ARTICLE OPEN ACCESS

Targeted RNA Sequencing of Formalin-FixedParaffin-Embedded Temporal Arteries FromGiant Cell Arteritis Cases Reveals ViralSignaturesAndrew N Bubak PhD Teresa Mescher BS Michael Mariani PhD Seth E Frietze PhD

James E Hassell Jr PhD Christy S Niemeyer PhD Christina N Como BA Anna M Burnet BS

Prem S Subramanian MD PhD Randall J Cohrs PhD Ravi Mahalingam PhD and Maria A Nagel MD

Neurol Neuroimmunol Neuroinflamm 20218e1078 doi101212NXI0000000000001078

Correspondence

Dr Nagel

marianagelucdenveredu

AbstractBackground and ObjectivesVaricella zoster virus (VZV) antigen has been detected in temporal arteries (TAs) of individualswith giant cell arteritis (GCA) the most common systemic vasculitis in older adults Thus weexplored the contribution of VZV to GCA pathogenesis

MethodsFormalin-fixed paraffin-embedded TA sections from biopsy-positiveGCA participants with VZVantigen (GCAVZV-positive n = 20) and without (GCAVZV-negative n = 20) and fromnormal participants with VZV antigen (controlVZV-positive n = 11) and without (controlVZV-negative n = 20) were analyzed by targeted RNA sequencing of the whole human tran-scriptome (BioSpyder TempO-Seq) Ingenuity pathway analysis and R-computational programwere used to identify differentially expressed genes and pathways between groups

ResultsCompared with controlVZV-negative TAs GCAVZV-negative and GCAVZV-positive TAswere significantly enriched for human transcripts specific for pathways involved in viral in-fections including viral entry nuclear factor kappa B activation by viruses and other pathogen-related immune activation pathways Similarly human gene sets supporting viral infection werefound in controlVZV-positive TAs that showed no morphological signs of inflammationsuggesting that the enriched pathways were not nonspecific signatures of infiltrating immunecells All GCA TAs and controlVZV-positive TAs showed enrichment of transcripts involvedin vascular remodeling including smooth muscle cell migration

DiscussionThe detection of viral and immune activation pathways in GCA TAs supports a role for virusinfection in GCA pathogenesis In addition the detection of viral pathways in controlVZV-positive TAs along with vascular remodeling pathways suggests that these samples mayrepresent early infection with progression to clinical disease depending on host and otherenvironmental factors

From the Department of Neurology (ANB TM JEH CSN CNC AMB PSS RJC RM MAN) University of Colorado School of Medicine Aurora Department of MedicalLaboratory Sciences (MM SEF) University of Vermont Burlington Department of Ophthalmology (PSS MAN) University of Colorado School of Medicine Aurora Departmentof Neurosurgery (PSS) University of Colorado School of Medicine Aurora and Department of Immunology and Microbiology (RJC) University of Colorado School of MedicineAurora

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

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 2021 The Author(s) Published by Wolters Kluwer Health Inc on behalf of the American Academy of Neurology 1

Giant cell arteritis (GCA) the most common systemic vas-culitis in the elderly has been proposed to be a form of ex-tracranial varicella zoster virus (VZV) vasculopathy1 becauseboth VZV vasculopathy and GCA (1) affect the elderly (2)can present with vision loss headaches and stroke2-5 and (3)produce transmural arterial inflammation with giantepithelioid cells and medial damage in temporal cerebralandor other systemic arteries6-9 Importantly 70 of biopsy-positive GCA temporal arteries (TAs) contain VZV antigencompared with 18 of controls10 Follow-up studies by othergroups yielded variable frequencies of VZV antigen11-15

raising the question of whether virus contributes to GCApathogenesis Previously gene expression analyses of GCATAs that would identify viral signatures were challengingbecause large numbers of fresh TAs were difficult to obtainand formalin-fixed paraffin-embedded (FFPE) TAs had de-graded RNA Recently a targeted RNA sequencing assay forthe whole human transcriptome (TempO-Seq BioSpyderTechnologies Carlsbad CA) was developed that (1) can di-rectly amplify fragmented human RNA in FFPE slides with-out a reverse transcription step and (2) has results equivalentto the analysis of fresh or frozen tissue16 This technologyadvancement has allowed us to determine the transcriptionalprofiles of specific VZV antigenndashcontaining regions withinsingle FFPEGCATA slides and compare with control TAs toelucidate the role of virus infection in GCA pathogenesis

MethodsStandard Protocol Approvals Registrationsand Patient ConsentsArchived patient samples were deidentified and deemed ex-empt not human research as defined by the Colorado Mul-tiple Institutional Review Board (COMIRB) policies andcurrent regulations and in accordance with Office for HumanResearch Protection and Food and Drug Administrationguidelines (Protocol no 13-2550 COMIRB 303-724-1055)

Patient SamplesForty FFPE TA biopsies from patients with GCA older than50 years were used all GCA TAs were biopsy-positive withtransmural inflammation giantepithelioid cells and medialdamage Controls were 31 normal TAs (without zoster di-abetes cancer substance abuse or immunosuppression)removed postmortem from participants gt50 years of age atthe University of Colorado Hospital Arapahoe CountyCoronerrsquos Office and Denver Office of the Medical Exam-iner These TAs were previously obtained and analyzed by

immunohistochemistry (IHC) and quantitative PCR for thepresence of VZV antigen and DNA respectively110 Pre-viously published data showed that of the 40 GCA TAs 16contained VZV antigen with corresponding DNA from thescraped VZV antigen-positive fields and 4 contained VZVantigen only (20 GCAVZV-positive TAs) 20 did notcontain VZV antigen (20 GCAVZV-negative TAs) Of 31control TAs 3 contained VZV antigen with correspondingDNA and 8 contained VZV antigen only (11 controlVZV-positive TAs) 20 did not contain VZV antigen (20 controlVZV-negative TAs)1 Patient sample information is in eTa-ble 1 (linkslwwcomNXIA574)

