ARTICLE OPEN ACCESS

Motor cortex transcriptome reveals microglialkey events in amyotrophic lateral sclerosisOriol Dols-Icardo PhD Vıctor Montal MSc Sonia Sirisi PhD Gema Lopez-Pernas MSc

Laura Cervera-Carles PhD Marta Querol-Vilaseca MSc Laia Muntildeoz BSc Olivia Belbin PhD

Daniel Alcolea MD PhD Laura Molina-Porcel MD PhD Jordi Pegueroles MSc Janina Turon-Sans MD

Rafael Blesa MD PhD Alberto Lleo MD PhD Juan Fortea MD PhD Ricard Rojas-Garcıa MD PhD and

Jordi Clarimon PhD

Neurol Neuroimmunol Neuroinflamm 20207e829 doi101212NXI0000000000000829

Correspondence

Dr Clarimon

jclarimonsantpaucat

or Dr Dols-Icardo

odolssantpaucat

AbstractObjectiveTo identify transcriptomic changes neuropathologic correlates and cellular subpopulations inthe motor cortex of sporadic amyotrophic lateral sclerosis (ALS)

MethodsWe performed massive RNA sequencing of the motor cortex of patients with ALS (n = 11) andhealthy controls (HCs n = 8) and analyzed gene expression alterations differential isoformusage and gene coexpression networks Furthermore we used cell type deconvolution algo-rithms with human single-nucleus RNA sequencing data as reference to identify perturbationsin cell type composition associated with ALS We performed immunohistochemical techniquesto evaluate neuropathologic changes in this brain region

ResultsWe report extensive RNA expression alterations at gene and isoform levels characterized by theenrichment of neuroinflammatory and synaptic-related pathways The assembly of genecoexpression modules confirmed the involvement of these 2 major transcriptomic changeswhich also showed opposite directions related to the disease Cell type deconvolution revealedan overrepresentation of microglial cells in ALS compared with HC Notably microgliosis wasdriven by a subcellular population presenting a gene expression signature overlapping with therecently described disease-associated microglia (DAM) Using immunohistochemistry wefurther evidenced that this microglial subpopulation is overrepresented in ALS and that thedensity of pTDP43 aggregates negatively correlates with the proportion of microglial cells

ConclusionsDAM has a central role in microglia-related neuroinflammatory changes in the motor cortex ofpatients with ALS and these alterations are coupled with a reduced expression of postsynaptictranscripts

From the Memory Unit (OD-I VM SS GL-P LC-C MQ-V LM OB DA JP RB AL JF JC) Neurology Department and Sant Pau Biomedical Research Institute Hospitalde la Santa Creu i Sant Pau Universitat Autonoma de Barcelona Network Center for Biomedical Research in Neurodegenerative Diseases (CIBERNED) (OD-I VM SS GL-P LC-C MQ-V LM OB DA JP RB AL JF JC) Madrid Neurological Tissue Bank of the Biobanc-Hospital Clınic-IDIBAPS (LM-P) Barcelona Alzheimerrsquos Disease and OtherCognitive Disorders Unit Neurology Department (LM-P) Hospital Clınic Institut drsquoInvestigacions Biomediques August Pi i Sunyer University of Barcelona Network Center forBiomedical Research in Rare Diseases (CIBERER) (JT-S RR-G) Madrid andNeuromuscular Disorders Unit (JT-S RR-G) Department of Neurology Hospital de la Santa Creu i SantPau Universitat Autonoma de Barcelona Spain

Go to NeurologyorgNN for full disclosures Funding information is provided at the end of the article

The Article Processing Charge was funded by the authors

This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 40 (CC BY-NC-ND) which permits downloadingand sharing the work provided it is properly cited The work cannot be changed in any way or used commercially without permission from the journal

Copyright copy 2020 The Author(s) Published by Wolters Kluwer Health Inc on behalf of the American Academy of Neurology 1

Amyotrophic lateral sclerosis (ALS) is a neurodegenerativedisease neuropathologically characterized by the aberrantcytoplasmic aggregation and phosphorylation of the 43-kDatransactive response DNA-binding protein (pTDP43) inthe majority of ALS cases at postmortem evaluation whichis known to occur in a sequential manner starting in themotor cortex1 Although the causes that lead to nongeneticforms of ALS (sporadic ALS) have yet to be fully de-termined the presence of neuroinflammation is a consistentfeature in the CNS of affected patients2 Microglia aremacrophage-like innate immune cells of the CNS that asa result of disease conditions change their gene expressionprofile to adapt and acquire a reactive state3 Recentlya study using single-cell RNA sequencing from the CNS ofAlzheimer disease and ALS mouse models reported thata specific microglial subpopulation known as disease-associated microglia (DAM) was the cell type responsiblefor microgliosis and that triggering receptor expressed onmyeloid cells 2 (TREM2) is a major driver of DAM acti-vation4 In support of this discovery some chitinase pro-teins which are known markers of microglial activation56

have been consistently shown to be increased in the CSF ofpatients with ALS78

The pivotal role of TDP43 and other genes known to causeALS (such as FUS hnRNPA2B1 hnRNPA1 TAF15 orTIA1) in the metabolism of RNA has implicated RNAdyshomeostasis as a crucial event in the pathophysiologyof the disease9 An unbiased resource to obtain a compre-hensive signature of gene and isoform expression changesassociated with the whole tissue response to disease con-ditions is total RNA sequencing (RNAseq) in bulk tissueFurthermore the advent of novel sequencing tools such assingle-nucleus RNAseq (snRNAseq) has opened a newwindow to better explore the transcriptome at a cell leveland has unraveled a complex and huge variety of humancell types with unique expression profiles in the humanbrain10ndash12 These transcriptomic signatures can be appliedto perform cell type deconvolution of bulk RNAseq datausing novel deconvolution methods These methods takeinto account cross-subject and cross-cell variability of geneexpression profiles without relying on preselected mark-ers thereby providing more realistic cell type proportionestimations and yielding crucial information about thecellular heterogeneity that is associated with a pathologiccondition13

