic within 9–12 months. Our hypothesis was that in this model, progressive central atrophy accompanies the demyelinating condition. This hypothesis was based on our preliminary observations of ventricular enlargement in chronically TMEV infected and significantly disabled SJL mice. The goal of this pilot study was to determine whether a neurodegenerative component leading to brain atrophy is truly detectable in this model, amd to correlate functional outcome measures to quantitative MRI metrics.
Eight TMEV infected SJL mice and 4 controls were scanned at 0, 1, 2, 3, 4, 6 and 12 months post infection, using volume acquisition T1 weighted (FLASH sequence, TRl: 15ms, TE:4.5ms, NEX:2, FOV: 3.2×1.92×1.92 cm, matrix: 256×128×128) and T2 weighted (RARE sequence, TR:1500ms, TE:70ms, RARE factor: 16, FOV: 3.2×1.92×1.92 cm, matrix: 256×128×128) sequences in a Bruker Biospec 7 Tesla horizontal bore small animal MRI system (Bruker Biospin, Ettlingen, Germany). All protocols have been approved by the local IACUC committee. Inhalational anesthesia was used for the imaging procedure as described by us previously (Pirko et al., 2008). The animals’ heart and respiratory rate as well as core temperature were monitored continuously using a SA Instruments system while in the scanner (SA Instruments Inc, Stony Brook, NY). There was no animal loss during the imaging procedure. We lost 4 mice due to the natural course of this model after the 6 month time point, and one control mouse.
Slice extraction for visualization and presentation, total brain and ventricular volume measurements were performed using the coregistration, slice extraction, and the semiautomated volumetric 3D ROI and 3D Scan tools in Analyze 8.0 (Robb, 1999). For determination of atrophy, we used lateral ventricle volumes, as reported previously in the literature (Berg et al., 2000; Brex et al., 2000; Dalton et al., 2002; Redmond et al., 2000). Of note, while brain parenchymal fraction (BPF) is the preferred method to determine brain atrophy in human MS, as it compensates for head and brain size variations, such interventions are not needed when imaging genetically identical inbred mice, as skull and brain size variations are negligible. In addition, the murine model does not include the presence of periventricular lesions, therefore the measurement of ventricular volumes are not compromised by “pseudo-enlargement” due to periventricular lesions that otherwise may show the same signal intensity as CSF on T2 weighted images.
After signal intensity normalization across the different time points, a thersholding-based segmentation was applied to the ventricular system. From the segmented T2 weighted images, non-ventricular areas with the same high signal intensity were manually removed. These mostly included non-CNS structures such as the corpus vitreum and subcutaneous fat. On the segmented images, the 3D Scan Tool in Analyze was applied to calculate the total CSF volume per mouse. Due to the semi-automated nature of this process, two investigators contributed to this process, both trained in test datasets derived from previously completed projects. All datasets were analyzed twice by each investigator. Their intra-and inter-rater correlation was monitored and found to be superior. The intra-class correlation coefficient of the 2 raters based on 8 infected mice was 0.92 (95% CI 0.85 to 0.99) and 0.91 (95% CI 0.86 to 0.97). The inter-rater comparison of the volumetric analysis did not show any statistically significant difference between the raters (P = .94).
We monitored disability by monthly Rotamex rotarod assay (Columbus Instruments, Columbus, OH), as reported earlier (Pirko et al., 2009). Rotarod scores provide an objective and more refined spectrum of disability measurement compared to graded analogue disability scales. This assay determines the time the animals are able to walk on a constantly accelerating rotating rod, and is a commonly used measure of disability in mice, capturing motor, sensory and coordination/balance-related components. 2 attempts were given to each animal at each time point, and the better score (the score suggestive of less disability) was used at each time point for data analysis.
Intergroup differences were analyzed using Student’s t-test and Pearson’s correlation statistics in SPSS (SAS Institute, Cary, NC).
Significant brain atrophy, resulting in ventricular enlargement was found as early as 3 months (p=0.005). A 32.4% increase in the lateral ventricular volumes was seen at when compared to controls. Brain atrophy reached its peak by 6 months post infection when a 101.8% increase was seen compared to controls (p= 0.00003), and the overall increase compared to the initial time point was 111.8% (Figure 1). The increase of atrophy between 4–6 months was again significant (p=0.03), whereas between months 6–12, we did not see a significant change, although the volume at 12 months is higher than at 6 months (p=0.3). Also of note is that due to animal loss naturally occurring in the studied mice, our control group had only 3 animals whereas the cases were reduced to 4 at the 12 month time point. In our age and gender matched, strain identical control mice, statistical analysis revealed no significant difference (p=0.5) between the 1 and 12 month brain atrophy data. It would not be unexpected to see some age-related atrophy in controls, and it is possible that by using a larger number of control mice, this “natural” atrophy would also reach significance.