FFPE Tissue Sequencing and AnalysisFor whole human transcriptome analysis FFPE slides fromTAs were assayed using TempO-Seqndashtargeted RNA se-quencing plates reagents protocols and software (BioSpyderTechnologies) TempO-Seq exclusively detects human tran-scripts viral transcripts will not be detected For slides withVZV antigen the same antigen-containing regions on un-stained adjacent slides were measured scraped (10 mm2

pooled regions of interest per participant) and placed intoPCR tubes containing 1X lysis buffer For slides without VZVantigen equivalent areas of arteries from unstained slides wereanalyzed Using a thermocycler samples were lysed codedadjacent primer pairs for each specific human transcript wereannealed to sample RNA and primer pairs for each transcriptligated and then amplified per manufacturerrsquos instructions16

If a transcript is present the adjacent 25-nucleotide primerpairs anneal to their specific target ligate together and pro-duce a 50-nucleotide transcript-specific coded ampliconAmplified PCR products were pooled into a single libraryconcentrated using a PCR cleanup kit (Macherey-NagelDuren Germany) and run on the Illumina NextSeq 500 se-quencing platform (Illumina Inc San Diego CA) Mappedreads were generated by TempO-SeqR for the alignment ofdemultiplexed FASTQ files from the sequencer to the ligateddetector oligomer gene sequences using Bowtie17 allowingfor up to 2mismatches in the 50-nucleotide target sequence16

Counts were assessed using SARTools18 Within this Rpackage edgeR is used for normalization and quality controlof count data19 Rawnormalized counts and multidimen-sional scaling between the 4 groups are shown in eFigure 1(linkslwwcomNXIA571) Differential expression betweengroups was assessed by the TempO-SeqR software whichused the DESeq2 method for differential analysis of countdata20 A significantly differentially expressed gene is definedas having an adjusted p value lt 005 with no fold-changethreshold

GlossaryCOMIRB = Colorado Multiple Institutional Review Board FFPE = formalin-fixed paraffin-embedded GCA = giant cellarteritis gE = glycoprotein E GSEA = Gene Set Enrichment Analysis HSV = herpes simplex virus IFN = interferon IHC =immunohistochemistry IL = interleukin TA = temporal artery TLR = toll-like receptor TNF = tumor necrosis factor VZV =varicella zoster virus

2 Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 6 | November 2021 NeurologyorgNN

Pathway enrichment analysis was performed using gene setsand pathways defined in the Ingenuity Pathway Analysissoftware (Qiagen Germantown MD) and the ClusterProfilerpackage2122 in R with default parameters23 ClusterProfilersupports enrichment analysis of Gene Ontology and KyotoEncyclopedia of Genes and Genomes databases with GeneSet Enrichment Analysis (GSEA) to identify biologicalthemes of a collection of genes These functional enrichmentanalyses use computational approaches to identify groups ofexperimentally observed human genes that are over-represented or depleted in a curated disease or biologicalfunction-specific gene set Additional figures were createdusing Prism 9 (GraphPad Software San Diego CA)

IHC AnalysisTo confirm sequencing results 5-μm FFPE sections from 10GCAVZV-positive TAs and 10 controlVZV-negative TAswere immunostained for interleukin (IL)-8 and neutrophils(CD15) After deparaffinization antigens were retrieved incitrate buffer at 95degC for 10 minutes Slides were immunos-tained as described1 using primary antibodies against IL-8(mouse anti-IL-8 1500 dilution Abcam Cambridge MAcatalog ab18672) and CD15 (mouse anti-CD15 150 di-lution Abcam catalog ab188610) IL-8 and CD15 positivecontrol was normal human tonsil Slides were imaged by lightmicroscopy

Data AvailabilityThe datasets generated andor analyzed are available on theNCBI Gene Expression Omnibus database (GSE174694)

ResultsGene Expression Analysis of ControlVZV-Negative and ControlVZV-Positive TAsThe 31 control TAs were previously immunostained for VZVglycoprotein E (gE) (50 sections from each artery) 20 werenegative for VZV gE and 11 were positive for VZV gE (rep-resentative staining Figure 1A) Compared with controlVZV-negative TAs controlVZV-positive TAs had 78 dif-ferentially expressed genes (padj lt 005 41 upregulated [red]and 37 downregulated [blue] Figure 1B gene list in eTable 2linkslwwcomNXIA575) ControlVZV-positive TAs wereenriched for pathways involved in viral infections fibro-blast proliferation and vascular cell migrationspreading(Figure 1C full gene list in eTable 3 linkslwwcomNXIA576) For example the viral infection pathway genesIFITM2 and IFITM3 encoding interferon (IFN)-inducibletransmembrane proteins that function in viral resistance24

were upregulated in controlVZV-positive TAs Multiple ac-tivated (HRAS and ESR1 red) and repressed (MYCN blue)upstream regulators were identified in controlVZV-positive

Figure 1 Gene Expression Analysis of Control TAs Without and With VZV Antigen

(A) TAs were analyzed by immunohistochemistry for VZV antigen representative arteries show the absence and presence of VZV antigen staining (left andright panels controlVZV-negative and controlVZV-positive respectively magnification times600) Targeted RNA sequencing of the whole human transcriptomeand gene expressionpathway analysis of controlVZV-positive compared with controlVZV-negative TAs revealed distinct differences (B) Scatterplot analysisshowed significantly upregulated (red) and downregulated (blue) genes (C) Comparedwith controlVZV-negative TAs controlVZV-positive TAswere enrichedfor pathways involved in viral infection cell proliferation and migration (D) Upstream regulators (transcription regulators enzymes ligand-dependentnuclear receptors and growth factors) and corresponding number of downstream genes were identified in controlVZV-positive comparedwith controlVZV-negative TAs (red signifies activated blue signifies repressed and gray signifies unknown activation pattern of transcripts) gE = glycoprotein E TA = temporalartery VZV = varicella zoster virus

NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 6 | November 2021 3

TAs (Figure 1D) Overall these results support a transcrip-tional profile indicative of an active viral infection in controlTAs that contained VZV antigen corroborating IHC de-tection of VZV gE in these samples

Gene Expression Analysis of ControlVZV-Negative and GCAVZV-Positive TAsThe 40 GCA TAs were previously analyzed for VZV gE an-tigen and VZV DNA All GCA TAs displayed classic biopsy-positive GCA features transmural inflammation medialdamage and multinucleated giantepithelioid cells(Figure 2A left panel) Twenty TAs contained VZV gE(GCAVZV-positive) of which 16 were confirmed to haveamplifiable VZV DNA (Figure 2A representative GCA TAwith VZV gE) The remaining 20 TAs did not contain VZV gE(GCAVZV-negative) Compared with controlVZV-negative TAs GCAVZV-positive TAs had 1353 differen-tially expressed genes (padj lt 005 1095 upregulated [red]and 258 downregulated [blue] Figure 2B gene list in eTable2 linkslwwcomNXIA575) Top canonical pathwaysrevealed known pathways associated with GCA2526 includingTh1 and Th2 pathways CD28 signaling in T-helper cellsT-cell exhaustion IFN signaling neuroinflammation signal-ing and toll-like receptor (TLR) signaling (Figure 2C) Inaddition defined canonical pathways that included virus entrythrough endocytic pathways were enriched in GCAVZV-positive samples (Figure 2C red signifies activation bluesignifies inhibition and gray signifies unavailable activationpattern) This involved a number of genes encoding endocyticcomponents such as clathrin (CLTA) and the actin-relatedproteins (ARPC2 ACTR3 and ACTR1B) Multiple cytokineswere identified as upstream regulators for GCAVZV-positiveTAs compared with controlVZV-negative TAs includingIFNγ tumor necrosis factor (TNF) IL-1B IL-2 IL-15 IL-6and IL-8 (Figure 2D) In addition several growth factors(TGFB1 and KITLG) transcription regulators (STAT1STAT3 IRF7 and SIRT1) TLRs (TLR3 TLR4 and TLR9)and the transferrin receptor (TFRC) were also linked as keyupstream regulators The major histocompatibility complexclass II gene human leukocyte antigen (HLA)-DRB1 ex-pression was higher in GCAVZV-positive TAs comparedwith controlVZV-negative TAs (351 plusmn 121 log2Fold-Change plusmn standard error p-adj lt 005) which has been pre-viously reported as a GCA risk factor (reviewed by Carmonaet al27)

Viral Infection and Immune Responses AreEnriched in GCAVZV-Positive TAs ComparedWith ControlVZV-Negative TAsFurther investigations into GCAVZV-positive TAs revealedenrichment of vascular and arterial disease gene sets associ-ated with activated viral infection and neutrophil de-granulation pathways (Figure 3A full gene list in eTable 3linkslwwcomNXIA576) Differentially expressed humantranscripts associated with advanced-stage peripheral arterial

disease degranulation of neutrophils and viral infectionshowed similar positive expression directionality (Figure 3B)and gene set overlap (Figure 3C) Running enrichment scoresin the GSEA showed an increase in neutrophil degranulationvirus diseases and vasculitis in GCAVZV-positive TAs com-pared with controlVZV-negative TAs and a decrease inmuscle structure development (Figure 3D) The presence ofIL-8 (a chemoattractant of neutrophils) and CD15 (marker ofneutrophils) was confirmed in all 10 GCAVZV-positive TAsexamined (representative TA second row eFigure 2 linkslwwcomNXIA572) No controlVZV-negative TAs were posi-tive for IL-8 (third row eFigure 2) and only 1 was positive forCD15 in the 10 samples examined Human tonsils were apositive control for IL-8 and CD15 (first row eFigure 2) Nostaining was observed for IL-8 or CD15 when primary anti-bodies were substituted with mIgG1 isotype control antibody(second and fourth columns eFigure 2)

Viral Infection and Immune Responses AreEnriched in GCAVZV-Negative TAs ComparedWith ControlVZV-Negative TAsA subset of TAs from patients with GCA were negative forVZV antigen (GCAVZV-negative) We investigated whetherthis group is reflective of a nonspecific inflammatory responseor potentially represents TAs where detection of VZV antigen(or another virus) was missed because of skip lesions1

Compared with controlVZV-negative TAs GCAVZV-negative TAs had 1784 differentially expressed genes (padjlt 005 1378 upregulated [red] and 406 downregulated[blue] Figure 4A gene list in eTable 2 linkslwwcomNXIA575) Top canonical pathways of GCAVZV-negative andGCAVZV-positive TAs were similar compared withcontrolVZV-negative TAs GCAVZV-negative TAs wereenriched for Th1 and Th2 pathways CD28 signaling inT-helper cells T-cell exhaustion IFN signaling and neuro-inflammation signaling (Figure 4B red signifies activationblue signifies inhibition and gray signifies unavailable acti-vation pattern) As was seen in GCAVZV-positive TAsGCAVZV-negative TAs also showed enrichment forpathways associated with viral infections including nuclearfactor kappa B activation by viruses and virus entry throughendocytic pathways (Figure 4B) Furthermore analysis ofenrichment scores through GSEA revealed activated genesets reflective of viral infections including defense responseto virus and TLR signaling (Figure 4C) This includedupregulated expression of genes JAK1 JAK2 STAT1 andSTAT2 of the JAK-STAT signaling pathway in GCAVZV-negative TAs Upstream regulators of differentiallyexpressed genes in GCAVZV-negative TAs were largelysimilar to GCAVZV-positive TAs such as cytokines (IFNγTNF IL-1B IL-2 IL-15 IL-6 and IL-8) the growth factorTGFB1 transcription regulators (STAT1 STAT3 andIRF7) TLRs (TLR3 and TLR4) and the iron transporterTFRC (eFigure 3 linkslwwcomNXIA573 red signifiesactivation and blue signifies repression)

