Giant cell arteritis (GCA) is a systemic vasculitis with 2

Transcription

1 Th17 and Th1 T-Cell Responses in Giant Cell Arteritis Jiusheng Deng, PhD; Brian R. Younge, MD; Richard A. Olshen, PhD; Jörg J. Goronzy, MD, PhD; Cornelia M. Weyand, MD, PhD Downloaded from by guest on June 19, 2018 Background In giant cell arteritis (GCA), vasculitic damage of the aorta and its branches is combined with a syndrome of intense systemic inflammation. Therapeutically, glucocorticoids remain the gold standard because they promptly and effectively suppress acute manifestations; however, they fail to eradicate vessel wall infiltrates. The effects of glucocorticoids on the systemic and vascular components of GCA are not understood. Methods and Results The immunoprofile of untreated and glucocorticoid-treated GCA was examined in peripheral blood and temporal artery biopsies with protein quantification assays, flow cytometry, quantitative real-time polymerase chain reaction, and immunohistochemistry. Plasma interferon- and interleukin (IL)-17 and frequencies of interferon- producing and IL-17 producing T cells were markedly elevated before therapy. Glucocorticoid treatment suppressed the Th17 but not the Th1 arm in the blood and the vascular lesions. Analysis of monocytes/macrophages in the circulation and in temporal arteries revealed glucocorticoid-mediated suppression of Th17-promoting cytokines (IL-1, IL-6, and IL-23) but sparing of Th1-promoting cytokines (IL-12). In human artery severe combined immunodeficiency mouse chimeras, in which patient-derived T cells cause inflammation of engrafted human temporal arteries, glucocorticoids were similarly selective in inhibiting Th17 cells and leaving Th1 cells unaffected. Conclusions Two pathogenic pathways mediated by Th17 and Th1 cells contribute to the systemic and vascular manifestations of GCA. IL-17 producing Th17 cells are sensitive to glucocorticoid-mediated suppression, but interferon- producing Th1 responses persist in treated patients. Targeting steroid-resistant Th1 responses will be necessary to resolve chronic smoldering vasculitis. Monitoring Th17 and Th1 frequencies can aid in assessing disease activity in GCA. (Circulation. 2010;121: ) Key Words: glucocorticoids inflammation interferons interleukin-17 Th1 cells vasculitis Giant cell arteritis (GCA) is a systemic vasculitis with 2 disease components: vessel wall inflammation inducing arterial stenosis/occlusion and a systemic inflammation leading to polymyalgias, anemia, failure to thrive, and malaise. 1 Prototypical vessel wall inflammation preferentially affects the upper extremity and extracranial branches of the aorta, causing blindness, stroke, aortic arch syndrome, aortic aneurysm, or dissection. 1,2 Clinical Perspective on p 915 Glucocorticoids remain the gold standard of therapy; attempts to introduce steroid-sparing agents, including antitumor necrosis factor blockers, have not been successful. 3,4 Methotrexate may have minor benefits when given over prolonged periods. 5 In a double-blind, placebo-controlled study employing glucocorticoid pulse therapy, patients pulsed at diagnosis had fewer disease flares and discontinued therapy earlier than patients without initial pulse therapy. However, benefits of pulse glucocorticoids were delayed until 12 to 18 months into treatment. 6 Acetylsalicylic acid has a role as adjuvant therapy, partially by inhibiting inflammation and partially through antiplatelet effects. 7,8 Steroid withdrawal after 2 years of therapy is often accompanied by recurrence of inflammatory markers, and vascular inflammation persists despite glucocorticoid treatment In essence, despite their prompt, impressive therapeutic effects, glucocorticoids fail to abrogate vasculitis. Improved disease management and avoidance of severe glucocorticoid side effects will require a better understanding of underlying immune abnormalities. Both the innate and adaptive immune system contribute to GCA pathogenesis. 12 Granulomatous inflammatory infiltrates composed of CD4 T cells, activated macrophages, and multinucleated giant cells induce intimal hyperplasia and luminal compromise. 13 In cell-depletion studies, dendritic cells (DCs) have emerged as the predominant antigen-presenting cell. 14,15 DCs are now recognized as an indigenous cell Received April 15, 2009; accepted December 24, From the Lowance Center for Human Immunology and Rheumatology, Emory University, Atlanta, Ga (J.D., J.J.G., C.M.W.); Department of Ophthalmology, Mayo Clinic, Rochester, Minn (B.R.Y.); and Divisions of Biostatistics (R.A.O.) and Immunology and Rheumatology (J.D., J.J.G., C.M.W.), Stanford University, Stanford, Calif. The online-only Data Supplement is available with this article at /DC1. Reprint requests to Cornelia M. Weyand, MD, PhD, Division of Immunology and Rheumatology, Stanford University School of Medicine, 269 W Campus Dr, Stanford, CA cweyand@stanford.edu 2010 American Heart Association, Inc. Circulation is available at DOI: /CIRCULATIONAHA

