Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Immunohistochemical and Molecular Investigations Show Alteration in the Inflammatory Profile of Multiple System Atrophy Brain

Immunohistochemical and Molecular Investigations Show Alteration in the Inflammatory Profile of... Key Words: CX3CL1, CX3CR1, Cytokine, Microglia, Multiple Multiple system atrophy (MSA) is an adult-onset neurodegenerative system atrophy, NanoString, Neuroinflammation, Neuropathology. disease characterized by aggregation of a-synuclein in oligodendro- cytes to form glial cytoplasmic inclusions. According to the distribution of neurodegeneration, MSA is subtyped as striatonigral degeneration INTRODUCTION (SND), olivopontocerebellar atrophy (OPCA), or as combination of Multiple system atrophy (MSA) is a sporadic adult- these 2 (mixed MSA). In the current study, we aimed to investigate re- onset progressive neurodegenerative disease characterized gional microglial populations and gene expression in the 3 different clinically by parkinsonism, cerebellar ataxia, and autonomic MSA subtypes. Microscopy with microglial marker Iba-1 combined failure. The disease is confirmed pathologically by accumu- with either proinflammatory marker CD68 or anti-inflammatory marker lated a-synuclein protein within oligodendrocytes forming Arginase-1 was analyzed in control, SND, and OPCA cases (n¼ 5) glial cytoplasmic inclusions as well as variable numbers of using paraffin embedded sections. Western immunoblotting and neuronal nuclear or cytoplasmic inclusions and regional neu- cytokine array were used to determine protein expression in MSA rodegeneration. Neuropathological classification of MSA and control brain regions. Gene expression was investigated using depends on the brain regions most severely affected by neuro- the NanoString nCounter Human Inflammation panel v2 mRNA nal loss: striatonigral degeneration (MSA SND), olivoponto- Expression Assay. Analysis of neuropathological subtypes of MSA cerebellar atrophy (MSA OPCA), or as a mixed form (mixed demonstrated a significant increase in microglia in the substantia MSA). Rare neuropathological subtypes are also described, nigra of OPCA cases. There was no difference in the microglial ac- known as minimal change MSA and atypical MSA (1, 2). tivation state in any region. Cytokine expression in MSA was com- Neuroinflammation is a dynamic response, resulting in parable with controls. Decreased expression of CX3CL1 precursor changes in glial cell function and morphology, upregulated protein and significantly greater CX3CR1 protein was found in transcription of inflammation linked genes and cytokine pro- MSA. NanoString analysis revealed the >2-fold greater expression duction. Neuroinflammation has been identified as a hallmark of ARG1, MASP1, NOX4, PTGDR2,and C6 in MSA. of neurodegenerative diseases including Alzheimer disease (3, 4), Parkinson disease (5), and dementia with Lewy bodies (6) among others. Proinflammatory neuroinflammation has also been shown to be a normal process of nondiseased aging From the Queen Square Brain Bank for Neurological Disorders, Department (reviewed in [7]). We have previously shown that microglia, of Molecular Neuroscience, UCL Institute of Neurology, London, UK (APK, CEM, SCF, BCB, RC, CS, TL, JLH). the key cellular component of the neuroinflammatory re- Send correspondence to: Janice L. Holton, PhD, MRCPath, Queen Square sponse, are increased in number in a group of MSA cases with Brain Bank for Neurological Disorders, UCL Institute of Neurology, mixed neuropathological subtypes (8). Others have reported London WC1N 1PJ, UK; E-mail: [email protected] This study was made possible due to funding awarded by the Multiple System similar upregulation of astrocytes along with microglial acti- Atrophy Trust (UK) which covered A.P.K.’s salary and consumable costs vation in MSA brain (9, 10). Microglia had generally been for the study excluding NanoString. Additional funding was awarded by proposed to have a bipolar activation style and in the absence the British Neuropathological Society to facilitate the NanoString compo- of insult, toxin, or disease, they exist in a ramified state. In nent of the investigation. T.L. is supported by an Alzheimer’s Research their ramified state, microglia have constantly motile pro- UK Senior Fellowship. C.E.M. is supported by an Alzheimer’s Research UK PhD Studentship. B.C.B. is supported by the Leonard Wolfson Centre cesses surveying their local area (11). Microglial processes for Experimental Neurology. The Queen Square Brain Bank is supported may then detect different inducers that can stimulate the cells by the Reta Lila Weston Institute for Neurological Studies, the Progressive to become activated either to a proinflammatory, phagocytic Supranuclear Palsy Association and the Medical Research Council UK. (M1 state), an anti-inflammatory, prohealing (M2 state), The authors have no duality or conflicts of interest to declare. Supplementary Data can be found at http://www.jnen.oxfordjournals.org. or a “resolution phase macrophage” (rM), which has hybrid V C 598 2018 American Association of Neuropathologists, Inc. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/ by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. Downloaded from https://academic.oup.com/jnen/article/77/7/598/5025079 by DeepDyve user on 19 July 2022 J Neuropathol Exp Neurol � Volume 77, Number 7, July 2018 Inflammatory Profile of MS expression of both M1 and M2 phenotypic markers (12). In this way, we aimed to gain a broad understanding of the In recent years, general consensus has begun to move away neuroinflammatory characteristics of MSA to gain insights from the bipolar theory of microglial activation in favor of a into disease mechanisms and provide information contributing spectrum of activation (13) in which active microglia may to the development of disease biomarkers and therapeutic fluctuate rapidly through multiple degrees of activation targets. expressing a continuum of variable markers. Interestingly, a recent study in a mouse model of trau- MATERIALS AND METHODS matic brain injury has shown gradual upregulation of M1 marker RNA expression postinjury, while M2 marker expres- Case Selection sion peaked rapidly postinjury before declining in expression Cases were selected for each part of the study (immuno- (14). Although this study examined the influence of an acute fluorescence, cytokine arrays, Western immunoblotting, and insult on the dynamics of the M1 to M2 microglial ratio, it NanoString analysis) based on diagnosis, tissue availability, might be anticipated that in neurodegeneration an acute re- gender and age pairing and postmortem delay, we sought to sponse will give way to either a long-term chronic or dysfunc- exclude cases that had concomitant pathologies that might ex- tional response. Indeed, both nitrated and aggregated forms of acerbate any inflammatory response. Due to these restrictions, a-synuclein protein, which are present in MSA (15, 16), have we were unable to use exactly the same cohort for the entire been shown to stimulate a pro-inflammatory microglial pheno- investigation (Supplementary Data Tables S1 and S2). type (17) and may maintain this activation during the disease course of MSA. We hypothesize that microglia in MSA have Human Brain Tissue become aberrantly activated, potentially exacerbating the de- We used brain tissue from cases donated to the Queen velopment of pathological changes. Square Brain Bank for Neurological Disorders, UCL Institute We aimed first to investigate whether differences in of Neurology. The brain donation program and protocols have microglial number, which we previously reported in MSA received ethical approval for research by the NRES Commit- brain in a group of cases unselected for neuropathological tee London—Central and tissue is stored for research under a subtype, would be reproduced when analyzing MSA cases by license issued by the Human Tissue Authority (No. 12198). disease subtype. For this, we selected MSA-SND and MSA- According to routine protocols, one-half of the fresh OPCA as the 2 disease subtypes showing the widest differen- brain was sliced in the coronal plane, flash frozen, and stored tial in pathological involvement of different brain regions. We chose substantia nigra because it is severely involved in at 80 C. The remaining half brain was fixed in 10% buffered MSA-SND and cerebellar white matter as a region with exten- formalin before being sliced in the coronal plane. Tissue sive pathology in MSA-OPCA. We also included the posterior blocks were selected for paraffin wax embedding and histol- frontal lobe because this region shows a moderate degree of ogy. This material was used for multi-immunofluorescence a-synuclein pathology in both subtypes of MSA. Next, we studies. Frozen tissue was dissected from the posterior frontal wished to investigate the activation state of these microglia, cortex and the cerebellar white matter and either homogenized whether they be in a pro- or anti-inflammatory state. for Western immunoblotting or lysed and processed for total To investigate additional aspects of the inflammatory re- RNA using the RNeasy mini prep kit according to the manu- sponse in MSA, we selected cases of mixed MSA to minimize facturer’s instructions (Qiagen, Manchester, UK). any influence of brain regional variability in pathology. The Pathological confirmation of MSA subtype diagnosis functional output of the inflammatory response can be quanti- based on the Ozawa criteria (25) was performed using paraffin fied and assessed by assaying the expression of cytokines and embedded sections (8 lm) stained for relevant markers chemokines. To understand whether the functional inflamma- (including hematoxylin and eosin and Luxol fast blue/cresyl tory output differs between control and MSA, we compared violet and a-synuclein). Multi-immunofluorescence was per- the expression of 36 cytokines and chemokines using commer- formed using isotype specific antirabbit IgG, antimouse IgG, cially available assay systems. CX3CL1 or fractalkine is a or antigoat IgG secondary antibodies conjugated with either chemokine which, together with its specific receptor Alexa 488 or 594 fluorescent dyes (1:400) (Life Technologies, CX3CR1, is abundantly expressed in the CNS, mediating Paisley, UK) followed by quenching of autofluorescence with neuron-microglial interaction (18). The glycosylated form of 0.1% Sudan black/70% ethanol (Sigma-Aldrich, Dorset, UK) the neuronal CX3CL1 can bind microglial CX3CR1 and solution for 10 minutes and mounting with glass coverslips us- prevent the activation of potentially harmful inflammatory ing VECTAshield mounting media with 4’,6-diamidino-2- microglial phenotypes (19). The expression of CX3CL1 phenylindole (DAPI) nuclear stain (Vector Laboratories, and CX3CR1 can be altered in the presence of injury or insult Peterborough, UK). Images were visualized using fluores- (20–22) and impaired CX3CR1 function may exacerbate neu- cence microscopy (Leica DM5500 B). rodegeneration (23). To investigate whether similar changes contribute to the inflammatory process in MSA, we deter- Microglial Cell Counting mined the expression of CX3CL1 and CX3CR1 in brain We selected age matched control, SND, and OPCA tissue. Finally, to provide a wider understanding of the inflam- cases (n¼ 5 each) for analysis. For each case, we chose se- matory response in MSA, we examined the expression of 249 genes associated with inflammation in disease cases compared verely affected brain regions: subcortical white matter of the with controls using NanoString nCounter technology (24). frontal lobe, substantia nigra, and cerebellar white matter. 