4 Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 6 | November 2021 NeurologyorgNN

Predicted Upstream Regulators and PathwaysGenerated From Unique DifferentiallyExpressed Genes in GCAVZV-Negative andGCAVZV-Positive TAs Compared With ControlVZV-Negative TAsAlthough GCAVZV-positive and GCAVZV-negative TAsdisplayed similar pathways and upstream regulators com-pared with controlVZV-negative TAs we investigated theextent of differentially expressed genes that were shared orunique between GCA TAs without or with VZV antigencompared with controlVZV-negative TAs GCAVZV-negative and GCAVZV-positive TAs shared 1111 genesthat were differentially expressed compared with controlVZV-negative TAs (Figure 5A middle) Conversely GCA

VZV-negative vs controlVZV-negative TAs had 673 uniquedifferentially expressed genes that were not found in GCAVZV-positive vs controlVZV-negative TAs (Figure 5Aleft) A total of 249 genes were unique to GCAVZV-positive vs controlVZV-negative TAs that were not foundin GCAVZV-negative vs controlVZV-negative TAs(Figure 5A right) Upstream regulators for GCAVZV-negative unique genes included IL-4 EIF4E and TGFB1(Figure 5B left red signifies activated) upstream regulatorsfor GCAVZV-positive unique genes included IFNγ TNFand myocardin-related transcription factor A (MRTFA)(Figure 5B right red signifies activated and blue signifiesinhibited) Effects of decreased MRTFA in our samples re-main to be determined because increased MRTFA has been

Figure 2 Gene Expression Analysis of Control TAs Without VZV Antigen and GCA TAs With VZV Antigen

(A) GCAVZV-positive TAs that contained VZV antigen by immunohistochemistry were previously stained with hematoxylin and eosin to confirm GCApathology1 A representative GCAVZV-positive TA is shown (left panel magnification times100) with transmural inflammation medial damage and multinu-cleated giant cells (left panel arrow and inset) as well another GCAVZV-positive TA containing VZV antigen (right panel red) Targeted RNA sequencing of thewhole human transcriptome and gene expressionpathway analysis of GCAVZV-positive TAs compared with controlVZV-negative TAs revealed distinctdifferences (B) Scatterplot analysis showed significantly upregulated (red) and downregulated (blue) genes in GCAVZV-positive TAs compared with controlVZV-negative TAs (C) Top canonical pathways identified in GCAVZV-positive vs controlVZV-negative TAs indicated that GCAVZV-positive TAs seem toactivate immune responses and viral pathways (D) Key upstream regulators (cytokines growth factors transcription regulators transmembrane receptorsand transporters) and corresponding number of downstream genes were identified in GCAVZV-positive TAs compared with controlVZV-negative TAs (redsignifies activated and blue signifies repressed) GCA = giant cell arteritis gE = glycoprotein E IFN = interferon IL = interleukin TA = temporal artery TNF =tumor necrosis factor VZV varicella zoster virus

NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 6 | November 2021 5

associated with cell migration and pathologic vascular remod-eling in mice28 Enrichment analysis of GCAVZV-negativeunique differentially expressed genes (DEGs) revealed path-ways and gene sets associated with cell movement of leukocytesdifferentiation of mononuclear leukocytes replication of Her-pesviridae and viral infection (Figure 5C left full gene list ineTable 3 linkslwwcomNXIA576) Associated pathways andgene sets for GCAVZV-positive unique DEGs included in-flammatory response peripheral vascular disease systemic au-toimmune syndrome and viral infection (Figure 5C right fullgene list in eTable 2 linkslwwcomNXIA575) Examples of

inflammatory genes upregulated included JAK1 TNFRSF10ANFATC1 and NFKB1

DiscussionVZV vasculopathy was initially described as virus infection ofcerebral arteries leading to ischemic or hemorrhage strokeaneurysm or other cerebrovascular abnormalities that couldoccur with or without associated rash Subsequently VZVvasculopathy expanded to include the extracranial circulation

Figure 3 ViralInflammatory Pathways in GCA TAs With VZV Antigen Compared With Control TAs Without VZV Antigen

(A) Human gene sets involved in vascular and arterial disease viral infections and neutrophil movement and degranulation were significantly enriched in GCAVZV-positive TAs compared with controlVZV-negative TAs (red signifies activated and gray signifies unknown activity pattern) (B) Volcano and density plotshowing distribution and direction of differentially expressed genes involved in advanced stage peripheral arterial disease (gray) degranulation of neutrophils(gold) and viral infection (blue)measured in GCAVZV-positive TAs compared with controlVZV-negative TAs (C) Venn diagram showing independent and shareddifferentially expressed genes between viral infection (blue) degranulation of neutrophils (gold) and advanced stage peripheral arterial disease (gray) genesetsmeasured inGCAVZV-positive TAs (D) InGCAVZV-positiveTAs gene setswerepositively enriched forneutrophil degranulation virusdiseases and vasculitisbut negatively enriched for muscle structure development The y-axis represents ES and RM (significant genes ranked by fold-change) x-axis are genes (verticalblack lines) represented in the ranked order gene sets The green line connects points of ES and genes (NES) ES = enrichment score FDR = false discovery rateGCA = giant cell arteritis NES = normalized enrichment score RM = ranked metric TA = temporal artery VZV varicella zoster virus

