You are here

Article Text

Article menu
Basic and translational research
Extended report
DNA methylation analysis of the temporal artery microenvironment in giant cell arteritis
  1. Patrick Coit1,
  2. Lindsey B De Lott2,
  3. Bin Nan3,
  4. Victor M Elner2,4,
  5. Amr H Sawalha1,5
  1. 1Division of Rheumatology, Department of Internal Medicine, University of Michigan, Ann Arbor, Michigan, USA
  2. 2Department of Ophthalmology and Visual Sciences, University of Michigan, Ann Arbor, Michigan, USA
  3. 3Department of Biostatistics, School of Public Health, University of Michigan, Ann Arbor, Michigan, USA
  4. 4Department of Pathology, University of Michigan, Ann Arbor, Michigan, USA
  5. 5Center for Computational Medicine and Bioinformatics, University of Michigan, Ann Arbor, Michigan, USA
  1. Correspondence to Dr Amr H Sawalha, Division of Rheumatology, Department of Internal Medicine, University of Michigan, 5520 MSRB-1, SPC 5680, 1150 W. Medical Center Drive, Ann Arbor, MI 48109, USA; asawalha{at}umich.edu

Abstract

Objective To investigate the inflammatory response in giant cell arteritis (GCA) by characterising the DNA methylation pattern within the temporal artery microenvironment.

Methods Twelve patients with non-equivocal histological evidence for GCA and 12 age-matched, sex-matched and ethnicity-matched controls with normal biopsies were studied. DNA was extracted from the affected portions of temporal artery tissue in patients with GCA and from histologically confirmed normal arteries in controls. Genome-wide DNA methylation status was evaluated using the Illumina Infinium HumanMethylation450 BeadChip Array. Differentially methylated loci between affected and unaffected arterial tissues were identified, and subsequent bioinformatic analysis performed. Immunohistochemistry was used to examine tissue expression patterns in temporal artery biopsies.

Results We identified 1555 hypomethylated CG sites (853 genes) in affected temporal artery tissue from patients with GCA compared with normal controls. Gene ontology enrichment analysis of hypomethylated genes revealed significant representation in T cell activation and differentiation pathways, including both TH1 and TH17 signatures. Our DNA methylation data suggest a role for increased activity of the calcineurin/nuclear factor of activated T cells (NFAT) signalling pathway in GCA, confirmed by immunohistochemistry showing increased expression and nuclear localisation of NFAT1. NFAT signalling downstream targets such as interleukin (IL)-21/IL-21R and CD40L were overexpressed in GCA-affected arteries. Further, proinflammatory genes including TNF, LTA, LTB, CCR7, RUNX3, CD6, CD40LG, IL2, IL6, NLRP1, IL1B, IL18, IL21, IL23R and IFNG were hypomethylated in the cellular milieu of GCA arteries.

Conclusions We characterised the inflammatory response in GCA-affected arteries using ‘epigenetic immunophenotyping’ and identified molecules and pathways relevant to disease pathogenesis in GCA.

  • Systemic vasculitis
  • Giant Cell Arteritis
  • Cardiovascular Disease

https://doi.org/10.1136/annrheumdis-2014-207116

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Introduction

    Giant cell arteritis (GCA) is a systemic vasculitis of the large and medium arteries characterised by granulomatous involvement. The disease can present with ischaemic complications, including vision loss, jaw claudication, scalp tenderness and necrosis.1 GCA is found almost exclusively in people over 50 years of age with a peak incidence from 70 to 79 years and is more prevalent in women than men (3:1).2 GCA has the highest incidence among people of European descent, particularly Northern Europeans (17 cases per 100 000 people ≥50 years of age), and the lowest incidence among people of African descent (0.4 cases per 100 000 people ≥50 years of age). As life expectancy continues to improve, it is projected that by the year 2050, the number of patients with GCA will approach three million in North America and Europe alone.1

    The immunological microenvironment of the inflamed temporal artery in patients with GCA is complex. Recent studies have shown the presence of indigenous populations of dendritic cells (DC) positioned at the media-adventitia junction of major arteries that function as pathogen detectors and directors of T cell stimulation.3 The presence of mature CCR7+ DCs, which are normally seen in lymphoid organs, within the inflammatory infiltrate in affected arteries of patients with GCA suggests that these antigen-presenting cells are a potent trigger of the T cell response and key to the formation of granulomatous lesions.4 TH1 and TH17 cells are two arms of the adaptive immune response that are highly represented in the temporal artery of patients with GCA.5 Their response to glucocorticoid therapy differs. Interleukin 17 (IL17)-producing TH17 cells show a distinct reduction while the interferon-γ (IFN-γ)-producing TH1 cells are less affected.6