2 Deng et al Immunomodulatory Effects of Glucocorticoids in GCA 907 Downloaded from by guest on June 19, 2018 population in human macrovessels, where they have important surveillance functions and sense pathogen-derived motifs. 16,17 The nature of T cells driving GCA and the manner in which they are affected by immunosuppression are unresolved. The finding of identical T cells in the right and left temporal artery has strongly supported antigen as an instigator. 18 Functional selection and mechanisms of action of wall-infiltrating T cells remain to be elucidated. Interferon (IFN)- is a dominant cytokine in the arteries, whereas Th2-derived cytokines are consistently absent. 19,20 Whether interleukin (IL)-17 producing Th17 cells have a role in GCA is unknown. GCA is a treatable yet incurable disease. In the present study, we have made use of the therapeutic effects of glucocorticoids to decipher which immunopathways maintain systemic and vascular inflammation. By studying patients before and after initiation of glucocorticoids, immune abnormalities responsive to this immunosuppressive treatment could be dissected from those resistant to glucocorticoids. Glucocorticoids inhibit the production of cytokines that promote and sustain CD4 T-cell differentiation into Th17 cells. Conversely, cytokines regulating the differentiation of Th1 cells and Th1 responses persist even in chronically treated patients. Differential susceptibility of Th1 and Th17 cells to steroid-mediated suppression provides important pathogenic clues and encourages the development of therapeutic strategies that can disrupt both arms of vasculitic T-cell responses. Methods Study Cohorts Demographic and clinical characteristics of patients and controls enrolled into this study are listed in Table I in the online-only Data Supplement. Eleven patients donated samples for studies before therapy was initiated and were reexamined on steroid therapy. In addition, 15 patients were tested only before therapy initiation, and 23 patients were recruited while they were already on therapy. Thirty-four healthy volunteers aged 61 to 80 years served as controls. The age distribution of patients and controls was not significantly different (Table I in the online-only Data Supplement). Prior or current diagnosis of cancer, autoimmune disease other than GCA, or chronic infection was an exclusion criterion. Emory University and the Mayo Clinic review boards approved the protocol, and all donors gave informed consent. Glucocorticoid Treatment Patients with a positive temporal artery biopsy started on prednisone 60 mg daily. Patients were monitored monthly for clinical signs of disease, and glucocorticoid was tapered by 10% every 2 weeks. Adjustment of glucocorticoid therapy was left to the decision of the treating physician and was based on clinical and laboratory findings. Reagents and Antibodies Reagents and antibodies are listed in Table II in the online-only Data Supplement. Cell Preparation, Culture, and Flow Cytometry Peripheral blood mononuclear cells (PBMCs) were isolated from healthy donors and GCA patients. After 4-hour stimulation with phorbol 12-myristate 13-acetate (PMA) and ionomycin in the presence of brefeldin A, cells were stained with anti-cd3 (PE), anti-cd4 (PerCP), anti-ifn- (FITC), and anti-il-17 (APC) antibodies, or, alternatively, they were stained with anti-cd3 (FITC) and anti- Foxp3 (PE). Flow cytometry was performed on an LSR II (BD Biosciences, San Jose, Calif), and data were analyzed with FlowJo software. Naive CD4 T cells and CD14 monocytes were purified as described 16 and cocultured at a 2:1 ratio for 4 days with lipopolysaccharide, anti-cd3, and anti-il-1, anti-il-6, or anti-il-23p19 neutralizing antibodies. Relevant antibody isotype served as control. For enrichment of Th17 cells, CD45RO memory CD4 T cells were isolated from PBMCs and cocultured with CD14 monocytes (2:1) for 7 days with lipopolysaccharide (1 g/ml), anti-cd3 (1 g/ml), anti-human IFN- (2 g/ml), and anti-human IL-4 (2 g/ml) antibodies. Accumulation of Th17 and Th1 cells in such T-cell lines was monitored by flow cytometry. Quantitative Real-Time Polymerase Chain Reaction Total RNA was isolated from human tissues and cells. IL-17, IFN-, Foxp3, IL-1, IL-6, IL-23p19, IL-12p35, IL-12p40, matrix metalloproteinase-9, and -actin gene transcripts were quantified by real-time polymerase chain reaction (PCR) and adjusted to copies of the housekeeping gene -actin as described. 10,16,17 Primer sequences and PCR conditions are listed in Table III in the online-only Data Supplement. ELISArray Plasma cytokine levels (IL-1, IL-6, IL-12, IL-17, IFN- ) were analyzed with Human Autoimmune Response Multi-Analyte ELISArray kits (SABiosciences Corp, Frederick, Md). Plasma samples (1:2 dilutions) were incubated in 96-well plates precoated with individual cytokine capture antibodies for 2 hours. After they were washed, biotin-conjugated cytokine detection antibodies were added for 30 minutes. Bound antibodies were detected with avidin-hrp and development solution. Color reactions were quantified at OD 450. Immunohistochemical Staining Sections of paraffin-embedded arteries were processed and stained with anti-cd3, anti-cd4, anti-cd8, anti-cd14, anti-cd15, anti-il- 17, or anti-ifn- antibodies as described. 16,21 24 Purified mouse or rabbit IgG control antibodies were used instead of the primary antibody as a control. Human Temporal Artery Severe Combined Immunodeficiency Mouse Chimeras Human temporal artery severe combined immunodeficiency (SCID) mouse chimeras were generated as described previously Mice implanted with artery tissues from the same donor were assigned to either of 2 treatment arms (dexamethasone or vehicle) and a control arm (noninflamed artery). At day 7 after implantation, mice were injected intraperitoneally with lipopolysaccharide (3 g per mouse) or PBS (control arm). On day 8, T-cell lines enriched for Th17 cells (10 7 cells per mouse) were adoptively transferred into the chimeras intravenously. On day 9, chimeras were treated with dexamethasone (4 mg/kg; Sigma-Aldrich, St Louis, Mo) as described. 11 Fifteen days after implantation, arterial grafts were harvested for RNA extraction or OCT embedded for immunostaining. The protocol was approved by the Emory University Institutional Animal Care and Use Committee. Data Analysis A detailed description of the statistical tools and study cohorts is given in Statistical Methods and Table IV in the online-only Data Supplement. Analysis of data in Figure 1 required the development of a customized approach because some patients were tested only before treatment, some were tested only while on treatment, and some patients were tested twice, before initiation of glucocorticoid and while on glucocorticoid. Analysis of data sets in Figures 2, 3, 4, and 5 involved 1-sample t or Wilcoxon test, 2-sample t or Wilcoxon test, or Behrens-Fisher-Welch testing, as appropriate. Friedman rank sum test was applied to compare results from antibody blocking

3 908 Circulation February 23, 2010 Downloaded from by guest on June 19, 2018 Figure 1. Circulating Th1 and Th17 cells are markedly expanded in GCA and differentially suppressed by glucocorticoids. PBMCs were isolated from untreated patients (No Tx; n 26) and glucocorticoid-treated patients (Tx; n 34) or age-matched healthy controls (Con; n 34). Cells were stimulated with PMA/ionomycin in the presence of brefeldin A for 4 hours and stained with PE anti-cd3, PerCP anti- CD4, FITC anti-ifn-, and APC anti-il-17 antibodies. Alternatively, PBMCs were stained with FITC anti-cd3, PerCP anti-cd4, and PE anti-foxp3 antibodies and analyzed by flow cytometry. Frequencies of IL-17 Th17 (A) and IFN- Th1 cells (B) among CD4 T cells and Foxp3 CD4 Treg (C) among CD3 T cells are presented. Representative cytometric dot plots of Th17 and Th1 cells from a healthy individual (D), a GCA patient before therapy (E), and a GCA patient on treatment (F) are shown. Plasma concentrations of IL-17 (G) and IFN- (H) protein were measured in 6 GCA patients before and after initiation of glucocorticoid therapy and 6 age-matched healthy controls with ELISArray kits. Results are presented as absorbance value at 450 nm. assays with anti-il-1, anti-il-6, and anti-il-23. P 0.05 was considered significant. Data are presented as mean SEM. The authors had full access to and take full responsibility for the integrity of the data. All authors have read and agreed to the manuscript as written. Results Circulating Th1 and Th17 Cells Are Markedly Expanded in GCA and Are Differentially Affected by Glucocorticoid Therapy To identify functionally relevant T-cell subsets in GCA, frequencies of IL-17 producing Th17 and IFN- producing Th1 and Foxp3 regulatory CD4 T cells were compared in patients with untreated GCA (n 26) and age-matched healthy controls (n 34). In GCA patients, frequencies of circulating Th17 cells among CD4 T cells were expanded 8-fold, ranging from 1.1% to 5.3% and averaging 2.2% (Figure 1A). In healthy controls, only 0.03% to 0.59% of T cells produced IL-17 (Figure 1A). Th1 cells accounted for 20.6% of CD4 T cells in GCA patients in contrast to only 11.8% in controls (Figure 1B). The patients also carried a population of IL-17 IFN- double-positive cells (Figure 1E), which were essentially lacking in controls (Figure 1D). Frequencies of CD4 Foxp3 Treg were indistinguishable in patients and controls (Figure 1C and Figure I in the onlineonly Data Supplement). In essence, untreated GCA was associated with a marked expansion of IFN- and IL-17 producing T-cell subsets. To explore the effect of glucocorticoid therapy on cytokine-producing T-cell subsets, frequencies of circulating Th1

4 Deng et al Immunomodulatory Effects of Glucocorticoids in GCA 909 Downloaded from by guest on June 19, 2018 Figure 2. Steroid sensitivity of tissue IL-17 and steroid resistance of tissue IFN- in temporal artery lesions. Total RNA was isolated from temporal arteries harvested before initiation of glucocorticoid therapy (No Tx) (n 8) and from a second-side biopsy 3 to 9 months after initiation of therapy (Tx). Arteries without GCA served as controls (Con) (n 8). Transcripts of IL-17 (A), IFN- (B), and Foxp3 (C) in the tissues were quantified by real-time PCR. Paraffin-embedded tissue sections of temporal arteries were stained with anti-cd3 (E) or anti-il-17 (F through H) antibodies. D, IL-17 cells in arteries harvested before (G) and on treatment (H) were quantified in 10 randomly selected high-powered fields (HPF). I, IFN- staining showed IFN- cells in the inner adventitia of a temporal artery on treatment. The images presented are representative of 6 biopsy samples. J, Purified rabbit IgG served as isotype control antibody. Int indicates intima; Med, media; and Adv, adventitia. Magnification 200 in E, F, and J; 400 in H and I; 600 in G.