599 Downloaded from https://academic.oup.com/jnen/article/77/7/598/5025079 by DeepDyve user on 19 July 2022 Kiely et al J Neuropathol Exp Neurol � Volume 77, Number 7, July 2018 Sections (8 lm) were probed with Iba-1, CD68, and arginase- RNA Isolation 1. In each region, 5 areas were randomly selected according to Total RNA was isolated from posterior frontal lobe and a predesigned zig-zag pattern and we counted the number of cerebellar white matter homogenates using RNAeasy mRNA cells that stained positively for Iba-1 as a marker of total kit (Qiagen) according to the manufacturer’s instructions. microglia and were of a suitable size and morphology, present RNA quantity and quality was assessed using a BioSpectrom- in the field of view at 20 magnification. Area of field of eter (Eppendorf, Stevenage, UK). view at 20 using an eyepiece magnification of 25 was cal- culated to be 1.19 mm and all cell counts were expressed as NanoString nCounter Analysis number of cells counted/mm . In addition, we assessed the To examine the inflammatory environment of MSA number of cells per field at 20 that showed colocalization by mRNA expression was determined using the NanoString multi-immunofluorescence microscopy of either Iba-1 and nCounter Human Inflammation panel v2 mRNA Expression CD68 as a marker of pro-inflammatory microglia or Iba-1 and Assay (NanoString Technologies). Total mRNA was isolated Arg-1 as a marker of anti-inflammatory microglia. from control (n¼ 6) and mixed MSA (n¼ 6) posterior frontal lobe and cerebellar white matter (Supplementary Data Table Human Cytokine Array S2) The protocol was carried out by evaluation project techni- cians at the Seattle NanoString facility as a proof of concept To analyze cytokine expression in control (n¼ 6) and study using 150 ng of total RNA from each sample following mixed MSA (n¼ 6) we selected posterior frontal cortex as a their commercial protocol. Data were collected using the region showing pathological involvement and used the human nCounter Digital Analyser. cytokine array A (R&D Systems, Abingdon, UK) according to the manufacturer’s instructions. Briefly, nitrocellulose mem- brane spotted in duplicate with 36 cytokine and chemokine Data Analysis capture antibodies was washed and blocked in blocking buffer Microglial cell and activated microglial cell counts of for 1 hour at room temperature. For each case, 300 lg of fron- all 5 fields of view were taken for each region and each case, tal cortex sample was used combined with blocking buffer and GraphPad Prism software was used to perform nonparametric human cytokine array A detection antibody cocktail. The Kruskal-Wallis ANOVA with Dunn’s multiple comparison membrane was incubated in the sample mixture overnight at post hoc test to compare control and disease cases and deter- 4 C. The membrane was then washed and incubated with Od- mine whether any changes between groups reached significant yssey fluorescent conjugated secondary antibody (IRDye difference when p value was adjusted for multiple comparison 800CW Streptavidin 926-32230) for 30 minutes at room tem- to an alpha of 0.05. Intrarater reliability of all microglial perature and visualized by LiCOR Odyssey. counts was assessed using an intraclass correlation coefficient (ICC). A minimum of 10% of cases was recounted by the rater and the ICC was calculated as 0.98, which indicates excellent Western Immunoblotting repeatability. ICC was calculated using a free online calculator Frozen tissue was dissected from control (n¼ 7) and http://www.danielsoper.com/statcalc/calculator.aspx?id¼42. mixed MSA (n¼ 7) posterior frontal lobe and homogenized in Human cytokine array data were analyzed using Image J a high salt buffer pH 7.4 (50 mM Tris–HCL, 175 mM NaCl, plug in protein array analyzer which was used according to the 1% Triton-X with protease, and phosphatase inhibitors developer’s instructions (http://image.bio.methods.free.fr/ [Roche, Burgess Hill, UK]) using the Precellys 24 with ImageJ/?Protein-A.). Densitometry of Western immunoblot- CKmix ceramic beads (Bertin Technologies, Paris, France). ting results was performed using Image J gel analyzer. Nor- The resulting homogenate was then spun at 1000g for malized densitometry results for each of the 36 cytokines and 10 minutes at 4 C, the supernatant of which was aliquoted, chemokines were analyzed using multiple t tests using the 2- and protein concentration was determined by BCA assay stage linear set-up procedure of Benjamini, Krieger, and (Pierce, Thermo Fischer Scientific, Rockford, IL). Samples Yekutieli, with Q=1%. Each row was analysed individually, were run at 120 V through a 7%–12% Bis–Tris gel using the without assuming a consistent standard deviation in order to XCell SureLock Mini-Cell Electrophoresis system (Thermo- determine any significant difference between MSA and Fisher Scientific) and transferred onto nitrocellulose mem- control. brane (GE). Membranes were blocked for 1 hour at room All analysis of NanoString data was performed using temperature using Odyssey blocking buffer (LI-COR, Lincoln, NanoString nSolver software. Data obtained from NanoString NE) and probed overnight at 4 C with either CX3CL1, nCounter expression counts were inspected and any samples CX3CR1 (Proteintech, Manchester, UK), Iba-1 (Wako, Ja- determined to be outliers relative to negative controls were ex- pan), or bIII-tubulin antibodies. Membranes were then washed cluded from further analysis. According to the manufacturer’s 3 times in 1 phosphate buffered saline solution containing instructions, the count data were normalized to negative con- 0.1% Tween and probed with the appropriate fluorescent con- trols and to positive controls that corrected for differences in jugated secondary antibody (LI-COR, US0041) for 1 hour at hybridization efficiency and processing variables, including room temperature. Membranes were washed 3 times in 1 purification and RNA/reporter complex immobilization. In ad- phosphate buffered saline solution without tween and results dition, normalization of mRNA content was performed using were visualized using Odyssey imaging system (LI-COR). housekeeping genes (GAPDH, GUSB, HPRT1, PGK1, 600 Downloaded from https://academic.oup.com/jnen/article/77/7/598/5025079 by DeepDyve user on 19 July 2022 J Neuropathol Exp Neurol � Volume 77, Number 7, July 2018 Inflammatory Profile of MS TUBB). Two-tailed Student t tests were then used to test for a CX3CR1 and CX3CL1 Expression in MSA significant difference between gene expression in MSA com- Compared with Control Brain pared with control posterior frontal lobe and cerebellar white Using Western immunoblotting, we compared the ex- matter. In order to correct for false discovery rate, the Benja- pression of CX3CL1 and CX3CR1 in mixed MSA (n¼ 6) and mini-Yekutieli method was used to exclude any false positive control (n¼ 6) frontal lobe (Supplementary Data Table S2 and results. Fig. 3). To control for neuronal and microglial tissue content, we normalized CX3CL1 (expressed by neurons) to the neuro- nal marker bIII-tubulin and CX3CR1 (expressed by microglia) RESULTS to the microglial marker Iba-1. We did not observe a signifi- cant difference in the expression levels of Iba-1 in MSA com- Microglia in MSA pared with control when normalized to b-actin (Fig. 3A, B). For each brain region the number of Iba-1-positive At the predicted molecular weight of 90–100 kDa, membrane- microglia in each case in controls, MSA-SND, and MSA- bound, glycosylated CX3CL1 expression was not significantly OPCA was counted. Details of cases used for this part of the altered in MSA, while an additional band representing the sol- study can be found in Supplementary Data Table S1. Repre- uble nonglycosylated form of the protein detected at 60 kDa sentative field of view images of Iba-1 immunoreactivity with was found to be significantly decreased in MSA (Fig. 3C, either CD68 or arginase-1 immunoreactivity can be seen in p¼ 0.03). Expression of CX3CR1 showed a highly significant Supplementary Data Figure S1, arrows indicate cells which increase in expression in MSA compared with control brain were counted as double immunoreactive. We demonstrated a analyzed by Student t test (Fig. 3C,p¼ 0.0002). significant increase in total microglial number in the substan- tia nigra of OPCA cases (p¼ 0.009), the number of microglia Inflammatory Gene Expression Determined in SND substantia nigra approached a significant increase in Using NanoString MSA but the p value lacked significance when adjusted for To characterize further the inflammatory environment multiple comparisons (p¼ 0.08). There was no significant dif- in mixed MSA brain compared with controls, we used Nano- ference in the number of microglia in MSA frontal lobe (SND String nCounter technology to determine the expression of p¼ 0.26, OPCA p¼ 0.43) or cerebellar white matter com- 249 human inflammation associated mRNAs. Supplementary pared with control (SND p¼ 0.99, OPCA p¼ 0.87) when the Data Figure S2 shows a heat map diagram using hierarchical p values were adjusted for multiple comparisons (Fig. 1A-C). clustering of differential gene expression in control and MSA In order to determine whether pro- or anti-inflammatory frontal lobe (A) and cerebellar white matter (B). Of the total microglia were differentially distributed in brain regions in 249 genes detected by the panel, 9 genes showed significant MSA subtypes and controls the number of cells coexpressing alteration in mRNA expression in MSA frontal lobe compared Iba-1 and CD68 (proinflammatory) and those coexpressing with control (Supplementary Data Table S3). Of these, 5 genes Iba-1 and Arginase-1 (anti-inflammatory) was assessed. Acti- were found to have a fold change of expression >2 when com- vated microglia with a proinflammatory and anti- pared with control. In the cerebellar white matter 1 gene was inflammatory profile were identified in controls and both found to have a fold change of expression >2(Table). We MSA subtypes in all brain regions examined. No significant chose to use a 2-fold change in either upregulation or downre- differences were found in the number of pro- or anti- gulation of a gene as the cut-off point to identify genes of in- inflammatory microglia in any of the brain regions or MSA terest. By reporting both the significance in terms of p value subtype when compared with controls (Fig. 1D–I). Differen- and the fold change in gene expression, we intended to repre- ces between groups were analyzed using Kruskal-Wallis sent both the statistical significance and variability in terms of ANOVA with Dunn’s post hoc test. p value and biological significance of the change which is more accurately represented by fold change (26).We interpret that the altered expression of ARG1, MASP1, NOX4, Chemokine and Cytokine Expression in MSA PTGDR2,and C6 is likely to be relevant compared with Next, we aimed to understand how cytokine and chemo- control. kine expression, as a measure of inflammatory activity, might be altered in MSA. Using a human cytokine array, we ana- DISCUSSION lyzed the expression of 36 cytokine and chemokine targets of We have shown that microglia, the key signaling cells in interest in MSA (n¼ 6) compared with control (n¼ 6). To the neuroinflammatory response, are of significantly greater minimize any influence of MSA subtype, we chose mixed number in the substantia nigra of OPCA cases. Microglial acti- MSA cases and investigated frontal lobe tissue (Supplemen- vation to a pro- or anti-inflammatory phenotype was not in- tary Data Table S2), using multiple t tests using the 2-stage creased above controls. Consistent with these results, we did linear step-up procedure of Benjamini, Krieger, and Yekutieli, not observe significant alteration in the expression of micro- with Q¼ 1%. Each row was analyzed individually, without as- glial marker Iba-1 protein or in the expression of cytokines suming a consistent standard deviation. The array showed no and chemokines in MSA compared with control. Investigating significant alteration in the expression of cytokines and che- the CX3CL1/CX3CR1 neuronal pathway of microglial regula- mokines analyzed between control and MSA cases (Fig. 2). tion, we found that CX3CR1 was increased in MSA indicating 601 Downloaded from https://academic.oup.com/jnen/article/77/7/598/5025079 by DeepDyve user on 19 July 2022 Kiely et al J Neuropathol Exp Neurol � Volume 77, Number 7, July 2018 FIGURE 1. Quantification of microglia in MSA. Analysis of microglial (Iba-1þ) cells, activated (CD68þ/Arginase-1þ), proinflammatory (CD68þ) and anti-inflammatory (Arginase-1þ) microglia in control and MSA frontal lobe (A, D, G), substantia nigra (B, E, H) and cerebellar white matter (C, F, I). No significant difference was found in SND and OPCA frontal lobe or cerebellum compared with control or between disease subtypes (A, C). A significantly greater number of microglia were detected in the substantia nigra of OPCA cases compared with control (B). The number of proinflammatory microglia (CD68þ) showed no difference in frontal lobe (D) substantia nigra (E) or cerebellar white matter (F) compared with control. No significant difference was found in SND or OPCA in the number of anti-inflammatory (Arginase-1þ) microglia in frontal lobe (G), substantia nigra (H), or cerebellar white matter (I) compared with control using Kruskal-Wallis nonparametric tests with Dunn’s post hoc test for multiple comparisons. a neuron-mediated anti-inflammatory response. By analyzing and MSA-SND cases separately to determine whether regional expression of genes associated with