Significant association between clinical characteristics and changes in peripheral immuno-phenotype in large vessel vasculitis

Patients with newly diagnosed LVV who visited Keio University Hospital and fulfilled the American College of Rheumatology criteria for GCA [11] and TAK [12] between May 2016 and May 2019 were consecutively enrolled. Patients with secondary LVV that could mimic GCA/TAK (for example, Cogan syndrome, sarcoidosis, Kawasaki disease, Behçet disease, IgG4-related disease, syphilis, tuberculosis, Ehlers-Danlos syndrome, Marfan syndrome, and neurofibromatosis) were excluded based on the medical chart at screening. We confirmed that the healthy controls (HCs) did not have an autoimmune disease, severe allergic disorder, malignancy, or infection.

This study was approved by the research ethics committee of the Keio University School of Medicine (#20140335) and was conducted according to the Declaration of Helsinki. Informed consent was obtained from all patients and HCs.

Clinical information was obtained from patients’ records. We collected information on age; gender; time from symptom onset to diagnosis; body mass index at diagnosis; smoking habit; comorbidities, including hypertension, diabetes mellitus, dyslipidemia, chronic kidney disease, polymyalgia rheumatica (PMR) [13], and inflammatory bowel disease (IBD) [14]; laboratory data on erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP) levels at diagnosis and each visit; and treatment during follow-up. Arterial involvement was evaluated using histological and/or radiological examinations (any or all of ultrasonography, computed tomography (CT), magnetic resonance imaging, and positron emission tomography CT).

Achievement of remission was defined as the disappearance of clinical symptoms with normal CRP with prednisolone (PSL) < 10 mg/day [5‐10]. Relapse was defined as the reappearance of vasculitis-related manifestations accompanied by elevated levels of acute phase reactants requiring an increase in GC dose or additional immunosuppressive agents [5‐10].

We collected clinical data and peripheral blood samples from patients with GCA and TAK at the time of diagnosis (week 0) and subsequently at weeks 4, 12, 24, and 52 of treatment. Twenty microliters of heparinized blood samples was collected from patients with LVV (GCA, n = 12; TAK, n = 8), and FACS analysis for immuno-phenotyping was carried out without delay after collecting the samples. FACS and data analyses were conducted on a FACS Aria II (BD Biosciences) and using FlowJo v.7.6.4 Software (Tree Star, Stanford University, CA, USA), according to the methods recommended by the manufacturers of the antibodies used (BD Biosciences and BioLegend: Additional file 4: Table S1). The phenotypes of immune cell subsets were defined based on the Human Immunology Project protocol [15]. Micro-sized cells, such as fragments of dead cells, were gated out when we analyzed the cells. Details of the gating strategy are shown in Additional file 4: Table S2 and Additional file 1: Figure S1. The mean number of each immune cell phenotype from patients with GCA and TAK was compared with that from age-matched HCs, and the fold change was calculated by dividing the cell number from patients by the corresponding average cell number from HCs. Fold change values for each immune cell phenotype were then compared between the groups.

Baseline data was analyzed using one-way analysis of variance (ANOVA) and post hoc test. Then, chronological data was analyzed using correlation analysis to identify the immune cell subsets associated with disease relapse. We also examined whether the use of biologic agents reduced the immune cells associated with disease relapse.

We collected FACS data and clinical profiles from 20 Japanese LVV patients who were followed longitudinally across a total of 90 visits. None of the patients had been previously treated with GC or any biologic agent. All patients received GC therapy at an initial dose equivalent to 0.6–1.0 mg PSL/kg/day, which was tapered by the attending physician based on previously reported clinical trials [5‐10].

We evaluated the patients at each visit based on their symptoms, ESR, CRP, and PSL dose. Baseline characteristics, treatment, and effects of treatment in patients with LVV and HCs are summarized in Table 1. TAK was younger (GCA vs TAK, 71 vs 47 years), and the time from onset to diagnosis was longer compared with GCA (2.5 vs 3.6 months). Fifty percent (6/12) of GCA were with PMR, and 25% (2/8) of TAK were with IBD. Laboratory tests of ESR and CRP were not different. The proportion of relapse and/or surgery was higher in TAK than in GCA (33% [4/12] vs 63% [5/8]).