    DNA methylation is a component of the epigenetic and transcriptional regulatory system in the cell. It refers to the addition of a methyl group to the 5th carbon of cytosines in CG dinucleotides producing 5-methylcytosine residues. These methylation marks are relatively stable and are maintained from cell to cell by DNA methyltransferases, specifically DNA methyltransferase 1 (DNMT1). DNMT1 recognises hemi-methylated CG dinucleotides in newly synthesised genomic DNA and catalyses the addition of methyl groups to unmethylated cytosines.7 Proteins containing methyl-CpG-binding domains, such as MeCP2, recognise and bind to symmetrically methylated DNA and form complexes with histone deacetylase proteins, resulting in de-acetylation of histone tails and thus the formation of condensed heterochromatin.8 This tightly bound form of chromatin blocks the access of transcriptional machinery and effectively ‘silences’ gene transcription. In this way, the hypomethylation of CG loci and the availability of the required transcription factors are permissive to gene expression in the cellular environment.

    No previous study has explored the epigenomic microenvironment of affected arteries in GCA. Since DNA methylation patterns are relatively stable, are cell-type specific and reflect cellular transcriptional capacity, we performed DNA methylation profiling of the temporal artery microenvironment in GCA to infer and characterise the inflammatory response in this disease.

    Methods

    Patient and temporal artery tissue selection

    We retrospectively reviewed patient charts from all temporal artery biopsies performed at the University of Michigan between 1 April 2002 and 20 November 2012. Data abstracted included age at biopsy, sex, ethnicity, date of steroid initiation and temporal artery biopsy (TAB) procedure, evidence of new headache, temporal artery abnormality, erythrocyte sedimentation rate, temporal artery tenderness or decreased pulse, other clinical symptoms (eg, fever, vision loss) and histopathology. We identified 12 patients (four male, eight female) who met American College of Rheumatology classification criteria for GCA with histological evidence of florid GCA and 12 age-matched (±5 years), sex-matched and ethnicity-matched non-GCA patients presenting with similar symptoms and normal biopsies (table 1). Each tissue specimen included in this study was reviewed by a pathologist to confirm the pathologic diagnosis. All patients with GCA were treated with corticosteroids with an average time between initiation of treatment and biopsy of 1.5±0.9 days. There was no difference in the average age of patients versus controls (81.7±6.4 vs 81.1±5.5, p=0.8). The University of Michigan Institutional Review Board approved this study.

    View this table:
    Table 1

    Demographics and clinical characteristics of patients with GCA and controls included in this study

    Nucleic acid isolation and quality control

    DNA was isolated from each formalin-fixed paraffin-embedded (FFPE) TAB sample. Tissue sections under 25 mg were excised and extraneous paraffin removed using a sterile scalpel and placed in a nuclease-free 1.5 mL tube. Tissue was ground using a sterile, nuclease-free polystyrene pestle. Tissue was deparaffinised using two washes of 1 mL of xylene followed by two washes with 1 mL of 100% ethanol to remove residual xylene. Residual ethanol was allowed to evaporate before processing tissue with AllPrep DNA/RNA FFPE kit (Qiagen, Limburg). The Infinium HD FFPE QC Assay (Illumina, USA) was used to assess the quality of FFPE-extracted DNA. This qPCR-based assay compares isolated DNA from FFPE tissues with a QC standard provided by the manufacturer. Briefly, each sample was quantified using the Qubit Fluorometer (Life Technologies, USA) and 2 ng of DNA was mixed with SYBR Select Master Mix (Life Technologies, USA). Each reaction was performed in triplicate and included a no-template control and DNA standard on the ViiA7 Real-time PCR System (Life Technologies, USA) using manufacturer's recommended thermocycler settings. A sample was determined suitable for use on the Infinium HumanMethylation450 array if the average quantification cycle (Cq) deviated less than five cycles from the QC standard. All samples used in this study had ΔCq values of <5.

    DNA methylation

    250 ng of DNA from each sample was bisulfite converted using the EZ DNA Methylation Kit (Zymo, USA).9 The bisulfite-converted DNA was hybridised to the Infinium HumanMethylation450 BeadChips (Illumina, USA) using the Infinium HD FFPE Recovery Kit (Illumina, USA) and modified protocol to be compatible with bisulfite-converted FFPE DNA as recommended by Illumina. The Infinium HumanMethylation450 array interrogates over 485 000 methylation sites across the human genome, covering 96% of CpG islands, shores and flanking regions as well as 99% of RefSeq genes with an average of 17 CG sites per gene region.

