Advertisement

CSF β-amyloid predicts early cerebellar atrophy and is associated with a poor prognosis in multiple sclerosis

Published:October 21, 2019DOI:https://doi.org/10.1016/j.msard.2019.101462

      Highlights

      • Reduced cerebellar volume were found in patients with low CSF Aβ levels.
      • CSF Aβ levels were the best predictor of cerebellar GM fraction at baseline and over time.
      • Disability was predicted by baseline cerebellar parenchymal fraction.
      • CSF Aβ levels correlated with EDSS score at baseline and follow-up.

      Abstract

      Background

      Neurodegeneration is present from the earliest stages of multiple sclerosis (MS) and is critically involved in MS related clinical disability. Aim of the present study was to assess the connection between amyloid burden and early cerebellar grey matter (GM) atrophy compared to early brain GM atrophy in MS patients.

      Methods

      Forty newly diagnosed relapsing-remitting (RR-) MS patients were recruited. β-amyloid1-42 (Aβ) levels were determined in cerebrospinal fluid (CSF) samples from all subjects. All participants underwent neurological examination and brain magnetic resonance imaging (MRI) at baseline. Twenty-nine out of 40 patients repeated a brain MRI at 1-year follow-up. T1-weighted scans were segmented using the Voxel-Based Morphometry (VBM) protocol and the Spatially Unbiased Infratentorial Toolbox (SUIT) from Statistical Parametric Mapping (SPM12).

      Results

      Between-group comparison of cerebellar parenchymal fraction (GM+WM/total cerebellar volume%) showed significant differences between Aβhigh and Aβlow at baseline (p < 0.0001) and follow-up (p = 0.02). Similarly, a between-group comparison of cerebellar GM fraction (GMF) showed significant differences between Aβhigh and Aβlow at baseline (p = 0.002) and follow-up (p = 0.04). The multiple regression analysis showed CSF Aβ concentration as the best predictor of GMF both at baseline and over time (β = 0.505, β=0.377; p < 0.05). No significant results were found regarding global brain atrophy and CSF Aβ concentration.

      Conclusions

      Early cerebellar atrophy seems to be crucial in predicting a poor prognosis in MS, more than early global brain atrophy.