inflammation, we con- microglial burden and activation is altered in these 2 disease firmed that ARG1, MASP1, NOX4, PTGDR2, and C6 were sig- subtypes that display the most marked regional differences in nificantly altered in MSA. These findings indicate subtle pathological changes. We chose substantia nigra (severely af- modulation of the regulation and expression of inflammatory fected in MSA-SND), cerebellar white matter (severely af- activity in MSA but these functional changes may not always fected in MSA-OPCA), and frontal lobe, which is less be reflected by analysis of microglial populations. severely affected in both subtypes. The results of the current It has previously been documented that microglia are in- study show a significantly greater number of microglia in the creased in MSA (8, 27). Our previous study using a case substantia nigra of OPCA cases but no significant increase in cohort, which included MSA-mixed, MSA-SND, and MSA- other brain regions or in SND cases after stringent statistical OPCA cases, showed a significantly larger number of micro- correction for multiple comparisons. As our previous study glia in MSA cerebellar and frontal white matter compared pooled together the MSA subtypes, this indicates that caution with control (8). In this study, we investigated MSA-OPCA should be exercised when analyzing regional pathological 602 Downloaded from https://academic.oup.com/jnen/article/77/7/598/5025079 by DeepDyve user on 19 July 2022 J Neuropathol Exp Neurol � Volume 77, Number 7, July 2018 Inflammatory Profile of MS FIGURE 2. Cytokine expression. Statistical analysis of cytokine array data from a panel of 36 cytokines by multiple t tests using the 2-stage linear set-up procedure of Benjamini, Krieger, and Yekutieli, with Q=1%, showed no significant alteration in cytokine expression in MSA compared with control. FIGURE 3. Quantification of Iba-1, CX3CL1, and CX3CR1 protein expression in MSA. (A) Representative Western immunoblots in control [C] and MSA [M] frontal cortex. (B) No significant alteration was detected in the expression of Iba-1 in MSA compared with control. CX3CL1 was detected at the predicted molecular weight of 100 kDa (A), which showed a trend toward upregulation without significance (C). The lower molecular weight band representing the soluble form of the protein was detected at 60 kDa (A) and was significantly decreased in MSA when normalized to bIII-tubulin (C, p¼ 0.03). CX3CR1 was detected at the predicted molecular weight of 47 kDa (A) and showed a highly significant increase in expression in MSA brain compared with control when normalized to Iba-1, determined by Student t test of MSA compared with control (C,p¼ 0.0002). 603 Downloaded from https://academic.oup.com/jnen/article/77/7/598/5025079 by DeepDyve user on 19 July 2022 Kiely et al J Neuropathol Exp Neurol � Volume 77, Number 7, July 2018 TABLE. Frontal Lobe mRNA Expression Determined to be Significantly Altered with >2-Fold Change in MSA Compared with Control Gene Name Upregulated Downregulated Significance Fold Protein Encoded and Function in MSA in MSA (p value) Change Frontal lobe ARG1 – Y 0.0215 3.6 Arginase-1: Anti-inflammatory (57) MASP1 Y – 0.0437 2.64 Mannan-binding lectin serine protease 1; Proinflammatory (58) NOX4 Y – 0.0472 2.28 NADPH oxidase 4: Proinflammatory (54) PTGDR2 – Y 0.0186 2.1 Prostaglandin D2 Receptor 2: Proinflammatory (59) Cerebellar white matter C6 Y – 0.0436 2.88 Complement component 6: Proinflammatory features in MSA. We did not observe any significant alteration expression of cytokines and chemokines in MSA. These in the expression of Iba-1 protein, as a marker of microglia, in results are in accordance with our multi-immunofluorescence MSA. Although we only investigated Iba-1 protein expression data which has shown no significant increase in activated in the frontal lobe, this data further corroborate our evidence microglia in MSA. When examining and interpreting cytokine that there is no significant increase in microglial number in the and chemokine expression we were conscious of several key majority of brain regions analyzed in MSA. Methodological factors that affect the stability of cytokines measured in post- differences may also contribute to variability between studies. mortem tissue (34). We selected cases that had the shortest In the current study, we employed a multi- possible postmortem delay in order to minimize the degree of immunofluorescence method as we also wished to determine cytokine degradation by proteases during this time; however, the activation status of the microglia, unamplified immunoflu- length of time in 80 C storage may also contribute to their orescence is a less sensitive method than DAB based immuno- degradation. In addition, we had sought to exclude cases that histochemistry (28) and so these results are not directly had concomitant pathologies that might exacerbate any in- comparable with the previous data. The substantia nigra flammatory response. Both cases and controls may have been showed the largest population of microglia in OPCA cases, exposed to over-the-counter anti-inflammatory medication this is consistent with the work of others who have shown which may not have been recorded in medical notes and could microglial upregulation to be variable and region specific have effected cytokine and chemokine expression. The cyto- (27). A key aim of this study was to determine whether micro- kine/chemokine response is a fluctuating and rapid process glia activated to have pro- or anti-inflammatory functions are which is quickly modulated by regulatory signaling. These altered in MSA. We found no differences in the numbers of factors may have influenced the study findings. microglia in each of these activation states when compared CX3CL1 or Fractalkine is a novel type of chemokine with controls. This supports our finding that there is little alter- with a unique CX3C motif (35). CX3CL1 differs from other ation in total microglial numbers as microglia have generally chemokines as the molecule exists as a membrane-bound gly- been shown to proliferate when they become activated (29). coprotein with the chemokine attached to the membrane via a Our findings suggest that microglial proliferation and activa- mucin-like stalk. CX3CL1 was found to bind with high affin- tion might not be a key component of the inflammatory re- ity to a microglial orphan chemokine receptor that was subse- quently renamed CX3CR1 (36). CX3CL1 has 2 mechanisms sponse in MSA. This is consistent with recent work that used the MSA mouse model to inhibit microglial myeloperoxidase, by which it can bind to CX3CR1, either while attached to the although this resulted in less microglial activation it did not cell membrane (100 kDa) or after detaching from the mucin- ameliorate motor symptoms or prevent neuronal loss (30). like stalk to provide a soluble form (60 kDa). Binding of Nevertheless, intervention to modulate microglial activation CX3CL1 to CX3CR1 triggers PI3K-dependent Ca2þ influx states and promote an anti-inflammatory environment may and activates MAPK and Akt pathways, the binding of mem- have therapeutic potential in MSA. brane bound CX3CL1 has been shown to produce more potent To minimize any influence of MSA subtype, we used effects than those of soluble CX3CL1 (37). Several reports mixed MSA cases for the remainder of the study in which we have shown that CX3CL1 has an inhibitory effect on micro- investigated cytokine and chemokine protein expression, the glial activation (38–42). Loss of this dampening down of CX3CL1/CX3CR1 microglial regulatory pathway and inflam- microglial activation via the CX3CL1/CX3CR1 interaction matory gene expression. Tissue availability for such studies may account for the increase in neurotoxicity in mouse models precluded the use of small brain regions and, as the frontal of both Alzheimer disease, Parkinson disease, prion disease, lobe is moderately affected in MSA, it was chosen for this part and amyotrophic lateral sclerosis, which lack CX3CR1 (23, of the study. We were also able to include the cerebellum for 39, 43–45). Our findings have shown a significant decrease in gene expression analysis. Previous studies have successfully the soluble 60 kDa form of CX3CL1 in MSA, the form of the used postmortem human brain tissue to demonstrate alteration molecule with less potent anti-inflammatory properties. This of cytokine expression in various diseases (31–33). Our cyto- altered expression may reflect modulation of microglial re- kine array data did not show any significant change in the sponse via alteration of the CX3CL1/CX3CR1 axis in MSA 604 Downloaded from https://academic.oup.com/jnen/article/77/7/598/5025079 by DeepDyve user on 19 July 2022 J Neuropathol Exp Neurol � Volume 77, Number 7, July 2018 Inflammatory Profile of MS leading to limited microglial activation and proliferation in the adenine dinucleotide phosphate oxidase 4 (NOX4), which has surrounding region (23, 46). ADAM10 is required for the been shown to be upregulated in neurons and endothelial cells cleavage of CX3CL1 and it is released as a soluble molecule after stroke, and its inhibition has been shown to be neuropro- (47). This lower level of soluble CX3CR1 expression com- tective (54, 55). pared with control may indicate a failure of ADAM10 to per- In conclusion, our study has shown greater numbers of mit the release of the molecule in MSA brain (48). The microglia in the substantia nigra of OPCA cases with no sig- slighter high level of the full-length, membrane-bound nificant difference in number or degree of activation in other 100 kDa CX3CL1 molecule in MSA may indicate this reduced brain regions or MSA disease subtypes. Consistent with this membrane release, however, this increase did not reach signif- data, we see no significant change in cytokines, one of the icance. Further indication that modulation of microglial acti- functional outputs of activated microglia. However, analysis vation by CX3CL1/CX3CR1 in MSA comes from our of mRNA shows a significant difference in the expression of a demonstration of increased expression of CX3CR1 in MSA. subset of inflammation associated genes, suggesting that some In the context of data from a model of glaucoma in which an level of inflammatory response is initiated in MSA. Here, we anti-inflammatory agent, crocin, successfully reduced micro- indicate a possible neuronal-mediated regulatory control on glial activation via CX3CR1 upregulation, this might indicate the activation and proliferation of microglia. We have shown an anti-inflammatory effect in MSA (49). These results sup- reduction in the level of CX3CL1 soluble protein and an in- port our findings that there is no significant alteration in Iba-1 crease in the microglial receptor of the glycosylated form protein in the frontal lobe and inflammatory cytokines were CX3CR1. These data implicate altered neuronal feedback on unchanged. microglial activation in MSA, while the increase in CX3CR1 To gain further insight into the inflammatory response expression may represent a persistent, potentially protective in MSA, we analyzed the mRNA expression of a panel of (56) effort to maintain monocyte survival. genes related to inflammation. The genes showing a fold Neuroinflammation has many times been described as a change >2 were ARG1, MASP1, NOX4,and PTGDR2 in the double-edged sword: too much leads to catastrophic damage, frontal lobe and C6 in the cerebellar white matter. Of these, while too little permits the accumulation of misfolded protein ARG1, with an anti-inflammatory function and PTGDR2, debris and fails to repair injured tissue. Here, we have with a proinflammatory function, were downregulated. highlighted alteration in transcription of genes involved in in- MASP1, NOX4, and C6 were upregulated and all have a proin- flammation and we have provided evidence that the neuronal flammatory influence. control of microglial activation may be altered in MSA. These Arginase-1 (ARG1) mRNA was significantly downregu- findings support the hypothesis that inflammatory mecha- lated in MSA, this correlated with our finding that Arginase-1- nisms may be targets for therapeutic intervention in MSA. positive microglia form a minority of the activated microglial population. Also, downregulated in MSA was Prostaglandin ACKNOWLEDGMENTS D2 receptor 2 (PTGDR2/CRTH2), a G-protein coupled recep- Nathan Elliot and his team at NanoString Technolo- tor that, when activated, causes chemotaxis of eosinophils, gies, Inc., Seattle, Washington. Tissue stored in the QSBB is basophils, and T helper type cells in the periphery. Little is un- under license 12198 from the