5 910 Circulation February 23, 2010 Downloaded from by guest on June 19, 2018 and Th17 cells were measured in steroid-treated patients (n 34) who were on an average dose of 16.7 mg of prednisone daily. All patients initially received prednisone at 60 mg/d with continuous dose reduction by 10% every 2 weeks, unless the patients had recurrence of vasculitic symptoms. Glucocorticoid therapy markedly suppressed circulating Th17 cells, normalizing the average frequencies to 0.46% of CD4 T cells (Figure 1A). In treated patients, Th17 frequencies were indistinguishable from those in control donors (range, 0.11% to 0.89%) (P 0.101). Glucocorticoid also ablated the IL-17 IFN- double-positive cells (Figure 1F). However, glucocorticoid essentially left unaffected frequencies of Th1 cells (Figure 1B and 1F), with a mean of 20.2% and 18.4% before and after therapy, respectively. In addition, frequencies of Foxp3 Treg were unaltered by glucocorticoid therapy (Figure 1C and Figure I in the onlineonly Data Supplement). In essence, in steroid-treated patients, Th17 frequencies returned to normal, whereas the expansion of Th1 cells persisted. Measurement of circulating IL-17 and IFN- protein confirmed elevation of both inflammatory cytokines before glucocorticoid therapy and effective suppression of IL-17 but not IFN- in treated patients (Figure 1G and 1H). To assess the time frame of glucocorticoid-mediated suppression of Th17 cells and understand whether Th17 cells rebound once steroid doses were tapered, 11 patients were examined before the initiation of prednisone therapy and Figure 3. Glucocorticoids suppress Th17- promoting cytokines and spare Th1-promoting cytokines in the circulation. The proinflammatory cytokines IL-1 (A), IL-6 (B), and IL-12 (C) were measured in plasma samples of GCA patients before (No Tx) and after initiation of glucocorticoid therapy (Tx) (n 6) and agematched controls (Con) (n 6) by ELISArrays. Data are presented as absorbance values at 450 nm. D to H, CD14 monocytes were purified from PBMCs of GCA patients before steroid therapy (n 8) and after steroid therapy (n 8) or GCA-negative donors (n 8). Transcripts of IL-1 (D), IL-6 (E), IL-23p19 (F), IL-12p40 (G), and IL-12p35 (H) in cell extracts were quantified by quantitative real-time PCR. reexamined at different time points while on steroids (Table; Table V in the online-only Data Supplement). Treatment for 10 weeks was sufficient to bring Th17 frequencies into the normal range. With decreasing steroid doses, the suppressive effect was maintained and was still observed after treatment for 9 months. There was no evidence that tapering steroids below 10 mg/d resulted in a rebound of Th17 cells. Glucocorticoid Therapy Selectively Suppresses IL-17 Producing Cells and Spares IFN- Secreting Cells in Vasculitic Lesions Temporal artery samples with typical changes of GCA were collected from 8 untreated patients, and the second-side biopsy was obtained after 3 to 9 months of glucocorticoid therapy. All second-side biopsies examined in this study were read positive for vasculitis by histomorphology. Temporal artery specimens free of GCA served as controls. Extracts from the tissue samples were analyzed for the expression of IL-17, IFN-, and Foxp3-specific sequences by real-time PCR, and IL-17 and IFN- producing cells were identified by immunohistochemical staining of tissue sections. The control samples were essentially negative for IL-17, IFN-, and Foxp3 transcripts, and no cells staining positive for IL-17 or IFN- were detected. The samples from untreated patients expressed high transcript levels for IL-17, IFN-, and Foxp3 (Figure 2A to 2C). Glucocorticoid therapy resulted in marked suppression of IL-17 production (Figure 2A). In contrast,

6 Deng et al Immunomodulatory Effects of Glucocorticoids in GCA 911 Downloaded from by guest on June 19, 2018 Figure 4. Effects of glucocorticoids on the expression of Th17- and Th1-promoting cytokines in inflamed temporal arteries. RNA was extracted from temporal artery biopsies collected as described in Figure 2 before steroid therapy (first side) (No Tx) and after steroid therapy (second side) (Tx). Arteries without GCA served as controls (Con). Transcripts of IL-1 (A), IL-6 (B), IL-23p19 (C), IL-12p40 (D), and IL-12p35 (E) were quantified by quantitative real-time PCR. IFN- and Foxp3 expression was indistinguishable between untreated and treated arteries (Figure 2B and 2C and Figure IC in the online-only Data Supplement). In the temporal arteries of untreated GCA patients, immunostaining for CD3 localized tissue-infiltrating T cells in all wall layers, with preferential clustering in the medial smooth muscle layer (Figure 2E). IL-17 producing cells were abundant (Figure 2F) and followed the distribution pattern of CD3 T cells (Figure 2F and 2G). IL-17 cells were also found around adventitial vasa vasorum and diffusely spread through the hyperplastic intima. Mean numbers of IL-17 producing cells reached 36 per high-powered field (Figure 2D). Almost all lymphocytic cells in the vessel wall infiltrates are CD4, with only very few CD8 T cells represented, assigning IL-17 production mostly to CD4 T cells (Figure II in the online-only Data Supplement). IFN- producing cells were represented among all wall-infiltrating T cells but displayed a preference for the adventitial layer. In arteries harvested from treated patients, the infiltrates continued to include abundant IFN- producing cells (Figure 2I), mainly in the inner adventitia. However, only few cells stained positive for IL-17, and the intensity of staining was reduced (Figure 2H), suggesting reduced production by individual cells. Glucocorticoids Inhibit Th17-Promoting Cytokines in Peripheral Monocytes and in Vasculitic Lesions T-cell differentiation into either Th1 or Th17 cells is directed by the cytokine environment that is shaped by antigen-presenting cells, such as DCs and monocytes/macrophages. IL- 1, IL-6, and IL-23 contribute to Th17 differentiation; IL-12 supports the development of Th1 cells. 25 To quantify the contribution of different cytokines to the expansion of Th17 cells in GCA, naive CD4 T cells were cocultured with autologous CD14 monocytes in the presence of neutralizing antibodies against IL-1, IL-6, or IL-23. Monocytes from GCA patients effectively drove Th17 and Th1 differentiation (Figure III in the online-only Data Supplement). Th17 differentiation was suppressed by anti-il-1, anti-il-6, and anti- IL-23, implicating all 3 cytokines in the process. Blocking IL-1 or IL-6 had no effect on the development of Th1 cells. Removing IL-23 activity, however, enhanced the outgrowth of Th1 cells markedly. The data assigned a critical role to IL-1, IL-6, and IL-23 in promoting the shift toward Th17 cells in GCA. To understand whether glucocorticoid functions by targeting Th1- and Th17-promoting cytokines, we analyzed protein levels in the plasma and gene expression in circulating monocytes and temporal artery tissue samples. Protein concentrations of circulating IL-1, IL-6, and IL-12 were elevated significantly in untreated patients (Figure 3A to 3C). In response to therapy, IL-1 and IL-6 were markedly reduced (Figure 3A and 3B), but IL-12 remained unaffected (Figure 3C). In circulating monocytes, transcript levels for IL-1, IL-6, IL-12p35, IL-12p40, and IL-23p19 were all increased in GCA patients compared with healthy individuals (Figure 3D to 3H). On glucocorticoid therapy, IL-1 and IL-6 gene expression levels in GCA monocytes were completely inhib-