The brain motor cortex is affected at the most early stages ofthe disease and is one of the most vulnerable regions in ALSHowever studies of this critical region both at the transcrip-tional and immunohistochemical level are lacking In thepresent work we aim to characterize the motor cortex ofsporadic ALS cases through total RNAseq analyses We alsouse cell type deconvolution using human snRNAseq data asreference and immunohistochemical analyses to resolvea signature of neuropathologic changes associated with ALS inthis particular brain region

MethodsHuman samplesThe study included human motor cortex (Brodmann area 4)samples provided by the Neurological Tissue Bank (NTB) ofthe Biobanc-Hospital Clınic-IDIBAPS None of the braintissues presented any infarcts in the motor cortex Diagnosisof ALS complied with the El Escorial criteria during life14 andnone of the patients had a family history of ALS or dementianor did show any sign of cognitive impairment All patientspresented pTDP43 inclusions in the motor cortex at post-mortem examination None of the samples carried theC9orf72 hexanucleotide repeat expansion or mutations in theTBK1 gene the most common genetic causes related to adult-onset ALS in Spain1516

RNA extraction and sequencingUsing a mortar and liquid nitrogen the tissue (60 mg) wasgrinded to powder and transferred to a solution of 600 μL ofTRIzol reagent (Thermo Fisher Scientific Waltham MA)We used standard recommendations and procedures to ex-tract RNAwith TRIzol the RNeasyMini Kit (Qiagen HildenGermany) and the Rnase-Free Dnase Set (Qiagen) Qubitwas used to measure RNA concentration whereas RNA in-tegrity (RIN) was verified on an Agilent 2100 Bioanalyzer(Agilent Technologies Santa Clara CA) Only samples witha quantity threshold of 5 μg and RIN ge65 were used for totalRNAseq The final study group included 11 ALS cases and 8healthy controls (HCs) (table 1) Paired-end sequencing li-braries were prepared using the TruSeq Stranded Total RNALibrary Preparation kit (Illumina San Diego CA) and se-quenced on the Illumina HiSeq 2500 platform by the CentroNacional de Analisis Genomicos (CNAG Barcelona Spain)with 101-bp paired-end reads to achieve at least 100 million

GlossaryALS = amyotrophic lateral sclerosis DAM = disease-associated microglia FFPE = formalin-fixed paraffin-embedded GO =gene ontology HC = healthy control IHC = immunohistochemistry KEGG = Kyoto Encyclopedia of Genes and GenomesMuSiC = Multi-Subject Single Cell deconvolution NTB = Neurological Tissue Bank qPCR = quantitative real-time PCRRIN = RNA integrityRNAseq = RNA sequencingROI = region of interest ROSMAP = Religious Orders Study andMemoryand Aging Project snRNAseq = single-nucleus RNAseq TREM2 = triggering receptor expressed on myeloid cells 2WGCNA = weighted gene coexpression network analyses

2 Neurology Neuroimmunology amp Neuroinflammation | Volume 7 Number 5 | September 2020 NeurologyorgNN

paired-end reads for each sample (ranging from 103447000to 138990000 paired-end reads)

Data processingWe aligned FASTQ files to the Grch38p12 genome assemblyusing STAR v261a17 We then used GATK Best Practices18

for variant calling and FeatureCounts (within the Subreadv162 package19) to assign fragments to each gene featureincluded in the Grch38v94 gene transfer format file For dif-ferential isoform usage we used the STAR v261 package inquantMode TranscriptomeSAM mode and used Salmon toobtain isoform expression values20 For differential gene ex-pression and differential isoform usage analyses we usedDESeq2 v12421 which applies the Benjamini-Hochberg for pvalue adjustment We used a false discovery rate threshold of5 To perform gene ontology (GO) and Kyoto Encyclo-pedia of Genes and Genomes (KEGG) pathway enrichmentanalyses we used Metascape 30 (metascapeorg)22 applyingdefault parameters (minimum overlap = 3 p value cutoff =001 and minimum enrichment = 15) For the synaptic en-richment analyses we used the web server SynGO (syngo-portalorg) which provides an expert-curated resource forsynaptic function and gene enrichment analysis23

Validation of transcriptome changes byquantitative real-time PCRWe performed quantitative real-time PCR (qPCR) on thesame brain-derived RNA samples used for total RNAseq (11sALS cases and 8 controls) A total of 500 ng of RNAwas usedto generate cDNA with the RevertAid First Strand cDNASynthesis Kit (Thermo Fisher Scientific) as per the manu-facturerrsquos instructions All qPCRs were conducted using FastSYBR Green Master Mix (Thermo Fisher Scientific) and runon an ABI Prism 7900HT Fast Real-Time PCR System(Applied Biosystems Foster City CA) All primer pairs usedare listed in table e-1 linkslwwcomNXIA280 For relativequantification we applied the DDCt method using RPL13 asan endogenous control to normalize the data We decided touse RPL13 as a housekeeping gene as it showed the smallestcoefficient of variation (lt35) compared with the 2 otherputative housekeeping genes that we assessed (RPL0 andGAPDH)

Cell type deconvolutionWe used the recently developed Multi-Subject Single Celldeconvolution (MuSiC)13 method to perform cell typedeconvolution of our bulk RNAseq data As a reference we

Table 1 Demographic clinical and quality control data

Sample ID Group Age at death y Sex Age at onset y Duration mo Site of onset RIN PMI h