      Keywords

      1. Introduction

      Multiple sclerosis (MS) is the most common chronic inflammatory demyelinating disease of the central nervous system (CNS) (
      • Reich D.S.
      • Lucchinetti C.F.
      • Calabresi P.A.
      Multiple sclerosis.
      ). Although its pathologic hallmark is myelin loss, the neurodegenerative component is now considered remarkably relevant, particularly as it is related to long-term disability (
      • Filippi M.
      • Preziosa P.
      • Copetti M.
      • Riccitelli G.
      • Horsfield M.A.
      • Martinelli V.
      • et al.
      Gray matter damage predicts the accumulation of disability 13 years later in MS.
      ). Histological studies have demonstrated that neurodegeneration is reflected by imaging-derived grey matter (GM) atrophy, as assessed by magnetic resonance imaging (MRI) (
      • Filippi M.
      • Rocca M.A.
      • Barkhof F.
      • Brück W.
      • Chen J.T.
      • Comi G.
      • et al.
      Association between pathological and MRI findings in multiple sclerosis.
      ). As expected, global GM atrophy develops at a faster rate in patients with MS than healthy control subjects. Interestingly, GM atrophy is not uniform across the brain, and some regions are more susceptible than others (
      • Steenwijk M.D.
      • Geurts J.J.G.
      • Daams M.
      • Tijms B.M.
      • Wink A.M.
      • Balk L.J.
      • et al.
      Cortical atrophy patterns in multiple sclerosis are non-random and clinically relevant.
      ;
      • Preziosa P.
      • Pagani E.
      • Mesaros S.
      • Riccitelli G.C.
      • Dackovic J.
      • Drulovicet J.
      • et al.
      Progression of regional atrophy in the left hemisphere contributes to clinical and cognitive deterioration in multiple sclerosis: a 5- year study.
      ), suggesting that it occurs largely in a nonrandom manner and develops according to distinct anatomical patterns. At present, however, the biological mechanisms underlying GM atrophy in MS are still largely unknown. Consequently, evaluating neurodegeneration using novel MRI techniques, and its possible link with predictive biomarkers, has become a crucial area of research in MS (
      • Pietroboni A.M.
      • Schiano di Cola F.
      • Scarioni M.
      • Fenoglio C.
      • Spanò B.
      • Arighi A.
      • et al.
      CSF beta-amyloid as a putative biomarker of disease progression in multiple sclerosis.
      ,
      • Pietroboni A.M.
      • Caprioli M.
      • Carandini T.
      • Scarioni M.
      • Ghezzi L.
      • Arighi A.
      • et al.
      CSF β-amyloid predicts prognosis in patients with multiple sclerosis.
      ).
      Cerebellar impairment is frequent in MS and traditionally considered predictive of a negative outcome (
      • Weinshenker B.G.
      • Rice G.P.
      • Noseworthy J.H.
      • Carriere W.
      • Baskerville J.
      • Ebers G.C
      The natural history of multiple sclerosis: a geographically based study. 4. applications to planning and interpretation of clinical therapeutic trials.
      ). The cerebellum is, in fact, a predilection site for lesion development and both GM and WM demyelination are observed (
      • Parmar K.
      • Stadelmann C.
      • Rocca M.A.
      • Langdon D.
      • D'Angelo E.
      • D'Souza M.
      • et al.
      The role of the cerebellum in multiple sclerosis-150 years after Charcot.
      ). In contrast to late stage disease (
      • Kutzelnigg A.
      • Faber-Rod J.C.
      • Bauer J.
      • Lucchinetti C.F.
      • Sorensen P.S.
      • Laursen H.
      • et al.
      Widespread demyelination in the cerebellar cortex in multiple sclerosis.
      ), cerebellar cortical demyelination in early MS has been described as sporadic, whereas cerebellar atrophy has been detected earlier in patients with relapse-onset disease compared to patients with primary-progressive MS (
      • Eshaghi A.
      • Marinescu R.V.
      • Young A.L.
      • Firth N.C.
      • Prados F.
      • Jorge Cardoso M.
      • et al.
      Progression of regional grey matter atrophy in multiple sclerosis.
      ).
      Several structural and functional imaging techniques have been applied to investigate the extent of cerebellar involvement in MS and its contribution to the clinical manifestations of the disease (
      • Parmar K.
      • Stadelmann C.
      • Rocca M.A.
      • Langdon D.
      • D'Angelo E.
      • D'Souza M.
      • et al.
      The role of the cerebellum in multiple sclerosis-150 years after Charcot.
      ). Macro- and microscopic alterations in the cerebellar GM and WM were already found early in the disease course and even in absence of focal lesions (
      • Deppe M.
      • Tabelow K.
      • Krämer J.
      • Tenberge J.G.
      • Schiffler P.
      • Bittner S.
      • et al.
      Evidence for early, non-lesional cerebellar damage in patients with multiple sclerosis: DTI measures correlate with disability, atrophy, and disease duration.
      ).
      Cerebellar volume loss has not been the primary focus of MRI research for a long time, partly due to technical challenges concerning correct segmentation of the thin cerebellar gyri, sulci and cerebellar nuclei, and extraction of cerebellar tissue from adjacent infratentorial structures (
      • Parmar K.
      • Stadelmann C.
      • Rocca M.A.
      • Langdon D.
      • D'Angelo E.
      • D'Souza M.
      • et al.
      The role of the cerebellum in multiple sclerosis-150 years after Charcot.
      ). Only recently, automated segmentation tools focusing on the cerebellum were developed (
      • Diedrichsen J.
      • Balsters J.H.
      • Flavell J.
      • Cussans E.
      • Ramnani N.
      A probabilistic MR atlas of the human cerebellum.
      ), even if results seem heterogeneous.
      β-amyloid1–42 (Aβ) has been recently thouroughly investigated as a putative biomarker of neurodegeneration in MS disability. Amyloid precursor protein (APP) has been detected in MS plaques with a higher APP immunoreactivity in actively demyelinating compared to chronic lesions, thus indicating a modification of APP metabolism across disease stages (
      • Gehrmann J.
      • Banati R.B.
      • Cuzner M.L.
      Amyloid precursor protein (APP) expression in multiple sclerosis lesions.
      ). Reduced cerebrospinal fluid (CSF) Aβ levels have also been reported in MS patients (
      • Pietroboni A.M.
      • Schiano di Cola F.
      • Scarioni M.
      • Fenoglio C.
      • Spanò B.
      • Arighi A.
      • et al.
      CSF beta-amyloid as a putative biomarker of disease progression in multiple sclerosis.
      ;
      • Augutis K.
      • Axelsson M.
      • Portelius E.
      Cerebrospinal fluid biomarkers of beta-amyloid metabolism in multiple sclerosis.
      ) although the interpretation of these findings is still partially unclear (
      • Gentile A.
      • Mori F.
      • Bernardini S.
      Role of amyloid-beta CSF levels in cognitive deficit in MS.
      ). In addition, two recent studies revealed that lower baseline levels of CSF Aβ are predictive of a more severe disease progression over a 3–5 years follow-up (
      • Pietroboni A.M.
      • Schiano di Cola F.
      • Scarioni M.
      • Fenoglio C.
      • Spanò B.
      • Arighi A.
      • et al.
      CSF beta-amyloid as a putative biomarker of disease progression in multiple sclerosis.
      ,
      • Pietroboni A.M.
      • Caprioli M.
      • Carandini T.
      • Scarioni M.
      • Ghezzi L.
      • Arighi A.
      • et al.
      CSF β-amyloid predicts prognosis in patients with multiple sclerosis.
      ). Moreover, reduced florbetapir uptake, using amyloid-PET imaging, has been found in WM lesions compared to normal appearing white matter in all clinical phenotypes of MS (
      • Matías-Guiu J.
      • Cabrera-Martín M.
      • Matías-Guiu J.
      • Oreja-Guevara C.
      • Riola-Parada C.
      • Moreno-Ramos T.
      • Arrazola J.
      • Carreras J.L.
      Amyloid PET imaging in multiple sclerosis: an (18)F-florbetaben study.
      ). In fact, amyloid tracers used in PET imaging bind extensively to WM and uptake decreases significantly with demyelination.
      In this scenario, we aimed to investigate how early cerebellar volume loss interacts with cerebrospinal fluid (CSF) Aβ levels in patients with MS in order to better understand the relationship between poor prognosis in MS patients, early cerebellar damage and amyloid pathology.

