Pathophysiology of large vessel vasculitis and utility of interleukin-6 inhibition therapy

Abstract

Takayasu arteritis (TAK) and giant cell arteritis (GCA) affect mainly large- and medium-sized arteries. In refractory cases, vascular remodeling progresses and leads to serious outcomes. Studies have demonstrated that cytokines such as interleukin (IL)-6 play crucial roles in the pathophysiology of TAK and GCA. Recently, randomized controlled trials on IL-6 inhibition therapy using tocilizumab (TCZ) were performed, and significant effects were exhibited. The purposes of conventional treatments have been to improve symptoms and decrease the levels of inflammatory markers. Arterial changes have been considered as damages. However, after TCZ came into practical use, establishment of treat to target is desired to prevent vascular remodeling. In contrast, a combination therapy of glucocorticoids (GCs) and TCZ notably increases the risk of infections. When TCZ is used, careful attention must be paid to possible infections, and dose of GC should be tapered as much as possible. Future tasks are to establish indication and dosage of TCZ, indication for discontinuation of TCZ due to remission, efficacy of TCZ monotherapy, and protocols of TCZ for pediatric cases.

KEYWORDS:

• Giant cell arteritis
• IL-6
• pathology
• Takayasu arteritis
• tocilizumab

1. Introduction

Takayasu arteritis (TAK) and giant cell arteritis (GCA) are categorized in large vessel vasculitis [1] affecting mainly large- and medium-sized arteries. Studies have revealed that cytokines such as interleukin (IL)-6 [2–5] are involved in the pathophysiology of TAK and GCA. Recently, the efficacy of IL-6 inhibition therapy using tocilizumab (TCZ, monoclonal antibody against IL-6 receptor) was demonstrated in clinical trials [6–8]. In 2017, the use of TCZ to TAK and GCA was approved in Japan, and the use to GCA was approved in the US and Europe. In this article, the pathophysiology of TAK and GCA is reviewed, and the efficacy and issues of IL-6 inhibition therapy are discussed.

1.1. Clinical features of TAK and GCA

TAK often affects women, and the onset age is usually below 40 years [9]. It can be explained by the association of TAK with HLA-B*52 [10] that TAK is relatively frequent in Asia [11] and that ulcerative colitis is complicated with TAK [12–14]. Enhanced expression of cytokines such as IL-6 [2,3] in TAK leads to fever, easy fatigability, and anemia. TAK causes inflammation, stenosis, and dilatation of the aorta and large arteries (Figure 1). Stenosis of the carotid artery causes neck pain, syncope, and ischemic retinopathy [15], whereas stenosis of the subclavian artery causes arm claudication and pulselessness. TAK has been conventionally treated with glucocorticoids (GCs) and immunosuppressive drugs such as methotrexate (MTX) and azathioprine. However, relapses often occur [16]. Advanced arterial lesions lead to aortic regurgitation, visual loss, cerebral infarction, and aortic aneurysms, some of which require cardiovascular surgery. Ishikawa has reported that 15-year survival rate was 82.9% [17].

Figure 1 Progression of large arterial lesions in TAK and GCA.

GCA predominantly affects women, and the onset age is usually above 50 years [1,18]. It can be explained by the association of GCA with HLA-DRB1*04 [19] that GCA is distributed in high-latitude areas [20] and that GCA is complicated by polymyalgia rheumatica [21,22]. Expression of cytokines such as IL-6 [4,5] is enhanced, resulting in fever, easy fatigability, and anemia. GCA causes inflammation and stenosis of the cranial arteries with a predilection [1], leading to temporal headache, jaw claudication, and ischemic optic neuropathy [23]. The standard treatment is GC therapy, and marginal but significant effects of MTX have been demonstrated by three randomized controlled trials (RCT) [24]. However, relapses frequently occur [25]. Large arterial lesions are observed in 10–83% of GCA cases [21] and cause aneurysms and stenosis. In a cohort study of GCA in the US [26], thoracic aortic aneurysms were observed in 11% of patients. Aortic dissection occurred in half of them, and 78% of the dissected patients died.

1.2. Histopathology of TAK and GCA

Arteries consist of three layers, namely, adventitia, media, and intima (Figures 1–3). There are external and internal elastic laminas between the three layers (Figures 2 and 3).

Figure 2 Histopathology of large arterial lesions in TAK and GCA.