AF5214 ALS 64 Male 62 33 Limb 82 7

AF5215 ALS 50 Male 46 58 Limb 88 13

AF5216 ALS 60 Female 54 72 Limb 83 6

AF5218 ALS 53 Female 50 32 Limb 84 9

AF5220 ALS 82 Male 81 1 Bulbar 77 16

AF5222 ALS 54 Female 53 18 Limb 9 7

AF5224 ALS 70 Female 68 24 Limb 84 14

AF5227 ALS 78 Male 76 24 Limb 81 12

AF5228 ALS 77 Female 75 31 Bulbar 7 16

AF5229 ALS 57 Female 54 24 Limb 68 5

AF5232 ALS 52 Male 51 12 Limb 74 18

AF5219 HC 64 Male mdash mdash mdash 82 10

AF5221 HC 78 Male mdash mdash mdash 79 6

AF5226 HC 81 Female mdash mdash mdash 75 23

AF5231 HC 83 Female mdash mdash mdash 73 7

AF5234 HC 58 Male mdash mdash mdash 75 4

AF5235 HC 76 Male mdash mdash mdash 75 11

AF5236 HC 83 Female mdash mdash mdash 77 7

AF5237 HC 68 Female mdash mdash mdash 65 13

Abbreviations ALS = amyotrophic lateral sclerosis HC = healthy control PMI = postmortem interval RIN = RNA integrity number

NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 7 Number 5 | September 2020 3

used the most comprehensive human snRNAseq data setavailable This data set includes the RNAseq from 70634 cellsin the frontal cortex (Brodmann area 10) of 24 Alzheimerdisease cases and 24 HCs from the Religious Orders Studyand Memory and Aging Project (ROSMAP)12 grouped into8 major cellular populations (excitatory neurons inhibitoryneurons microglia astrocytes oligodendrocytes oligoden-drocyte precursor cells endothelial cells and pericytes) Tocheck consistency and validate our results we used 2 in-dependent snRNAseq repositories as a reference to estimatemajor cell type proportions To this aim we downloaded themedial temporal gyrus snRNAseq data derived from the AllenBrain Atlas which includes 15928 cells from 8 human tissuedonors We selected this data set as it contains double thenumber of nuclei than the primary visual cortex and the an-terior cingulate cortex also contained in the database10 Thethird data set comprised 10319 cells derived from the frontalcortex of 4 individuals11 Cell type deconvolution of cellularsubpopulations was performed using the ROSMAP data asreference

Weighted gene coexpression network analysesWe accounted for genes with more than 10 counts across allindividuals DESeq2 was then used to normalize the inputdata to construct coexpression networks using the weightedgene coexpression network analyses (WGCNA) package24

We then constructed the signed weighted correlation net-work which takes into account both negative and positivecorrelations using the manual function in WGCNA applyingthe biweight midcorrelation selecting a power of 7 (thelowest possible power term where topology fits a scale freenetwork) and run in a single block analysis We defined allmodules by the hybrid treecutting option with deepsplit pa-rameter = 2 applying a minimal module size of 30 genes andmerging modules with a cutHeight le025 Module eigengenes(first principal component as a summary of each module)summarized the modules and correlations performed usingspearman correlation for continuous traits and Pearson cor-relation for binary traits (bicor and corPvalueStudent in-cluded in the WGCNA package) We used Cytoscape 372 tovisualize networks

ImmunohistochemistryWe used formalin-fixed paraffin-embedded (FFPE) motorcortex tissue sections of 5 μm in 11 patients with ALS and 7controls One control sample (AF5237) was excluded fromthe analysis due to lack of availability of FFPE motor cortexsections We dewaxed the FFPE sections by placing them inxylene and decreasing concentrations of ethanol followedby hydration 30 minutes in distilled water We then treatedtissue sections with a solution containing 10methanol 3H2O2 and PBS for 10 minutes to inhibit endogenous per-oxidase and boiled in citrate 1x solution to mediate antigenretrieval Once cooled we blocked all sections with 5bovine serum albumin for 1 hour at room temperature andincubated them overnight with the corresponding primaryantibody at 4degC and then with the anti-rabbit horse radish

peroxidase secondary antibody (1200) for 1 hour at roomtemperature Primary antibodies were Iba1 rabbit poly-clonal (1500 Wako Chemicals Osaka Japan) and ps409410-TDP rabbit polyclonal (12000 Cosmo Bio TokyoJapan) Following peroxidase development with 339-dia-minobenzidine tetrahydrochloride (Dako liquid DAB +substrate Chromogen System Dako Denmark GlostrupDenmark) sections were stained with hematoxylin (EnVi-sion FLEX Haematoxylin Dako Denmark) and dehydratedusing increasing concentrations of ethanol and thenmounted with DPX mounting medium (PanReac Appli-Chem ITW Reagents Chicago IL) For immunofluores-cence we used 3 ALS and 3 control motor cortex sectionswhich were incubated with primary antibodies overnight at4degC We used Iba1 rabbit polyclonal antibody as a markerfor the broad microglial population (1500 Wako Chem-icals) and human HLA-DP DQ DR antigen (MHC class IImarkers) mouse clone CR343 (120 AgilentM077501ndash2) as a marker for activated cells MHCII markergenes are specifically expressed in the Mic1 gene signa-ture12 Thus we decided to combine these primary anti-bodies to visualize Mic1 cells as performed in the study ofMathys et al12 as well as to estimate the proportion of cellsof this particular subpopulation Then we incubated sec-tions for 1 hour at room temperature with secondary anti-bodies (11000) which were conjugated with Alexa Fluor488 and Alexa Fluor 555 (Invitrogen Carlsbad CA) Cellnuclei visualized with Hoechst 33258 (Life Technologies)Sections were mounted with Shandon Immu-Mount(Thermo Fisher Scientific) We acquired all the imageswith a Leica TCS SP5 Confocal Laser Scanning Microscopewith a 63times objective

Quantification analysisWe obtained full-section immunohistochemistry (IHC)images with Pannoramic MIDI II (3DHistech) Blinded tothe neuropathologic diagnoses we delimited 6 gray matterareas of the motor cortex from each case and for each graymatter area we randomly generated 6 regions of interest(ROIs) with the same size using the multicropm script Weused the Iba1m script to quantify densities for Iba1 stainingin each ROI which represents a total of 36 images per in-dividual This algorithm binarizes the random images andcompute densities of protein expression to quantify thenumber of immunoreactive objects Both in-house algo-rithms can be freely accessed at githubcomMemo-ryUnitSantPau We used MATLAB R2017b software todevelop the algorithm (The MathWorks Inc Natick MA)For pTDP43 quantification 2 different researchers blindedto neuropathologic diagnoses quantified immunoreactivepositive objects manually We evaluated immunofluores-cence using FIJI imaging software We estimated an auto-mated threshold to create binary images Following the noiseremoval with the despeckle and the remove outliersrsquo filterswe determined the number of Iba1+ microglial cellsexpressing MHC class II markers using the analyze particlesfunction