7 912 Circulation February 23, 2010 Downloaded from by guest on June 19, 2018 Figure 5. Dexamethasone prevents Th17- but not Th1-mediated vessel wall inflammation in human artery SCID chimeras. Normal human arteries were engrafted into SCID mice. One week later, mice were injected with lipopolysaccharide or saline as control (Con). Twenty-four hours later, T-cell lines established from the blood of GCA patients were adoptively transferred (10 7 cells per mouse). Chimeras were treated with either dexamethasone (4 mg/kg) (Tx) or saline (No Tx), and temporal arteries were explanted. T-cell lines from the GCA patients had similar proportions of Th17, Th1, and Th17/Th1 cells; a representative cytometric dot plot is shown (A). Tissue extracts from explanted arteries were analyzed for IL-17 (B) and IFN- (C) by quantitative real-time PCR. Data are presented as mean SEM from 3 independent experiments. Tissue sections from explanted arteries were stained with anti-il-17 antibodies (D to F). IL-17 T cells (dark brown) were quantified in a minimum of 12 high-powered fields (HPF) for each arterial cross section (G). Transcript levels of IL-6 (H) and matrix metalloproteinase (MMP)-9 (I) in the implants were measured by quantitative real-time PCR as well. Magnification 600 in D, E, and F. Table. Effect of Glucocorticoid Treatment on IL-17 Producing T Cells Th17 Cells One-Sample t Test P Patients Median, % Range, % Mean, % SD Untreated (n 11) Treated (n 11) ited (Figure 3D and 3E). IL-23p19 gene expression was suppressed by 52% in CD14 monocytes from treated patients (Figure 3F). The treatment did not affect IL-12 gene expression (Figure 3G and 3H). A similar pattern emerged for the tissue expression of Th17- and Th1-promoting cytokines. In untreated cases, IL-1, IL-6, IL-23p19, IL-12p40, and IL-12p35 gene expression levels were markedly upregulated (Figure 4A to 4E). Glucocorticoid treatment resulted in suppression of IL-1, IL-6, and IL-23p19 transcription (Figure 4A to 4C), but production of IL-12p35 and IL-12p40 transcripts was maintained in the arteries (Figure 4D and 4E). Glucocorticoid Selectively Inhibits Recruitment and Survival of Th17 Cells in Vessel Wall Inflammation in Human Artery SCID Mouse Chimeras To verify the differential effects of corticosteroid treatment on the accumulation of T-cell effectors in vasculitic lesions, T-cell lines generated from GCA patients were adoptively transferred into artery-scid chimeras engrafted with human temporal arteries (Figure 5). Expanded T-cell lines contained a high frequency of Th17 cells as well as Th17/Th1 cells, confirming a bias toward the induction of the Th17 lineage (Figure 5A). Under similar conditions, T-cell cultures from healthy controls yielded 2% Th17 cells (data not shown). After injection of the patient-derived T-cell lines, Th17 as well as Th1 cells accumulated in the arteries, as evidenced by the levels of IL-17 and IFN- specific sequences (Figure

8 Deng et al Immunomodulatory Effects of Glucocorticoids in GCA 913 Downloaded from by guest on June 19, B and 5C). Treatment with dexamethasone (4 mg/kg) inhibited accumulation of IL-17 messenger RNA but did not affect production of IFN-. Immunohistochemical analysis confirmed dense infiltrates of IL-17 cells in the arteries from untreated chimeras (Figure 5E and 5G). Conversely, after corticosteroid therapy, the density of IL-17 cells in the vessel wall was markedly diminished (Figure 5F and 5G). Reduction in the frequency of wall-infiltrating IL-17 cells was associated with significantly lower levels of IL-6 sequences (Figure 5H). In addition, matrix metalloproteinase-9 transcript levels, a marker for tissue-destructive macrophages, declined, suggesting that preventing the infiltration/ survival of IL-17 cells in the infiltrates may have tissueprotective effects (Figure 5I). In essence, glucocorticoids selectively regulate the composition of the inflammatory T-cell infiltrate; Th17 cells are highly sensitive to the inhibitory effects, whereas the recruitment and enrichment of Th1 cells and their cytokine-producing functions are unaffected. Discussion Glucocorticoids are highly effective in treating the constitutional symptoms of GCA, including myalgias, malaise, fever, anorexia, and headaches. However, most patients require therapy for years to avoid disease flares, suggesting only partial responsiveness to this immunosuppressive approach. In the present study, we report that glucocorticoids act by selectively suppressing Th17 responses in the vasculitic lesions and in the periphery, whereas Th1 responses are spared. The anti-inflammatory actions of glucocorticoid result from modulation of monocytes/macrophages; Th17- promoting cytokines decline markedly, and the Th1- promoting IL-12 continues to be produced. In essence, glucocorticoids effectively treat the acute manifestations of GCA, which appears to be closely associated with excessive production of Th17-promoting cytokines by innate immune cells and IL-17 by T cells. Glucocorticoids fail to abrogate vasculitis because Th1-driving cytokines and Th1 cells are refractory to glucocorticoid-mediated suppression. To reset immune abnormalities in GCA, glucocorticoid must be combined with additional therapies. GCA is considered a prototypical steroid-responsive disease. Patients have fast and profound improvement of constitutional symptoms, such as malaise, anorexia, fevers, and myalgias. Steroids are believed to be vision protective, and headaches and scalp tenderness often respond promptly However, vascular lesions persist, and disease flares are frequent on tapering of glucocorticoids. The present study has made use of comparing abnormalities in the innate and adaptive immune system before and after therapy. T cells and macrophages are critical players in the vasculitic process sustaining the granulomatous inflammation in the arteries. 22,29 Highly activated macrophages not only occupy the wall infiltrates but also circulate in the blood High levels of IL-6 have been associated with disease activity and therapeutic response. 11,33 Besides IL-6, circulating IL-1 and IL-12 proteins were elevated in untreated patients (Figure 3A to 3C). The cytokine production pattern found in the vasculitic lesions is essentially identical to that detected in circulating monocytes. On the side of the adaptive immune system, GCA patients have marked expansions of Th1 and Th17 cells. Th1 and Th17 cells are now considered to drive 2 distinct inflammatory pathways Only 1 of these pathways in GCA is amendable to glucocorticoid-mediated suppression in both patients and mouse chimeras. GCA patients utilize IL-1, IL-6, and IL-23 to induce Th17 cells, and these cytokines are effectively suppressed by glucocorticoids. From these data, it appears that the patients immune defects originate in the innate immune system, where excessive amounts of IL-1, IL-6, and IL-23 are produced. Multiple cell populations, including DCs, monocytes, and macrophages, are the main sources of IL-1, IL-6, and IL-23. The present study assigns a top position to these cells in the pathogenic events leading to GCA, supporting the disease model in which vessel wall residing DCs have checkpoint function and are the first to respond to disease-inducing danger signals. 16,17 Th17 cells have been linked to a number of autoimmune diseases, 37 cardiac allograft rejection, and vasculopathy. 38 How IL-17 facilitates vascular tissue damage is less well understood, but IL-17 receptors are expressed on fibroblasts, endothelial cells, and vascular smooth muscle cells, all instrumental in the arterial wall structure (data not shown). With 8-fold enrichment of Th17 cells in the blood, IL-17 may even have a role in the typical myalgias and muscle stiffness associated with GCA. A model system in which vessel wall inflammation can be induced in normal human arteries allowed analysis of some of the mechanisms related to Th1- and Th17-mediated disease. T-cell lines from GCA patients were tested for their inflammation-inducing capacity. T-cell subset analysis revealed a strong bias toward the enrichment of Th17 and Th17/Th1 cells. In normal donors, Th17 cells account for only a minor fraction of effector T cells. Conversely, T-cell lines expanded from GCA patients fell into 3 equally proportioned subsets: Th1 cells, Th17 cells, and Th1/Th17 cells. Whether the expansion of Th17 cells is solely a consequence of disturbed innate immune function, producing excess amounts of Th17-promoting cytokines, or whether the patients also have restructuring of the T-cell repertoire needs to be addressed in further studies. Corticosteroids essentially depleted Th17 cells from the lesions either by inhibiting recruitment or by jeopardizing survival in the vessel wall environment. Recent data suggest that tissue invasion of human arteries is a feature of CCR6 T cells. 16 The CCR6 receptor is typically found on Th17 cells (data not shown). CCL20, the ligand for CCR6, has been implicated in facilitating selective recruitment of T cells, causing panarteritis as opposed to a periarteritic architecture of the vascular infiltrates. 16 After dexamethasone treatment, Th1 cells continued to infiltrate the wall, recapitulating findings in the arteries from the treated patients. The present study has multiple implications for our understanding of GCA pathogenesis and management. The data suggest that at least 2 arms of the adaptive immune system contribute to GCA. Th17 immunity seems to be more important for acute manifestations, both systemically and in the blood vessels; Th1 immunity is associated with chronically persistent vascular lesions. From a mechanistic point of