Figure 3 Histopathology of medium-sized arterial lesions in GCA.

The progression of arterial lesions in TAK is shown in Figure 1. First, inflammatory reaction causes local pain and edema of the arterial walls. Next, the media is eroded by inflammatory cells, and it is followed by the two events: (1) damaged arteries are dilated by inner pressure or sheer stress to develop aneurysms or (2) intimal hyperplasia and adventitial fibrosis occur, causing stenosis of the lumen and leading to ischemia or infarction. In the histopathology of TAK (Figure 2, left), the increase of vasa vasorum and their thickened walls are observed. Inflammatory cell infiltration starts around the vasa vasorum. Granuloma formation accompanying multinucleated giant cells is usually observed in the outer layer of the media [27]. Occasionally, phagocytosed elastic fibers are observed in the giant cells [28]. Erosion of the media progresses from its outer rim [28]. Colocalized with the eroded part of the media, intimal hyperplasia, and adventitial fibrosis are observed [28].

Histopathology of medium-sized arteries in GCA is shown in Figure 3. In the early phase, inflammatory cell infiltration and edema are observed around the internal elastic lamina. It begins to be destroyed and surrounded by granulomas with multinucleated giant cells. Occasionally, phagocytosis of elastic fibers by the giant cells is also observed [28]. In the late phase, these findings expand to all layers of the artery. Destruction of the internal elastic lamina causes intimal hyperplasia and adventitial fibrosis, which bring out stenosis or occlusion of the lumen, leading to ischemic symptoms. Progression of large arterial lesions in GCA is similar to that in TAK (Figure 1). In histopathology (Figure 2, right), granulomatous erosion starts from the middle or inner layers of the media [27,28] and consequently leads to aneurysmal formation and dissection [26,29].

1.3. Differences between TAK and GCA

The characteristics of the two diseases are shown in Table 1. Although ages of onset differ between TAK and GCA, note that the age of onset is not a necessary condition in the classification criteria of the American College of Rheumatology (1990) for both TAK [9] and GCA [18] and the statements of Chapel Hill classification nomenclature (2012) [1]. As described above, the main histopathological loci of aortic lesions are the outer layers of the media in TAK and inner layers in GCA [27,28]. Aortic dissection is occasionally complicated with GCA [26,29,30], whereas it is rare in TAK. It might be explained by the difference in depth of aortic layers or frequencies of atherosclerosis correlated with age.

Table 1. Characteristics of TAK and GCA.

	TAK	GCA	References
Onset age	Usually ≤40	Usually ≥50	[9,18]
Associated HLA	HLA-B*52 (class I)	HLA-DRB1*04 (class II)	[10,19]
Geoepidemiology	Asia, South America	Europe, North America	[11,57]
Comorbidity	Ulcerative colitis	Polymyalgia rheumatica	[12,13,21]
Dominantly affected arteries	Mesenteric artery, carotid artery	Cranial arteries, axillary artery	[57–60]
Common ophthalmopathy	Ischemic retinopathy	Ischemic optic neuropathy	[15,23,28]
Main histopathological locus of aortic lesion	Outer layer of the media, adventitia	Middle or inner layer of the media	[27,28]
Form of aortic lesion	Aneurysm, stenosis	Aneurysm	[57,60]
Aortic dissection	Rare	Occasional	[26,30]
Regions of associated SNPs	IL12B, FCGR2A/FCGR3A, MLX, IL6, RSP9/LILRB3	REL, PTPN22, LRRC32 (below significant level)	[31–34]
Th1 and Th17 cytokines	Both Th1 and Th17 prominent	Relatively less enhanced (except high IL-23)	[45]
Response to TNF inhibitors	Yes (without RCT)	No	[37–42]
Response to abatacept	No	Yes	[35,36]

HLA: human leukocyte antigen; RCT: randomized controlled trial; SNP: single-nucleotide polymorphism; TNF: tumor necrosis factor.

Genome-wide association studies of TAK [31–33] have revealed single-nucleotide polymorphisms (SNP) in several genetic regions such as IL12B. In contrast, a genome-wide association study of GCA [34] did not reveal any SNPs that reached significant levels, other than HLA. Taken together, the contribution of genetic backgrounds might be stronger in TAK.

