In a cross-sectional study, fresh blood samples were obtained from 30 GCA patients. Fifteen patients were receiving glucocorticoids (GC), i.e., prednisolone. Four of these patients were also receiving methotrexate. The remaining 15 patients were not receiving GC or methotrexate at the time of blood-withdrawal (Table 1). Blood samples were obtained before noon and all donors were non-fasted/took their prednisolone in the morning. GCA diagnosis was either confirmed by a positive temporal artery biopsies (TABs) and/or positive 18F-fluorodeoxyglucose-positron emission tomography-computed tomography (FDG-PET/CT) (Supplementary Table S1A). As controls, we obtained fresh blood samples from 18 sex- and age-matched healthy controls (HCs) who were screened for past or actual morbidities and pharmacological treatments (Supplementary Table S1B).

At time of diagnosis, patients were examined for both cranial and/or systemic symptoms. For cranial symptoms, patients had to present one of the following: new headache, temporal artery abnormality, scalp tenderness, jaw/tongue claudication, vision loss (ischemic symptom), amaurosis fugax (ischemic symptom), transient ischemic attack (TIA), or cerebrovascular accident (CVA). The following systemic symptoms reflecting GCA disease activity were taken into account: arm/leg claudication or PMR clinic, and/or two of the following symptoms: fever, weight loss, malaise or night sweats. Symptoms were scored only if they could not be explained by other causes (i.e., infection).

Written informed consent was obtained from all study participants. All procedures were in compliance with the declaration of Helsinki. The study was approved by the institutional review board of the UMCG (METc 2012/375 for HC and METc 2010/222 for GCA patients).

GCA patients were initially treated with 40–60 mg/day of prednisone. Tapering of prednisone treatment started after 2–4 weeks, based on normalization of clinical signs and symptoms accompanied by normalization of the erythrocyte sedimentation rate (ESR) and/or C-reactive protein (CRP). After 3 months, the median prednisone dose was 15 mg/day. Seventeen GCA patients were in remission after 3 months of GC treatment. Remission was defined based on the absence of clinical signs and symptoms and normal ESR (<30 mm/h) and/or C-reactive protein (CRP) <5 mg/L.

Fresh blood samples were collected in EDTA anticoagulant tubes and processed within 2 h. The samples were stained in two separate multi-color flow cytometry panels. The first panel (Panel 1) included the following monoclonal antibodies: anti-CD3, anti-CD4, anti-CD45RA, anti-CD25, anti-CD28, anti-CTLA-4, anti-CD279 (PD-1), and anti-VISTA (Supplementary Table S2A). The second multi-color flow cytometry panel (Panel 2) included the following monoclonal antibodies: anti-CD3, anti-CD16, anti-CD14, anti-CD274 (PD-L1), anti-CD273 (PD-L2), anti-CD80, anti-CD86, and anti-VISTA (Supplementary Table S2B). After this, cells were fixed and erythrocytes were lysed using FACS Lysing solution (BD Biosciences, Durham, NC, USA) according to instructions of the manufacturer. Immediately thereafter, samples were measured on a BD LSR-II flow cytometer. Data were collected for at least 1 × 10⁵ cells and analyzed with Kaluza Analysis Software (Beckman Coulter). Proportions of marker positive/negative cell populations were calculated by quadrant dot-plot analysis based on proper isotype controls. In panel 1, CD4+ T cell subsets were identified and gated as previously described (Figure 4A) (26). For gating strategy see (Supplementary Figure S1A). In panel 2, classical (CD14^brightCD16^neg), intermediate (CD14^brightCD16+), and non-classical (CD14^dimCD16+) monocytes subsets were gated as previously described (Supplementary Figure S1B) (27). In parallel, the absolute numbers of CD3, CD4, and CD8 T-cells and monocytes were determined according to the MultiTest TruCount method (BD), as described by the manufacturer. Data were acquired on a FACS Canto-II (BD Biosciences) and analyzed with FACS Canto Clinical Software (BD).