4 Neurology Neuroimmunology amp Neuroinflammation | Volume 7 Number 5 | September 2020 NeurologyorgNN

Statistical analysisWe used the Shapiro-Wilk test to test for deviation froma standard distribution For correlation analyses we de-termined the Pearson or Spearman correlation coefficientsusing the ggscatter function within the ggpubr package in RWe performed mean comparison analyses through the Stu-dent t test or Student test with Welch correction (using ttestfunction in R) depending on data distribution Statisticalsignificance for all tests was set at 5 (α = 005) and allstatistical tests were 2 sided We used R version 362 toperform all statistical analyses

Standard protocol approvals registrationsand patient consentsThe Sant Pau Hospital Ethics Committee approved the studywith all experiments with human tissue performed in accor-dance with the Declaration of Helsinki The NTB of theBiobanc-Hospital Clınic-IDIBAPS supplied all tissue follow-ing approval from their Scientific Advisory Committee

Data availability statementWe have deposited all raw sequencing data (FASTQ files) atthe European Genome-phenome Archive (EGA) which ishosted by the EBI and the CRG under accession numberEGAS00001004286

ResultsDemographics clinical features andsample compositionWe performed RNAseq analysis of postmortem motor cor-tex samples from 11 ALS cases (6 females) and 8 individualswithout neurologic disease (4 females) Postmortem intervaland RIN did not differ between groups (p = 068 and p =0089 respectively) The mean age at death was lower forALS cases compared with HC (634 years range 54ndash82years in ALS and 739 years range 58ndash83 years in HC p =004) Patients had an average disease onset of 609 years(range 46ndash81 years) and a disease duration of 299 months(range 1ndash72 months) Nine of the 11 cases (818) pre-sented a limb disease onset (table 1) We used RNAseq datato rule out the presence of other ALS-disease causingmutations confirming the presumably sporadic nature of ourALS group of cases

Gene expression alterations and differentialisoform usageDifferential gene expression between patients with ALS andHCs disclosed a total of 108 upregulated and 16 down-regulated genes (adjusted p lt 005) (figure 1A and table e-2linkslwwcomNXIA280) We validated 9 of the 10 genesselected by qPCR as we did not confirm the CHI3L2 geneoverexpression in patients with ALS (p = 0085) (figure 1B)Correlation analyses between counts derived from totalRNAseq data and qPCR expression values indicated a highand significant concordance between both expression

measures in these 10 genes (table e-3 linkslwwcomNXIA280) GO and KEGG pathway analyses revealed an enrichedinvolvement of the inflammatory response in ALS (figure1C) The assessment of changes at the isoform expressionlevel resulted in 167 upregulated and 40 downregulated iso-forms of which 181 were from unique genes (table e-4 linkslwwcomNXIA280) Enrichment analyses revealed the in-volvement of postsynaptic density and immune response asdominant pathways (figure e-1 linkslwwcomNXIA279)To further characterize and confirm the synaptic involvementwe used the SynGO database which provides an expert-curated resource for synaptic function and gene enrichmentanalysis23 This analysis confirmed the enrichment of post-synaptic components in our list of differentially expressedisoforms (table e-5 linkslwwcomNXIA280) Among thelist of altered genes identified by the differential isoform ex-pression and gene expression models 32 of them overlappedbetween both approaches (table e-6 linkslwwcomNXIA280) The involvement of synaptic-related changes wasconfirmed from the differential isoform usage analysis (figuree-2 linkslwwcomNXIA279) whereas the enrichment ofinflammatory markers was evidenced in the gene model ap-proach (figure e-3 linkslwwcomNXIA279)

Cell type deconvolutionTo characterize the variability of cell types and compare cel-lular proportions between groups we applied the recentlydeveloped MuSiC algorithm As reference we used data fromthe first and most comprehensive human snRNAseq studyperformed in brain tissue of individuals with a neurodegen-erative disease which includes 24 patients with Alzheimerdisease and 24 HCs12 Cell type deconvolution revealed anoverrepresentation of microglial cells in the ALS motor cortexcompared with HC (p = 0025 fold change = 165) anda reduction in excitatory neurons (p = 0101 fold change =-085) (figure 2) We confirmed microglial upregulation using2 additional independent reference data sets1011 In bothcases we found the same pattern of downregulation of ex-citatory neurons (figures e-4 and e-5 linkslwwcomNXIA279) To further gain insight into which microglial sub-population might be driving this effect we estimated theproportion of each subpopulation identified in the ROSMAPstudy Our results showed that a specific microglial sub-population (Mic1 as named in the study performed byMathys and collaborators and recognized as the humanDAM12) is disproportionately presented in the ALS motorcortex (p = 66 times 10minus3 fold change = 433) (figure 2) Amongthe 77 marker genes which characterize the Mic1 subclusteras DAM (Mic1 unique genes or those overlapping withDAM)12 22 of them (286) are genes differentiallyexpressed in our bulk RNAseq analysis (unadjusted p lt 005(table e-7 linkslwwcomNXIA280) Of interest the pro-portion of Mic1 cells showed a positive correlation with theexpression of TREM2 (p = 16 times 10minus3 R = 067 figure e-6linkslwwcomNXIA279) a receptor required for the tran-sition from the homeostatic microglia to DAM

NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 7 Number 5 | September 2020 5