Human Tissue Authority and derstood about the role of PTGDR2 in brain; however, has been donated for research according to protocols ap- agonist-mediated activation of the receptor was found to exac- proved by the NRES Committee London-Central. The data erbate glutamate neurotoxicity (50). Of the genes that we de- sets used and analyzed during the current study are available termined to be upregulated in MSA, the greatest difference from the corresponding author on reasonable request. compared with control was in the expression of Complement component 6 (C6) in the cerebellar white matter; this was the REFERENCES only gene that was found to be significantly altered in the cere- 1. Ahmed Z, Asi YT, Sailer A, et al. The neuropathology, pathophysiology bellar white matter. C6 joins a complex with other comple- and genetics of multiple system atrophy. Neuropathol Appl Neurobiol ment components; this complex forms a pore on the cell 2012;38:4–24 2þ þ surface and allows the influx of Ca ,Na , small molecules, 2. Aoki N, Boyer P, Lund C, et al. Atypical multiple system atrophy is a new subtype of frontotemporal lobar degeneration: Frontotemporal lobar and water, which results in cell lysis (51). It is currently degeneration associated with a-synuclein. Acta Neuropathol 2015;130: unclear as to why the cerebellar white matter, which is no less 93–105 affected by the disease than the other areas examined, should 3. Wirth M, Madison CM, Rabinovici GD, et al. Alzheimer’s disease neuro- show less alteration in the expression of inflammation-linked degenerative biomarkers are associated with decreased cognitive func- tion but not b-amyloid in cognitively normal older individuals. J genes as well as altered expression of a gene which is unaf- Neurosci 2013;33:5553–63 fected in frontal lobe. The cerebellum in MSA may have a dif- 4. Sperling RA, Aisen PS, Beckett LA, et al. Toward defining the preclini- ferent mechanism of response to disease and may, as has been cal stages of Alzheimer’s disease: Recommendations from the National suggested, employ a specialized melatonin-mediated response Institute on Aging-Alzheimer’s Association workgroups on diagnostic to inflammation that reduces the inflammatory response (52). guidelines for Alzheimer’s disease. Alzheimers Dement 2011;7:280–92 5. Hirsch EC, Vyas S, Hunot S. Neuroinflammation in Parkinson’s disease. Of the genes found to be upregulated in the frontal lobe, Parkinsonism Relat Disord 2012;18(Suppl 1):S210–2 Mannan-binding lectin serine protease 1 (MASP1) showed the 6. Surendranathan A, Rowe JB, O’Brien JT. Neuroinflammation in Lewy highest fold change. MASP1 acts via protease-activated recep- body dementia. Parkinsonism Relat Disord 2015;21:1398–406 tor signaling to cause recruitment of leukocytes (53). Also 7. Finch CE. Neurons, glia, and plasticity in normal brain aging. Neurobiol upregulated in the MSA frontal lobe was Nicotinamide Aging 2003;24(Suppl 1):S123–7 605 Downloaded from https://academic.oup.com/jnen/article/77/7/598/5025079 by DeepDyve user on 19 July 2022 Kiely et al J Neuropathol Exp Neurol � Volume 77, Number 7, July 2018 8. Ahmed Z, Asi YT, Lees AJ, et al. Identification and quantification of oli- Parkinson’s disease brain is region dependent. Brain Pathol 2014;24: godendrocyte precursor cells in multiple system atrophy, progressive 584–98 supranuclear palsy and Parkinson’s disease. Brain Pathol 2013;23:263–73 34. Bienvenu J, Monneret G, Fabien N, et al. The clinical usefulness of the 9. Salvesen L, Winge K, Brudek T, et al. Neocortical neuronal loss in measurement of cytokines. Clin Chem Lab Med 2000;38:267–85 patients with multiple system atrophy: A stereological study. Cereb Cor- 35. Bazan JF, Bacon KB, Hardiman G, et al. A new class of membrane- tex 2017;27:400–10 bound chemokine with a CX(3)C motif. Nature 1997;385:640–4 10. Wakabayashi K, Yoshimoto M, Tsuji S, et al. Alpha-synuclein immuno- 36. Imai T, Hieshima K, Haskell C, et al. Identification and molecular char- reactivity in glial cytoplasmic inclusions in multiple system atrophy. acterization of fractalkine receptor CX(3)CR1, which mediates both leu- Neurosci Lett 1998;249:180–2 kocyte migration and adhesion. Cell 1997;91:521–30 11. Nimmerjahn A, Kirchhoff F, Helmchen F. Resting microglial cells are 37. Berangere Re D, Przedborski S. Fractalkine: Moving from chemotaxis to highly dynamic surveillants of brain parenchyma in vivo. Science 2005; neuroprotection. Nat Neurosci 2006;9:859–61 308:1314–8 38. Cardona AE, Sasse ME, Liu L, et al. Scavenging roles of chemokine 12. Gilroy D, De Maeyer R. New insights into the resolution of inflamma- receptors: Chemokine receptor deficiency is associated with increased tion. Semin Immunol 2015;27:161–8 levels of ligand in circulation and tissues. Blood 2008;112:256–63 13. Mosser DM, Edwards JP. Exploring the full spectrum of macrophage ac- 39. Cho S-H, Sun B, Zhou Y, et al. CX3CR1 protein signaling modulates tivation. Nat Rev Immunol 2008;8:958–69 microglial activation and protects against plaque-independent cognitive 14. Wang G, Zhang J, Hu X, et al. Microglia/macrophage polarization dy- deficits in a mouse model of Alzheimer disease. J Biol Chem 2011;286: namics in white matter after traumatic brain injury. J Cereb Blood Flow 32713–22 Metab 2013;33:1864–74 40. Liu Z, Condello C, Schain A, et al. CX3CR1 in microglia regulates brain 15. Giasson BI, Duda JE, Murray IV, et al. Oxidative damage linked to neu- amyloid deposition through selective protofibrillar amyloid-b phagocy- rodegeneration by selective a-synuclein nitration in synucleinopathy tosis. J Neurosci 2010;30:17091–101 lesions. Science 2000;290:985–9 41. Raoul W, Feumi C, Keller N, et al. Lipid-bloated subretinal microglial 16. Duda JE, Giasson BI, Chen Q, et al. Widespread nitration of pathological cells are at the origin of drusen appearance in CX3CR1-deficient mice. inclusions in neurodegenerative synucleinopathies. Am J Pathol 2000; Ophthalmic Res 2008;40:115–9 157:1439–45 42. Rogers JT, Morganti JM, Bachstetter AD, et al. CX3CR1 deficiency 17. Reynolds A, Glanzer J, Kadiu I, et al. Nitrated alpha-synuclein-activated leads to impairment of hippocampal cognitive function and synaptic plas- microglial profiling for Parkinson’s disease. J Neurochem 2008;104: ticity. J Neurosci 2011;31:16241–50 1504–25 43. Grizenkova J, Akhtar S, Brandner S, et al. Microglial Cx3cr1knockout 18. Tran PB, Miller RJ. Chemokine receptors: Signposts to brain develop- reduces prion disease incubation time in mice. BMC Neurosci 2014;15: ment and disease. Nat Rev Neurosci 2003;4:444–55 19. Biber K, Neumann H, Inoue K, et al. Neuronal ‘On’ and ‘Off’ signals 44. Dauer W, Kholodilov N, Vila M, et al. Resistance of alpha-synuclein null control microglia. Trends Neurosci 2007;30:596–602 mice to the parkinsonian neurotoxin MPTP. Proc Natl Acad Sci USA 20. Limatola C, Lauro C, Catalano M, et al. Chemokine CX3CL1 protects rat hippocampal neurons against glutamate-mediated excitotoxicity. J 2002;99:14524–9 Neuroimmunol 2005;166:19–28 45. Gurney M, Pu H, Chiu A, et al. Motor neuron degeneration in mice that 21. Harrison JK, Jiang Y, Chen S, et al. Role for neuronally derived fractal- express a human Cu, Zn superoxide dismutase mutation. Science 1994; kine in mediating interactions between neurons and CX3CR1-expressing 264:1772–5 microglia. Proc Natl Acad Sci USA 1998;95:10896–901 46. Thome AD, Standaert DG, Harms AS. Fractalkine signaling regulates 22. Chapman GA, Moores K, Harrison D, et al. Fractalkine cleavage from the inflammatory response in an a-synuclein model of Parkinson disease. neuronal membranes represents an acute event in the inflammatory re- PLoS One 2015;10:e0140566 sponse to excitotoxic brain damage. J Neurosci 2000;20:RC87 47. Hundhausen C, Misztela D, Berkhout TA, et al. The disintegrin-like met- 23. Cardona AE, Pioro EP, Sasse ME, et al. Control of microglial neurotoxic- alloproteinase ADAM10 is involved in constitutive cleavage of CX3CL1 ity by the fractalkine receptor. Nat Neurosci 2006;9:917–24 (fractalkine) and regulates CX3CL1-mediated cell-cell adhesion. Blood 24. Reis P, Waldron L, Goswami R, et al. mRNA transcript quantification in 2003;102:1186–95 archival samples using multiplexed, color-coded probes. BMC Biotech- 48. O’Sullivan SA, Gasparini F, Mir AK, et al. Fractalkine shedding is medi- nol 2011;11:46 ated by p38 and the ADAM10 protease under pro-inflammatory condi- 25. Ozawa T, Paviour D, Quinn NP, et al. The spectrum of pathological in- tions in human astrocytes. J Neuroinflammation 2016;13:189 volvement of the striatonigral and olivopontocerebellar systems in multiple 49. Lv B, Huo F, Zhu Z, et al. Crocin upregulates CX3CR1 expression by system atrophy: Clinicopathological correlations. Brain 2004;127:2657–71 suppressing NF-jB/YY1 signaling and inhibiting lipopolysaccharide- 26. McCarthy DJ, Smyth GK. Testing significance relative to a fold-change induced microglial activation. Neurochem Res 2016;41:1949–57 threshold is a TREAT. Bioinformatics 2009;25:765–71 50. Grill M, Heinemann A, Hoefler G, et al. Effect of endotoxin treatment on 27. Ishizawa K, Komori T, Sasaki S, et al. Microglial activation parallels sys- the expression and localization of spinal cyclooxygenase, tem degeneration in multiple system atrophy. J Neuropathol Exp Neurol prostaglandin synthases, and PGD2 receptors. J Neurochem 2008;104: 2004;63:43–52 1345–57 28. Krell J, Rabin BS. Comparison of an immunohistochemical and immuno- 51. Aleshin AE, Schraufstatter IU, Stec B, et al. Structure of complement C6 fluorescence procedure to detect antibody to pancreatic islet cells. Diabe- suggests a mechanism for initiation and unidirectional, sequential assem- tes 1984;33:709–11 bly of membrane attack complex (MAC). J Biol Chem 2012;287: 29. Ajami B, Bennett JL, Krieger C, et al. Local self-renewal can sustain 10210–22 CNS microglia maintenance and function throughout adult life. Nat Neu- 52. Pinato L, da Silveira Cruz-Machado S, et al. Selective protection of the rosci 2007;10:1538–43 cerebellum against intracerebroventricular LPS is mediated by local mel- 30. Kaindlstorfer C, Sommer P, Georgievska B, et al. Failure of neuroprotec- atonin synthesis. Brain Struct Func 2015;220:827–40 tion despite microglial suppression by delayed-start myeloperoxidase in- 53. Yousef GM, Kishi T, Diamandis EP. Role of kallikrein enzymes in the hibition in a model of advanced multiple system atrophy: Clinical central nervous system. Clin Chim Acta 2003;329:1–8 implications. Neurotox Res 2015;28:185–94 54. Kleinschnitz C, Grund H, Wingler K, et al. Post-stroke inhibition of in- 31. Warford J, Lamport AC, Kennedy B, et al. Human brain chemokine and duced nadph oxidase type 4 prevents oxidative stress and neurodegenera- cytokine expression in sepsis: A report of three cases. Can J Neurol Sci tion. PLoS Biol 2010;8:e1000479 2017;44:96–104 55. Vallet P, Charnay Y, Steger K, et al. Neuronal expression of the NADPH 32. Alcendor DJ, Charest AM, Zhu WQ, et al. Infection and upregulation of oxidase NOX4, and its regulation in mouse experimental brain ischemia. proinflammatory cytokines in human brain vascular pericytes by human Neuroscience 2005;132:233–8 cytomegalovirus. J Neuroinflammation 2012;9:95 56. Asi YT, Simpson JE, Heath PR, et al. Alpha-synuclein mRNA expression 33. Garcia-Esparcia P, Llorens F, Carmona M, et al. Complex deregulation and expression of cytokines and mediators of the immune response in in oligodendrocytes in MSA. Glia 2014;62:964–70 606 Downloaded from https://academic.oup.com/jnen/article/77/7/598/5025079 by DeepDyve user on 19 July 2022 J Neuropathol Exp Neurol � Volume 77, Number 7, July 2018 Inflammatory Profile of MS 57. Ohashi K, Parker JL, Ouchi N, et al. Adiponectin promotes macrophage 59. Hirai H, Tanaka K, Yoshie O, et al. Prostaglandin D2 selectively induces polarization toward an anti-inflammatory phenotype. J Biol Chem 2010; chemotaxis in t helper type 2 cells, eosinophils, and basophils via seven- 285:6153–60 transmembrane receptor Crth2. J Exp Med 2001;193:255–62 58. Megyeri M, Mako V, Beinrohr L, et al. Complement protease 60. Kim WS, Weickert CS, Garner B. Role of ATP-binding cassette trans- MASP-1 activates human endothelial cells: pAR4 activation is a porters in brain lipid transport and neurological disease. J Neurochem link between complement and endothelial function. J Immunol 2009; 2008;104:1145–66 183:3409–16 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Neuropathology & Experimental Neurology Oxford University Press

Immunohistochemical and Molecular Investigations Show Alteration in the Inflammatory Profile of Multiple System Atrophy Brain

Download PDF
10 pages

Loading next page...