    Statistical analysis

    Methylation data analysis was conducted using the Illumina GenomeStudio platform (V.2011.1) and Methylation module software (V.1.9.0) (Illumina, USA) as previously described.9 ,10 Briefly, channel intensities were normalised using internal control probes included on each BeadChip. Likewise, background signals were subtracted from each channel using internal negative control beads. Per cent methylation (β) for each probe was calculated as β=Methylated/(Methylated+Unmethylated+100). Differential DNA methylation was calculated by Illumina Custom method with Benjamini Hochberg False Discovery Rate (FDR) of 5% to correct for multiple testing. Probes with a single nucleotide polymorphism within 10 base pairs of the 3′ end of the probe sequence and probes with a detection p value (detection above array background) of >0.05 were removed from the analysis. Remaining probes with differential DNA methylation between GCA-positive and GCA-negative samples (FDR corrected p value of <0.01) were identified.

    The variation in the estimate of β is a function of β and was estimated by Illumina for all values of β by repeatedly measuring loci with known methylation fractions ranging from 0 to 1 and then fitting a parabola to the SD as a function of β (GenomeStudio Methylation Module User Guide, Illumina, USA). The SD estimate was then calculated as s=Aβ2+Bβ+C, where A=−0.1511, B=0.1444 and C=0.01646.

    P values for differential DNA methylation were calculated using the formula: Embedded Image where z is the two-sided tail probability of the standard normal distribution.

    Gene ontology

    Gene ontologies (GO) associated with significantly differentially methylated CG sites were found using the Database for Annotation, Visualization and Integrated Discovery (DAVID V.6.7)11 ,12 using a minimum gene group membership of 2, a modified Fisher exact p value (EASE Score) maximum of 0.1 and the human genome as background. The GO Biological Process FAT Ontology and Kyoto Encyclopedia of Genes and Genomes (KEGG) Pathway databases were used to find GO terms and pathways associated with genes containing differentially methylated CG sites.

    Immunohistochemistry

    Three GCA-positive and three GCA-negative FFPE temporal artery tissue samples were cut in 5 μm sections and rehydrated in water. Heat-induced epitope retrieval was performed for all samples using FLEX TRS Low pH Retrieval Buffer (pH 6.10) (Dako, USA) for 20 min. After peroxidase blocking, manual staining was performed for CD40L (goat polyclonal) (1:200) (LS-B7421, LifeSpan Biosciences, USA), IL-21 (rabbit polyclonal) (1:100) (AHP1845, AbD Serotec, USA) and IL-21R (goat polyclonal) (1:25) (AHP1819, AbD Serotec, USA). Staining for CD40L was performed at room temperature for 2 h. IL-21 and IL-21R were stained overnight at 4°C. NFAT1 (mouse monoclonal; 25A10.D6.D2) (1:200) (ab2722, AbCam, USA) staining was performed using the Dako Autostainer Link (Dako, USA) at room temperature for 20 min. Detection was performed for CD40L and IL-21R using the Biocare Goat-on-Rodent HRP Polymer system (Biocare, USA). Detection of IL-21 was performed using the Dako FLEX+ Rabbit EnVision System (Dako, USA) and for NFAT1 using the Dako FLEX+ Mouse EnVision System (Dako, USA). 3,3'-Diaminobenzidine chromagen was then applied to each slide for 10 min. All slides were then counterstained using Harris haematoxylin for 5 s, dehydrated and cover slipped. Staining for each antibody was also performed in lymph node tissue and adipose tissue, as a positive and negative control, respectively.

    Results

    Differential DNA methylation of CG sites between GCA and control temporal artery biopsies

    Of the 485 577 probes included on the Illumina HumanMethylation 450 BeadChip array, 484 984±167 probes had a detection p value of <0.05 in our samples and were included in the analysis. To reduce the amount of noise introduced by the complex environment being sampled, we initially applied an average methylation difference (Δβ) threshold of ≥|0.20|. Applying this threshold, we found 1555 (36.1%) CG sites representing 853 genes to be hypomethylated and 2752 (63.9%) CG sites representing 1471 genes to be hypermethylated when comparing GCA tissue with controls (see online supplementary figure S1 and supplementary table S1). The most hypomethylated locus was 1500 bp upstream of the transcription start site in the gene encoding Runt-related transcription factor 3, RUNX3 (Δβ=−0.43; p=3.06×10−36) (table 2) and the most hypermethylated CG locus was located in the body of the gene SLFN12L (Δβ=0.39; p=1.64×10−36).