Gene coexpression network analysesTo identify gene clusters with varying coexpression patternsthat could behave differently between patients and healthyindividuals and elucidate possible biological mechanismsdriving pathologic processes we performed a gene coex-pression network analyses using the WGCNA package Theanalysis disclosed a total of 8 modules Among them 3modules associated with disease status (MEblack p = 0003 R= 064 MEyellow p = 0025 R = 057 and MEpink p =0011 R = minus051 figure 3) Moreover MEblack and MEyel-low were enriched for inflammatory responses whereasMEpink had an overrepresentation of genes involved insynaptic and neuronal functions (table e-8 linkslwwcomNXIA280) The SynGO database confirmed the post-synaptic signature of this enrichment in MEpink (table e-9linkslwwcomNXIA280) Of interest the 2 inflammatorygene coexpression modules (MEblack and MEyellow)showed a strong negative correlation with the postsynapticmodule (MEpink) (p = 8 times 10minus5 R = minus078 and p = 3 times 10minus4R = minus074 respectively) (figure 3) suggesting that both

inflammation and synaptic alterations are interconnected inALS pathologic processes Furthermore the proportion ofMic1 cells showed a strong direct correlation with the 2 in-flammatory gene coexpression modules associated with ALS(MEblack p = 27 times 10minus8 R = 092 and MEyellow p = 14 times10minus3 R = 068) and a negative correlation with the synapticmodule (MEpink p = 24 times 10minus4 R = minus075) (figure 3)Overall our results from differential gene and isoform ex-pression gene coexpression modules and cell type decon-volution comparisons strongly indicate that microglia-relatedinflammatory changes mainly driven by the Mic1 sub-population (also known as DAM) are central in ALS path-ophysiology and that these inflammatory processes are closelyrelated to the synaptic disturbances that are present in themotor cortex

ImmunohistochemistryTo further investigate the inflammatory response in post-mortem brain tissue we performed IHC analyses using Iba1a proteinwhose expression is primarily restricted to homeostatic

Figure 1 Differential regulation of gene expression in ALS and HC

(A) Volcano plot displaying differentially expressed genes between the ALS and the HC motor cortex The vertical axis (y-axis) corresponds to the minuslog10adjusted p value and the horizontal axis (x-axis) represents the log2 fold change value obtained when comparing ALS with HC gene expressionSignificantly differential expressed genes are depicted with blue circles (adjusted p value lt005) whereas gray circles display the nonsignificant genesGene names are shown for the 10 genes selected for qPCR validation (B) Box plot showing the relative RNA expression values in ALS and HC obtainedthrough qPCR for the 10 genes selected for validation Dashed line corresponds to reference relative gene expression p lt 005 p lt 001 (C) Top 5 geneontology and KEGG pathway enrichment analyses obtained from the list of genes showing a significant differential expression ALS = amyotrophic lateralsclerosis HC = healthy control

6 Neurology Neuroimmunology amp Neuroinflammation | Volume 7 Number 5 | September 2020 NeurologyorgNN

microglia to examine the broad population of microglial cellsand MHC class II a marker responsible for antigen recognitionand the activation of the adaptive immune system which isexpressed in the Mic1 cell subpopulation Positive immunos-taining was increased in the ALS motor cortex for Iba1 how-ever differences between patients and controls did not reachstatistical significance (p = 0106 figure e-7 linkslwwcomNXIA279) Notwithstanding co-immunofluorescence of Iba1andMHC class II markers revealed an increase in the number ofmicroglial cells expressing MHCII markers in ALS comparedwith HC (701 vs 10 respectively p = 0021) (figure 4 andfigure e-8 linkslwwcomNXIA279) strengthening ourresults and suggesting thatMic1 is driving neuroinflammation inthe ALS motor cortex

We finally assessed the frequency of pTDP43 immunoreac-tive structures in the motor cortex As expected patients withALS showed a higher density of pTDP43 inclusions

compared with HCs (p = 0017) (figure e-9A linkslwwcomNXIA279) Of interest a high degree of variability on thenumber of pTDP43 aggregates was noted in ALS and was notexplained by age at death age at onset or disease durationThe burden of pTDP43 inclusions inversely correlated withthe proportion of microglial cells in our group of patients withALS (p = 00026 R = minus081) (figure e-9B linkslwwcomNXIA279) and the same pattern was found with themicroglial subpopulation Mic1 (p = 0019 R = minus069)

DiscussionThe motor cortex is one of the major vulnerable and earlyaffected regions in ALS and represents a target region todisentangle key pathologic processes in this neurodegenera-tive disorder To date few studies have performed a wholetranscriptomic assessment of the CNS in ALS25ndash27 Through

Figure 2 Cell type composition in the ALS and HC motor cortex

Box plot showingmajor cell type andMic1 proportions for HC and ALS estimated byMuSiC using the ROSMAP human single-nucleus RNAseq data p lt 005p lt 001 ALS = amyotrophic lateral sclerosis HC = healthy control MuSiC = Multi-Subject Single Cell deconvolution ROSMAP = Religious Orders Study andMemory and Aging Project

NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 7 Number 5 | September 2020 7

Figure 3WGCNAvisualization and correlationmap constructedwithWGCNAmodule eigengenes and cell type proportions

Network diagrams showing the 50most connected genes for each of the 3modules associatedwith ALS (MEblackMEyellow andMEpink) are depicted on thetop of the figure Node darkness and size are proportional to the number of connections within the module Correlation matrix showing the correlationcoefficients between cell type proportions is depicted for the 6 major cell types and the Mic1 (DAM) cell subtype The 3 significant modules (MEblackMEyellow and MEpink) each of one sharing unique groups of coexpressed genes that are differentially expressed between patients with ALS and controlsare also included in the correlogram Only significant correlations are depicted with a colored circle (blue for direct and red for inverse correlations) The sizeof the circle is proportional to the correlation significance The plot indicates the high correlation between the 3modules and how thesemodules are relatedto the proportion of specific cell types ALS = amyotrophic lateral sclerosis DAM = disease-associated microglia HC = healthy control WGGNA = weightedgene coexpression correlation network analyses

8 Neurology Neuroimmunology amp Neuroinflammation | Volume 7 Number 5 | September 2020 NeurologyorgNN

an unbiased transcriptomic analysis using high-throughputRNA sequencing we have elucidated gene and isoform ex-pression alterations gene coexpression networks and celltype proportions associated with ALS For the first time in theALS field we have performed cell type deconvolution usinghuman brain single-nucleus RNA sequencing data from 3independent sources as reference thus providing highly reli-able results that have been reinforced through our immuno-histochemical analyses