Ninety-six-well flat-bottom plates were coated with purified anti-CD3 (clone OKT3,) at 2.5 μg/mL mixed together with 10 μg/ ml VISTA-Ig (R&D Systems) or control-Ig protein (110-HG, R&D Systems) in PBS for 2 h at 37°C and stored at 4°C overnight. Wells were washed twice with RPMI 1640 before adding cells. Peripheral blood mononuclear cells (PBMC) were isolated from heparinized blood with Lymphoprep (Axis-Shield). Immediately after, CD4+ T cells were isolated by negative selection using the MagniSortTM Human CD4+ T cell Enrichment Kit (Thermo Fisher Scientific, Waltham, MA., USA) according to instructions of the manufacturer. CD4+T cells were plated at 1 × 10⁵ cells per well in RPMI1640 with 10% heat-inactivated fetal bovine serum and gentamycin. Next, soluble anti-CD28 (clone L293, BD Biosciences) at 250 ng/mL was added to the culture media. After 5 days in culture at 37°C, cells were stained with zombie NIR fixable viability kit (Biolegend) according to instructions of the manufacturer and surface marker CD4. Next, cells were fixed and permeabilized with FoxP3 staining buffer set (eBiosciences) followed by intracellular staining for transcription factors: FoxP3, T-bet, RORgt, BCL6, GATA3 (Supplementary Table S2C).

Temporal artery biopsies (TABs) were obtained from 5 GCA patients before start of GC treatment. The tissue was fixed in formalin and paraffin embedded. Sections were deparaffinized in xylene and rehydrated in graded ethanol. After antigen retrieval in 1 mM EDTA (pH 8) and endogenous peroxidase blocking, sections were incubated with a rabbit monoclonal anti-VISTA antibody (Clone: D1L2G; Dilution: 1:200; Cell Signaling, Danvers, USA) or an isotype control antibody. The primary antibody was visualized using an EnVision+ Kit (HRP Rabbit AEC+, Dako K4009) and sections were counterstained with hematoxylin. Stainings for PD-1 (Clone: MRQ-22; Dilution: ready to use (R.T.U); Ventana) and PD-L1 (Clone: SP263; Dilution: R.T.U; Ventana) were performed in a Benchmark Ultra automated immunostainer (Ventana, Tucson, Arizona) using pre-diluted antibodies and following the manufacturer's protocols (Supplementary Table S2D). Non-GCA TABs (n = 5) and human tonsil tissue were used as controls (Temporal arteries were considered negative for GCA if there were no typical histological findings). Stained sections were scanned using a Nanozoomer Digital Pathology Scanner (NDP Scan U10074–01, Hamamatsu Photonics K.K., Hamamatsu, Japan).

Detection of VISTA, PD-1 and PD-L1-expressing cells (irrespective of intensity) in affected temporal artery biopsy areas was semi-quantitatively scored on a five-point scale (0–4): 0 = no positive cells, 1 = occasional positive cells (0–1% estimated positive), 2 = low numbers of positive cells (>1–20%), 3 = moderate numbers of positive cells (>20–50%), 4 = high numbers of positive cells (more than 50%). Affected regions containing infiltrating cells were scored. Scoring was performed by two independent investigators, trained by a pathologist and average scores were calculated.

Double staining was performed in limited number of samples to confirm the co-localization of VISTA with T cells using CD3 or with macrophages using pu.1 (29). MultiVision Polymer Detection system (MultiVision anti-rabbit/AP + anti-mouse/HRP polymers; Thermo Fisher Scientific, Fremont, USA) was used according to the manufacturer's instructions. Binding of primary antibodies was visualized with the chromogens of the MultiVision kit, LVBlue, and LVRed.

Non-parametric tests were used for data analysis. The Mann-Whitney U-test was used to compare data of patients with HC. Paired samples were analyzed with the Wilcoxon signed rank test. Analyses were performed with GraphPad Prism 7.0 software. P-values of <0.05 (2-tailed) were considered statistically significant.

We first assessed the numbers of monocytes in peripheral blood (PB) of GCA patients and HCs. Total monocyte counts were comparable between groups. Next, we analyzed the expression of CD80/86, as ligands of CD28 and Cytotoxic T-Lymphocyte-associated antigen-4 (CTLA-4) and the expression of PD-L1/PD-L2 as ligands for PD-1 on the surface of monocytes. As VISTA has been reported to behave both as ligand and receptor to suppress T cell activation (30, 31), we measured VISTA+ cells in both panels (panel 1 and 2, Supplementary Tables S2A,S2B). The frequencies of PD-L1 and PD-L2-expressing cells were not different between GCA patients and HCs. Interestingly, the proportions of circulating monocytes expressing CD80/CD86 and VISTA were decreased in GCA patients (Figure 1). As GC treatment has a major effect on immune cells, especially monocytes (27, 32), we analyzed expression in GCA patients who were not receiving GC (n = 15) and GCA patients receiving GC (n = 15) separately. Indeed, the proportional decrease of monocytes expressing CD80/CD86 and VISTA was treatment related (Figure 2).

