Background
Juvenile idiopathic arthritis (JIA) is the most common childhood rheumatic disease affecting approximately 1:1000 children [
1]. The polyarticular JIA (polyJIA) subtype is the most clinically similar to adult rheumatoid arthritis (RA), yet its pathogenesis remains poorly understood. Untreated, JIA can cause pain, lead to joint damage, and result in disability. Fortunately, therapeutics for JIA have greatly expanded, and there are now options that span a variety of mechanisms. This has made inactive disease the treatment goal for all patients [
2]. Data and expert consensus support achieving the goal of inactive disease, regardless of the specific choice of medication(s) or their mechanism [
3]. While many patients respond promptly and well to the initiation of treatment, this approach can, and does, lead to instances of failure of a cascade of empirically selected agents prior to arriving at the therapeutic that allows for development of clinically inactive disease. However, there are currently insufficient data to guide selection amongst therapeutic options in a directed manner [
4]. Identification of biomarkers that could be detected at diagnosis, distinguish molecularly between subsets of patients with clinically similar disease, and allow for targeted therapeutic recommendations is the apex of arthritis treatment.
One targetable inflammatory pathway of interest that might stratify patients into molecular groups is the interleukin(IL)-6/signal transducer and activator of transcription 3 (STAT3) pathway, for which an anti-cytokine pathway biologic therapy has been approved for polyJIA [
5]. IL-6/IL-6 receptor (R) engagement and activation of gp130 signaling leads to Janus kinase (JAK)-mediated activation of STAT3 [
6]. The importance of this pathway in human immunobiology has been clearly demonstrated by rare inborn errors of immunity, wherein defects of IL-6/STAT3 signaling result in phenotypically similar primary immune deficiencies [
7], and gain-of-function leads to early-onset, multiorgan autoimmunity, including inflammatory arthritis indistinguishable from JIA [
8]. Detection of elevated IL-6 in the serum of patients with RA [
9,
10] was an important initial observation on the way to the development of IL-6/IL-6R inhibition [
11]. Increases in STAT3 and activated (phosphorylated) STAT3 are correspondingly reported in peripheral blood and synovial fluid of patients with RA at baseline [
12,
13]. IL-6 is also elevated in the serum of patients with polyJIA [
14], although similar STAT3 studies have not been reported.
The IL-6/STAT3 inflammatory pathway is an important factor in the differentiation of different subsets from naïve CD4 + T cells [
11]. IL-6 driven STAT3 activation influences the development of T helper 17 (Th17) cells [
15], at the expense of T regulatory (Treg) cell development. Th17 cells are key for host defense against extracellular fungi and play an important role in immune homeostasis, the understanding of the depth of which is continually evolving [
16]. Despite their beneficial role, Th17 cells are also a contributing factor in chronic inflammation, including in RA and JIA [
17,
18]. The balance of pro- and anti-inflammatory T cells can be influenced by targeted IL-6 pathway blockade [
19‐
22], even in the setting of genetically driven IL-6/STAT3 pathway activation [
23].
Detection of polyJIA patients with enhanced IL-6/STAT3 pathway activation may distinguish a subset for whom application of directed biologic therapy may provide targeted treatment. We aimed to profile STAT3 activation and T cell subsets in polyJIA patients before and during treatment.
Discussion
Activity of the IL-6/STAT3 inflammatory pathway may be able to stratify patients with polyJIA and distinguish a subset for which targeted anti-cytokine therapy is indicated. We evaluated peripheral blood from treatment-naïve polyJIA patients and found significantly elevated frequencies of Th17 cells compared to healthy pediatric controls; Treg cell frequencies did not differ. We also detected a decreased response to ex vivo inflammatory cytokine stimulation in PMBCs from treatment-naïve patients compared to controls, although this trend was not significant. Finally, evaluation of samples collected on therapy after clinical remission revealed, on average, significantly elevated post-treatment frequencies of both Th17/1 and ex-Th17 cells in polyJIA patients compared to their matched treatment-naïve samples. Elevations in post-treatment Th17, Th17/1 and ex-Th17 cells were notable in two patients who remain therapy-bound years after remission.