We report a list of 124 genes differentially expressed in theALS motor cortex which reflect the RNA expression changesin this brain region Among them chitinase-related genesCHI3L1 and CHI3L2 presented 2 of the most prominentshifts in gene expression whereas CHIT1 showed a nominalsignificance Of interest these chitinases are neuro-inflammatory biomarkers which have been shown to beconsistently increased in the CSF of patients with ALScompared with neurologically healthy individuals78 Theseresults indicate that among the unbiased signature of RNAalterations provided herein some of them might lead to thediscovery of novel promising biomarkers

An exacerbated innate immune response withmicrogliosis hasbeen recently described in the ALS motor cortex28 A recentstudy that has investigated the whole transcriptome of theALS motor cortex suggested that among ALS cases a sub-group of them is characterized by a molecular signature

related to glial activation and inflammation27 Our resultsclearly reinforce the idea of these inflammatory-relatedchanges in the human motor cortex and emphasize the roleof DAM as a key factor in this process First differential geneand isoform expression data point toward an inflammatoryresponse as a major event that is strongly intensified in theALS motor cortex Second the assembly of weighted genecoexpression networks resulted in 2 significant modules as-sociated with ALS (MEblack and MEyellow) both highlyenriched with genes related to inflammatory functions Thirdcell type deconvolution of the bulk RNAseq data demon-strated an increased proportion of microglial cells comparedwith the motor cortex of healthy cases In this context studiesin mice have recently evidenced DAM as the microglial sub-population with a more prominent role in ALS4 Our resultsshow that Mic1 the human microglial subpopulation thatharbors the majority of markers found in the previously de-scribed DAM transcriptomic signature12 is the main micro-glial subpopulation that drives microgliosis in ALS In factalmost a third of Mic1 (DAM) marker genes are deregulatedin the motor cortex of our group of patients with ALS Ofnote Mic1 proportion highly correlated with the expressionof TREM2 which is required to enhance the proinflammatorystage of DAM4

Our immunohistochemical analyses did not show a relevantincreased density of Iba1 (a widely used marker of microglialprocesses) in ALS-related brain tissue nor was the expression

Figure 4 Immunohistochemistry of Mic1 markers

Immunohistochemistry with anti-IBA1 (red) and anti-MHC class II (green) antibodies in the ALS and HC motor cortex Scale bar corresponds to 50 μm ALS =amyotrophic lateral sclerosis HC = healthy control

NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 7 Number 5 | September 2020 9

of its coding gene (AIF1) altered in our RNASeq data setPrevious studies have frequently provided contradictoryresults when using this marker to assess microgliosis29 Thissomewhat unexpected result could be explained by the factthat some marker genes might downregulate on microglialactivation Also whereas cell type deconvolution is performedusing a complete catalogue of gene expression profilesobtained from brain-derived snRNAseq immunohistochem-istry only uses a single marker and does not reflect the com-plexity and heterogeneity of this cell type making it anunderpowered method to detect differences in the proportionof cell subpopulations That being said we did validate theincreased proportion of Mic1 cells in the ALS motor cortexthrough co-immunofluorescence of 2 markers (Iba1 andMHC class II) previously shown to characterize this micro-glial subpopulation12 This finding further establishes Mic1(the human disease-associated microglia) as the microglialsubpopulation that drives microgliosis in the ALS motorcortex

We observed a high degree of variability in the density ofpTDP43 inclusions across patients which could not beexplained by any of the demographic or clinical featuresavailable in this study Notably we found a striking inversecorrelation between the proportion of microglial cells and theamount of pTDP43 aggregates These results are in line withrecent in vivo and in vitro studies suggesting that althoughmicrogliosis arises as a phagocytic response to pTDP43aggregates at some point these cells lose their ability to clearthese neuropathologic insults and are downregulated3031

Whether biofluid levels of microglial markers such asTREM2 could be used as a proxy of pTDP43 density in themotor cortex is an avenue worth pursuing and would bea valuable addition to the biomarker arsenal for use in de-signing clinical trials and assessing therapeutic efficacy

Synaptic dysfunction is an early pathogenic event in ALS32

Our data point toward an underrepresentation of post-synaptic markers and a decrease of excitatory neurons inpatients with ALS Together these results are consistentwith upper motor neuron degeneration occurring in thisbrain region A limitation of our approach is the lack of anavailable motor cortex snRNAseq data set precluding anyfirm and more detailed conclusion related to the alteration ofBetz cells a unique class of motor neurons expressed in thisspecific brain area Recent studies in other neurodegenera-tive diseases have suggested that microglia are a key and earlymediator of synapse loss through phagocytosis induced bythe complement cascade3334 Our gene expression datashow an increased expression of some key complementcascade-related genes and the most significant enrichedpathway corresponds to the complement and coagulationcascades reinforcing this hypothesis Furthermore ourresults indicate that the 2 neuroinflammatory-related mod-ules of gene expression (MEblack and MEyellow) and theproportion of Mic1 cells inversely correlate with the pres-ence of synaptic markers

Overall our study strongly suggests that DAM plays a key rolein driving neuroinflammatory changes and synapse loss in theALS motor cortex The identification of specific microglialpopulations with well-defined transcriptional signatures willcontribute to disentangle new mechanisms and novel thera-peutic targets to fight against this devastating disorder

AcknowledgmentThe authors are indebted to the IDIBAPS Biobank for sampleand data procurement

Study fundingThe authors are indebted to the ldquoFundacion Espantildeola para elFomento de la Investigacion de la Esclerosis LateralAmiotrofica (FUNDELA)rdquo for funding the present study ODols-Icardo is a recipient of a grant by The Association forFrontotemporal Degeneration (Clinical Research Post-doctoral Fellowship AFTD 2019ndash2021) This work wassupported by research grants from Institute of Health CarlosIII (ISCIII) Spain PI1800326 to JC PI1800435 to DAand INT1900016 to DA and by the Department of HealthGeneralitat de Catalunya PERIS program SLT00617125 toDA This work was also supported in part by Generalitat deCatalunya (2017 SGR 00547) to the ldquoGrup de Recerca enDemencies Sant Paurdquo