      2. Materials and methods

      2.1 Subjects

      Forty patients with a new diagnosis of relapsing-remitting (RR-) MS, according to the 2010 revised McDonald criteria (
      • Polman C.H.
      • Reingold S.C.
      • Banwell B.
      • et al.
      Diagnostic criteria for multiple sclerosis: 2010 revisions to the McDonald criteria.
      ) have been recruited from January 2015 to January 2018. All patients underwent clinical assessment, brain MRI and lumbar puncture (LP) at baseline. LP was always performed in the acute phase of the disease (i.e. clinical relapse or the presence of new and/or gadolinium enhancement lesions at MRI) and before starting any treatment, including corticosteroids.
      According to a previous study (
      • Pietroboni A.M.
      • Caprioli M.
      • Carandini T.
      • Scarioni M.
      • Ghezzi L.
      • Arighi A.
      • et al.
      CSF β-amyloid predicts prognosis in patients with multiple sclerosis.
      ), MS patients were split into two groups based on CSF Aβ levels (Aβlow under 813 pg/ml and Aβhigh over 813 pg/ml).
      Twenty-nine out of 40 patients underwent extra MRI scan at 1-year follow-up.
      The main demographic and clinical characteristics of all subjects are summarized in Tables 1 and 2.
      Table 1Subjects’ demographic, laboratory, and clinical characteristics.
      VariableSM-RR (n = 40)
      Age at LP, years (mean, SD)37 ± 12.38
      Sex, female% (n)65% (26)
      Disease duration at LP, years (mean, SD)1.9 ± 4.3
      EDSS baseline, total score (median; IQ1-IQ3)2.0;2.0–3.0
      Aβ, pg/mL (mean, SD)799 ± 266
      Intrathecal IgG synthesis% (n)67.5% (27)
      Table 2Subjects’ demographic, laboratory, and clinical characteristics classified according to their CSF Aβ levels (high: >813 pg/ml; low: <813 pg/ml).
      Variablehighlowp
      Subjects (n)2020ns
      Mann–Whitney U test.
      Age at LP, years (mean, SD)33 ± 11.340 ± 12.5ns
      Mann–Whitney U test.
      Disease duration at LP, years (mean, SD)2.3 ± 5.031.5 ± 3.5ns
      Mann–Whitney U test.
      EDSS baseline, total score (median; IQ1-IQ3)2.0; 2.0–2.52.0; 2.0–3.5ns
      Median test.
      EDSS follow-up, total score (median; IQ1-IQ3)1.0; 1.0–1.02.0; 2.0–3.00.02
      Median test.
      Aβ, pg/mL (mean, SD)1021 ± 136576± 149<0.0001
      Mann–Whitney U test.
      Cerebellar GMF baseline (mean, SD)65.06 ± 2.0362.06 ± 3.40.002
      Mann–Whitney U test.
      Cerebellar GMF follow-up (mean, SD)64.65 ± 2.162.50 ± 3.40.04
      Mann–Whitney U test.
      Cerebellar WMF baseline (mean, SD)33.92 ± 1.733.43 ± 3.1ns
      Mann–Whitney U test.
      Cerebellar WMF follow-up (mean, SD)33.60 ± 2.332.70 ± 3.0ns
      Mann–Whitney U test.
      a Mann–Whitney U test.
      b Median test.
      The current study was approved by the Institutional Review Board of the Fondazione IRCCS Ca' Granda Ospedale Maggiore Policlinico (Milan, Italy). All MS patients gave their written informed consent for this research before entering the study.

      2.2 CSF collection and Aβ determination

      CSF samples were collected by LP in the L3/L4 or L4/L5 interspace. The LP was performed between 8 and 10 a.m. after one-night fasting. Then, CSF samples were centrifuged in 8000 rpm for 10 min. The supernatants were aliquoted in polypropylene tubes and stored at −80°C until use. Volume of the aliquots produced for the analyses was 500 µl. CSF cell counts, glucose, and proteins were determined. CSF Aβ levels were measured using a commercially available sandwich enzyme-linked immunosorbent assay (ELISA) kit (Fujirebio, Ghent, Belgium).