Next, differences in responses to treatments between TAK and GCA are discussed. RCTs of the same protocol using abatacept in both TAK and GCA were performed [35,36]. Abatacept demonstrated significant relapse-prevention effects in GCA [35], whereas it showed disappointing results for TAK [36]. In three single-arm studies using tumor necrosis factor (TNF) inhibitors for TAK [37–39], favorable effects of remission induction were observed in refractory cases, although second relapses were partially observed. Future validation by RCT is awaited. On the contrary, the three RCTs using TNF inhibitors in GCA did not show clinical effects [40–42]. These results support the differences in immunological status between TAK and GCA.

In an analysis of peripheral blood mononuclear cells (PBMC) in TAK [43], production of IL-12 (Th1) was significantly more enhanced than in healthy controls (HC), whereas no difference was observed in the production of IL-23 (Th17) between TAK and HCs. In a study using biopsied specimens of GCA [44], expression of both Th1 and Th17 cytokines was significantly more enhanced than that in HCs. Saadoun et al. investigated PBMC of TAK and GCA in the same procedure [45]. In TAK, production of both Th1 and Th17 cytokines was prominently enhanced. In GCA, production of Th1 and Th17 cytokines was relatively less enhanced, whereas production of IL-23 (Th17) was prominent. In addition, there was a significant difference in TNF-α expression between TAK and GCA (higher in TAK). The roles of Th1 and Th17 in TAK and GCA are shown in Figure 4. Enhanced Th1 cytokine expression in TAK can be explained by its association with HLA class I [10] and infiltration of NK cells and cytotoxic T cells in histopathology [46]. However, Th1 activation leads to Th17 activation via IL-6. In GCA, IL-23 expression may be predominant, according to the direct comparison by Saadoun et al., although both Th1 and Th17 are important. Activation of Th17 leads to further activation of Th17 via IL-6, forming a vicious cycle of continuous inflammation.

Figure 4 (a) IL-12 and IL-23 share p40, which is encoded by IL12B, a candidate susceptibility gene in two genome-wide association studies of TAK [31,32]. (b) Roles of Th1 and Th17 in TAK and GCA. IL-6 and Th17 form a vicious cycle.

1.4. Indication, efficacy, and safety of IL-6 inhibition

Increased production of IL-6 has been reported in TAK [2,3] and GCA [4,5], and IL-6 has been recognized as an important molecule in the pathophysiology of continuous inflammation in TAK and GCA (Figure 4). Administration of TCZ, an IL-6 receptor inhibitor, has been attempted in the treatments of TAK [47,48] and GCA [49,50] and has shown promising results (Table 2). Three RCTs were performed, namely, Swiss RCT for GCA [6], RCT for GCA in the US and Europe (GiACTA trial) [7], and Japanese RCT for TAK (TAKT study) [8]. In all the trials, TCZ was combined with GC therapy, and the primary endpoint was the inhibition of relapses during the tapering of GC. As a result, significant relapse-prevention effects were observed in the Swiss RCT and GiACTA, and favorable effects were observed in TAKT (Table 2). In GiACTA, adverse events probably attributed to the effects of GC were significantly reduced by TCZ because of its reduction effects on cumulated GC dose.

Table 2. Trials and case series using TCZ for TAK and GCA.

Disease	Author, year	Design	Case No.	Dosage	Route	Main outcomes	p-Value
TAK	Nakaoka et al., 2013 [47]	Single arm	4	8 mg/kg/4 wks.	div.	Regression of inflamed arteries, reduction of GC dose	–
TAK	Mekinian et al., 2018 [48]	Retrospective	46	8 mg/kg/mo.	div.	Higher relapse-free survival rate than in the DMARD group	.02
TAK	Nakaoka et al., 2018 [8]	RCT	36	162 mg/wk.	s.c.	Higher relapse-free survival rate than in the placebo group	.0596
GCA	Seitz et al., 2011 [49]	Single arm	7	8 mg/kg/2–4 wks.	div.	Improvement of symptoms, reduction of GC dose	–
GCA	Loricera et al., 2015 [50]	Single arm	22	8 mg/kg/mo.	div.	Improvement of symptoms, reduction of GC dose	–
GCA	Villiger et al., 2016 [6]	RCT	30	8 mg/kg/4 wks.	div.	Higher relapse-free survival rate than in the placebo group	.0005
GCA	Stone et al., 2017 [7]	RCT	250	162 mg/wk.	s.c.	Higher relapse-free survival rate than in the placebo group	<.001

div.: intravenous drip; DMARD: disease-modifying antirheumatic drug; GC: glucocorticoid; RCT: randomized controlled trial; s.c.: subcutaneous.