9 914 Circulation February 23, 2010 Downloaded from by guest on June 19, 2018 view, the ability to suppress IL-1, IL-6, and IL-23 while IL-12 production is essentially unaffected suggests that 2 different types of antigen-presenting cells or 2 different danger signals may promote T-cell differentiation in GCA. It is important to realize, however, that the approach taken in this study, analyzing the different blood and tissue cytokines as dependent parameters, does not entirely reflect the complex and interrelated biology of the disease, yet this is the only approach currently available. The potency of glucocorticoids to normalize proinflammatory Th17 cells is consistent with the clinical experience that this therapy is prompt and effective. In most patients, constitutional symptoms (myalgias, fever, malaise, anorexia) and signs of vascular inflammation (headaches, temporal tenderness) are markedly improved. However, vessel wall infiltrates, and thus continuous wall damage, can persist over extended periods. The resistance of circulating and tissue Th1 cells indicates partial response, even on high doses of glucocorticoids. 9,10 Current data support the concept that many of the acute disease manifestations are mediated by IL-17, but further work is necessary to separate the distinct biological and clinical consequences of Th1- and Th17- dependent disease components. A recent double-blind study exploring the therapeutic benefits of pulse glucocorticoids demonstrated that patients who received high doses of pulse steroids at the beginning of therapy had fewer disease flares and could discontinue therapy earlier. 6 Thus, inhibiting the Th17 arm of immunity may have long-lasting effects. One possible explanation is that IL-17 functions mainly as an inflammatory amplifier. Advancing from partial to more comprehensive disease control in GCA would require combining glucocorticoid therapy with Th1-targeted interventions. The most effective approach may be to inhibit Th1- promoting cytokines, in particular IL-12. The tight correlation between blood and tissue T-cell subsets opens the door for blood-based biomarkers that can be used to closely follow the extent of T-cell and monocyte abnormalities and to appropriately tailor suppression of the multiple pathways driving this vasculitis. Acknowledgments The authors thank all patients for participating in this study and Linda Arneson for editorial assistance. Sources of Funding This work was funded in part by the National Institutes of Health (PO1 HL , U19 AI 57266, R01 AR , R01 AG , ULI RR025744, R01 EY 11916) and the Vasculitis Foundation. None. Disclosures References 1. Weyand CM, Goronzy JJ. Giant-cell arteritis and polymyalgia rheumatica. Ann Intern Med. 2003;139: Evans JM, O Fallon WM, Hunder GG. Increased incidence of aortic aneurysm and dissection in giant cell (temporal) arteritis: a population-based study. Ann Intern Med. 1995;122: Hoffman GS, Cid MC, Rendt-Zagar KE, Merkel PA, Weyand CM, Stone JH, Salvarani C, Xu W, Visvanathan S, Rahman MU. Infliximab for maintenance of glucocorticosteroid-induced remission of giant cell arteritis: a randomized trial. Ann Intern Med. 2007;146: Salvarani C, Macchioni P, Manzini C, Paolazzi G, Trotta A, Manganelli P, Cimmino M, Gerli R, Catanoso MG, Boiardi L, Cantini F, Klersy C, Hunder GG. Infliximab plus prednisone or placebo plus prednisone for the initial treatment of polymyalgia rheumatica: a randomized trial. Ann Intern Med. 2007;146: Mahr AD, Jover JA, Spiera RF, Hernandez-Garcia C, Fernandez- Gutierrez B, Lavalley MP, Merkel PA. Adjunctive methotrexate for treatment of giant cell arteritis: an individual patient data meta-analysis. Arthritis Rheum. 2007;56: Mazlumzadeh M, Hunder GG, Easley KA, Calamia KT, Matteson EL, Griffing WL, Younge BR, Weyand CM, Goronzy JJ. Treatment of giant cell arteritis using induction therapy with high-dose glucocorticoids: a double-blind, placebo-controlled, randomized prospective clinical trial. Arthritis Rheum. 2006;54: Stone JH. Antiplatelet versus anticoagulant therapy in patients with giant cell arteritis: which is best? Nat Clin Pract Rheumatol. 2007;3: Weyand CM, Kaiser M, Yang H, Younge B, Goronzy JJ. Therapeutic effects of acetylsalicylic acid in giant cell arteritis. Arthritis Rheum. 2002;46: Achkar AA, Lie JT, Hunder GG, O Fallon WM, Gabriel SE. How does previous corticosteroid treatment affect the biopsy findings in giant cell (temporal) arteritis? Ann Intern Med. 1994;120: Brack A, Rittner HL, Younge BR, Kaltschmidt C, Weyand CM, Goronzy JJ. Glucocorticoid-mediated repression of cytokine gene transcription in human arteritis-scid chimeras. J Clin Invest. 1997;99: Weyand CM, Fulbright JW, Hunder GG, Evans JM, Goronzy JJ. Treatment of giant cell arteritis: interleukin-6 as a biologic marker of disease activity. Arthritis Rheum. 2000;43: Weyand CM, Goronzy JJ. Medium- and large-vessel vasculitis. N Engl J Med. 2003;349: Weyand CM, Younge BR, Goronzy JJ. T cells in arteritis and atherosclerosis. Curr Opin Lipidol. 2008;19: Krupa WM, Dewan M, Jeon MS, Kurtin PJ, Younge BR, Goronzy JJ, Weyand CM. Trapping of misdirected dendritic cells in the granulomatous lesions of giant cell arteritis. Am J Pathol. 2002;161: Ma-Krupa W, Jeon MS, Spoerl S, Tedder TF, Goronzy JJ, Weyand CM. Activation of arterial wall dendritic cells and breakdown of self-tolerance in giant cell arteritis. J Exp Med. 2004;199: Deng J, Ma-Krupa W, Gewirtz AT, Younge BR, Goronzy JJ, Weyand CM. Toll-like receptors 4 and 5 induce distinct types of vasculitis. Circ Res. 2009;104: Pryshchep O, Ma-Krupa W, Younge BR, Goronzy JJ, Weyand CM. Vessel-specific Toll-like receptor profiles in human medium and large arteries. Circulation. 2008;118: Weyand CM, Schonberger J, Oppitz U, Hunder NN, Hicok KC, Goronzy JJ. Distinct vascular lesions in giant cell arteritis share identical T cell clonotypes. J Exp Med. 1994;179: Weyand CM, Hicok KC, Hunder GG, Goronzy JJ. Tissue cytokine patterns in patients with polymyalgia rheumatica and giant cell arteritis. Ann Intern Med. 1994;121: Weyand CM, Tetzlaff N, Bjornsson J, Brack A, Younge B, Goronzy JJ. Disease patterns and tissue cytokine profiles in giant cell arteritis. Arthritis Rheum. 1997;40: Kaiser M, Younge B, Bjornsson J, Goronzy JJ, Weyand CM. Formation of new vasa vasorum in vasculitis: production of angiogenic cytokines by multinucleated giant cells. Am J Pathol. 1999;155: Wagner AD, Bjornsson J, Bartley GB, Goronzy JJ, Weyand CM. Interferon-gamma-producing T cells in giant cell vasculitis represent a minority of tissue-infiltrating cells and are located distant from the site of pathology. Am J Pathol. 1996;148: Nordborg C, Larsson K, Nordborg E. Stereological study of neovascularization in temporal arteritis. J Rheumatol. 2006;33: Rodriguez-Pla A, Bosch-Gil JA, Rossello-Urgell J, Huguet-Redecilla P, Stone JH, Vilardell-Tarres M. Metalloproteinase-2 and -9 in giant cell arteritis: involvement in vascular remodeling. Circulation. 2005;112: Chen Z, O Shea JJ. Th17 cells: a new fate for differentiating helper T cells. Immunol Res. 2008;41: Cantini F, Niccoli L, Nannini C, Bertoni M, Salvarani C. Diagnosis and treatment of giant cell arteritis. Drugs Aging. 2008;25: Langford CA. Vasculitis in the geriatric population. Rheum Dis Clin North Am. 2007;33: Seo P, Stone JH. Large-vessel vasculitis. Arthritis Rheum. 2004;51:

10 Deng et al Immunomodulatory Effects of Glucocorticoids in GCA Weyand CM, Goronzy JJ. Arterial wall injury in giant cell arteritis. Arthritis Rheum. 1999;42: Rittner HL, Hafner V, Klimiuk PA, Szweda LI, Goronzy JJ, Weyand CM. Aldose reductase functions as a detoxification system for lipid peroxidation products in vasculitis. J Clin Invest. 1999;103: Wagner AD, Goronzy JJ, Weyand CM. Functional profile of tissueinfiltrating and circulating CD68 cells in giant cell arteritis: evidence for two components of the disease. J Clin Invest. 1994;94: Weyand CM, Wagner AD, Bjornsson J, Goronzy JJ. Correlation of the topographical arrangement and the functional pattern of tissue-infiltrating macrophages in giant cell arteritis. J Clin Invest. 1996;98: Roche NE, Fulbright JW, Wagner AD, Hunder GG, Goronzy JJ, Weyand CM. Correlation of interleukin-6 production and disease activity in polymyalgia rheumatica and giant cell arteritis. Arthritis Rheum. 1993;36: Boniface K, Blom B, Liu YJ, de Waal Malefyt R. From interleukin-23 to T-helper 17 cells: human T-helper cell differentiation revisited. Immunol Rev. 2008;226: Dong C. TH17 cells in development: an updated view of their molecular identity and genetic programming. Nat Rev Immunol. 2008; 8: Luger D, Silver PB, Tang J, Cua D, Chen Z, Iwakura Y, Bowman EP, Sgambellone NM, Chan CC, Caspi RR. Either a Th17 or a Th1 effector response can drive autoimmunity: conditions of disease induction affect dominant effector category. J Exp Med. 2008;205: Kebir H, Kreymborg K, Ifergan I, Dodelet-Devillers A, Cayrol R, Bernard M, Giuliani F, Arbour N, Becher B, Prat A. Human TH17 lymphocytes promote blood-brain barrier disruption and central nervous system inflammation. Nat Med. 2007;13: Yuan X, Paez-Cortez J, Schmitt-Knosalla I, D Addio F, Mfarrej B, Donnarumma M, Habicht A, Clarkson MR, Iacomini J, Glimcher LH, Sayegh MH, Ansari MJ. A novel role of CD4 Th17 cells in mediating cardiac allograft rejection and vasculopathy. J Exp Med. 2008;205: Downloaded from by guest on June 19, 2018 CLINICAL PERSPECTIVE Giant cell arteritis (GCA), an inflammatory vasculopathy of the aorta and its branches, causes blindness, stroke, and aortic arch syndrome combined with a syndrome of intense systemic inflammation. Pathognomic findings are granulomatous vessel wall infiltrates of activated T cells and macrophages in a temporal artery biopsy. Corticosteroids are highly effective in suppressing the systemic manifestations of disease. In the present study, patients with GCA were studied before and after corticosteroid treatment to identify which immunopathways mediate acute and chronic vasculitis and the manner in which therapy affects the immune-inflammatory abnormalities underlying GCA. In temporal artery biopsies and in the blood of untreated patients, 2 distinct types of effector T cells were expanded: interleukin-17 producing Th17 cells and interferon- producing Th1 cells. Corticosteroids effectively suppressed Th17 cells but left Th1 cells unaffected. Corticosteroids functioned by inhibiting macrophages that produce Th17-promoting cytokines, including interleukin-1, interleukin-6, and interleukin-23. Th1-promoting macrophages escaped from corticosteroid suppression, chronically sustaining one arm of vasculitic T-cell responses. In essence, the systemic and vascular inflammation of GCA involves abnormally activated innate and adaptive immune responses, with both Th17 and Th1 cells contributing to disease. Corticosteroids disrupt abnormal Th17 responses but fail to induce disease remission because Th1 cells are refractory and continue to support vascular inflammation. Thus, successful management of GCA requires combination therapy targeting several arms of the immune system. The identified immune abnormalities emerge as potential biomarkers to optimize monitoring of disease activity in large-vessel vasculitis.

11 Th17 and Th1 T-Cell Responses in Giant Cell Arteritis Jiusheng Deng, Brian R. Younge, Richard A. Olshen, Jörg J. Goronzy and Cornelia M. Weyand Downloaded from by guest on June 19, 2018 Circulation. 2010;121: ; originally published online February 8, 2010; doi: /CIRCULATIONAHA Circulation is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX Copyright 2010 American Heart Association, Inc. All rights reserved. Print ISSN: Online ISSN: The online version of this article, along with updated information and services, is located on the World Wide Web at: Data Supplement (unedited) at: Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published in Circulation can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office. Once the online version of the published article for which permission is being requested is located, click Request Permissions in the middle column of the Web page under Services. Further information about this process is available in the Permissions and Rights Question and Answer document. Reprints: Information about reprints can be found online at: Subscriptions: Information about subscribing to Circulation is online at:

12 Figures 1A and 1B have identical structures but, of course, they differ in what was measured. In each case there were three comparisons. The least standard of them was comparing subjects, before and after treatment. This comparison was complicated because while 11 subjects were measured both before and after treatment, 15 subjects were measured only before; and 23 only after. We compared all subjects measured before with all subjects measured after. Thus, the specification is that all subjects before are independent and identically distributed. Likewise, all subjects measured after are assumed independent and identically distributed. However, 11 subjects in each of the two groups were measured both times. Their measurements are assumed independent of those from the other 38 subjects, but their before and after treatment measurements are not independent. We denote the (population) mean value of each before measurement by µ B and the corresponding variance as σ 2 B. Parameters for individual measurements taken after treatment are µ A and σ 2 A. A principal null hypothesis is that µ A = µ B. The respective variances are examples of nuisance parameters Sampling distributions of ordinary two-sample t-tests are predicated not only on the assumption that all data are normally distributed, but additionally on the assumption that the two variances are equal. There is the additional assumption regarding normality of both sets of distributions. For data in Figures 1A and 1B there are inferences on comparisons of (26) before and (34) after treatment measurements with (the 34) control measurements. These comparisons involve ordinary two-sample t test statistics. Always, control measurements are assumed independent of each other, and of measurements taken before or after treatment. They are assumed to have common mean value µ C and common variance σc 2. Thus, when before measurements are compared with control, the null hypothesis is that µ B = µ C. Obviously analogous statements can be made regarding a comparison of measurements taken after treatment and controls. The assumption of normality matters little to any of the three cited comparisons owing to the central limit effect (Bickel and Doksum, Section A.2.7, page 464). This is 1