      2.3 MRI acquisition

      All patients underwent an MRI examination on Achieva 3T scanner (Philips, The Netherlands) at baseline. Twenty-nine out of 40 patients repeated an MRI scan at 1-year follow-up.
      The acquisition protocol included: 1) a 3D T1-weighted scan (TR 9.90 ms; TE 4.61 ms; Flip angle 8°; slices thickness 1 mm; gap 0); 2) a T2-weighted scan (TR 2492 ms; TE 78 ms; Flip angle 90°; slices thickness 4 mm; gap 0); 3) a Fluid attenuated inversion recovery (FLAIR) scan (TR 11,000 ms; TE 125 ms; Flip angle 90°; slices thickness 2 mm; gap 0); 4) a double inversion recovery (DIR) scan (TR 14,109 ms; TE 25 ms; Flip angle 90°; slices thickness 3 mm; gap 0).

      2.3.1 Cerebellar volumetrics

      All 3D T1-weighted scans (n = 69) were first visually inspected to exclude the presence of macroscopic artefacts. Then, all MRI scans were preprocessed using the Spatially Unbiased Infratentorial Atlas Template (SUIT) toolbox (
      • Diedrichsen J.
      A spatially unbiased atlas template of the human cerebellum.
      ), implemented in the Statistical Parametric Mapping toolbox version 12 (SPM12; http://www.fil.ion.ucl.ac.uk/spm/). Images were then segmented into GM, white WM, and CSF using the unified segmentation algorithm (
      • Ashburner J.
      • Friston K.J.
      Unified segmentation.
      ). For each scan, GM and WM fractions (GMF and WMF) were calculated, respectively, as the ratio of total WM and GM volume to the total cerebellar volume (TCV). Additionally, longitudinal changes were assessed for each patient (Δ = 1-year follow-up minus baseline volume/TCV%).

      2.3.2 Brain volumetrics

      All 3D T1-weighted scans (n = 69) were first visually inspected to exclude the presence of macroscopic artefacts. Then, all MRI scans were processed using an optimized voxel-based morphometry (VBM) protocol in SPM 12 (Wellcome Department of Imaging Neuroscience; www.fil.ion.ucl.ac.uk/spm/). Segmentation and normalization produced a GM probability map (
      • Ashburner J.
      • Friston K.J.
      Unified segmentation.
      ) in Montreal Neurological Institute (MNI) coordinates. To compensate for compression or expansion during warping of images to match the template, GM maps were modulated by multiplying the intensity of each voxel by the local value derived from the deformation field (
      • Ashburner J.
      • Friston K.J.
      Why voxel-based morphometry should be used.
      ). All data were smoothed using a 8-mm full width half maximum (FWHM) Gaussian kernel.
      We derived for each scan the WMF and GMF fractions, which were calculated, respectively, as the ratio of total WM and GM volume to the total intracranial volume (TIV). Additionally, longitudinal changes were assessed for each patient (Δ = 1-year follow-up minus baseline volume/TIV%).

      2.4 . Statistical analyses

      All statistical analyses were performed using SPSS 21.0 for Windows (SPSS Inc., Chicago, IL, USA). Due to the non-normal distribution of data, all between-group comparisons were tested by non-parametric inferential statistical analyses (Median test and unpaired Mann–Whitney U test).
      Hierarchical multiple regressions analyses between cerebellum and brain volumes as dependent variables and CSF Aβ levels and EDSS scores as explanatory variables were performed using SPSS. The regression model was adjusted in order to control for the potential effect of age, gender and disease duration.
      Spearman correlation coefficient between CSF Aβ levels and EDSS scores was assessed in all patients.
      For all statistical analyses the significance threshold was set to p ≤ 0.05.

      3. Results

      3.1 Clinical variables and CSF biomarkers

      CSF Aβ levels correlated with the EDSS score at baseline (r = −0.443, p = 0.004) and at 1-year follow-up (r = −0.371, p = 0.01).