    View this table:
    Table 2

    Top 20 hypomethylated CG sites in giant cell arteritis temporal arteries

    Functional annotation and pathway analysis of hypomethylated genes reveals and characterises the T cell signature in GCA

    Pathway and functional annotation analysis was performed using the separated hypomethylated and hypermethylated gene groups as input for DAVID Functional Annotation Tool, using the human genome as a background. The most significant GO term found among hypomethylated loci was ‘immune response’ (n=105 genes; p=3.78×10−26). A clear enrichment for T cell-related GO terms was seen among the top 20 most significant terms, including ontologies for ‘positive regulation of T cell activation’ (n=22 genes; p=1.63×10−9), ‘T cell differentiation’ (n=17 genes; p=2.10×10−6) and ‘regulation of T cell activation’ (n=29 genes; p=5.88×10−11) (see online supplementary table S2). In addition, DAVID KEGG Pathway13 ,14 analysis revealed the most significant pathway associated with hypomethylated loci to be ‘T cell receptor signalling’ (n=29 genes; p=3.73×10−10) (figure 1 and see online supplementary table S3). ‘Natural killer cell mediated cytotoxicity’ was the following most significant pathway (n=28 genes; p=1.89×10−7). The significance of T cell receptor signalling in GCA is made apparent from the hypomethylation of several genes that encode proteins in the T cell receptor complex including CD3E, CD3G, CD3D, and CD3Z, costimulatory protein gene CD28 and the downstream kinase gene ZAP70. Among the most hypomethylated sites in GCA are the genes, NFATC1 (Δβ=−0.35; p=3.06×10−36), which encodes a transcription factor for several T cell-associated cytokines, and PPP3CC, a calcineurin-encoding gene, which is an upstream activator of nuclear factor of activated T cells (NFAT) signalling proteins. NFAT is a master transcription factor of several important T cell cytokine genes, including TNF and IFNG which are both significantly hypomethylated in the TABs of patients with GCA.

    Figure 1
    Figure 1

    T cell receptor signalling KEGG Pathway (hsa04660) depicting hypomethylated (Δβ≤−0.20) genes in GCA-affected temporal artery biopsies compared with normal controls (red dots).

    CaN, calcineurin; GCA, giant cell arteritis; IL, interleukin; IFN-γ, interferon-γ; KEGG, Kyoto Encyclopedia of Genes and Genomes; MAPK, Map-kinase; MHC, major histocompatibility complex; NFAT, nuclear factor of activated T cells; NF-κB, nuclear factor-κB; TNF α, tumour necrosis factor α.

    The transcriptionally repressive environment detailed by the GO and KEGG pathway analysis of hypermethylated genes did not present any apparent source of pathology in GCA TABs (see online supplementary table S4). The most significant GO term found among hypermethylated genes was ‘actin cytoskeleton organisation’ (n=44 genes; p=3.86×10−5) and the most significant KEGG pathway was ‘dilated cardiomyopathy’ (n=26 genes; p=2.78×10−6).

    Cytokine activity signature in GCA

    To further characterise the inflammatory response in GCA, we explored the hypomethylation of cytokine genes by comparing the ‘cytokine activity’ (GO:0005125) molecular function GO and array probes with a methylation difference of ≤−0.10 between GCA-positive and GCA-negative TABs (see online supplementary table S5). Genes containing hypomethylated sites associated with cytokine activity in GCA samples included the proinflammatory cytokine genes TNF, IL2, IL1B, IL18 and the TH1 key cytokine gene IFNG. LTA and LTB, both members of the tumour necrosis factor (TNF) family, and CD40LG were also demethylated. A clear TH17 signature was also present in our data set, and is represented by demethylation of IL6, IL21, IL23R and IL17RA.

    Immunohistochemistry

    To determine if the DNA methylome profile we obtained in GCA reflects the inflammatory response in GCA-affected arteries, we examined the protein expression levels of four key molecules identified in our study using immunohistochemistry in GCA-positive and GCA-negative TABs. Our data showed clear expression of NFAT1, CD40L, IL-21 and IL-21R in GCA-positive TABs (figures 2 and 3). Furthermore, there was evidence for nuclear localisation of NFAT1, suggesting dephosphorylation and activation of NFAT1 in GCA-affected tissue (figure 2B).