6 Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 6 | November 2021 NeurologyorgNN

with viral antigen found in systemic arteries including TAsand the aorta6-9 GCA is a systemic vasculitis of the elderlythat produces a constellation of symptoms and signs includingmalaise headache TA tenderness vision loss elevatedC-reactive protein platelet count andor sedimentation ratethat can progress to aortitis and stroke TA biopsies show atransmural vasculitis with medial damage and giantepithelioid cells Treatment is with corticosteroids to reducethe vascular inflammation however the exact cause of GCA isunknown Previous studies demonstrating that VZV vascul-opathy and GCA share clinicohistopathological features andthat VZV antigen is found in a large number of GCA TAssuggest that a subset or all of GCA is a form of extracranialVZV vasculopathy110 Several other studies did not find VZVantigen in TAs or found it at much lower frequencies11-15these differences were attributed to nonspecific antigenstaining missed antigen detection because of skip lesionsdifferences in methodology andor insufficient sample sizesanalyzed In addition it was speculated that if VZV antigen aswell as confirmatory VZV DNA were present in the GCATAs it was merely a bystander and uninvolved in the path-ogenesis Thus we examined the gene expression profiles of

GCA TAs to determine if human transcripts supportive ofvirus infection were present Specifically we completed wholehuman transcriptome and pathway analysis of GCA TAs withand without VZV antigen (n = 20 in each group) and controlTAs with and without VZV antigen (n = 11 and 20 re-spectively) Our results showed activation of viral and in-flammatory pathways in GCA TAs supporting a pathogenicrole for a virus most likely VZV in this vasculitis

The TempO-Seq targeted RNA sequencing assay used in ouranalysis provides a valuable tool to investigate human tran-scriptional pathways from archived FFPE tissues without theneed for RNA extraction and conversion to complementaryDNA of note this assay exclusively detects human tran-scripts not viral transcripts The results of TempO-seq anal-ysis of FFPE tissue are comparable with that of fresh andfrozen tissue16 using RNA samples TempO-seq generatesresults consistent with Affymetrix microarrays and Illuminawhole transcriptome RNA-Seq29 Furthermore this assay hasbeen used by other laboratories to identify biomarkers foraggressiveness of clear cell renal cell carcinoma and prognosisin FFPE tumor samples that were then verified by RNA-seq

Figure 4 ViralInflammatory Pathways in GCA TAs Without VZV Antigen Compared With Control TAs Without VZV Antigen

(A) Scatterplot depicting significantly upregulated (red) and downregulated (blue) genes in GCAVZV-negative TAs compared with controlVZV-negative TAs(B) Top canonical pathways identified in GCAVZV-negative TAs vs controlVZV-negative TAs showed signatures of viral infections and immune responses (C)InGCAVZV-negative TAs gene setswere positively enriched for defense response to virus viral life cycle and toll-like receptor signaling The y-axis representsES and RM (significant genes ranked by fold-change) x-axis are genes (vertical black lines) represented in the ranked order gene sets The green line connectspoints of ES and genes (NES) ES = enrichment score FDR = false discovery rate GCA = giant cell arteritis IL = interleukin NES = normalized enrichment scoreNFAT = nuclear factor of activated T-cells NF-KB = nuclear factor kappa B RM = ranked metric TA = temporal artery VZV varicella zoster virus

NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 6 | November 2021 7

data from The Cancer Genome Atlas Kidney Renal Clear CellCarcinoma30 Without the limitation of access to fresh orfrozen GCA TA tissues we were able to sequence FFPEsamples that have been archived up to 16 years Importantlyslides can be immunostained for regions of pathology(10 mm2 5ndash7 μm thickness) then scraped and analyzed Inthis study adjacent sections can be used in IHC to confirmtranscriptional pathways such as the IL-8 and neutrophil in-filtration confirmation Multiple findings of novel human geneexpression pathways supporting a viral component in GCAand findings consistent with previous reports emerged fromthis study

Our most salient finding was the detection of robust viralsignatures in GCAVZV-positive TAs compared withcontrolVZV-negative TAs supporting previous IHC find-ings suggesting a viral etiologycontribution to GCApathogenesis1031 Specifically we found human gene ex-pression pathways associated with viral entry and replication

and several shared genes in gene sets associated with viralinfections arterial disease and immune cell activation sup-porting a pathogenic role for virus in GCA TAs that containVZV antigen Like GCAVZV-positive TAs GCAVZV-negative TAs showed transcriptional signatures associatedwith viral pathways and robust immune response pathwaysThese 2 groups were similar sharing a majority of top ca-nonical pathways It is possible that individuals with GCAVZV-negative TAs have cleared viral antigen and now rep-resent a cascade of pathology independent of active viralreplication or another pathogen is involved alternativelyVZV antigen on IHC may have been missed because of skiplesions Indeed a comparison of unique transcripts betweencontrol and GCA-positive TAs with or without VZV antigenrevealed gene sets associated with generalized Herpesviridaeinfection potentially representing a contribution of patho-gens beyond VZV Additional epidemiological studies sup-port an association between VZV and to a lesser extentherpes simplex virus (HSV) and GCA (1) among

Figure 5 Predicted Upstream Regulators and Pathways in GCA TAs Without or With VZV Antigen

(A) In the analysis of GCAVZV-nega-tive TAs vs controlVZV-negative TAscompared with the analysis of GCAVZV-positive TAs vs controlVZV-neg-ative TAs a Venn diagram showedshared differentially expressedgenes as well as unique differentiallyexpressed genes between the 2analyses suggesting that GCA TAswithout or with VZV antigen mayrepresent different stages of diseaseor different virus profiles that may inpart be determined by host re-sponses (B) Enrichment of uniqueGCAVZV-negative genes identified IL-4 EIF4E and TGFB1 as key upstreamregulators (left red signifies activa-tion) Upstream regulators for GCAVZV-positive unique genes were IFNγTNF and MRTFA (right red signifiesactivation and blue signifies re-pression) (C) Volcano and density plotof enriched gene sets show distribu-tion and direction of genes involved incell movement of leukocytes differ-entiation of mononuclear leukocytesreplication of Herpesviridae and viralinfection measured in GCAVZV-neg-ative unique genes (left) Volcano anddensity plot of enriched gene setsshows distribution and direction ofgenes involved in inflammatory re-sponse peripheral vascular diseasesystemic autoimmune syndrome andviral infection measured in GCAVZV-positive unique genes (right) GCA =giant cell arteritis IFN = interferon IL= interleukin TA = temporal arteryTNF = tumor necrosis factor VZVvaricella zoster virus