      3.2 Cerebellar volumes

      Between-group comparison of cerebellar parenchymal fraction (GM + WM/TCV%) showed significant differences between Aβhigh and Aβlow at baseline (98.98 ± 1.63 vs 95.5 ± 3.48; p < 0.0001; Fig. 1) and follow-up (98.26 ± 3.19 vs 95.22 ± 3.49; p = 0.02; Fig. 1).
      Fig. 1
      Fig. 1(A) Mean grey matter fraction (GMF) at baseline (T0) and follow-up (T1) in the Aβhigh and Aβlow subgroups. (B) Mean cerebellar parenchymal fraction at baseline (T0) and follow-up (T1) in the Aβhigh and Aβlow subgroups. [Aβhigh: >813 pg/ml; Aβlow: ≤813 pg/ml].
      Similarly, a between-group comparison of cerebellar GMF showed significant differences between Aβhigh and Aβlow at baseline (65.06 ± 2.03 vs 62.06 ± 3.04; p = 0.002; Fig. 1) and follow-up (64.6 ± 2.1 vs 62.5 ± 3.4; p = 0.04; Fig. 1).
      Then, a multiple regression analysis (adjusted for the potential effect of age and gender) was conducted to test whether CSF Aβ baseline concentration could predict cerebellar GMF and WMF at baseline. CSF Aβ concentration significantly predicted cerebellar GMF at baseline (β = 0.505, p = 0.001; Fig. 2), but not cerebellar WMF.
      Fig. 2
      Fig. 2Scatter plot of multiple regression analysis of GMF at baseline (T0) in function of CSF Aβ levels.
      Lastly, a multiple regression analysis was conducted to test whether cerebellar volumes (GMF, WMF, ΔGMF, Δ WMF, parenchymal fraction) could predict patients’ EDSS latest score (mean disease duration 1.9 years ±4.3). EDSS was found to be significantly predicted by baseline cerebellar parenchymal fraction (β = −0.406, p = 0.025). Moreover, subjects were divided into two subgroups according to their EDSS score (≥ or < 3) at follow-up. We then performed a MANOVA analysis and found a statistically significant difference in cerebellar parenchymal fraction at follow-up based on EDSS score at follow-up. In particular, the subgroup with the highest EDSS score (EDSS ≥ 3) displayed lower overall cerebellar parenchymal fraction, F (9, 17) = 3.960, p = 0.05; Wilk's Λ = 0.377, partial η2 = 0.623.

      3.3 Brain volumes

      A multiple regression analysis (adjusted for the potential effect of age and gender) was conducted to test whether CSF Aβ baseline concentrations could predict global brain volumes at baseline and at 1-year imaging follow-up. This analysis did not show any statistically significant result (p > 0.05).