    Figure 2
    Figure 2

    Immunohistochemistry staining of NFAT1 in GCA temporal artery tissue at 100× magnification (A) and 400× magnification (B) from two representative patients. There is evidence of increased NFAT1 expression and nuclear localisation, indicating NFAT1 dephosphorylation and activation in GCA (magnified view in the left lower corner of Panel B).

    GCA, giant cell arteritis; NFAT, nuclear factor of activated T cells.

    Figure 3
    Figure 3

    Immunohistochemistry staining of CD40L, interleukin (IL)-21 and IL-21R in giant cell arteritis temporal artery tissue at 100× magnification (A) and 400× magnification (B) from two representative patients.

    Discussion

    We conducted the first DNA methylome analysis of TABs from 12 patients with GCA and 12 controls and highlighted the central role of T cells in the pathology of GCA. Indeed, the most hypomethylated gene in GCA was RUNX3, which encodes a transcription factor that plays an important role in T cell maturation and activation.15 Our approach characterised the inflammatory response in GCA, confirmed a central role for T cells in the disease and identified key novel molecules and pathways that might play a role in the pathogenesis of this large vessel vasculitis.

    We have shown hypomethylation of several genetic risk loci associated with GCA, such as IFNG, TNF, NLRP1 and PTPN22,16 ,17 suggesting a possible role for genetic–epigenetic interaction in this disease. TNF encodes TNF-α, which is primarily expressed by activated macrophages, and localises in the media and intima of inflamed temporal artery in patients with GCA.18 TNF-α was found to be highly expressed in GCA19 and has been the target of several clinical trials using TNF-α blocking compounds though none have so far shown an ability to reduce the duration of corticosteroid treatment in patients with GCA.20–22

    A consistently hypomethylated gene that highlights the presence of mature DCs in the TAB microenvironment in GCA is CCR7. The CCR7 gene encodes C-C chemokine receptor type 7 (CCR7), which is a marker of mature DCs and has previously been found to be expressed in DCs located at the adventitia-media border of TABs from patients with GCA.4 Previous studies have elucidated the role that vascular DCs play in antigen sensing and presentation as well as T cell activation and inflammatory response modulation in large blood vessels, the vessel-specific pattern of toll-like receptor expression and subsequent recognition of specific antigen motifs.23 In fact, it has been proposed that the loss of antigenic tolerance of DCs and their subsequent activation is an important event in the early stages of GCA pathology.24 We also found hypomethylation of CCL18 (Δβ=−0.27; p=3.06×10−36) at a locus in the gene body. CCL18 is a chemokine specific to mature DCs and is known to attract naïve T cells for activation.25

    DCs play a role in antigen presentation, T cell recruitment, activation and differentiation. As seen in figure 1, several genes that encode proteins that are initiators of T cell activation and differentiation are hypomethylated in the TABs of patients with GCA. These included the TCR genes CD3E, CD3G, CD3D, and CD3Z, costimulatory protein gene CD28 and the downstream kinase gene ZAP70. The TCR complex and CD28 work synergistically to amplify the activation signal received by the T cell through their interaction with ZAP-70.26 These findings support the importance of TCR/CD28 stimulation of T cells in the temporal artery microenvironment in GCA.

    An interesting finding in our study of GCA temporal arteries was the significant hypomethylation of genes in the calcineurin (CaN)/NFAT pathway, specifically the genes PPP3CC, NFATC1 and NFATC2. The CaN/NFAT pathway begins with the activation of the calmodulin protein by Ca2+ transported into the cell and released from cellular calcium stores into the cytoplasm downstream of TCR engagement. Calmodulin, an upstream activator of CaN, binds these Ca2+ stores and activates CaN, which docks with a phosphorylated NFAT protein. CaN binding is accomplished by its recognition of a highly conserved motif PXIXIT (X being any amino acid) on NFAT. The CaN dephosphorylates serine residues present on NFAT and exposes its nuclear localisation signal, allowing NFAT to translocate into the nucleus to promote gene expression.27 Indeed, our immunohistochemistry studies showed increased NFAT1 expression in the affected portions of GCA artery tissue and clear evidence for nuclear localisation of NFAT1, indicating dephosphorylation and NFAT1 activation (figure 2B).