We compared the absolute number of circulating immune cells among GCA, TAK, and HCs using one-way ANOVA and post hoc test. We revealed that the number of helper T (Th), follicular helper T (Tfh), CD8⁺ T, CD8⁺ Tem, CD14⁺⁺ CD16⁺ monocytes, and neutrophils was higher in patients with GCA and/or TAK than HCs. Among them, the number of CD8⁺ T and CD8⁺ Tem was higher in patients with TAK than GCA (Table 2).

Fold change in the number of immune cells in patients with GCA and TAK was determined by dividing the cell number in patients by the average cell number of the corresponding immuno-phenotype in age-matched HCs (Fig. 1). Then, fold changes of each cell subset in patients with GCA or TAK, GCA with or without PMR, and TAK with or without IBD were compared. The proportion of Tfh, Tfh1, Tfh17, CD8⁺ T, naive CD8⁺ T, CD8⁺ Tem, CD19⁺ B, CD27⁻ IgD⁺ naive B, and CD27⁺ IgD⁻ memory B was higher in TAK than in GCA, which was consistent with ANOVA (Fig. 1a). There was no statistical difference in the proportion of cell subsets between patients with or without PMR/IBD (Fig. 1b, c).

Because of the higher proportion of relapse/surgery in patients with TAK than GCA [3, 16, 17], we hypothesized that the immuno-phenotype profile may differ after treatment between the two groups of the patients. Thus, we compared the immuno-phenotyping data between GCA (n = 8) and TAK (n = 5) patients without relapse. Compared to patients with GCA (Fig. 2a), memory CD4⁺ T and CD8⁺ T cells in patients with TAK remained high even in the remission phase (Fig. 2b). To assess the effect of treatments, we analyzed the fold changes in immune cell subsets in patients successfully treated with GC and biologic agents (GCA, n = 5; TAK, n = 3). As a result, we found that memory CD4⁺ T and CD8⁺ T cells in patients with TAK also remained high levels even after treatment with biologic agents (Additional file 2: Figure S2A and S2B).

To assess changes in immuno-phenotype at relapse, we longitudinally profiled the immuno-phenotyping data of patients who achieved remission then relapsed. We followed the immuno-phenotyping data of seven patients (GCA, n = 4; TAK, n = 3) who achieved remission then relapsed until week 52 of treatment and calculated the correlation between the fluctuation in each immune cell subset and change in disease activity.

The results of matrix correlation analysis for each subset and the fluctuation in disease activity are shown in Additional file 3: Figure S3. Given that Th1 and Th17 have been implicated in the pathogenesis of GCA and TAK [18‐20], we examined the changes in relevant selected subsets including Th1, Th17, Tfh, CD8⁺ T, and CD19⁺ B cells and laboratory data of ESR and CRP (Fig. 3). Th1, Th17, and Tfh cells were strongly correlated with disease activity in GCA and TAK (Fig. 3A), as well as ESR and CRP levels (Fig. 3B). Notably, CD8⁺ T cells were strongly correlated with disease activity in TAK, but not in GCA. In addition, the fold changes of CD8⁺ T cells at onset (TAK vs GCA, 1.9 ± 0.8 vs 0.3 ± 0.3, p = 0.032) and at relapse (2.0 ± 0.9 vs 0.4 ± 0.1, p = 0.011) were significantly higher in patients with TAK than those in GCA.

Eleven patients (GCA, n = 5; TAK: n = 6) were treated with TCZ (n = 9) and IFX (n = 2). To determine whether these drugs cause immunological changes, we compared the number of immune cells of selected subsets before and after treatment with biologic agents (Fig. 4). It is noteworthy that treatment with biologic agents did not decrease the proportion of CD8⁺ T cells (Fig. 4d), while the treatment decreased the proportion of Th1, Th17, and Tfh cells (Fig. 4a–c).