DisclosureThe authors report no disclosures relevant to the manuscriptGo to NeurologyorgNN for full disclosures

Publication historyReceived by Neurology Neuroimmunology amp NeuroinflammationFebruary 21 2020 Accepted in final form May 15 2020

Appendix Authors

Name Location Contribution

Oriol Dols-Icardo PhD

Sant Pau BiomedicalResearch Institute Hospitalde la Santa Creu i Sant PauBarcelona Spain

Major role in the designand conceptualization ofthe study drafted themanuscript and analyzedthe data

VıctorMontal MSc

Sant Pau BiomedicalResearch Institute Hospitalde la Santa Creu i Sant PauBarcelona Spain

Analyzed the data andparticipated in manuscriptdrafting

Sonia SirisiPhD

Sant Pau BiomedicalResearch Institute Hospitalde la Santa Creu i Sant PauBarcelona Spain

Analyzed the data andparticipated in manuscriptdrafting

GemaLopez-Pernas MSc

Sant Pau BiomedicalResearch Institute Hospitalde la Santa Creu i Sant PauBarcelona Spain

Analyzed the data andperformed a part of theexperiments

LauraCervera-Carles PhD

Sant Pau BiomedicalResearch Institute Hospitalde la Santa Creu i Sant PauBarcelona Spain

Participated in manuscriptdrafting and datainterpretation

10 Neurology Neuroimmunology amp Neuroinflammation | Volume 7 Number 5 | September 2020 NeurologyorgNN

References1 Brettschneider J Del Tredici K Toledo JB et al Stages of pTDP-43 pathology in

amyotrophic lateral sclerosis Ann Neurol 20137420ndash382 Liu J Wang F Role of neuroinflammation in amyotrophic lateral sclerosis cellular

mechanisms and therapeutic implications Front Immunol 2017810053 Prinz M Jung S Priller J Microglia biology one century of evolving concepts Cell

2019179292ndash3114 Keren-Shaul H Spinrad A Weiner A et al A unique microglia type Associated with

restricting development of Alzheimerrsquos disease Cell 20171691276ndash1290

5 Steinacker P Verde F Fang L et al Chitotriosidase (CHIT1) is increased in microgliaand macrophages in spinal cord of amyotrophic lateral sclerosis and cerebrospinalfluid levels correlate with disease severity and progression J Neurol NeurosurgPsychiatry 201889239ndash247

6 Thompson AG Gray E Thezenas ML et al Cerebrospinal fluid macrophage bio-markers in amyotrophic lateral sclerosis Ann Neurol 201883258ndash268

7 Illan-Gala I Alcolea D Montal V et al CSF sAPPβ YKL-40 and NfL along the ALS-FTD spectrum Neurology 201891e1619ndashe1628

8 Thompson AG Gray E Bampton A Raciborska D Talbot K Turner MR CSFchitinase proteins in amyotrophic lateral sclerosis J Neurol Neurosurg Psychiatry2019901215ndash1220

9 Ling SC PolymenidouM Cleveland DW Converging mechanisms in ALS and FTDdisrupted RNA and protein homeostasis Neuron 201379416ndash438

10 Hodge RD Bakken TEMiller JA et al Conserved cell types with divergent features inhuman versus mouse cortex Nature 201957361ndash68

11 Lake BB Chen S Sos BC Integrative single-cell analysis of transcriptional and epi-genetic states in the human adult brain Nat Biotechnol 20183670ndash80

12 Mathys H Davila-Velderrain J Peng Z et al Single-cell transcriptomic analysis ofAlzheimerrsquos disease Nature 2019570332ndash337

13 Wang X Park J Susztak K Zhang NR Li M Bulk tissue cell type deconvo-lution with multi-subject single-cell expression reference Nat Commun 201910380

14 Brooks BR Miller RG Swash M Munsat TL World Federation of Neurology Re-search Group on Motor Neuron Diseases El Escorial revisited revised criteria for thediagnosis of amyotrophic lateral sclerosis Amyotroph Lateral Scler Other MotorNeuron Disord 20001293ndash299

15 Garcıa-Redondo A Dols-Icardo O Rojas-Garcıa R et al Analysis of the C9orf72 genein patients with amyotrophic lateral sclerosis in Spain and different populationsworldwide Hum Mutat 20133479ndash82

16 Dols-Icardo O Garcıa-Redondo A Rojas-Garcıa R et al Analysis of known amyo-trophic lateral sclerosis and frontotemporal dementia genes reveals a substantial ge-netic burden in patients manifesting both diseases not carrying the C9orf72 expansionmutation J Neurol Neurosurg Psychiatry 201889162ndash168

17 Dobin A Davis CA Schlesinger F et al STAR ultrafast universal RNA-seq alignerBioinformatics 20132915ndash21

18 McKenna A Hanna M Banks E et al The Genome Analysis Toolkit a MapReduceframework for analyzing next-generation DNA sequencing data Genome Res 2010201297ndash1303

19 Liao Y Smyth GK Shi W The R package Rsubread is easier faster cheaper and betterfor alignment and quantification of RNA sequencing reads Nucleic Acids Res 201947e47

20 Patro R Duggal G Love MI Irizarry RA Kingsford C Salmon provides fast andbias-aware quantification of transcript expression Nat Methods 201714417ndash419

21 LoveMI HuberW Anders S Moderated estimation of fold change and dispersion forRNA-seq data with DESeq2 Genome Biol 201415550

22 Zhou Y Zhou B Pache L et al Metascape provides a biologist-oriented resource forthe analysis of systems-level datasets Nat Commun 2019101523

23 Koopmans F van Nierop P Andres-Alonso M et al SynGO an evidence-basedexpert-curated knowledge base for the synapse Neuron 2019103217ndash234e4

24 Langfelder P Horvath S WGCNA an R package for weighted correlation networkanalysis BMC Bioinformatics 20089559

25 Prudencio M Belzil VV Batra R et al Distinct brain transcriptome profiles inC9orf72-associated and sporadic ALS Nat Neurosci 2015181175ndash1182