 
/lp/oxford-university-press/immunohistochemical-and-molecular-investigations-show-alteration-in-TS98I0SlbP

References (62)

Publisher
Oxford University Press
Copyright
Copyright © 2022 American Association of Neuropathologists
ISSN
0022-3069
eISSN
1554-6578
DOI
10.1093/jnen/nly035
Publisher site
See Article on Publisher Site

Abstract

Key Words: CX3CL1, CX3CR1, Cytokine, Microglia, Multiple Multiple system atrophy (MSA) is an adult-onset neurodegenerative system atrophy, NanoString, Neuroinflammation, Neuropathology. disease characterized by aggregation of a-synuclein in oligodendro- cytes to form glial cytoplasmic inclusions. According to the distribution of neurodegeneration, MSA is subtyped as striatonigral degeneration INTRODUCTION (SND), olivopontocerebellar atrophy (OPCA), or as combination of Multiple system atrophy (MSA) is a sporadic adult- these 2 (mixed MSA). In the current study, we aimed to investigate re- onset progressive neurodegenerative disease characterized gional microglial populations and gene expression in the 3 different clinically by parkinsonism, cerebellar ataxia, and autonomic MSA subtypes. Microscopy with microglial marker Iba-1 combined failure. The disease is confirmed pathologically by accumu- with either proinflammatory marker CD68 or anti-inflammatory marker lated a-synuclein protein within oligodendrocytes forming Arginase-1 was analyzed in control, SND, and OPCA cases (n¼ 5) glial cytoplasmic inclusions as well as variable numbers of using paraffin embedded sections. Western immunoblotting and neuronal nuclear or cytoplasmic inclusions and regional neu- cytokine array were used to determine protein expression in MSA rodegeneration. Neuropathological classification of MSA and control brain regions. Gene expression was investigated using depends on the brain regions most severely affected by neuro- the NanoString nCounter Human Inflammation panel v2 mRNA nal loss: striatonigral degeneration (MSA SND), olivoponto- Expression Assay. Analysis of neuropathological subtypes of MSA cerebellar atrophy (MSA OPCA), or as a mixed form (mixed demonstrated a significant increase in microglia in the substantia MSA). Rare neuropathological subtypes are also described, nigra of OPCA cases. There was no difference in the microglial ac- known as minimal change MSA and atypical MSA (1, 2). tivation state in any region. Cytokine expression in MSA was com- Neuroinflammation is a dynamic response, resulting in parable with controls. Decreased expression of CX3CL1 precursor changes in glial cell function and morphology, upregulated protein and significantly greater CX3CR1 protein was found in transcription of inflammation linked genes and cytokine pro- MSA. NanoString analysis revealed the >2-fold greater expression duction. Neuroinflammation has been identified as a hallmark of ARG1, MASP1, NOX4, PTGDR2,and C6 in MSA. of neurodegenerative diseases including Alzheimer disease (3, 4), Parkinson disease (5), and dementia with Lewy bodies (6) among others. Proinflammatory neuroinflammation has also been shown to be a normal process of nondiseased aging From the Queen Square Brain Bank for Neurological Disorders, Department (reviewed in [7]). We have previously shown that microglia, of Molecular Neuroscience, UCL Institute of Neurology, London, UK (APK, CEM, SCF, BCB, RC, CS, TL, JLH). the key cellular component of the neuroinflammatory re- Send correspondence to: Janice L. Holton, PhD, MRCPath, Queen Square sponse, are increased in number in a group of MSA cases with Brain Bank for Neurological Disorders, UCL Institute of Neurology, mixed neuropathological subtypes (8). Others have reported London WC1N 1PJ, UK; E-mail: [email protected] This study was made possible due to funding awarded by the Multiple System similar upregulation of astrocytes along with microglial acti- Atrophy Trust (UK) which covered A.P.K.’s salary and consumable costs vation in MSA brain (9, 10). Microglia had generally been for the study excluding NanoString. Additional funding was awarded by proposed to have a bipolar activation style and in the absence the British Neuropathological Society to facilitate the NanoString compo- of insult, toxin, or disease, they exist in a ramified state. In nent of the investigation. T.L. is supported by an Alzheimer’s Research their ramified state, microglia have constantly motile pro- UK Senior Fellowship. C.E.M. is supported by an Alzheimer’s Research UK PhD Studentship. B.C.B. is supported by the Leonard Wolfson Centre cesses surveying their local area (11). Microglial processes for Experimental Neurology. The Queen Square Brain Bank is supported may then detect different inducers that can stimulate the cells by the Reta Lila Weston Institute for Neurological Studies, the Progressive to become activated either to a proinflammatory, phagocytic Supranuclear Palsy Association and the Medical Research Council UK. (M1 state), an anti-inflammatory, prohealing (M2 state), The authors have no duality or conflicts of interest to declare. Supplementary Data can be found at http://www.jnen.oxfordjournals.org. or a “resolution phase macrophage” (rM), which has hybrid V C 598 2018 American Association of Neuropathologists, Inc. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/ by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. Downloaded from https://academic.oup.com/jnen/article/77/7/598/5025079 by DeepDyve user on 19 July 2022 J Neuropathol Exp Neurol � Volume 77, Number 7, July 2018 Inflammatory Profile of MS expression of both M1 and M2 phenotypic markers (12). In this way, we aimed to gain a broad understanding of the In recent years, general consensus has begun to move away neuroinflammatory characteristics of MSA to gain insights from the bipolar theory of microglial activation in favor of a into disease mechanisms and provide information contributing spectrum of activation (13) in which active microglia may to the development of disease biomarkers and therapeutic fluctuate rapidly through multiple degrees of activation targets. expressing a continuum of variable markers. Interestingly, a recent study in a mouse model of trau- MATERIALS AND METHODS matic brain injury has shown gradual upregulation of M1 marker RNA expression postinjury, while M2 marker expres- Case Selection sion peaked rapidly postinjury before declining in expression Cases were selected for each part of the study (immuno- (14). Although this study examined the influence of an acute fluorescence, cytokine arrays, Western immunoblotting, and insult on the dynamics of the M1 to M2 microglial ratio, it NanoString analysis) based on diagnosis, tissue availability, might be anticipated that in neurodegeneration an acute re- gender and age pairing and postmortem delay, we sought to sponse will give way to either a long-term chronic or dysfunc- exclude cases that had concomitant pathologies that might ex- tional response. Indeed, both nitrated and aggregated forms of acerbate any inflammatory response. Due to these restrictions, a-synuclein protein, which are present in MSA (15, 16), have we were unable to use exactly the same cohort for the entire been shown to stimulate a pro-inflammatory microglial pheno- investigation (Supplementary Data Tables S1 and S2). type (17) and may maintain this activation during the disease course of MSA. We hypothesize that microglia in MSA have Human Brain Tissue become aberrantly activated, potentially exacerbating the de- We used brain tissue from cases donated to the Queen velopment of pathological changes. Square Brain Bank for Neurological Disorders, UCL Institute We aimed first to investigate whether differences in of Neurology. The brain donation program and protocols have microglial number, which we previously reported in MSA received ethical approval for research by the NRES Commit- brain in a group of cases unselected for neuropathological tee London—Central and tissue is stored for research under a subtype, would be reproduced when analyzing MSA cases by license issued by the Human Tissue Authority (No. 12198). disease subtype. For this, we selected MSA-SND and MSA- According to routine protocols, one-half of the fresh OPCA as the 2 disease subtypes showing the widest differen- brain was sliced in the coronal plane, flash frozen, and stored tial in pathological involvement of different brain regions. We chose substantia nigra because it is severely involved in at 80 C. The remaining half brain was fixed in 10% buffered MSA-SND and cerebellar white matter as a region with exten- formalin before being sliced in the coronal plane. Tissue sive pathology in MSA-OPCA. We also included the posterior blocks were selected for paraffin wax embedding and histol- frontal lobe because this region shows a moderate degree of ogy. This material was used for multi-immunofluorescence a-synuclein pathology in both subtypes of MSA. Next, we studies. Frozen tissue was dissected from the posterior frontal wished to investigate the activation state of these microglia, cortex and the cerebellar white matter and either homogenized whether they be in a pro- or anti-inflammatory state. for Western immunoblotting or lysed and processed for total To investigate additional aspects of the inflammatory re- RNA using the RNeasy mini prep kit according to the manu- sponse in MSA, we selected cases of mixed MSA to minimize facturer’s instructions (Qiagen, Manchester, UK). any influence of brain regional variability in pathology. The Pathological confirmation of MSA subtype diagnosis functional output of the inflammatory response can be quanti- based on the Ozawa criteria (25) was performed using paraffin fied and assessed by assaying the expression of cytokines and embedded sections (8 lm) stained for relevant markers chemokines. To understand whether the functional inflamma- (including hematoxylin and eosin and Luxol fast blue/cresyl tory output differs between control and MSA, we compared violet and a-synuclein). Multi-immunofluorescence was per- the expression of 36 cytokines and chemokines using commer- formed using isotype specific antirabbit IgG, antimouse IgG, cially available assay systems. CX3CL1 or fractalkine is a or antigoat IgG secondary antibodies conjugated with either chemokine which, together with its specific receptor Alexa 488 or 594 fluorescent dyes (1:400) (Life Technologies, CX3CR1, is abundantly expressed in the CNS, mediating Paisley, UK) followed by quenching of autofluorescence with neuron-microglial interaction (18). The glycosylated form of 0.1% Sudan black/70% ethanol (Sigma-Aldrich, Dorset, UK) the neuronal CX3CL1 can bind microglial CX3CR1 and solution for 10 minutes and mounting with glass coverslips us- prevent the activation of potentially harmful inflammatory ing VECTAshield mounting media with 4’,6-diamidino-2- microglial phenotypes (19). The expression of CX3CL1 phenylindole (DAPI) nuclear stain (Vector Laboratories, and CX3CR1 can be altered in the presence of injury or insult Peterborough, UK). Images were visualized using fluores- (20–22) and impaired CX3CR1 function may exacerbate neu- cence microscopy (Leica DM5500 B). rodegeneration (23). To investigate whether similar changes contribute to the inflammatory process in MSA, we deter- Microglial Cell Counting mined the expression of CX3CL1 and CX3CR1 in brain We selected age matched control, SND, and OPCA tissue. Finally, to provide a wider understanding of the inflam- cases (n¼ 5 each) for analysis. For each case, we chose se- matory response in MSA, we examined the expression of 249 genes associated with inflammation in disease cases compared verely affected brain regions: subcortical white matter of the with controls using NanoString nCounter technology (24). frontal lobe, substantia nigra, and cerebellar white matter. 