      4. Discussion

      To explore the contribution of early cerebellar damage to disease progression in MS, we first compared cerebellar volumes between Aβhigh and Aβlow patients, divided accordingly to their risk of progression based on CSF Aβ levels. As reduced CSF Aβ concentrations have been recently suggested as a prognostic biomarker in MS (
      • Pietroboni A.M.
      • Schiano di Cola F.
      • Scarioni M.
      • Fenoglio C.
      • Spanò B.
      • Arighi A.
      • et al.
      CSF beta-amyloid as a putative biomarker of disease progression in multiple sclerosis.
      ,
      • Pietroboni A.M.
      • Caprioli M.
      • Carandini T.
      • Scarioni M.
      • Ghezzi L.
      • Arighi A.
      • et al.
      CSF β-amyloid predicts prognosis in patients with multiple sclerosis.
      ;
      • Pietroboni A.M.
      • Carandini T.
      • Colombi A.
      • Mercurio M.
      • Ghezzi L.
      • Giulietti G.
      • et al.
      Amyloid PET as a marker of normal-appearing white matter early damage in multiple sclerosis: correlation with CSF β-amyloid levels and brain volumes.
      ;
      • Mattsson N.
      • Axelsson M.
      • Haghighi S.
      Reduced cerebrospinal fluid BACE1 activity in multiple sclerosis.
      ;
      • Matías-Guiu J.A.
      • Oreja-Guevara C.
      • Cabrera-Martín M.N.
      Amyloid proteins and their role in multiple sclerosis. Considerations in the use of amyloid-PET imaging.
      ), we decided to focus on this topic.
      To the best of our knowledge, there is no other study evaluating early cerebellar GM atrophy related to amyloid burden in MS.
      The Aβlow subgroup displayed higher EDSS score at follow-up and had significantly lower cerebellar GMF both at baseline and at follow-up compared to the Aβhigh subgroup. Moreover, CSF Aβ levels proved to be a significant predictor of cerebellar GMF already at the very early stages of the disease. Interestingly, cerebellar atrophy seems to be an earlier predictor of disease activity than global brain atrophy, which appears to occur at a relatively later stage.
      In the present study, we could not find any significant association between baseline global brain volumes and CSF Aβ concentration nor EDSS. It could be argued that cerebellum atrophy might precede altogether cerebral atrophy. Being CSF Aβ levels associated with neurodegeneration and inflammation and being the cerebellum particularly susceptible to both GM and WM demyelination, this could explain why CSF Aβ concentrations were able to predict cerebellar volumes very early.
      The role of inflammation as a major contributor to cerebellar GM damage is usually accepted and, traditionally, the cerebellum is considered to be particularly susceptible to inflammatory damage from the CSF (
      • Parmar K.
      • Stadelmann C.
      • Rocca M.A.
      • Langdon D.
      • D'Angelo E.
      • D'Souza M.
      • et al.
      The role of the cerebellum in multiple sclerosis-150 years after Charcot.
      ). Therefore, in the cerebellum, overlying inflammation may play a role and amplify other pathological mechanisms, such as retrograde neurodegeneration secondary to WM lesions (
      • Gilmore C.P.
      • Donaldson I.
      • Bö L.
      • Owens T.
      • Lowe J.
      • Evangelou N.
      Regional variations in the extent and pattern of grey matter demyelination in multiple sclerosis: a comparison between the cerebral cortex, cerebellar cortex, deep grey matter nuclei and the spinal cord.
      ). The presence of cerebellar volume loss from the very early stages of the disease, together with an increased risk of disease progression, might be explained by a more inflammatory milieu (
      • Eshaghi A.
      • Marinescu R.V.
      • Young A.L.
      • Firth N.C.
      • Prados F.
      • Jorge Cardoso M.
      • et al.
      Progression of regional grey matter atrophy in multiple sclerosis.
      ). However, histopathological studies showing significant widespread axonal degeneration suggest that cerebellar pathology in MS is at least in part independent of global cerebellar lesion load (
      • Pomeroy I.M.
      • Jordan E.K.
      • Frank J.A.
      • Matthews P.M.
      • Esiri M.M.
      Focal and diffuse cortical degenerative changes in a marmoset model of multiple sclerosis.
      ). Hence, there is a pressing need to identify reliable biomarkers of early cerebellar GM atrophy.
      Concerning β-amyloid and MS, the real issue is: why do CSF Aβ levels decrease in patients with early progressing disability? As already speculated (
      • Pietroboni A.M.
      • Schiano di Cola F.
      • Scarioni M.
      • Fenoglio C.
      • Spanò B.
      • Arighi A.
      • et al.
      CSF beta-amyloid as a putative biomarker of disease progression in multiple sclerosis.
      ), lower CSF Aβ levels may be associated with the activation of more aggressive pathogenic mechanisms that are already in place since the early stages of disease. By referring only to the cerebellum, we argue that a precocious more inflammatory milieu could explain the strong association between CSF Aβ levels and early cerebellar volume loss in patients with higher disability.
      Astrocytes are key regulators of the brain's inflammatory response. Astrogliosis occurs when astrocytes respond to injuries to the CNS by undergoing a spectrum of molecular and morphological changes. The impact of reactive astrogliosis in diseases is complex, and so for MS: reactive astrocytes can be both harmful and beneficial to surrounding cells and may worsen or resolve the initial CNS injury. Several studies have demonstrated that astrocytes express β-secretase (BACE1) at sufficient levels to generate Aβ, and that expression can be increased by inflammation (
      • Zhao J.
      • O'Connor T.
      • Vassar R.
      The contribution of activated astrocytes to Aβ production: implications for Alzheimer's disease pathogenesis.
      ;
      • Jin S.M.
      • Cho H.J.
      • Kim Y.W.
      • Hwang J.Y.
      • Mook-Jung I.
      Aβ-induced Ca(2+) influx regulates astrocytic BACE1 expression via calcineurin/NFAT4 signals.
      ). Specifically, proinflammatory cytokines upregulate BACE1 activity (
      • Sastre M.
      • Dewachter I.
      • Landreth G.E.
      • Willson T.M.
      • Klockgether T.
      • van Leuven F.
      • Heneka M.T.
      Nonsteroidal anti-inflammatory drugs and peroxisome proliferator-activated receptor-gamma agonists modulate immunostimulated processing of amyloid precursor protein through regulation of beta-secretase.
      ). Astrocytic expression of APP has also been demonstrated. Additionally, inflammation can upregulate APP expression and, therefore, Aβ secretion (
      • Frost G.R.
      • Li Y.M.
      The role of astrocytes in amyloid production and Alzheimer's disease.
      ). Aβ itself can be considered a proinflammatory mediator due to its ability to induce inflammation (
      • Maezawa I.1.
      • Zimin P.I.
      • Wulff H.
      • Jin L.W.
      Amyloid-beta protein oligomer at low nanomolar concentrations activates microglia and induces microglial neurotoxicity.
      ). Additionally, Aβ can stimulate proinflammatory cytokine release from astrocytes (
      • Craft J.M.
      • Watterson D.M.
      • Frautschy S.A.
      • Van Eldik L.J.
      Aminopyridazines inhibit beta-amyloid-induced glial activation and neuronal damage in vivo.
      ).
      In line with these findings and hypotheses, the correlation we describe between CSF Aβ concentration and early cerebellar volume loss suggests the cerebellum may represent a site particularly susceptible both to inflammation and neurodegeneration, and seems to confirm the prognostic role of amyloid in MS.
      Some limitations should be considered. First, this represents an exploratory study and a larger cohort of patients will be needed to confirm our findings. Second, it would be necessary to prolong the follow-up time to clarify cerebellar impairment, also in terms of WM volume loss and possible clinical manifestations. Third, our analysis refers to global cerebellar volume loss, but it would be interesting to extend our speculations investigating the cerebellum in its anatomical components, i.e. lobules and deep nuclei.