Differences in disability as assessed by rotarod revealed no significant difference at 1 months (p=0.88) and at 3 months (p=0.45), whereas at 4 months the difference reached significance (72% lower compared to controls, p=0.0005) and it remained highly significant throughout the experiment. The progression of motor disability was still significant when comparing 6 months to 12 months data (p=0.003) but beyond 9 months the changes were not significant (9 to 12 months p=0.1).
In this model, the development of brain atrophy preceded the development of rotarod detectable disability. When the first 4 months of data is analyzed, and rotarod disability scores are correlated with the volumetric MRI changes, a very strong (r=−0.88) negative correlation is seen between the disability scores and the relative increase in lateral ventricular volumes: the larger the ventricles get due to atrophy, the lower the rotarod scores become (Figure 4).
In this murine model of MS, we determined the rate of progressive brain atrophy and its relationship to disability. Our results demonstrated the development of highly significant brain atrophy, which correlated with the progression of disability. This functional outcome – MRI measure correlation is among the strongest correlations ever reported between disability and paraclinical markers in mouse models of MS. Of note, the development of atrophy slightly preceded the detectability of disability by the rotarod assay: atrophy was already statistically significant at 3 months, whereas the applied disability measure only reached significance at 4 months. This either means that atrophy truly precedes the onset of disability – and in human MS, there are observations suggestive of atrophy being present at early, non-disabling stages of the illness (Benedict et al., 2002; Benedict et al., 2004; Dalton et al., 2002) – or it may represent an assay sensitivity issue: rotarod assay may not capture all aspects of disability, and may not be sensitive enough to lower levels of disability. Other functional outcome measures, such as activity box, hanging wire assay and footprint analysis will also be performed in future experiments to fully understand the correlation between the MRI findings and clinical outcome.
The presence of cerebellar cortical atrophy using ex vivo MRI was reported in an experimental allergic encephalomyelitis (EAE) model of MS: A 7% decrease in cerebellar volume was found in C57BL/6 mice with myelin oligodendrocyte glycoprotein (MOG) induced EAE in the late disease stage (MacKenzie-Graham et al., 2006). The same group also recently reported substantial loss in the molecular cell layer of the cerebellum, and that Purkinje cell loss contributes to the observed atrophy (MacKenzie-Graham et al., 2009). In the same MOG induced EAE model, hippocampal atrophy was also recently demonstrated using microscopy-based measures. (Ziehn et al.) These publications did not address whether central atrophy, as seen in advanced MS cases and in our model, also accompanies the observed pathology in MOG induced EAE in C57BL/6 mice. In a marmoset-based EAE model, cerebral cortical atrophy was demonstrated using microscopy-based quantitative methods. The observed diffuse cortical thickness reduction was not related to the presence of cortical lesions.(Pomeroy et al., 2008) This paper also did not discuss whether central cerebral atrophy accompanies the development of the MS-like disease in this model.
It is also of special note that in MS mouse models in general, brain lesions are less frequently seen than spinal cord lesions, and our model is no exception to this rule. The fact that brain lesions are not commonly seen in this model, yet brain atrophy develops underlines the importance of changes occurring in the normal appearing gray and white matter areas. The importance of potential Wallerian degeneration giving rise to the above findings is the subject of a planned extension to our study. In addition to Wallerian degeneration, it is also possible that direct gray matter pathology is responsible for the observed atrophy. It is known that there is no substantial demyelination in the cerebral cortical and subcortical gray matter in the late stages of this model, but inflammatory infiltration is often seen. It is possible that similar to the above discussed marmoset model, a diffuse loss of gray matter volume unrelated to the presence or absence of actual demyelinated gray matter plaques may be the reason for the observed atrophy. In this model and in MS an “inside out”-type demyelination has also been suggested: a primary axonal and/or neuronal “hit” may lead to secondary demyelination, therefore a potential direct neuronal pathology is also to be investigated as a potential cause for the observed atrophy.(Tsunoda and Fujinami, 2002)
In addition to the above demonstrated brain atrophy, we earlier reported the development of another MS-related neurodegenerative feature in this MS model: the presence of T2 hypointensity in deep gray nuclei, most notably the thalamus (Pirko et al., 2009). The development of T2 hypointensity also showed a strong correlation with disability. This represents another overlap between known features of the human disease and our MS model, and emphasizes the relevance of this model to neurodegenerative aspects of MS. This finding also suggests that an as of yet poorly understood deep gray matter pathology may be at least partially responsible for the observed atrophy.
We demonstrated strong functional-structural correlations between MRI-based brain atrophy measurements and rotarod detectable disability in the TMEV induced MS model in SJL/J mice. In this pilot study we could not aim for the full determination of the pathomechanism(s) and we also could not address MRI-tissue correlation, as animals were not sacrificed at each time point. Future experiments in this model will allow us to characterize the substrates of atrophy and the pathomechanism responsible for the development of the demonstrated neurodegenerative features.
Funding source: This study was funded by the National Multiple Sclerosis Society and by the National Institutes of Health (R01NS058698).
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.