599 Downloaded from https://academic.oup.com/jnen/article/77/7/598/5025079 by DeepDyve user on 19 July 2022 Kiely et al J Neuropathol Exp Neurol � Volume 77, Number 7, July 2018 Sections (8 lm) were probed with Iba-1, CD68, and arginase- RNA Isolation 1. In each region, 5 areas were randomly selected according to Total RNA was isolated from posterior frontal lobe and a predesigned zig-zag pattern and we counted the number of cerebellar white matter homogenates using RNAeasy mRNA cells that stained positively for Iba-1 as a marker of total kit (Qiagen) according to the manufacturer’s instructions. microglia and were of a suitable size and morphology, present RNA quantity and quality was assessed using a BioSpectrom- in the field of view at 20 magnification. Area of field of eter (Eppendorf, Stevenage, UK). view at 20 using an eyepiece magnification of 25 was cal- culated to be 1.19 mm and all cell counts were expressed as NanoString nCounter Analysis number of cells counted/mm . In addition, we assessed the To examine the inflammatory environment of MSA number of cells per field at 20 that showed colocalization by mRNA expression was determined using the NanoString multi-immunofluorescence microscopy of either Iba-1 and nCounter Human Inflammation panel v2 mRNA Expression CD68 as a marker of pro-inflammatory microglia or Iba-1 and Assay (NanoString Technologies). Total mRNA was isolated Arg-1 as a marker of anti-inflammatory microglia. from control (n¼ 6) and mixed MSA (n¼ 6) posterior frontal lobe and cerebellar white matter (Supplementary Data Table Human Cytokine Array S2) The protocol was carried out by evaluation project techni- cians at the Seattle NanoString facility as a proof of concept To analyze cytokine expression in control (n¼ 6) and study using 150 ng of total RNA from each sample following mixed MSA (n¼ 6) we selected posterior frontal cortex as a their commercial protocol. Data were collected using the region showing pathological involvement and used the human nCounter Digital Analyser. cytokine array A (R&D Systems, Abingdon, UK) according to the manufacturer’s instructions. Briefly, nitrocellulose mem- brane spotted in duplicate with 36 cytokine and chemokine Data Analysis capture antibodies was washed and blocked in blocking buffer Microglial cell and activated microglial cell counts of for 1 hour at room temperature. For each case, 300 lg of fron- all 5 fields of view were taken for each region and each case, tal cortex sample was used combined with blocking buffer and GraphPad Prism software was used to perform nonparametric human cytokine array A detection antibody cocktail. The Kruskal-Wallis ANOVA with Dunn’s multiple comparison membrane was incubated in the sample mixture overnight at post hoc test to compare control and disease cases and deter- 4 C. The membrane was then washed and incubated with Od- mine whether any changes between groups reached significant yssey fluorescent conjugated secondary antibody (IRDye difference when p value was adjusted for multiple comparison 800CW Streptavidin 926-32230) for 30 minutes at room tem- to an alpha of 0.05. Intrarater reliability of all microglial perature and visualized by LiCOR Odyssey. counts was assessed using an intraclass correlation coefficient (ICC). A minimum of 10% of cases was recounted by the rater and the ICC was calculated as 0.98, which indicates excellent Western Immunoblotting repeatability. ICC was calculated using a free online calculator Frozen tissue was dissected from control (n¼ 7) and http://www.danielsoper.com/statcalc/calculator.aspx?id¼42. mixed MSA (n¼ 7) posterior frontal lobe and homogenized in Human cytokine array data were analyzed using Image J a high salt buffer pH 7.4 (50 mM Tris–HCL, 175 mM NaCl, plug in protein array analyzer which was used according to the 1% Triton-X with protease, and phosphatase inhibitors developer’s instructions (http://image.bio.methods.free.fr/ [Roche, Burgess Hill, UK]) using the Precellys 24 with ImageJ/?Protein-A.). Densitometry of Western immunoblot- CKmix ceramic beads (Bertin Technologies, Paris, France). ting results was performed using Image J gel analyzer. Nor- The resulting homogenate was then spun at 1000g for malized densitometry results for each of the 36 cytokines and 10 minutes at 4 C, the supernatant of which was aliquoted, chemokines were analyzed using multiple t tests using the 2- and protein concentration was determined by BCA assay stage linear set-up procedure of Benjamini, Krieger, and (Pierce, Thermo Fischer Scientific, Rockford, IL). Samples Yekutieli, with Q=1%. Each row was analysed individually, were run at 120 V through a 7%–12% Bis–Tris gel using the without assuming a consistent standard deviation in order to XCell SureLock Mini-Cell Electrophoresis system (Thermo- determine any significant difference between MSA and Fisher Scientific) and transferred onto nitrocellulose mem- control. brane (GE). Membranes were blocked for 1 hour at room All analysis of NanoString data was performed using temperature using Odyssey blocking buffer (LI-COR, Lincoln, NanoString nSolver software. Data obtained from NanoString NE) and probed overnight at 4 C with either CX3CL1, nCounter expression counts were inspected and any samples CX3CR1 (Proteintech, Manchester, UK), Iba-1 (Wako, Ja- determined to be outliers relative to negative controls were ex- pan), or bIII-tubulin antibodies. Membranes were then washed cluded from further analysis. According to the manufacturer’s 3 times in 1 phosphate buffered saline solution containing instructions, the count data were normalized to negative con- 0.1% Tween and probed with the appropriate fluorescent con- trols and to positive controls that corrected for differences in jugated secondary antibody (LI-COR, US0041) for 1 hour at hybridization efficiency and processing variables, including room temperature. Membranes were washed 3 times in 1 purification and RNA/reporter complex immobilization. In ad- phosphate buffered saline solution without tween and results dition, normalization of mRNA content was performed using were visualized using Odyssey imaging system (LI-COR). housekeeping genes (GAPDH, GUSB, HPRT1, PGK1, 600 Downloaded from https://academic.oup.com/jnen/article/77/7/598/5025079 by DeepDyve user on 19 July 2022 J Neuropathol Exp Neurol � Volume 77, Number 7, July 2018 Inflammatory Profile of MS TUBB). Two-tailed Student t tests were then used to test for a CX3CR1 and CX3CL1 Expression in MSA significant difference between gene expression in MSA com- Compared with Control Brain pared with control posterior frontal lobe and cerebellar white Using Western immunoblotting, we compared the ex- matter. In order to correct for false discovery rate, the Benja- pression of CX3CL1 and CX3CR1 in mixed MSA (n¼ 6) and mini-Yekutieli method was used to exclude any false positive control (n¼ 6) frontal lobe (Supplementary Data Table S2 and results. Fig. 3). To control for neuronal and microglial tissue content, we normalized CX3CL1 (expressed by neurons) to the neuro- nal marker bIII-tubulin and CX3CR1 (expressed by microglia) RESULTS to the microglial marker Iba-1. We did not observe a signifi- cant difference in the expression levels of Iba-1 in MSA com- Microglia in MSA pared with control when normalized to b-actin (Fig. 3A, B). For each brain region the number of Iba-1-positive At the predicted molecular weight of 90–100 kDa, membrane- microglia in each case in controls, MSA-SND, and MSA- bound, glycosylated CX3CL1 expression was not significantly OPCA was counted. Details of cases used for this part of the altered in MSA, while an additional band representing the sol- study can be found in Supplementary Data Table S1. Repre- uble nonglycosylated form of the protein detected at 60 kDa sentative field of view images of Iba-1 immunoreactivity with was found to be significantly decreased in MSA (Fig. 3C, either CD68 or arginase-1 immunoreactivity can be seen in p¼ 0.03). Expression of CX3CR1 showed a highly significant Supplementary Data Figure S1, arrows indicate cells which increase in expression in MSA compared with control brain were counted as double immunoreactive. We demonstrated a analyzed by Student t test (Fig. 3C,p¼ 0.0002). significant increase in total microglial number in the substan- tia nigra of OPCA cases (p¼ 0.009), the number of microglia Inflammatory Gene Expression Determined in SND substantia nigra approached a significant increase in Using NanoString MSA but the p value lacked significance when adjusted for To characterize further the inflammatory environment multiple comparisons (p¼ 0.08). There was no significant dif- in mixed MSA brain compared with controls, we used Nano- ference in the number of microglia in MSA frontal lobe (SND String nCounter technology to determine the expression of p¼ 0.26, OPCA p¼ 0.43) or cerebellar white matter com- 249 human inflammation associated mRNAs. Supplementary pared with control (SND p¼ 0.99, OPCA p¼ 0.87) when the Data Figure S2 shows a heat map diagram using hierarchical p values were adjusted for multiple comparisons (Fig. 1A-C). clustering of differential gene expression in control and MSA In order to determine whether pro- or anti-inflammatory frontal lobe (A) and cerebellar white matter (B). Of the total microglia were differentially distributed in brain regions in 249 genes detected by the panel, 9 genes showed significant MSA subtypes and controls the number of cells coexpressing alteration in mRNA expression in MSA frontal lobe compared Iba-1 and CD68 (proinflammatory) and those coexpressing with control (Supplementary Data Table S3). Of these, 5 genes Iba-1 and Arginase-1 (anti-inflammatory) was assessed. Acti- were found to have a fold change of expression >2 when com- vated microglia with a proinflammatory and anti- pared with control. In the cerebellar white matter 1 gene was inflammatory profile were identified in controls and both found to have a fold change of expression >2(Table). We MSA subtypes in all brain regions examined. No significant chose to use a 2-fold change in either upregulation or downre- differences were found in the number of pro- or anti- gulation of a gene as the cut-off point to identify genes of in- inflammatory microglia in any of the brain regions or MSA terest. By reporting both the significance in terms of p value subtype when compared with controls (Fig. 1D–I). Differen- and the fold change in gene expression, we intended to repre- ces between groups were analyzed using Kruskal-Wallis sent both the statistical significance and variability in terms of ANOVA with Dunn’s post hoc test. p value and biological significance of the change which is more accurately represented by fold change (26).We interpret that the altered expression of ARG1, MASP1, NOX4, Chemokine and Cytokine Expression in MSA PTGDR2,and C6 is likely to be relevant compared with Next, we aimed to understand how cytokine and chemo- control. kine expression, as a measure of inflammatory activity, might be altered in MSA. Using a human cytokine array, we ana- DISCUSSION lyzed the expression of 36 cytokine and chemokine targets of We have shown that microglia, the key signaling cells in interest in MSA (n¼ 6) compared with control (n¼ 6). To the neuroinflammatory response, are of significantly greater minimize any influence of MSA subtype, we chose mixed number in the substantia nigra of OPCA cases. Microglial acti- MSA cases and investigated frontal lobe tissue (Supplemen- vation to a pro- or anti-inflammatory phenotype was not in- tary Data Table S2), using multiple t tests using the 2-stage creased above controls. Consistent with these results, we did linear step-up procedure of Benjamini, Krieger, and Yekutieli, not observe significant alteration in the expression of micro- with Q¼ 1%. Each row was analyzed individually, without as- glial marker Iba-1 protein or in the expression of cytokines suming a consistent standard deviation. The array showed no and chemokines in MSA compared with control. Investigating significant alteration in the expression of cytokines and che- the CX3CL1/CX3CR1 neuronal pathway of microglial regula- mokines analyzed between control and MSA cases (Fig. 2). tion, we found that CX3CR1 was increased in MSA indicating 601 Downloaded from https://academic.oup.com/jnen/article/77/7/598/5025079 by DeepDyve user on 19 July 2022 Kiely et al J Neuropathol Exp Neurol � Volume 77, Number 7, July 2018 FIGURE 1. Quantification of microglia in MSA. Analysis of microglial (Iba-1þ) cells, activated (CD68þ/Arginase-1þ), proinflammatory (CD68þ) and anti-inflammatory (Arginase-1þ) microglia in control and MSA frontal lobe (A, D, G), substantia nigra (B, E, H) and cerebellar white matter (C, F, I). No significant difference was found in SND and OPCA frontal lobe or cerebellum compared with control or between disease subtypes (A, C). A significantly greater number of microglia were detected in the substantia nigra of OPCA cases compared with control (B). The number of proinflammatory microglia (CD68þ) showed no difference in frontal lobe (D) substantia nigra (E) or cerebellar white matter (F) compared with control. No significant difference was found in SND or OPCA in the number of anti-inflammatory (Arginase-1þ) microglia in frontal lobe (G), substantia nigra (H), or cerebellar white matter (I) compared with control using Kruskal-Wallis nonparametric tests with Dunn’s post hoc test for multiple comparisons. a neuron-mediated anti-inflammatory response. By analyzing and MSA-SND cases separately to determine