Investigators have reported elevated Treg cells in the synovial fluid of patients with oligoarticular JIA, not concurrently found in the peripheral blood [
34‐
37]. But, whether patients with polyJIA, particularly treatment-naïve patients, have altered peripheral Treg cells compared to pediatric controls has not been clear [
38‐
41]. We expected to find reduced frequencies of Treg cells, identified by CD25 and FOXP3, in peripheral samples from treatment-naïve polyJIA patients compared to pediatric controls, or between patients from their treatment-naïve to post-treatment samples, but we found no differences (Fig.
1). While Treg cells are classically identified by the expression of FOXP3, there are subsets of cells among FOXP3-expressing T cells, including cells that have developed inflammatory characteristics [
42,
43]. It was our goal to identify peripheral Treg cells in order to subset polyJIA patients, however, to identify and track peripheral cells which are genuinely regulatory will likely require additional stratification, if not functional evaluation. These types of evaluations are important to dissect pathogenesis, but less feasible as a real-time evaluation tool.
Peripheral Th17 cells have been evaluated in various cohorts of patients with JIA compared to pediatric controls. Elevated peripheral Th17 cells were detected in systemic-onset JIA and in mixed JIA populations (oligoarticular and polyJIA) [
41,
44‐
46], even though samples were collected during different levels of disease activity and on a spectrum of therapies. However, other investigations reported no differences in peripheral Th17 cells in oligoarticular JIA [
47], enthesitis-related arthritis [
48] and other mixed JIA cohorts [
37,
49,
50]. With respect to the local inflammatory environment, the preponderance of available evidence suggests that the synovial compartment of patients with JIA is enriched for Th17 cells when directly compared to the peripheral blood of the same individuals [
37,
41,
47‐
49,
51]. Depending on a variety of uncontrollable clinical factors, intra-articular sampling for diagnostics/therapeutics is not routinely done at the time of diagnosis of JIA, and rarely in patients with polyJIA. Since synovial fluid is not reliably available for analysis to guide therapeutic recommendations, we determined the peripheral levels of Th17 cells in a large, treatment-naïve poly JIA cohort of 17 patients (Fig.
1), and found, on the whole, Th17 cells to be significantly elevated. Further, there is a spectrum of frequencies of Th17 cells amongst treatment-naïve poly-JIA patients. This may provide an opportunity to distinguish patients with the strongest evidence of IL-6/STAT3 activation, to select anti-IL-6R therapeutics as a tailored approach to treatment.
One easily detected measure of STAT3 activation is phosphorylation at tyrosine 705. We assessed ex vivo STAT3 activation in CD4 + T cells using this parameter. We found that patients with polyJIA had a trend toward decreased newly activated STAT3, particularly after IL-6 stimulation, but this finding was not significant (Fig.
2). Others have also found no difference in phosphorylated STAT3 after IL-6 stimulation in samples from polyJIA patients compared to controls [
52]. Interestingly, samples from patients with the rare inborn error of immunity STAT3 gain-of-function, the result of genetic mutations in STAT3 that lead to higher transcriptional capacity than wild-type STAT3, also have decreased phosphorylation of STAT3 after ex vivo inflammatory cytokine stimulation [
23]. Since inflamed patients with polyJIA would be expected to have higher activation of the IL-6/STAT3 inflammatory pathway, we had expected to see increases and were surprised by our finding, although this has also been reported in patients with RA [
12,
53]. Decreases in newly phosphorylated STAT3 after stimulation may actually reflect higher baseline activation of STAT3 preserved in treatment-naïve samples. Further, decreased responsiveness of patient CD4 + T cells to ex vivo stimulus may result from compensatory mechanisms due to tonic STAT3 activation in vivo [
54]. Additionally, as this is only one measure of STAT3 activation, it may be that other ways to measure activation may reveal increases. However, the idea that the IL-6/STAT3 inflammatory pathway is at a higher state in vivo in our patients with polyJIA is support by the findings of increased frequencies of Th17 cells (Fig.