13 because there are = 26 measurements before, = 34 after, and 34 control measurements. Each comparison involves a difference in sample mean values divided by an estimate of the standard deviation of that difference. Our estimate of µ B is the mean of all measurements taken before treatment, and is denoted by x B ; it is computed from all available measurements, 26 for data cited. Our estimate of σ 2 B is the estimated sample variance of these 26 measurements, write s 2 B. Estimates x A and s 2 A, respectively of µ A and σ 2 A are defined analogously from the 34 measurements taken after treatment. The correlation between before and after measurements on the same individual is denoted by ρ AB. In our application regarding data in Figures 1A and 1B, the estimate computed from data, r AB, is based only 11 subjects. Discussion thus far and simple algebra entail that Var ( x A x B ) = σ2 B 26 + σ2 A 34 2( 11 (26)(34) )ρ AB σ A σ B. It follows from Slutkys theorem (Bickel and Doksum, Section A.2.7, page 460) that when the null hypothesis µ A = µ B is true, then x A x B z = s 2 B 26 + s2 A ( (26)(34) r A,B s A s B ) has, at least approximately, a standard normal distribution. By computing the probability that a standard normal Z exceeds the observed computed z in absolute value, one has an approximate attained significance value (p-value) for a two-sided test of the null hypothesis that µ A = µ B. Readers note that regarding data in Figures 1A and 1B, we are testing three separate hypotheses regarding pair-wise equality of the parameters µ A,µ B, and µ C. We should not employ a standard one-way analysis of variance, especially but not only because the three sets of observations that comprise x A, x B and x C are not independent. However, an overall p-value for the combined tests µ A = µ B ; µ A = µ C ; µ B = µ C can be accomplished 2

14 by applying the usual Bonferroni inequality (Bickel and Doksum, pages 288 and 439), and simply adding the p-values. While testing µ A = µ B can be accomplished by using the statistic z we have derived, testing the null hypothesis that µ B, alternatively µ A equals µ C can be accomplished by a two-sample t-test. Assumptions underlying the ordinary test may fail because of inequality of variances, in which case the Behrens-Fisher test, which is analogous to z (see can be used. One might also apply the two-sample Wilcoxon statistic for testing µ B = µ C or µ A = µ C (see Neither is entirely appropriate since in both cases, the sampling distributions are not identical up to a possible shift of one of them. However, readers please note that all three tests of both null hypotheses have exceedingly small p-values. So it is immaterial which test is employed, and by any reasonably criterion we reject not only the two just cited, but even all three null hypotheses. When it comes to data other than those of Figures 1A and 1B, issues described already are both simpler and more complex. They are simpler because in testing µ A = µ B all six subjects that figure in computing x A and x B were measured both before and after treatment. Therefore, z reduces to a simple one-sample t statistic. However, because it has only five degrees of freedom, for the null central t-distribution to apply reasonably accurately, the six differences, after minus before by subject, should appear to have normal distributions. We assessed normality by inspecting so-called q-q plots ( and also by application of the Wilk- Shapiro statistic ( When distributions appear normal to the extent we can infer that, then one-sample t might be trusted. When they are not, then we prefer the one-sample Wilcoxon statistic (URL already given), though any test of its assumption of symmetry about its center is not testable from so little data. Entirely analogous comments apply to testing µ B = µ C and µ A = µ C from cited data other than those of Figures 1A and 1B. Here two-sample t has 10 degrees of freedom, 3

15 so we are less concerned about normality of data than when we test µ A = µ C. Further, when distributions appear non-normal, we have both Behrens-Fisher-Welch t and the two sample Wilcoxon statistic for testing. Testing normality is, obviously, suspect with only six observations, and we are somewhat perplexed which inferences to draw when competitive p-values give somewhat different signals. Reference Bickel, P.J. and Doksum, K.A. (1977), MATHEMATICAL STATISTICS: Basic Ideas and Selected Topics, Holden-Day, Inc., San Francisco. 4

16 Supplemental Table 1. Demographic and clinical characteristics of GCA patients and control donors Items Con Untreated GCA Treated GCA No. of subjects Sex, female/male 26/8 19/7 28/6 Age, mean+/-sd years Ethnicity, % Cauc African American Headache, % 75 Jaw claudication, % 50 Scalp tenderness, % 50 Ischemic optic neuropathy, % 17 Fever, % 25 Weight loss, % 33 Fatigue, % 75 Anemia, % 50 Polymyalgia rheumatica, % 42 ESR+/-SD CRP+/-SD Platelets+/-SD

17 Supplemental Table 2. Reagents and antibodies utilized in cell culture and staining Reagents/Antibodies Company Usage Concentration or Dilution LPS Sigma-Aldrich, St. Louis, MO Cell culture 1µg/ml Goat anti-human IL-1β R&D Systems, Minneapolis, Cell culture 2µg/ml MN Goat anti-human IL-6 R&D Systems Cell culture 4µg/ml Goat anti-human IL-23p19 R&D Systems Cell culture 6µg/ml Purified goat IgG control R&D Systems Cell culture 6µg/ml Mouse anti-human IL-4 R&D Systems Cell culture 2µg/ml Mouse anti-human IFN-γ R&D Systems Cell culture 2µg/ml Anti-human CD3 (OKT) Ortho Biotech, Raritan, NJ Cell culture 1µg/ml Anti-human CD45RA microbeads Miltenyi Biotec, Auburn, CA Purification See company instruction Anti-human CD45RO microbeads Miltenyi Biotec Purification See company instruction phorbol 12-myristate 13-acetate (PMA) Sigma-Aldrich Cell culture 50µg/ml Ionomycin Sigma-Aldrich Cell culture 1µg/ml Brefeldin A ebioscience, San Diego, CA Cell culture 3µg/ml FITC-conjugated mouse anti-human CD3 BD, San Diego, CA Cell staining 1:100 PE-conjugated mouse anti-human CD3 BD Cell staining 1:100 PerCP-conjugated mouse anti-human CD4 BD Cell staining 1:100 FITC-conjugated mouse anti-human IFN-γ BD, Pharmingen, San Jose, Cell staining 1:100 CA APC-conjugated mouse anti-human IL-17 ebioscience, San Diego, CA Cell staining 1:50 PE-conjugated mouse anti-human Foxp3 BD, Pharmingen Cell staining 1:25 Mouse anti-human CD3 Dako, Carpinteria, CA Tissue 1:100 staining Mouse anti-human CD4 Neuromics, Edina, MN Tissue 1:100 staining Mouse anti-human CD8 Dako Tissue 1:100 staining Mouse anti-human CD14 Santa Cruz Biotechnology, Santa Cruz, CA Tissue staining 1:100 Mouse anti-human CD15 Dako Tissue staining 1:100 Purified mouse IgG control Innovative Research, Tissue 1:100 Southfield, MI staining Rabbit anti-human IL-17 Santa Cruz Biotechnology Tissue 1:100 staining Rabbit anti-human IFN-γ Santa Cruz Biotechnology Tissue 1:100 staining Purified Rabbit IgG isotype control IMGENEX, San Diego, CA Tissue staining 1:100