      5. Conclusion

      In conclusion, this study emphasizes the predictive role of CSF Aβ levels in MS, particularly in relation to early cerebellar involvement, and may allow an early prognosis in MS patients.

      Informed consent

      Informed consent was obtained from all individual participants included in the study.

      Funding

      This research was supported by the Italian Ministry of Health (Ricerca Corrente to ES and to FT).

      Declaration of Competing Interest

      The authors declare that they have no conflict of interest.

      References

        • Ashburner J.
        • Friston K.J.
        Why voxel-based morphometry should be used.
        Neuroimage. 2001; 14: 1238-1243
        • Ashburner J.
        • Friston K.J.
        Unified segmentation.
        Neuroimage. 2005; 26: 839-851
        • Augutis K.
        • Axelsson M.
        • Portelius E.
        Cerebrospinal fluid biomarkers of beta-amyloid metabolism in multiple sclerosis.
        Mult. Scler. 2013; 19: 543-552
        • Craft J.M.
        • Watterson D.M.
        • Frautschy S.A.
        • Van Eldik L.J.
        Aminopyridazines inhibit beta-amyloid-induced glial activation and neuronal damage in vivo.
        Neurobiol. Aging. 2004; 25: 1283-1292
        • Deppe M.
        • Tabelow K.
        • Krämer J.
        • Tenberge J.G.
        • Schiffler P.
        • Bittner S.
        • et al.
        Evidence for early, non-lesional cerebellar damage in patients with multiple sclerosis: DTI measures correlate with disability, atrophy, and disease duration.
        Mult Scler. 2016; 22: 73-84
        • Diedrichsen J.
        A spatially unbiased atlas template of the human cerebellum.
        Neuroimage. 2006; 33: 127-138
        • Diedrichsen J.
        • Balsters J.H.
        • Flavell J.
        • Cussans E.
        • Ramnani N.
        A probabilistic MR atlas of the human cerebellum.
        Neuroimage. 2009; 46: 39-46
        • Eshaghi A.
        • Marinescu R.V.
        • Young A.L.
        • Firth N.C.
        • Prados F.
        • Jorge Cardoso M.
        • et al.
        Progression of regional grey matter atrophy in multiple sclerosis.
        Brain. 2018; 141: 1665-1677
        • Filippi M.
        • Preziosa P.
        • Copetti M.
        • Riccitelli G.
        • Horsfield M.A.
        • Martinelli V.
        • et al.
        Gray matter damage predicts the accumulation of disability 13 years later in MS.
        Neurology. 2013; 81: 1759-1767
        • Filippi M.
        • Rocca M.A.
        • Barkhof F.
        • Brück W.
        • Chen J.T.
        • Comi G.
        • et al.
        Association between pathological and MRI findings in multiple sclerosis.
        Lancet Neurol. 2012; 11: 349-360
        • Frost G.R.
        • Li Y.M.
        The role of astrocytes in amyloid production and Alzheimer's disease.
        Open Biol. 2017; 7
        • Gehrmann J.
        • Banati R.B.
        • Cuzner M.L.
        Amyloid precursor protein (APP) expression in multiple sclerosis lesions.
        Glia. 1995; 15: 141-151
        • Gentile A.
        • Mori F.
        • Bernardini S.
        Role of amyloid-beta CSF levels in cognitive deficit in MS.
        Clin. Chim. Acta. 2015; 449: 23-30
        • Gilmore C.P.
        • Donaldson I.
        • Bö L.
        • Owens T.
        • Lowe J.
        • Evangelou N.
        Regional variations in the extent and pattern of grey matter demyelination in multiple sclerosis: a comparison between the cerebral cortex, cerebellar cortex, deep grey matter nuclei and the spinal cord.
        J. Neurol. Neurosurg. Psychiatry. 2009; 80: 182-187
        • Jin S.M.
        • Cho H.J.
        • Kim Y.W.
        • Hwang J.Y.
        • Mook-Jung I.
        Aβ-induced Ca(2+) influx regulates astrocytic BACE1 expression via calcineurin/NFAT4 signals.
        Biochem. Biophys. Res. Commun. 2012; 425: 649-655
        • Kutzelnigg A.
        • Faber-Rod J.C.
        • Bauer J.
        • Lucchinetti C.F.
        • Sorensen P.S.
        • Laursen H.
        • et al.
        Widespread demyelination in the cerebellar cortex in multiple sclerosis.
        Brain Pathol. 2007; 17: 38-44
        • Maezawa I.1.
        • Zimin P.I.
        • Wulff H.
        • Jin L.W.