8 Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 6 | November 2021 NeurologyorgNN

16686345 US participants of whom 5932 had GCA com-plicated and uncomplicated herpes zoster was associated witha significantly increased risk of GCA (hazard ratios 199 and142 respectively in the Medicare cohort 216 and 145 re-spectively in the MarketScan cohort)32 antiviral treatmentwas marginally associated with a decreased GCA risk in theMedicare cohort and (2) among 3026005 British Columbiaparticipants of which 4315 had GCA the prevalence of GCAin herpes zoster participants (034) was significantly higherthan in the general population (0143)33 the prevalence ofGCA in HSV was also higher with the authors concluding thatGCA seems to increase with herpetic infections yet moresignificantly with zoster

Aside from detecting novel viral and immune pathways ourresults were consistent with previous GCA reports Specificallywe uncovered pathways associated with Th1 activation25

macrophage-mediated tissue disruption through reactive oxy-gen species34 and neutrophil involvement3536 In additionTLRs have been implicated in GCA pathogenesis26 Denget al25 found that TLR4 activation causes a transmural pan-arteritis through T-cell recruitment and activation In this studyTLR4 is a predicted upstream regulator in all GCA TAs Wealso found other predicted upstream regulators that have pre-viously been associated with GCA risk or disease severity in-cluding INFγ3738 TNF3940 IL-1β3840 IL-23738 and IL-640-42

Finally multiple studies have shown an association withGCA susceptibilityseverity and HLA-class II regions namelyHLA-DRB1 (reviewed by Carmona et al)27 we also measuredhigher levels of HLA-DRB1 transcripts in GCAVZV-positiveTAs compared with controls

Previously we found that 18 of TAs from normal partici-pants contained VZV antigen with no associated GCA pa-thology (transmural inflammation medial disruption andgiant andor epithelioid cells)10 The significance of theseobservations was unclear was the presence of VZV antigenincidental not affecting the vascular environment or becauseof nonspecific staining with the anti-VZV antibodies used Inour TempO-Seq assay we found that compared with controlVZV-negative TAs controlVZV-positive TAs had pathwaysassociated with viral infection viral replication budding ofvirus and antigen presentation and crosstalk between den-dritic cells and natural killer cells supporting the ability ofVZV to still elicit an antiviral response within arterial wallsThese participants did not have clinical symptoms signs orGCA TA pathology Yet the presence of viral pathways raisesthe possibility that these participants may represent earlyGCA that may later progress to fulminant disease or that theseindividuals may have a greater ability to clear virus infection ofarteries andor more appropriately modulate the immuneresponse preventing aberrant inflammation leading to vas-culitis Importantly the detection of viral-associated pathwaysin these TAs without the presence of robust inflammationsuggests that the enriched pathways are unlikely nonspecificsignatures of infiltrating immune cells and argues againstnonspecific VZV antigen staining

Overall gene expression and pathway analysis indicated strongevidence for a viral contribution to the pathogenesis of biopsy-confirmed GCA Although this assay did not specifically identifythe virus among all published viral causes of GCA VZV is themost biologically plausible given the compelling clin-icohistopathological overlap between VZV vasculopathy andGCA The unbiased finding of viral contributions and host an-tiviral responses support the immunohistochemical observationsof the presence of VZV antigen inGCATAs (confirmed inmanycases by the presence of VZV DNA) and more specificallysupport the ability of virus to induce vascular inflammation that ischaracteristic of GCA pathogenesis These findings have thepotential to change the standard of care for patients with GCACurrently patients with GCA are treated with long-term corti-costeroids to reduce the vascular inflammation but not the un-derlying cause of inflammation however steroids have significantside effects particularly in the vulnerable elderly population suchas potentiation of virus infection accelerated osteoporosis hy-pertension hyperglycemia and irritability Furthermore ap-proximately 50 of patients on corticosteroids continue toworsen clinically (vision loss stroke) or have recurrent diseasewhen steroids are tapered43 The addition of an antiviral agentthat would treat the underlying cause of vascular inflammationmay improve patient outcomes thus warranting multicenterclinical trials to determine antiviral drug dose and duration as wellas concomitant administration with corticosteroids

AcknowledgmentThe authors would like to thank Cathy Allen for manuscriptpreparation and the BioFrontiers Institute Next-Gen SequencingCore Facility which performed the Illumina sequencingCoauthorDr Cohrs passed away prior tomanuscript publication

Study FundingThis work was supported by Grant AG032958 from the NIHto MA Nagel RJ Cohrs and R Mahalingam

DisclosureThe authors report no disclosures Go to NeurologyorgNNfor full disclosures

Publication HistoryReceived by Neurology Neuroimmunology amp NeuroinflammationApril 20 2021 Accepted in final form June 28 2021

Appendix Authors

Name Location Contribution

Andrew NBubak PhD

University ofColoradoAurora

Designed and conceptualized thestudy performed experimentsanalyzed the data discussed andinterpreted the findings and draftedand revised the article for intellectualcontent