Similarly, we measured the absolute numbers and frequencies of CD4+ T cells. No significant differences between GCA patients and HCs were observed (Supplementary Figure S2). Next, we analyzed the frequencies of expression of IC expressing-CD4+ T cells. Proportions of co-stimulatory molecule CD28+ and its outcompeting co-inhibitory molecule CTLA-4+ cells were similar between patients and HCs. We sought to confirm that frequencies of circulating PD-1+ Th cells were reduced in GCA patients (8). Indeed, proportions of these cells in GCA patients were lower than in HCs. In addition, proportions of VISTA+ Th cells were also decreased in GCA patients (Figure 3). Contrary to the outcome seen on IC expression on monocytes, no effect of GC treatment on IC expression by CD4+ T cells was detected. Decreased expression of the negative IC was seen in both patients receiving GCs and not receiving GCs (Supplementary Figure 3).

Next, we determined the frequencies of IC molecule-expressing cells within functionally distinct populations of naïve and memory CD4+ Th cells by CD45RA and CD25 expression (Figure 4A) according to Miyara et al. (28) and adapted by van der Geest et al. to add one age-associated subset (26). The memory fractions include Memory CD25- (fraction 5), Memory CD25^dim (fraction 4), Memory CD25^int (fraction 3), and Memory CD25^high Treg (fraction 2). Within fractions 5 and 4, frequencies of CD28+, PD-1+, and VISTA+ cells were decreased in GCA patients compared to HCs. Memory CD25^int (fraction 3) which has intermediate expression of CD25 but is considered as a non-regulatory T cell subset, showed a decrease in the proportions of PD-1+ and VISTA+ cells of GCA patients compared to HCs. Interestingly, within fraction 2, the activated Treg subset, we found unchanged frequencies of cells expressing the negative IC molecules measured, but a decrease in proportions of cells expressing the positive co-stimulator CD28 in GCA patients compared to HCs, thereby suggesting a potentially decreased Treg activity (Figure 4). Within the naïve fractions we identified Naïve CD25- (fraction 6), Naïve CD25^int Treg (fraction 1) from the Miyara classification plus the age-associated subset CD45RA+ CD25^dim (fraction 7) cells which have been shown to represent a broad and functional reservoir of naïve T cells in the elderly population (26). Within these naïve fractions, frequencies of VISTA+ cells were decreased in GCA patients compared to HCs. Proportions of PD-1–expressing cells were also significantly decreased in conventional naïve T cell fractions 6 and 7 but not in naïve CD25^int Treg (fraction 1). Frequencies of CTLA-4-expressing cells within the memory and naïve fractions were comparable between GCA patients and HC (Figure 4C).

As we found PD-1 and VISTA to be modulated in Th cells from GCA patients, we next assessed expression of these checkpoints, including the ligand PD-L1, at lesional sites using immunohistochemistry on GCA diagnostic TABs (Figure 5 and Supplementary Figure S4 for isotype control staining). We found VISTA-expressing cells to be increased within the infiltrates of the adventitia, media and intima layer of the vessel wall when compared to non-GCA biopsies (Figure 5C). PD-1-expressing cells were found mostly within the infiltrates of the medial layer of the vessel wall (Figure 5D). Finally, PD-L1-expressing cells were also increased in GCA biopsies compared with non-GCA TABs (Figure 5E). In addition, double staining for VISTA/CD3 and VISTA/pu.1 confirmed that T cells as well as macrophages in the vascular wall express VISTA (Supplementary Figure S5).

To evaluate whether the decrease in the frequencies of VISTA+ cells seen in GCA patients would translate into a functional effect, we performed experiments using recombinant human VISTA on human T cells. To this end, VISTA-Ig or control Ig was immobilized on plates with anti-CD3 (OKT3) and soluble anti-CD28 was added to provide appropriate co-stimulatory signaling. Unstimulated samples served as controls. Dead cells were excluded from the analysis and gates were set according to unstimulated samples (Supplementary Figure S6). The ability of CD4+ T cells to differentiate into Th subsets (Th1, Th17, Tfh, Th2, and Treg) was assessed by measuring the expression of intracellular transcription factors (T-bet, RORɤt, BCL6, GATA3, and FoxP3) (Supplementary Figure S7). The effect of VISTA-Ig engagement on CD4+ T cell differentiation in HCs showed a decrease of T-bet, RORɤt and BCL6-expressing cells when compared to Ig control engagement (Supplementary Figures S8, S9) The expression of GATA3 and FoxP3 remained unchanged. On the other hand, applying the same differentiation conditions to CD4+ T cells of GCA patients, we found that VISTA-Ig engagement failed to modulate any of these transcription factors, hence facilitating Th1, Th17, and Tfh lineage differentiation (Figure 6).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.