    The expression of cytokines in the inflamed temporal artery is important in GCA. We explored the hypomethylation of cytokine genes included in the ‘cytokine activity’ Gene Ontology (GO:0005125). We identified hypomethylation in key cytokine genes in GCA-positive TABs, including TNF, IL6 and IL1B. Among patients with GCA with greater clinical inflammatory signs, IL6 and IL1B were previously found to have increased transcription in temporal artery tissue compared with patients with a less severe inflammatory response. Further, TNF expression in the GCA artery tissue had a significant correlation with the time required for patients to reach a maintenance dose of prednisone.28 LTA and LTB, both members of the TNF family, were demethylated as well. Other demethylated genes that characterise the GCA inflammatory microenvironment in our study include IL18, IL1B, IL2, IL6, IL21, IL2RA, IL17RA and IL23R. IL18 encodes IL-18, a member of the IL-1 cytokine family, and is an important inducer of IFN-γ in TH1 cells as well as the production of TNF and IL-1.29 IL2 encodes the NFAT-regulated cytokine IL-2 that signals naïve CD4+ T cells to enter cell cycle and begin the process of transforming into mature TH1 and TH2 cells.27 IL-21 is pleiotropic cytokine that is vital for the development of TH17 cells and the production of IL-17.30

    The gene encoding CD40 ligand (CD40LG) is another hypomethylated NFAT-regulated T cell gene in GCA temporal arteries. The CD40/CD40 ligand (CD40L or CD154) interaction is an important step in the activation and development of stable TH cell lineages and their effector functions. CD40L allows mature T cells to interact with other arms of the immune system including CD40-expressing macrophages.31 The regulation of IL-21 and CD40L expression in the effector T cell response by CaN/NFAT signalling highlights the importance of this signalling pathway in GCA. Immunohistochemistry staining of GCA-affected arteries shows protein expression of NFAT1, CD40 L, IL-21 and IL-21R in our study, although confirmation at the mRNA level is warranted to validate these results.

    It remains unknown if methylation differences between patients and controls are present in circulating T cells or other peripheral blood mononuclear cell subsets. Evaluating the DNA methylome of specific cell subsets would be of value in demonstrating whether DNA methylation changes in specific cell subsets predispose to increased susceptibility to GCA and if age-related DNA methylation changes can explain increased frequency of the disease in older individuals. Further studies will be required to determine if differential DNA methylation in key genes involved in T cell lineages and the development of GCA precede the inflammatory response in the temporal artery. It should be noted that while DNA methylation profiles indicate chromatin accessibility in hypomethylated genetic loci, transcriptional activity of these identified genes was not directly examined in this study.

    In summary, using DNA methylation profiling in temporal artery tissue, we have characterised the inflammatory response in GCA-affected arteries. Our data confirmed a central role for T cells in GCA and uncovered key novel molecules and pathways in the pathogenesis of this disease. We provided evidence for the involvement of the calcineurin/NFAT pathway in GCA and suggest that specific inhibitors of this pathway or key downstream molecules, such as IL-21/IL-21R and CD40L, might be therapeutic considerations for GCA in the future.