26 DrsquoErchia AM Gallo A Manzari C et al Massive transcriptome sequencing of humanspinal cord tissues provides new insights into motor neuron degeneration in ALS SciRep 2017710046

27 Tam OH Rozhkov NV Shaw R et al Postmortem cortex samples identify distinctmolecular subtypes of ALS retrotransposon activation oxidative stress and activatedglia Cell Rep 2019291164ndash1177e5

28 Jara JH Genccedil B Stanford MJ et al Evidence for an early innate immune response inthe motor cortex of ALS J Neuroinflammation 201714129

29 Hopperton KE Mohammad D Trepanier MO Giuliano V Bazinet RP Markers ofmicroglia in post-mortem brain samples from patients with Alzheimerrsquos diseasea systematic review Mol Psychiatry 201823177ndash198

30 Spiller KJ Restrepo CR Khan T et al Microglia-mediated recovery from ALS-relevant motor neuron degeneration in a mouse model of TDP-43 proteinopathy NatNeurosci 201821329ndash340

31 Svahn AJ Don EK Badrock AP et al Nucleo-cytoplasmic transport of TDP-43 studied in real time impaired microglia function leads to axonal spreadingof TDP-43 in degenerating motor neurons Acta Neuropathol 2018136445ndash459

32 Henstridge CM Pickett E Spires-Jones TL Synaptic pathology a shared mechanismin neurological disease Ageing Res Rev 20162872ndash84

33 Hong S Beja-Glasser VF Nfonoyim BM et al Complement and microglia mediateearly synapse loss in Alzheimer mouse models Science 2016352712ndash716

34 Dejanovic B Huntley MA De Maziere A et al Changes in the synaptic proteome intauopathy and rescue of tau-induced synapse loss by C1q antibodies Neuron 20181001322ndash1336e7

Appendix (continued)

Name Location Contribution

MartaQuerol-VilasecaMSc

Sant Pau BiomedicalResearch Institute Hospitalde la Santa Creu i Sant PauBarcelona Spain

Participated in manuscriptdrafting and datainterpretation

Laia MuntildeozBSc

Sant Pau BiomedicalResearch Institute Hospitalde la Santa Creu i Sant PauBarcelona Spain

Contributed to samplepreparation andparticipated inexperimental procedures

OliviaBelbin PhD

Sant Pau BiomedicalResearch Institute Hospitalde la Santa Creu i Sant PauBarcelona Spain

Participated in the revisionand editing of themanuscript and datainterpretation

DanielAlcoleaMD PhD

Sant Pau BiomedicalResearch Institute Hospitalde la Santa Creu i Sant PauBarcelona Spain

Participated in datacollection analysis anddata interpretation

LauraMolina-Porcel MDPhD

Neurological Tissue Bankof the Biobanc-HospitalClınic-IDIBAPS BarcelonaSpain

Contributed to samplepreparation andparticipated inexperimental procedures

JordiPeguerolesMSc

Sant Pau BiomedicalResearch Institute Hospitalde la Santa Creu i Sant PauBarcelona Spain

Participated in dataanalysis

JaninaTuron-SansMD

Sant Pau BiomedicalResearch Institute Hospitalde la Santa Creu i Sant PauBarcelona Spain

Major role in theacquisition of samples anddata and reviewed andedited the manuscript

RafaelBlesa MDPhD

Sant Pau BiomedicalResearch Institute Hospitalde la Santa Creu i Sant PauBarcelona Spain

Major role in theacquisition of samples anddata and reviewed andedited the manuscript

AlbertoLleo MDPhD

Sant Pau BiomedicalResearch Institute Hospitalde la Santa Creu i Sant PauBarcelona Spain

Major role in theacquisition of samples anddata and reviewed andedited the manuscript

Juan ForteaMD PhD

Sant Pau BiomedicalResearch Institute Hospitalde la Santa Creu i Sant PauBarcelona Spain

Major role in theacquisition of samples anddata and reviewed andedited the manuscript

RicardRojas-Garcıa MDPhD

Sant Pau BiomedicalResearch Institute Hospitalde la Santa Creu i Sant PauBarcelona Spain

Major role in theacquisition of samples anddata and reviewed andedited the manuscript

JordiClarimonPhD

Sant Pau BiomedicalResearch Institute Hospitalde la Santa Creu i Sant PauBarcelona Spain

Major role in fundingacquisition design andconceptualization of thestudy drafted themanuscript and analyzedthe data

NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 7 Number 5 | September 2020 11

DOI 101212NXI000000000000082920207 Neurol Neuroimmunol Neuroinflamm

Oriol Dols-Icardo Viacutector Montal Sogravenia Sirisi et al sclerosis

Motor cortex transcriptome reveals microglial key events in amyotrophic lateral

This information is current as of July 15 2020

ServicesUpdated Information amp

httpnnneurologyorgcontent75e829fullhtmlincluding high resolution figures can be found at

References httpnnneurologyorgcontent75e829fullhtmlref-list-1

This article cites 34 articles 5 of which you can access for free at

Subspecialty Collections

httpnnneurologyorgcgicollectiongene_expression_studiesGene expression studies

httpnnneurologyorgcgicollectionamyotrophic_lateral_sclerosis_Amyotrophic lateral sclerosisfollowing collection(s) This article along with others on similar topics appears in the

Permissions amp Licensing

httpnnneurologyorgmiscaboutxhtmlpermissionsits entirety can be found online atInformation about reproducing this article in parts (figurestables) or in

Reprints

httpnnneurologyorgmiscaddirxhtmlreprintsusInformation about ordering reprints can be found online

Academy of Neurology All rights reserved Online ISSN 2332-7812Copyright copy 2020 The Author(s) Published by Wolters Kluwer Health Inc on behalf of the AmericanPublished since April 2014 it is an open-access online-only continuous publication journal Copyright

is an official journal of the American Academy of NeurologyNeurol Neuroimmunol Neuroinflamm