whether regional expression of genes associated with inflammation, we con- microglial burden and activation is altered in these 2 disease firmed that ARG1, MASP1, NOX4, PTGDR2, and C6 were sig- subtypes that display the most marked regional differences in nificantly altered in MSA. These findings indicate subtle pathological changes. We chose substantia nigra (severely af- modulation of the regulation and expression of inflammatory fected in MSA-SND), cerebellar white matter (severely af- activity in MSA but these functional changes may not always fected in MSA-OPCA), and frontal lobe, which is less be reflected by analysis of microglial populations. severely affected in both subtypes. The results of the current It has previously been documented that microglia are in- study show a significantly greater number of microglia in the creased in MSA (8, 27). Our previous study using a case substantia nigra of OPCA cases but no significant increase in cohort, which included MSA-mixed, MSA-SND, and MSA- other brain regions or in SND cases after stringent statistical OPCA cases, showed a significantly larger number of micro- correction for multiple comparisons. As our previous study glia in MSA cerebellar and frontal white matter compared pooled together the MSA subtypes, this indicates that caution with control (8). In this study, we investigated MSA-OPCA should be exercised when analyzing regional pathological 602 Downloaded from https://academic.oup.com/jnen/article/77/7/598/5025079 by DeepDyve user on 19 July 2022 J Neuropathol Exp Neurol � Volume 77, Number 7, July 2018 Inflammatory Profile of MS FIGURE 2. Cytokine expression. Statistical analysis of cytokine array data from a panel of 36 cytokines by multiple t tests using the 2-stage linear set-up procedure of Benjamini, Krieger, and Yekutieli, with Q=1%, showed no significant alteration in cytokine expression in MSA compared with control. FIGURE 3. Quantification of Iba-1, CX3CL1, and CX3CR1 protein expression in MSA. (A) Representative Western immunoblots in control [C] and MSA [M] frontal cortex. (B) No significant alteration was detected in the expression of Iba-1 in MSA compared with control. CX3CL1 was detected at the predicted molecular weight of 100 kDa (A), which showed a trend toward upregulation without significance (C). The lower molecular weight band representing the soluble form of the protein was detected at 60 kDa (A) and was significantly decreased in MSA when normalized to bIII-tubulin (C, p¼ 0.03). CX3CR1 was detected at the predicted molecular weight of 47 kDa (A) and showed a highly significant increase in expression in MSA brain compared with control when normalized to Iba-1, determined by Student t test of MSA compared with control (C,p¼ 0.0002). 603 Downloaded from https://academic.oup.com/jnen/article/77/7/598/5025079 by DeepDyve user on 19 July 2022 Kiely et al J Neuropathol Exp Neurol � Volume 77, Number 7, July 2018 TABLE. Frontal Lobe mRNA Expression Determined to be Significantly Altered with >2-Fold Change in MSA Compared with Control Gene Name Upregulated Downregulated Significance Fold Protein Encoded and Function in MSA in MSA (p value) Change Frontal lobe ARG1 – Y 0.0215 3.6 Arginase-1: Anti-inflammatory (57) MASP1 Y – 0.0437 2.64 Mannan-binding lectin serine protease 1; Proinflammatory (58) NOX4 Y – 0.0472 2.28 NADPH oxidase 4: Proinflammatory (54) PTGDR2 – Y 0.0186 2.1 Prostaglandin D2 Receptor 2: Proinflammatory (59) Cerebellar white matter C6 Y – 0.0436 2.88 Complement component 6: Proinflammatory features in MSA. We did not observe any significant alteration expression of cytokines and chemokines in MSA. These in the expression of Iba-1 protein, as a marker of microglia, in results are in accordance with our multi-immunofluorescence MSA. Although we only investigated Iba-1 protein expression data which has shown no significant increase in activated in the frontal lobe, this data further corroborate our evidence microglia in MSA. When examining and interpreting cytokine that there is no significant increase in microglial number in the and chemokine expression we were conscious of several key majority of brain regions analyzed in MSA. Methodological factors that affect the stability of cytokines measured in post- differences may also contribute to variability between studies. mortem tissue (34). We selected cases that had the shortest In the current study, we employed a multi- possible postmortem delay in order to minimize the degree of immunofluorescence method as we also wished to determine cytokine degradation by proteases during this time; however, the activation status of the microglia, unamplified immunoflu- length of time in 80 C storage may also contribute to their orescence is a less sensitive method than DAB based immuno- degradation. In addition, we had sought to exclude cases that histochemistry (28) and so these results are not directly had concomitant pathologies that might exacerbate any in- comparable with the previous data. The substantia nigra flammatory response. Both cases and controls may have been showed the largest population of microglia in OPCA cases, exposed to over-the-counter anti-inflammatory medication this is consistent with the work of others who have shown which may not have been recorded in medical notes and could microglial upregulation to be variable and region specific have effected cytokine and chemokine expression. The cyto- (27). A key aim of this study was to determine whether micro- kine/chemokine response is a fluctuating and rapid process glia activated to have pro- or anti-inflammatory functions are which is quickly modulated by regulatory signaling. These altered in MSA. We found no differences in the numbers of factors may have influenced the study findings. microglia in each of these activation states when compared CX3CL1 or Fractalkine is a novel type of chemokine with controls. This supports our finding that there is little alter- with a unique CX3C motif (35). CX3CL1 differs from other ation in total microglial numbers as microglia have generally chemokines as the molecule exists as a membrane-bound gly- been shown to proliferate when they become activated (29). coprotein with the chemokine attached to the membrane via a Our findings suggest that microglial proliferation and activa- mucin-like stalk. CX3CL1 was found to bind with high affin- tion might not be a key component of the inflammatory re- ity to a microglial orphan chemokine receptor that was subse- quently renamed CX3CR1 (36). CX3CL1 has 2 mechanisms sponse in MSA. This is consistent with recent work that used the MSA mouse model to inhibit microglial myeloperoxidase, by which it can bind to CX3CR1, either while attached to the although this resulted in less microglial activation it did not cell membrane (100 kDa) or after detaching from the mucin- ameliorate motor symptoms or prevent neuronal loss (30). like stalk to provide a soluble form (60 kDa). Binding of Nevertheless, intervention to modulate microglial activation CX3CL1 to CX3CR1 triggers PI3K-dependent Ca2þ influx states and promote an anti-inflammatory environment may and activates MAPK and Akt pathways, the binding of mem- have therapeutic potential in MSA. brane bound CX3CL1 has been shown to produce more potent To minimize any influence of MSA subtype, we used effects than those of soluble CX3CL1 (37). Several reports mixed MSA cases for the remainder of the study in which we have shown that CX3CL1 has an inhibitory effect on micro- investigated cytokine and chemokine protein expression, the glial activation (38–42). Loss of this dampening down of CX3CL1/CX3CR1 microglial regulatory pathway and inflam- microglial activation via the CX3CL1/CX3CR1 interaction matory gene expression. Tissue availability for such studies may account for the increase in neurotoxicity in mouse models precluded the use of small brain regions and, as the frontal of both Alzheimer disease, Parkinson disease, prion disease, lobe is moderately affected in MSA, it was chosen for this part and amyotrophic lateral sclerosis, which lack CX3CR1 (23, of the study. We were also able to include the cerebellum for 39, 43–45). Our findings have shown a significant decrease in gene expression analysis. Previous studies have successfully the soluble 60 kDa form of CX3CL1 in MSA, the form of the used postmortem human brain tissue to demonstrate alteration molecule with less potent anti-inflammatory properties. This of cytokine expression in various diseases (31–33). Our cyto- altered expression may reflect modulation of microglial re- kine array data did not show any significant change in the sponse via alteration of the CX3CL1/CX3CR1 axis in MSA 604 Downloaded from https://academic.oup.com/jnen/article/77/7/598/5025079 by DeepDyve user on 19 July 2022 J Neuropathol Exp Neurol � Volume 77, Number 7, July 2018 Inflammatory Profile of MS leading to limited microglial activation and proliferation in the adenine dinucleotide phosphate oxidase 4 (NOX4), which has surrounding region (23, 46). ADAM10 is required for the been shown to be upregulated in neurons and endothelial cells cleavage of CX3CL1 and it is released as a soluble molecule after stroke, and its inhibition has been shown to be neuropro- (47). This lower level of soluble CX3CR1 expression com- tective (54, 55). pared with control may indicate a failure of ADAM10 to per- In conclusion, our study has shown greater numbers of mit the release of the molecule in MSA brain (48). The microglia in the substantia nigra of OPCA cases with no sig- slighter high level of the full-length, membrane-bound nificant difference in number or degree of activation in other 100 kDa CX3CL1 molecule in MSA may indicate this reduced brain regions or MSA disease subtypes. Consistent with this membrane release, however, this increase did not reach signif- data, we see no significant change in cytokines, one of the icance. Further indication that modulation of microglial acti- functional outputs of activated microglia. However, analysis vation by CX3CL1/CX3CR1 in MSA comes from our of mRNA shows a significant difference in the expression of a demonstration of increased expression of CX3CR1 in MSA. subset of inflammation associated genes, suggesting that some In the context of data from a model of glaucoma in which an level of inflammatory response is initiated in MSA. Here, we anti-inflammatory agent, crocin, successfully reduced micro- indicate a possible neuronal-mediated regulatory control on glial activation via CX3CR1 upregulation, this might indicate the activation and proliferation of microglia. We have shown an anti-inflammatory effect in MSA (49). These results sup- reduction in the level of CX3CL1 soluble protein and an in- port our findings that there is no significant alteration in Iba-1 crease in the microglial receptor of the glycosylated form protein in the frontal lobe and inflammatory cytokines were CX3CR1. These data implicate altered neuronal feedback on unchanged. microglial activation in MSA, while the increase in CX3CR1 To gain further insight into the inflammatory response expression may represent a persistent, potentially protective in MSA, we analyzed the mRNA expression of a panel of (56) effort to maintain monocyte survival. genes related to inflammation. The genes showing a fold Neuroinflammation has many times been described as a change >2 were ARG1, MASP1, NOX4,and PTGDR2 in the double-edged sword: too much leads to catastrophic damage, frontal lobe and C6 in the cerebellar white matter. Of these, while too little permits the accumulation of misfolded protein ARG1, with an anti-inflammatory function and PTGDR2, debris and fails to repair injured tissue. Here, we have with a proinflammatory function, were downregulated. highlighted alteration in transcription of genes involved in in- MASP1, NOX4, and C6 were upregulated and all have a proin- flammation and we have provided evidence that the neuronal flammatory influence. control of microglial activation may be altered in MSA. These Arginase-1 (ARG1) mRNA was significantly downregu- findings support the hypothesis that inflammatory mecha- lated in MSA, this correlated with our finding that Arginase-1- nisms may be targets for therapeutic intervention in MSA. positive microglia form a minority of the activated microglial population. Also, downregulated in MSA was Prostaglandin ACKNOWLEDGMENTS D2 receptor 2 (PTGDR2/CRTH2), a G-protein coupled recep- Nathan Elliot and his team at NanoString Technolo- tor that, when activated, causes chemotaxis of eosinophils, gies, Inc., Seattle, Washington. Tissue stored in the QSBB is basophils, and T helper type cells in the