    References

    1. ↵
      1. De Smit E,
      2. Palmer AJ,
      3. Hewitt AW
      . Projected worldwide disease burden from giant cell arteritis by 2050. J Rheumatol 2015;42:119–25. doi:10.3899/jrheum.140318
      OpenUrlAbstract/FREE Full Text
    2. ↵
      1. Gonzalez-Gay MA,
      2. Vazquez-Rodriguez TR,
      3. Lopez-Diaz MJ, et al
      . Epidemiology of giant cell arteritis and polymyalgia rheumatica. Arthritis Rheum 2009;61:1454–61. doi:10.1002/art.24459
      OpenUrlCrossRefPubMedWeb of Science
    3. ↵
      1. Weyand CM,
      2. Ma-Krupa W,
      3. Pryshchep O, et al
      . Vascular dendritic cells in giant cell arteritis. Ann N Y Acad Sci 2005;1062:195–208. doi:10.1196/annals.1358.023
      OpenUrlCrossRefPubMedWeb of Science
    4. ↵
      1. Krupa WM,
      2. Dewan M,
      3. Jeon MS, et al
      . Trapping of misdirected dendritic cells in the granulomatous lesions of giant cell arteritis. Am J Pathol 2002;161:1815–23. doi:10.1016/S0002-9440(10)64458-6
      OpenUrlCrossRefPubMedWeb of Science
    5. ↵
      1. Terrier B,
      2. Geri G,
      3. Chaara W, et al
      . Interleukin-21 modulates Th1 and Th17 responses in giant cell arteritis. Arthritis Rheum 2012;64:2001–11. doi:10.1002/art.34327
      OpenUrlCrossRefPubMedWeb of Science
    6. ↵
      1. Deng J,
      2. Younge BR,
      3. Olshen RA, et al
      . Th17 and Th1 T-cell responses in giant cell arteritis. Circulation 2010;121:906–15. doi:10.1161/CIRCULATIONAHA.109.872903
      OpenUrlAbstract/FREE Full Text
    7. ↵
      1. Hermann A,
      2. Goyal R,
      3. Jeltsch A
      . The Dnmt1 DNA-(cytosine-C5)-methyltransferase methylates DNA processively with high preference for hemimethylated target sites. J Biol Chem 2004;279:48350–9. doi:10.1074/jbc.M403427200
      OpenUrlAbstract/FREE Full Text
    8. ↵
      1. Fuks F,
      2. Burgers WA,
      3. Brehm A, et al
      . DNA methyltransferase Dnmt1 associates with histone deacetylase activity. Nat Genet 2000;24:88–91. doi:10.1038/71750
      OpenUrlCrossRefPubMedWeb of Science
    9. ↵
      1. Coit P,
      2. Jeffries M,
      3. Altorok N, et al
      . Genome-wide DNA methylation study suggests epigenetic accessibility and transcriptional poising of interferon-regulated genes in naive CD4+ T cells from lupus patients. J Autoimmun 2013;43:78–84. doi:10.1016/j.jaut.2013.04.003
      OpenUrlCrossRefPubMedWeb of Science
    10. ↵
      1. Coit P,
      2. Yalavarthi S,
      3. Ognenovski M, et al
      . Epigenome profiling reveals significant DNA demethylation of interferon signature genes in lupus neutrophils. J Autoimmun 2015;58:59–66. doi:10.1016/j.jaut.2015.01.004
      OpenUrlCrossRefPubMed
    11. ↵
      1. Huang da W,
      2. Sherman BT,
      3. Lempicki RA
      . Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nat Protoc 2009;4:44–57. doi:10.1038/nprot.2008.211
      OpenUrlCrossRefPubMedWeb of Science
    12. ↵
      1. Huang da W,
      2. Sherman BT,
      3. Lempicki RA
      . Bioinformatics enrichment tools: paths toward the comprehensive functional analysis of large gene lists. Nucleic Acids Res 2009;37:1–13. doi:10.1093/nar/gkn923
      OpenUrlAbstract/FREE Full Text
    13. ↵
      1. Kanehisa M,
      2. Goto S
      . KEGG: kyoto encyclopedia of genes and genomes. Nucleic Acids Res 2000;28:27–30. doi:10.1093/nar/28.1.27
      OpenUrlAbstract/FREE Full Text
    14. ↵
      1. Kanehisa M,
      2. Goto S,
      3. Sato Y, et al
      . Data, information, knowledge and principle: back to metabolism in KEGG. Nucleic Acids Res 2014;42:D199–205. doi:10.1093/nar/gkt1076
      OpenUrlAbstract/FREE Full Text
    15. ↵
      1. Djuretic IM,
      2. Levanon D,
      3. Negreanu V, et al
      . Transcription factors T-bet and Runx3 cooperate to activate Ifng and silence Il4 in T helper type 1 cells. Nat Immunol 2007;8:145–53. doi:10.1038/ni1424
      OpenUrlCrossRefPubMedWeb of Science
    16. ↵
      1. Carmona FD,
      2. Gonzalez-Gay MA,
      3. Martin J
      . Genetic component of giant cell arteritis. Rheumatology (Oxford) 2014;53:6–18. doi:10.1093/rheumatology/ket231
      OpenUrlAbstract/FREE Full Text
    17. ↵
      1. Ly KH,
      2. Regent A,
      3. Tamby MC, et al
      . Pathogenesis of giant cell arteritis: More than just an inflammatory condition? Autoimmun Rev 2010;9:635–45. doi:10.1016/j.autrev.2010.05.002
      OpenUrlCrossRefPubMed
    18. ↵
      1. Field M,
      2. Cook A,
      3. Gallagher G