The approved indications of TCZ for TAK and GCA in Japan are (1) disease activity even after adequate treatments with GC and (2) difficulty in continuation of GC. Thus, GC must precede TCZ in the indications. In contrast, an initial combination therapy of GC and TCZ was administered in patients with new-onset GCA in the Swiss RCT and GiACTA. Future research is required on whether initial combination or add-on of TCZ is more reasonable. Successful monotherapy with TCZ has been reported in TAK [51] and GCA [49,52], although validation is needed. Combination of TCZ and immunosuppressive drugs might be an option. However, its efficacy and safety have not been established. Immunosuppressive drugs were washed out in TAKT, whereas some participants in GiACTA were treated with stable doses of MTX. A sub-analysis is awaited.

The approved dosage of TCZ for TAK and GCA in Japan is 162 mg subcutaneously (s.c.) every week (Q1W), twice the dosage for rheumatoid arthritis (RA):162 mg s.c. every other week (Q2W). Note that the dosage of TCZ in the past retrospective and single-arm studies has been 8 mg/kg intravenous drip every 4 weeks, identical to the standard dosage for RA. It implies that Q2W may be effective for TAK and GCA as well. Indeed, Q2W showed significant relapse prevention than placebo in GiACTA [7]. However, in patients included in GiACTA based on relapse, Q2W did not show significant effects, whereas Q1W was significantly effective [7]. Indications of Q1W and Q2W should be developed in the future. In addition, the indication for discontinuation of TCZ due to remission has not been established. A patient with TAK who presented with cytokine storm after discontinuation of TCZ has been reported [53]. When TCZ is withdrawn in TAK, meticulous observation is needed, or switch to MTX is an option [47,53].

There have been reports of pediatric cases with TAK successfully treated with TCZ [54,55]. However, the precise efficacy in pediatric cases cannot be assessed, because the TAKT study excluded participants below 12 years old. TCZ is desirable to prevent GC-induced growth depression, requiring future studies. Subcutaneous injection is disadvantageous in dose adjustment for children compared to intravenous drip. Adjustment of intervals between subcutaneous injections may be an option, as approved for systemic juvenile idiopathic arthritis by the US Food and Drug Administration.

Risk factors of infections in treatments of RA with TCZ were age older than 65 years, disease duration longer than 10 years, concomitant respiratory disease, and prednisolone dose more than 5 mg/day in a Japanese cohort [56]. The risk of TCZ in GCA is probably higher than that in RA, since patients with GCA are usually older than 65 years and treated with moderate or higher dose of GC. Moreover, it is difficult to detect infections because serum C-reactive protein levels are suppressed by TCZ. Treatment with TCZ requires initial screening of latent infections, appropriate prophylactic measures, regular monitoring of infectious signs, and utmost reduction of GC dose. It should be conducted for patients with reasonable indications by physicians with sufficient experience in treating vasculitis at hospitals that can provide appropriate treatments in case of emergency.

1.5. Revolution in treatments of TAK and GCA

Application of TCZ will change the treatments of TAK and GCA. The targets of conventional therapy have been improvement of symptoms and reduction of serum inflammatory markers. Arterial changes (Figure 1) have been recognized as damages, and advanced cases have required surgeries. However, the new target of treatments should be prevention of vascular remodeling. The following four milestones are proposed: (1) Early diagnosis: Education of residents, primary care physicians, and the society will be important. (2) Evaluation of disease activity: Accurate and convenient disease activity indices are required. Utility of imaging modalities should be improved. (3) Treat to target: A principle should be established for prevention of vascular remodeling. (4) Standardization of treatment protocols: Revision of guidelines based on application of TCZ is needed [28].

In conclusion, cytokines such as IL-6 are important in the pathophysiology of TAK and GCA. RCTs of IL-6 inhibition therapy with TCZ have shown significant effects. The new goal will be the establishment of treat to target to prevent vascular remodeling. However, attention should be paid to the risk of infections, and the protocols on TCZ use should be improved in consideration of efficacy and safety.

Conflict of interest

H. Yoshifuji has received speech fees from Chugai Pharmaceutical Co., Ltd.

Acknowledgements

A referred article [28] is written in Japanese but currently being translated into English.