TeresaMescher BS

University ofColoradoAurora

Performed experiments and revisedthe article for intellectual content

Continued

NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 6 | November 2021 9

References1 Gilden D White T Khmeleva N et al Prevalence and distribution of VZV in tem-

poral arteries of patients with giant cell arteritis Neurology 201584(19)1948-19552 Al-Abdulla NA Rismondo V Minkowski JS Miller NR Herpes zoster vasculitis

presenting as giant cell arteritis with bilateral internuclear ophthalmoplegia Am JOphthalmol 2002134(6)912-914

3 Al-Abdulla NA Kelley JS Green WR Miller NR Herpes zoster vasculitis presentingas giant cell arteritis with choroidal infarction Retina 200323(4)567-569

4 NagelMA Khmeleva N Boyer PJ Choe A Bert R GildenD Varicella zoster virus in thetemporal artery of a patient with giant cell arteritis J Neurol Sci 2013335(1-2)228-230

5 NagelMA Bennett JL Khmeleva N et al Multifocal VZV vasculopathy with temporalartery infection mimics giant cell arteritis Neurology 201380(22)2017-2021

6 Victor DI Green WR Temporal artery biopsy in herpes zoster ophthalmicus withdelayed arteritis Am J Ophthalmol 197682(4)628-630

7 Nagel MA Traktinskiy I Stenmark KR Frid MG Choe A Gilden D Varicella-zoster virusvasculopathy immune characteristics of virus-infected arteriesNeurology 201380(1)62-68

8 Nagel MA Lenggenhager D White T et al Disseminated VZV infection andasymptomatic VZV vasculopathy after steroid abuse J Clin Virol 20156672-75

9 Gilden D White T Boyer PJ et al Varicella zoster virus infection in granulomatousarteritis of the aorta J Infect Dis 2016213(12)1866-1871

10 Gilden D White T Khmeleva N Boyer PJ Nagel MA VZV in biopsy-positive and-negative giant cell arteritis analysis of 100+ temporal arteries Neurol NeuroimmunolNeuroinflamm 20163(2)e216 doi 101212NXI0000000000000216

11 Muratore F Croci S Tamagnini I et al No detection of varicella-zoster virus in temporalarteries of patients with giant cell arteritis Semin Arthritis Rheum 201747(2)235-240

12 Buckingham EM Foley MA Grose C et al Identification of herpes zoster-associatedtemporal arteritis among cases of giant cell arteritis Am J Ophthalmol 201818751-60

13 Solomon IH Docken WP Padera RF Jr Investigating the association of giant cellarteritis with varicella zoster virus in temporal artery biopsies or ascending aorticresections J Rheumatol 201946(12)1614-1618

14 Sammel AM Smith S Nguyen K et al Assessment for varicella zoster virus in patientsnewly suspected of having giant cell arteritis Rheumatology (Oxford) 202059(8)1992-1996

15 Verdijk RM Ouwendijk WJD Kruipers RWAM Verjans GMGM No evidence ofvaricella-zoster virus infection in temporal artery biopsies of anterior ischemic opticneuropathy patients with and without giant cell arteritis J Infect Dis 2021223(1)109-112

16 Trejo CL Babic M Imler E et al Extraction-free whole transcriptome gene expres-sion analysis of FFPE sections and histology-directed subareas of tissue PLoS One201914(2)e0212031

17 Langmead B Trapnell C Pop M Salzberg SL Ultrafast and memory-efficientalignment of short DNA sequences to the human genome Genome Biol 200910(3)R25

18 Varet H Brillet-Gueguen L Coppee JY Dillies MA SARTools a DESeq2-andEdgeRbased R pipeline for comprehensive differential analysis of RNA-Seq data PLoSOne 201611(6)e0157022

19 Robinson MD McCarthy DJ Smyth GK edgeR a Bioconductor package for dif-ferential expression analysis of digital gene expression data Bioinformatics 201026(1)139-140

20 LoveMI Huber W Anders S Moderated estimation of fold change and dispersion forRNA-Seq data with DESeq2 Genome Biol 201415(12)550

21 Yu G Wang L Han Y He Q clusterProfiler an R package for comparing biologicalthemes among gene clusters OMICS 201216(5)284-287

22 Yu G Wang LG Yan GR He QY DOSE an RBioconductor package for diseaseontology semantic and enrichment analysis Bioinformatics 201531(4)608-609

23 R Core Team R A Language and Environment for Statistical Computing R Foundationfor Statistical Computing 2020 Available at R-projectorg

24 Brass AL Huang I-C Benita Y et al The IFITM proteins mediate cellular resistanceto influenza A H1N1 virus West Nile virus and dengue virus Cell 2009139(7)1243-1254

25 Deng J Ma-Krupa W Gewirtz AT Younge BR Goronzy JJ Weyand CM Toll-likereceptors 4 and 5 induce distinct types of vasculitis Circ Res 2009104(4)488-495

26 OrsquoNeill L Molloy ES The role of toll like receptors in giant cell arteritis Rheumatology(Oxford) 201655(11)1921-1931

27 Carmona FD Gonzalez-Gay MA Martın J Genetic component of giant cell arteritisRheumatology (Oxford) 201453(1)6-18

28 Minami T Kuwahara K Nakagawa Y et al Reciprocal expression of MRTF-A andmyocardin is crucial for pathological vascular remodelling in mice EMBO J 201231(23)4428-4440

29 Bushel PR Paules RS Auerbach SS A comparison of the TempO-Seq S1500+platform to RNA-Seq and microarray using rat liver mode of action samples FrontGenet 20189485

30 Batai K Imler E Pangilinan J et al Whole-transcriptome sequencing identified geneexpression signatures associated with aggressive clear cell renal cell carcinoma GenesCancer 20189(5-6)247-256

31 Blackmon AM Como CN Bubak AN Mescher T Jones D Nagel MA Varicellazoster virus alters expression of cell adhesion proteins in human perineurial cells viainterleukin 6 J Infect Dis 2019220(9)1453-1461