      . Immuno-localisation of tumour necrosis factor and its receptors in temporal arteritis. Rheumatol Int 1997;17:113–18. doi:10.1007/s002960050019
      OpenUrlCrossRefPubMedWeb of Science
    19. ↵
      1. Visvanathan S,
      2. Rahman MU,
      3. Hoffman GS, et al
      . Tissue and serum markers of inflammation during the follow-up of patients with giant-cell arteritis—a prospective longitudinal study. Rheumatology (Oxford) 2011;50:2061–70. doi:10.1093/rheumatology/ker163
      OpenUrlAbstract/FREE Full Text
    20. ↵
      1. Hoffman GS,
      2. Cid MC,
      3. Rendt-Zagar KE, et al
      . Infliximab for maintenance of glucocorticosteroid-induced remission of giant cell arteritis: a randomized trial. Ann Intern Med 2007;146:621–30. doi:10.7326/0003-4819-146-9-200705010-00004
      OpenUrlCrossRefPubMedWeb of Science
    21. ↵
      1. Martinez-Taboada VM,
      2. Rodriguez-Valverde V,
      3. Carreno L, et al
      . A double-blind placebo controlled trial of etanercept in patients with giant cell arteritis and corticosteroid side effects. Ann Rheum Dis 2008;67:625–30. doi:10.1136/ard.2007.082115
      OpenUrlAbstract/FREE Full Text
    22. ↵
      1. Seror R,
      2. Baron G,
      3. Hachulla E, et al
      . Adalimumab for steroid sparing in patients with giant-cell arteritis: results of a multicentre randomised controlled trial. Ann Rheum Dis 2014;73:2074–81. doi:10.1136/annrheumdis-2013-203586
      OpenUrlAbstract/FREE Full Text
    23. ↵
      1. Pryshchep O,
      2. Ma-Krupa W,
      3. Younge BR, et al
      . Vessel-specific Toll-like receptor profiles in human medium and large arteries. Circulation 2008;118:1276–84. doi:10.1161/CIRCULATIONAHA.108.789172
      OpenUrlAbstract/FREE Full Text
    24. ↵
      1. Ma-Krupa W,
      2. Jeon MS,
      3. Spoerl S, et al
      . Activation of arterial wall dendritic cells and breakdown of self-tolerance in giant cell arteritis. J Exp Med 2004;199:173–83. doi:10.1084/jem.20030850
      OpenUrlAbstract/FREE Full Text
    25. ↵
      1. Adema GJ,
      2. Hartgers F,
      3. Verstraten R, et al
      . A dendritic-cell-derived C-C chemokine that preferentially attracts naive T cells. Nature 1997;387:713–17. doi:10.1038/42716
      OpenUrlCrossRefPubMed
    26. ↵
      1. Salojin KV,
      2. Zhang J,
      3. Delovitch TL
      . TCR and CD28 are coupled via ZAP-70 to the activation of the Vav/Rac-1-/PAK-1/p38 MAPK signaling pathway. J Immunol 1999;163:844–53.
      OpenUrlAbstract/FREE Full Text
    27. ↵
      1. Hermann-Kleiter N,
      2. Baier G
      . NFAT pulls the strings during CD4+ T helper cell effector functions. Blood 2010;115:2989–97. doi:10.1182/blood-2009-10-233585
      OpenUrlAbstract/FREE Full Text
    28. ↵
      1. Hernandez-Rodriguez J,
      2. Segarra M,
      3. Vilardell C, et al
      . Tissue production of pro-inflammatory cytokines (IL-1beta, TNFalpha and IL-6) correlates with the intensity of the systemic inflammatory response and with corticosteroid requirements in giant-cell arteritis. Rheumatology (Oxford) 2004;43:294–301. doi:10.1093/rheumatology/keh058
      OpenUrlAbstract/FREE Full Text
    29. ↵
      1. Dinarello CA
      . IL-18: A TH1-inducing, proinflammatory cytokine and new member of the IL-1 family. J Allergy Clin Immunol 1999;103:11–24. doi:10.1016/S0091-6749(99)70518-X
      OpenUrlCrossRefPubMedWeb of Science
    30. ↵
      1. Wei L,
      2. Laurence A,
      3. Elias KM, et al
      . IL-21 is produced by Th17 cells and drives IL-17 production in a STAT3-dependent manner. J Biol Chem 2007;282:34605–10. doi:10.1074/jbc.M705100200
      OpenUrlAbstract/FREE Full Text
    31. ↵
      1. Grewal IS,
      2. Flavell RA
      . The role of CD40 ligand in costimulation and T-cell activation. Immunol Rev 1996;153:85–106. doi:10.1111/j.1600-065X.1996.tb00921.x
      OpenUrlCrossRefPubMedWeb of Science

    Supplementary materials

    Footnotes

    • Handling editor Tore K Kvien

    • Contributors All authors made substantial contributions to the conception or design of the work; acquisition, analysis or interpretation of data for the work; drafting the work or revising it critically for important intellectual content; and gave final approval of the version to be published; and agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

    • Competing interests None declared.

    • Ethics approval The study was approved by the Institutional Review Board at the University of Michigan.

    • Provenance and peer review Not commissioned; externally peer reviewed.