        Amyloid-beta protein oligomer at low nanomolar concentrations activates microglia and induces microglial neurotoxicity.
        J. Biol. Chem. 2011; 286: 3693-3706
        • Matías-Guiu J.
        • Cabrera-Martín M.
        • Matías-Guiu J.
        • Oreja-Guevara C.
        • Riola-Parada C.
        • Moreno-Ramos T.
        • Arrazola J.
        • Carreras J.L.
        Amyloid PET imaging in multiple sclerosis: an (18)F-florbetaben study.
        BMC Neurol. 2015; 15: 243
        • Matías-Guiu J.A.
        • Oreja-Guevara C.
        • Cabrera-Martín M.N.
        Amyloid proteins and their role in multiple sclerosis. Considerations in the use of amyloid-PET imaging.
        Front Neurol. 2016; 7: 53
        • Mattsson N.
        • Axelsson M.
        • Haghighi S.
        Reduced cerebrospinal fluid BACE1 activity in multiple sclerosis.
        Mult. Scler. 2009; 15: 448-454
        • Parmar K.
        • Stadelmann C.
        • Rocca M.A.
        • Langdon D.
        • D'Angelo E.
        • D'Souza M.
        • et al.
        The role of the cerebellum in multiple sclerosis-150 years after Charcot.
        Neurosci Biobehav Rev. 2018; 89: 85-98
        • Pietroboni A.M.
        • Caprioli M.
        • Carandini T.
        • Scarioni M.
        • Ghezzi L.
        • Arighi A.
        • et al.
        CSF β-amyloid predicts prognosis in patients with multiple sclerosis.
        Mult. Scler. 2018 Aug 7; (1352458518791709)https://doi.org/10.1177/1352458518791709
        • Pietroboni A.M.
        • Carandini T.
        • Colombi A.
        • Mercurio M.
        • Ghezzi L.
        • Giulietti G.
        • et al.
        Amyloid PET as a marker of normal-appearing white matter early damage in multiple sclerosis: correlation with CSF β-amyloid levels and brain volumes.
        Eur. J. Nucl. Med. Mol. Imaging. 2018; 46: 280-287
        • Pietroboni A.M.
        • Schiano di Cola F.
        • Scarioni M.
        • Fenoglio C.
        • Spanò B.
        • Arighi A.
        • et al.
        CSF beta-amyloid as a putative biomarker of disease progression in multiple sclerosis.
        Mult. Scler. 2017; 23: 1085-1091
        • Polman C.H.
        • Reingold S.C.
        • Banwell B.
        • et al.
        Diagnostic criteria for multiple sclerosis: 2010 revisions to the McDonald criteria.
        Ann. Neurol. 2011; 69: 292-302
        • Pomeroy I.M.
        • Jordan E.K.
        • Frank J.A.
        • Matthews P.M.
        • Esiri M.M.
        Focal and diffuse cortical degenerative changes in a marmoset model of multiple sclerosis.
        Mult. Scler. 2010; 16: 537-548
        • Preziosa P.
        • Pagani E.
        • Mesaros S.
        • Riccitelli G.C.
        • Dackovic J.
        • Drulovicet J.
        • et al.
        Progression of regional atrophy in the left hemisphere contributes to clinical and cognitive deterioration in multiple sclerosis: a 5- year study.
        Hum. Brain Mapp. 2017; 38: 5648-5665
        • Reich D.S.
        • Lucchinetti C.F.
        • Calabresi P.A.
        Multiple sclerosis.
        N. Engl. J. Med. 2018; 378: 169-180
        • Sastre M.
        • Dewachter I.
        • Landreth G.E.
        • Willson T.M.
        • Klockgether T.
        • van Leuven F.
        • Heneka M.T.
        Nonsteroidal anti-inflammatory drugs and peroxisome proliferator-activated receptor-gamma agonists modulate immunostimulated processing of amyloid precursor protein through regulation of beta-secretase.
        J. Neurosci. 2003; 23: 9796-9804
        • Steenwijk M.D.
        • Geurts J.J.G.
        • Daams M.
        • Tijms B.M.
        • Wink A.M.
        • Balk L.J.
        • et al.
        Cortical atrophy patterns in multiple sclerosis are non-random and clinically relevant.
        Brain J. Neurol. 2016; 139: 115-126
        • Weinshenker B.G.
        • Rice G.P.
        • Noseworthy J.H.
        • Carriere W.
        • Baskerville J.
        • Ebers G.C
        The natural history of multiple sclerosis: a geographically based study. 4. applications to planning and interpretation of clinical therapeutic trials.
        Brain. 1991; 114: 1057-1067
        • Zhao J.
        • O'Connor T.
        • Vassar R.
        The contribution of activated astrocytes to Aβ production: implications for Alzheimer's disease pathogenesis.
        J. Neuroinflammation. 2011; 8: 150