18 Supplemental Table 3. Primer pairs utilized in quantitative RT-PCR assays* Gene Sense (5-3 ) Anti-sense (5-3 ) IL-17 AACCGATCCACCTCACCTTGGAAT TTCATGTGGTAGTCCACGTTCCCA IFN-γ ACTAGGCAGCCAACCTAAGCAAGA CATCAGGGTCACCTGACACATTCA Foxp3 TTCAAGTTCCACAACATGCGACCC GCACAAAGCACTTGTGCAGACTCA IL-1β AAGTACCTGAGCTCGCCAGTGAAA TTGCTGTAGTGGTGGTCGGAGATT IL-6 AGCCACTCACCTCTTCAGAACGAA AGTGCCTCTTTGCTGCTTTCACAC IL-23p19 ACTCAGCAGATTCCAAGCCTCAGT TGGAGATCTGAGTGCCATCCTTGA IL-12p35 TAACTAATGGGAGTTGCCTGGCCT AGGGCCTGCATCAGCTCATCAATA IL-12p40 TCATCAAACCTGACCCACCCAAGA TTTCTCTCTTGCTCTTGCCCTGGA MMP-9 TACCACCTCGAACTTTGACAGCGA GCCATTCACGTCGTCCTTATGCAA β-actin ACCAACTGGGACGACATGGAGAAA TAGCACAGCCTGGATAGCAACGTA * cdna (0.5µl) was mixed with 9µl of double distilled water, 10 µl of SYBR Green Master Mix (2 PCR buffer, 5 mm MgCl 2, 0.4 mm each dntp, 0.05% bovine serum albumin and 1:10000 SYBR Green) (Invitrogen), 0.2µl of each primer (0.1µM) and 0.1 µl of Platinum Taq (5 U/µl) (Invitrogen). Amplifications were performed in a Mx3000 PCR machine (Stratagene) under the following cycling conditions: denaturation at 95 C for 5 min, followed by 40 cycles at 95 C for 30 s, 60 C for 30 s, and 72 C for 30 s. For each sample, PCR reactions were performed in triplicate. The level of gene expression was determined by interpolation with a standard curve.

19 Supplemental Table 4. Summary of Statistical Analysis Analysis 1 Analysis 2 Analysis 3 Analysis 4 Analysis 5 Analysis 6 Experimental set Groups Endpoint Analysis performed Fig.1 A-C Controls a n=34 Frequencies Effect of glucocorticoids of cytokine on circulating Th1 and producing T Th17 cells cells Fig.1 G-H; Fig. 3 A-C Effect of glucocorticoids on plasma cytokines (IL- 17, IFN-γ, IL-1β, IL-6, IL-12) Fig.2 A-B; Fig.4 A-E Effect of glucocorticoids on tissue cytokines (IL-17, IFN-g, IL-1b, IL-6, IL-12p40, IL-12p35, IL-23p19) Fig.3 D-H Effect of glucocorticoids on cytokine production in peripheral monocytes (IL-1b, Il-6, IL-12p35, IL- 12p40, IL-23p19) Fig.5 Effect of glucocorticoids on tissue cytokines in human artery-scid chimeras. Suppl. Fig.3 In vitro testing of anticytokine antibodies during the induction of Th17 cells GCA Patients untreated b n=26 treated b n=34 Controls a n=6 GCA Patients untreated b n=6 treated b n=6 Controls a n=8 GCA Patients untreated b n=8 treated b n=8 Controls a n=8 GCA Patients untreated b n=8 treated b n=8 Control tissues c n=9 Tissues untreated d n=9 Tissues treated d n=9 Isotype control c n=3 Anti-IL-1b c n=3 Anti-IL-6 c n=3 Anti-IL-23 c n=3 Plasma levels of cytokines Transcript levels of tissue cytokines Transcript levels of cytokines Transcript levels of tissue cytokines Frequency of TH17 cells Customized t- like statistics (see supplement Statistical Methods) One-sample t, two-sample t, Behrens- Fisher-Welch, or one-sample Wilcoxon as appropriate One-sample t, two-sample t, Behrens- Fisher-Welch, or one-sample Wilcoxon as appropriate One-sample t, two-sample t, Behrens- Fisher-Welch, or one-sample Wilcoxon as appropriate One-sample t, two-sample t, Behrens- Fisher-Welch, or one-sample Wilcoxon as appropriate Friedman rank sum test Comments b 11 patients were enrolled when untreated and analyzed again on therapy. 15 patients were tested only as untreated and 23 patients were only analyzed while on therapy. b patients were enrolled before treatment and analyzed again on therapy b samples from untreated and treated patients were paired b samples from the untreated and treated patients were paired c, d artery tissues were engrafted into mice. Arterial wall inflammation was induced or not (control). Arteries with inflammation were treated with steroids or left untreated. c cells were isolated from untreated GCA patients and cocultured with either isotype control or anti-il-

20 Analysis 7 Table 1 Effect of glucocorticoids on the frequencies of Th17 cells GCA Patients untreated b n=11 treated b n=11 One-sample t a Healthy volunteers served as controls. Age distributions in controls, untreated and treated patients were not significantly different. 1b, or anti-il-6 or anti-il-23 antibody in tissue culture. Frequencies of induced Th17 cells were measured. b 11 patients were enrolled while untreated and were analyzed again while on therapy.

21 Supplemental Table 5. The influence of glucocorticoid therapy on the frequency of circulating Th17 cells in individual patients Untreated Treated Patient Th17 (%) Treatment duration (months) Current prednisone dose (mg/day) Th17 (%) No No No No No No No No No No No

22 Supplemental Figure 1 A. B. C. Con 2.92 No Tx 2.89 Tx 2.63 Foxp3 CD4 CD4 CD4 Supplemental Figure 1. Foxp3 + CD4 regulatory T-cells in GCA. PBMC isolated from agematched control donors (A), untreated (B) and GC-treated GCA patient (C) as described in Figure 1 were stained with FITC anti-cd3, PerCP anti-cd4 and PE anti-foxp3 antibodies and analyzed by flow cytometry. Representative cytometric dot plots are shown. Frequencies of Foxp3 + regulatory CD4 T cells amongst CD3 T cells were similar in controls, untreated and treated patients.

23 Supplemental Figure 2 A. CD4 B. CD8 Adv Adv Med Med C. CD14 D. CD15 Adv Med Adv Med Supplemental Figure 2. Representation of different cell types in the granulomatous lesions of GCA. Paraffin-embedded tissue sections of temporal arteries from GCA patients were de-waxed, and stained with mouse anti-human CD4 (A), mouse anti-human CD8 (B), mouse anti-human CD14 (monocyte marker) (C), anti-human CD15 (neutrophil marker) (D) antibodies or purified mouse IgG control (not shown). The images presented are representative of five different biopsy samples. Positive cells are marked by red arrows. Similar to anti-human CD15 staining (D), mouse IgG control showed complete negative. Magnification 600 in (A), (B), (C) and (D).

24 Supplemental Figure 3 A. B. Th17 cells (%) P< P<0.05 P< Th1 cells (%) P= n.s. P= n. s. P< Iso con anti- IL-1β anti- IL-6 anti- IL-23 0 Iso con anti- IL-1β anti- IL-6 anti- IL-23 C. D. E. F. Isotype IgG Anti-IL-1β Anti-IL-6 Anti-IL IL-17 IFN-γ IFN-γ IFN-γ IFN-γ Supplemental Figure 3. IL-1β, IL-6, and IL-23 promote Th17 cell differentiation in GCA. CD45RO - CD4 naïve T cells and CD14 + monocytes were purified from PBMC of GCA patients. Naïve T cells were co-cultured with monocytes at a 2:1 ratio in the presence of anti-cd3 antibodies and LPS for 4 days, supplemented with goat anti-human IL-1β, anti-human IL-6, or anti-human IL-23 neutralizing antibodies, or purified goat IgG isotype control antibody. Cells were stained with PE-conjugated anti-cd3, PerCP-conjugated anti-cd4, FITC-conjugated anti- IFN-γ, and APC-conjugated anti-il-17 antibodies, and analyzed by flow cytometry as described in Figure 1. Frequencies of Th17 (A) and Th1 cells (B) were analyzed by flow cytometry, and are presented from three independent experiments. (C-F) Representatives of cytometric dot