periphery. Little is un- under license 12198 from the Human Tissue Authority and derstood about the role of PTGDR2 in brain; however, has been donated for research according to protocols ap- agonist-mediated activation of the receptor was found to exac- proved by the NRES Committee London-Central. The data erbate glutamate neurotoxicity (50). Of the genes that we de- sets used and analyzed during the current study are available termined to be upregulated in MSA, the greatest difference from the corresponding author on reasonable request. compared with control was in the expression of Complement component 6 (C6) in the cerebellar white matter; this was the REFERENCES only gene that was found to be significantly altered in the cere- 1. Ahmed Z, Asi YT, Sailer A, et al. The neuropathology, pathophysiology bellar white matter. C6 joins a complex with other comple- and genetics of multiple system atrophy. Neuropathol Appl Neurobiol ment components; this complex forms a pore on the cell 2012;38:4–24 2þ þ surface and allows the influx of Ca ,Na , small molecules, 2. Aoki N, Boyer P, Lund C, et al. Atypical multiple system atrophy is a new subtype of frontotemporal lobar degeneration: Frontotemporal lobar and water, which results in cell lysis (51). It is currently degeneration associated with a-synuclein. Acta Neuropathol 2015;130: unclear as to why the cerebellar white matter, which is no less 93–105 affected by the disease than the other areas examined, should 3. Wirth M, Madison CM, Rabinovici GD, et al. Alzheimer’s disease neuro- show less alteration in the expression of inflammation-linked degenerative biomarkers are associated with decreased cognitive func- tion but not b-amyloid in cognitively normal older individuals. J genes as well as altered expression of a gene which is unaf- Neurosci 2013;33:5553–63 fected in frontal lobe. The cerebellum in MSA may have a dif- 4. Sperling RA, Aisen PS, Beckett LA, et al. Toward defining the preclini- ferent mechanism of response to disease and may, as has been cal stages of Alzheimer’s disease: Recommendations from the National suggested, employ a specialized melatonin-mediated response Institute on Aging-Alzheimer’s Association workgroups on diagnostic to inflammation that reduces the inflammatory response (52). guidelines for Alzheimer’s disease. Alzheimers Dement 2011;7:280–92 5. Hirsch EC, Vyas S, Hunot S. Neuroinflammation in Parkinson’s disease. Of the genes found to be upregulated in the frontal lobe, Parkinsonism Relat Disord 2012;18(Suppl 1):S210–2 Mannan-binding lectin serine protease 1 (MASP1) showed the 6. Surendranathan A, Rowe JB, O’Brien JT. Neuroinflammation in Lewy highest fold change. MASP1 acts via protease-activated recep- body dementia. Parkinsonism Relat Disord 2015;21:1398–406 tor signaling to cause recruitment of leukocytes (53). Also 7. Finch CE. Neurons, glia, and plasticity in normal brain aging. Neurobiol upregulated in the MSA frontal lobe was Nicotinamide Aging 2003;24(Suppl 1):S123–7 605 Downloaded from https://academic.oup.com/jnen/article/77/7/598/5025079 by DeepDyve user on 19 July 2022 Kiely et al J Neuropathol Exp Neurol � Volume 77, Number 7, July 2018 8. Ahmed Z, Asi YT, Lees AJ, et al. Identification and quantification of oli- Parkinson’s disease brain is region dependent. Brain Pathol 2014;24: godendrocyte precursor cells in multiple system atrophy, progressive 584–98 supranuclear palsy and Parkinson’s disease. Brain Pathol 2013;23:263–73 34. Bienvenu J, Monneret G, Fabien N, et al. The clinical usefulness of the 9. Salvesen L, Winge K, Brudek T, et al. Neocortical neuronal loss in measurement of cytokines. Clin Chem Lab Med 2000;38:267–85 patients with multiple system atrophy: A stereological study. Cereb Cor- 35. Bazan JF, Bacon KB, Hardiman G, et al. A new class of membrane- tex 2017;27:400–10 bound chemokine with a CX(3)C motif. Nature 1997;385:640–4 10. Wakabayashi K, Yoshimoto M, Tsuji S, et al. Alpha-synuclein immuno- 36. Imai T, Hieshima K, Haskell C, et al. Identification and molecular char- reactivity in glial cytoplasmic inclusions in multiple system atrophy. acterization of fractalkine receptor CX(3)CR1, which mediates both leu- Neurosci Lett 1998;249:180–2 kocyte migration and adhesion. Cell 1997;91:521–30 11. Nimmerjahn A, Kirchhoff F, Helmchen F. Resting microglial cells are 37. Berangere Re D, Przedborski S. Fractalkine: Moving from chemotaxis to highly dynamic surveillants of brain parenchyma in vivo. Science 2005; neuroprotection. Nat Neurosci 2006;9:859–61 308:1314–8 38. Cardona AE, Sasse ME, Liu L, et al. Scavenging roles of chemokine 12. Gilroy D, De Maeyer R. New insights into the resolution of inflamma- receptors: Chemokine receptor deficiency is associated with increased tion. Semin Immunol 2015;27:161–8 levels of ligand in circulation and tissues. Blood 2008;112:256–63 13. Mosser DM, Edwards JP. Exploring the full spectrum of macrophage ac- 39. Cho S-H, Sun B, Zhou Y, et al. CX3CR1 protein signaling modulates tivation. Nat Rev Immunol 2008;8:958–69 microglial activation and protects against plaque-independent cognitive 14. Wang G, Zhang J, Hu X, et al. Microglia/macrophage polarization dy- deficits in a mouse model of Alzheimer disease. J Biol Chem 2011;286: namics in white matter after traumatic brain injury. J Cereb Blood Flow 32713–22 Metab 2013;33:1864–74 40. Liu Z, Condello C, Schain A, et al. CX3CR1 in microglia regulates brain 15. Giasson BI, Duda JE, Murray IV, et al. Oxidative damage linked to neu- amyloid deposition through selective protofibrillar amyloid-b phagocy- rodegeneration by selective a-synuclein nitration in synucleinopathy tosis. J Neurosci 2010;30:17091–101 lesions. Science 2000;290:985–9 41. Raoul W, Feumi C, Keller N, et al. Lipid-bloated subretinal microglial 16. Duda JE, Giasson BI, Chen Q, et al. Widespread nitration of pathological cells are at the origin of drusen appearance in CX3CR1-deficient mice. inclusions in neurodegenerative synucleinopathies. Am J Pathol 2000; Ophthalmic Res 2008;40:115–9 157:1439–45 42. Rogers JT, Morganti JM, Bachstetter AD, et al. CX3CR1 deficiency 17. Reynolds A, Glanzer J, Kadiu I, et al. Nitrated alpha-synuclein-activated leads to impairment of hippocampal cognitive function and synaptic plas- microglial profiling for Parkinson’s disease. J Neurochem 2008;104: ticity. J Neurosci 2011;31:16241–50 1504–25 43. Grizenkova J, Akhtar S, Brandner S, et al. Microglial Cx3cr1knockout 18. Tran PB, Miller RJ. Chemokine receptors: Signposts to brain develop- reduces prion disease incubation time in mice. BMC Neurosci 2014;15: ment and disease. Nat Rev Neurosci 2003;4:444–55 19. Biber K, Neumann H, Inoue K, et al. Neuronal ‘On’ and ‘Off’ signals 44. Dauer W, Kholodilov N, Vila M, et al. Resistance of alpha-synuclein null control microglia. Trends Neurosci 2007;30:596–602 mice to the parkinsonian neurotoxin MPTP. Proc Natl Acad Sci USA 20. Limatola C, Lauro C, Catalano M, et al. Chemokine CX3CL1 protects rat hippocampal neurons against glutamate-mediated excitotoxicity. J 2002;99:14524–9 Neuroimmunol 2005;166:19–28 45. Gurney M, Pu H, Chiu A, et al. Motor neuron degeneration in mice that 21. Harrison JK, Jiang Y, Chen S, et al. Role for neuronally derived fractal- express a human Cu, Zn superoxide dismutase mutation. Science 1994; kine in mediating interactions between neurons and CX3CR1-expressing 264:1772–5 microglia. Proc Natl Acad Sci USA 1998;95:10896–901 46. Thome AD, Standaert DG, Harms AS. Fractalkine signaling regulates 22. Chapman GA, Moores K, Harrison D, et al. Fractalkine cleavage from the inflammatory response in an a-synuclein model of Parkinson disease. neuronal membranes represents an acute event in the inflammatory re- PLoS One 2015;10:e0140566 sponse to excitotoxic brain damage. J Neurosci 2000;20:RC87 47. Hundhausen C, Misztela D, Berkhout TA, et al. The disintegrin-like met- 23. Cardona AE, Pioro EP, Sasse ME, et al. Control of microglial neurotoxic- alloproteinase ADAM10 is involved in constitutive cleavage of CX3CL1 ity by the fractalkine receptor. Nat Neurosci 2006;9:917–24 (fractalkine) and regulates CX3CL1-mediated cell-cell adhesion. Blood 24. Reis P, Waldron L, Goswami R, et al. mRNA transcript quantification in 2003;102:1186–95 archival samples using multiplexed, color-coded probes. BMC Biotech- 48. O’Sullivan SA, Gasparini F, Mir AK, et al. Fractalkine shedding is medi- nol 2011;11:46 ated by p38 and the ADAM10 protease under pro-inflammatory condi- 25. Ozawa T, Paviour D, Quinn NP, et al. The spectrum of pathological in- tions in human astrocytes. J Neuroinflammation 2016;13:189 volvement of the striatonigral and olivopontocerebellar systems in multiple 49. Lv B, Huo F, Zhu Z, et al. Crocin upregulates CX3CR1 expression by system atrophy: Clinicopathological correlations. Brain 2004;127:2657–71 suppressing NF-jB/YY1 signaling and inhibiting lipopolysaccharide- 26. McCarthy DJ, Smyth GK. Testing significance relative to a fold-change induced microglial activation. Neurochem Res 2016;41:1949–57 threshold is a TREAT. Bioinformatics 2009;25:765–71 50. Grill M, Heinemann A, Hoefler G, et al. Effect of endotoxin treatment on 27. Ishizawa K, Komori T, Sasaki S, et al. Microglial activation parallels sys- the expression and localization of spinal cyclooxygenase, tem degeneration in multiple system atrophy. J Neuropathol Exp Neurol prostaglandin synthases, and PGD2 receptors. J Neurochem 2008;104: 2004;63:43–52 1345–57 28. Krell J, Rabin BS. Comparison of an immunohistochemical and immuno- 51. Aleshin AE, Schraufstatter IU, Stec B, et al. Structure of complement C6 fluorescence procedure to detect antibody to pancreatic islet cells. Diabe- suggests a mechanism for initiation and unidirectional, sequential assem- tes 1984;33:709–11 bly of membrane attack complex (MAC). J Biol Chem 2012;287: 29. Ajami B, Bennett JL, Krieger C, et al. Local self-renewal can sustain 10210–22 CNS microglia maintenance and function throughout adult life. Nat Neu- 52. Pinato L, da Silveira Cruz-Machado S, et al. Selective protection of the rosci 2007;10:1538–43 cerebellum against intracerebroventricular LPS is mediated by local mel- 30. Kaindlstorfer C, Sommer P, Georgievska B, et al. Failure of neuroprotec- atonin synthesis. Brain Struct Func 2015;220:827–40 tion despite microglial suppression by delayed-start myeloperoxidase in- 53. Yousef GM, Kishi T, Diamandis EP. Role of kallikrein enzymes in the hibition in a model of advanced multiple system atrophy: Clinical central nervous system. Clin Chim Acta 2003;329:1–8 implications. Neurotox Res 2015;28:185–94 54. Kleinschnitz C, Grund H, Wingler K, et al. Post-stroke inhibition of in- 31. Warford J, Lamport AC, Kennedy B, et al. Human brain chemokine and duced nadph oxidase type 4 prevents oxidative stress and neurodegenera- cytokine expression in sepsis: A report of three cases. Can J Neurol Sci tion. PLoS Biol 2010;8:e1000479 2017;44:96–104 55. Vallet P, Charnay Y, Steger K, et al. Neuronal expression of the NADPH 32. Alcendor DJ, Charest AM, Zhu WQ, et al. Infection and upregulation of oxidase NOX4, and its regulation in mouse experimental brain ischemia. proinflammatory cytokines in human brain vascular pericytes by human Neuroscience 2005;132:233–8 cytomegalovirus. J Neuroinflammation 2012;9:95 56. Asi YT, Simpson JE, Heath PR, et al. Alpha-synuclein mRNA expression 33. Garcia-Esparcia P, Llorens F, Carmona M, et al. Complex deregulation and expression of cytokines and mediators of the immune response in in oligodendrocytes in MSA. Glia 2014;62:964–70 606 Downloaded from https://academic.oup.com/jnen/article/77/7/598/5025079 by DeepDyve user on 19 July 2022 J Neuropathol Exp Neurol � Volume 77, Number 7, July 2018 Inflammatory Profile of MS 57. Ohashi K, Parker JL, Ouchi N, et al. Adiponectin promotes macrophage 59. Hirai H, Tanaka K, Yoshie O, et al. Prostaglandin D2 selectively induces polarization toward an anti-inflammatory phenotype. J Biol Chem 2010; chemotaxis in t helper type 2 cells, eosinophils, and basophils via seven- 285:6153–60 transmembrane receptor Crth2. J Exp Med 2001;193:255–62 58. Megyeri M, Mako V, Beinrohr L, et al. Complement protease 60. Kim WS, Weickert CS, Garner B. Role of ATP-binding cassette trans- MASP-1 activates human endothelial cells: pAR4 activation is a porters in brain lipid transport and neurological disease. J Neurochem link between complement and endothelial function. J Immunol 2009; 2008;104:1145–66 183:3409–16

Journal

Journal of Neuropathology & Experimental Neurology – Oxford University Press

Published: Jul 1, 2018

Keywords: olivopontocerebellar atrophies; nadph oxidase 4; cd68 antigen, human

There are no references for this article.