1).
It is increasingly clear that CD4 + T cells marked by expression of IL-17 are a heterogeneous population with both homeostatic and inflammatory properties [
16]. However, the true complexity of the roles of IL-17 expressing CD4 + T cells in human autoimmune disease, both in the blood and at local site of inflammation, are still being unraveled. The plasticity of Th17 cells results in the shifting of classic Th17 cells into dual-cytokine expressing cells (IL-17 + IFNγ +), termed Th17/1 cells, as well as ex-Th17 cells, which no longer express IL-17, but retain markers that identify them as prior Th17 cells [
18]. How these subtypes may function in the pathogenesis of JIA is not known, as little work thus far has evaluated these subtypes in JIA patients. In one report, investigators found that ex-Th17 cells were not elevated in the peripheral blood of patients with oligoarticular JIA compared to controls, but were elevated in synovial fluid compared to peripheral blood of these patients [
47]. Others reported elevated Th17/1 cells in the synovial fluid of patients with JIA [
55]. Noting the presence of dual-cytokine expressing CD4 + T cells in polyJIA patients not really detectable in controls (Fig.
3), lead us to evaluate further for Th17/1 and ex-Th17 cells in patient samples. We found that frequencies of both Th17/1 and ex-Th17 cells increased in samples from patients on treatment following remission compared to their treatment-naïve baseline (Fig.
4). This finding is contrary to prior work reporting decreases in ex-Th17 cells in patients with oligoarticular JIA on treatment [
56]. This discrepancy might be explained by differences in the inherent disease processes (peripheral analysis in oligoarticular versus polyJIA) or the methodology of the analysis, as the prior work compared different treatment-naïve and post-treatment individuals, but we were able to compare the changes in ex-Th17 cell frequencies in matched treatment-naïve and post-treatment samples from the same patients. The presence of circulating and intra-articular Th17/1 and ex-Th17 cells must be evaluated in additional cohorts of patients with JIA.
It was notable to us that two polyJIA patients had markedly high elevations of post-treatment IL-17 secreting cells: Th17 cells (Fig.
1B), dual-cytokine expressing cells (Fig.
3C), and Th17/1 cells (Fig.
4B). Review of the clinical course of these individuals compared to the other patients in the polyJIA cohort revealed that they have remained therapy-bound > 4 years after attaining clinical remission on medication, by which we mean repeated attempts to wean their medications have led to prompt return of arthritis (Supplementary Figure
1). Current guidelines for treatment of polyJIA do not include recommendations for treatment withdrawal [
57], as there is insufficient evidence that clinical parameters can guide these decisions and effective biomarker-based strategies have not yet been developed [
58‐
60]. A subset of patients with JIA are likely destined to remain on treatment for best outcomes, but sustained, drug-free, inactive disease is possible for some [
61]. The capacity to use a biomarker to distinguish between these subsets of patients would be a remarkable addition to the treatment arsenal of pediatric rheumatologists. We are interested in further pursuing the idea that a detectable peripheral rise in IL-17 expressing CD4 + T cells, either as Th17 cells or as a shift into dual-cytokine expressing Th17/1 cells, marks a subset of polyJIA patients who will remain therapy-bound in order to retain a clinically inactive disease state.
Our study has several limitations. It is a single-institution study of a small cohort of patients with one subtype of JIA, although it is a large cohort of treatment-naïve samples. The study was purely observational, as we had no influence on the recommendations for treatment or withdrawal made by the patients’ primary rheumatologists. Additionally, study of cells circulating in the periphery of patients does not address the tissue inflammatory environment, but it does allow for serial assessments in an accessible sample type.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.