32 England BR Mikuls TR Xie F Yang S Chen L Curtis JR Herpes zoster as a riskfactor for incident giant cell arteritis Arthritis Rheum 201769(12)2351-2358

33 Lee DH Iovieno A Sheldon CA Is there an association between herpetic infectionsand giant cell arteritis A population-based study J Clin Med 202110(1)63

34 Rittner HL Kaiser M Brack A Szweda LI Goronzy JJ Weyand CM Tissue-destructive macrophages in giant cell arteritis Circ Res 199984(9)1050-1080

35 Nadkarni S Dalli J Hollywood J Mason JC Dasgupta B Perretti M Investigationalanalysis reveals a potential role for neutrophils in giant-cell arteritis disease pro-gression Circ Res 2014114(2)242-248

36 Wang L Ai Z Khoyratty T et al ROS-producing immature neutrophils in giant cellarteritis are linked to vascular pathologies CI Insight 20205(20)e139163

37 Weyand CM Hicok KC Hunder GG Goronzy JJ Tissue cytokine patterns in pa-tients with polymyalgia rheumatic and giant cell arteritis Ann Intern Med 1994121(7)484-491

38 Weyand CM Tetzlaff N Bjornsson J Brack A Younge B Goronzy JJ Diseasepatterns and tissue cytokine profiles in giant cell arteritis Arthritis Rheum 199740(1)19-26

39 Mattey DL Hajeer AH Dababneh A et al Association of giant cell arteritis andpolymyalgia rheumatica with different tumor necrosis factor microsatellite polymor-phisms Arthritis Rheum 200043(8)1749-1755

40 Hernandez-Rodriguez J Segarra M Vilardell C et al Tissue production of pro-inflammatory cytokines (IL-1β TNFα and IL-6) correlates with the intensity of thesystemic inflammatory response and with corticosteroid requirements in giant-cellarteritis Rheumatology (Oxford) 200443(3)294-301

41 Dasgupta B Panayi GS Interleukin-6 in serum of patients with polymyalgia rheu-matica and giant cell arteritis Br J Rheumatol 199029(6)456-458

42 Roche NE Fulbright JW Wagner AD Hunder GG Goronzy JJ Weyand CM Cor-relation of interleukin‐6 production and disease activity in polymyalgia rheumaticaand giant cell arteritis Arthritis Rheum 199336(9)1286-1294

43 Fraser JA Weyand CM Newman NJ Biousse V The treatment of giant cell arteritisRev Neurol Dis 20085(3)140-152

Appendix (continued)

Name Location Contribution

MichaelMariani PhD

University ofVermontBurlington

Performed bioinformatics analysesand revised the article for intellectualcontent

Seth E FrietzePhD

University ofVermontBurlington

Analyzed the data and interpreted thefindings and revised the article forintellectual content

James EHassell Jr PhD

University ofColoradoAurora

Performed bioinformatics analysesand revised the article for intellectualcontent

Christy SNiemeyer PhD

University ofColoradoAurora

Performed experiments and revisedthe article for intellectual content

Christina NComo BA

University ofColoradoAurora

Performed experiments and revisedthe article for intellectual content

Anna MBurnet BS

University ofColoradoAurora

Performed experiments and revisedthe article for intellectual content

Prem SSubramanianMD PhD

University ofColoradoAurora

Analyzed the data and interpreted thefindings and revised the article forintellectual content

Randall JCohrs PhD

University ofColoradoAurora

Analyzed the data and interpreted thefindings and revised the article forintellectual content

RaviMahalingamPhD

University ofColoradoAurora

Analyzed the data and interpreted thefindings and revised the article forintellectual content

Maria A NagelMD

University ofColoradoAurora

Designed and supervised the studyanalyzed the data discussed andinterpreted the findings and revisedthe article for intellectual content

10 Neurology Neuroimmunology amp Neuroinflammation | Volume 8 Number 6 | November 2021 NeurologyorgNN

DOI 101212NXI000000000000107820218 Neurol Neuroimmunol Neuroinflamm

Andrew N Bubak Teresa Mescher Michael Mariani et al From Giant Cell Arteritis Cases Reveals Viral Signatures

Targeted RNA Sequencing of Formalin-Fixed Paraffin-Embedded Temporal Arteries

This information is current as of September 7 2021

Academy of Neurology All rights reserved Online ISSN 2332-7812Copyright copy 2021 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

ServicesUpdated Information amp

httpnnneurologyorgcontent86e1078fullhtmlincluding high resolution figures can be found at

References httpnnneurologyorgcontent86e1078fullhtmlref-list-1

This article cites 42 articles 6 of which you can access for free at

Subspecialty Collections

httpnnneurologyorgcgicollectionviral_infectionsViral infections

httpnnneurologyorgcgicollectionvasculitisVasculitis

httpnnneurologyorgcgicollectiongene_expression_studiesGene expression studies

httpnnneurologyorgcgicollectionall_immunologyAll Immunology

okehttpnnneurologyorgcgicollectionall_cerebrovascular_disease_strAll Cerebrovascular diseaseStrokefollowing 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 2021 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

ServicesUpdated Information amp

httpnnneurologyorgcontent86e1078fullhtmlincluding high resolution figures can be found at

References httpnnneurologyorgcontent86e1078fullhtmlref-list-1

This article cites 42 articles 6 of which you can access for free at

Subspecialty Collections

httpnnneurologyorgcgicollectionviral_infectionsViral infections

httpnnneurologyorgcgicollectionvasculitisVasculitis

httpnnneurologyorgcgicollectiongene_expression_studiesGene expression studies

httpnnneurologyorgcgicollectionall_immunologyAll Immunology

okehttpnnneurologyorgcgicollectionall_cerebrovascular_disease_strAll Cerebrovascular diseaseStrokefollowing 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 2021 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