Introduction
While the pathogenesis of juvenile idiopathic arthritis (JIA) is not completely understood, fibroblast-like synoviocytes (FLS) are thought to play a critical role in JIA progression and may contribute to increased inflammation associated with disease progression [
1,
2]. In this study, we delve deeper into understanding FLS as effector cells for disease progression and the impact of advanced therapeutics on this cell type.
The current classification scheme for JIA includes seven subtypes of disease based on onset [
3,
4]. In this study we focused on the persistent oligoarticular subtype (persistent), which is the most benign, and extended oligoarticular subtype, in which there is a greater chance of disability when joint involvement evolves from a persistent oligoarticular to a polyarticular disease course. Specifically, we examined FLS and synovial fluid from patients with a persistent disease course and compared them to patients prior to extension to a polyarticular course, termed extended-to-be (ETB), to observe differences in inflammatory profiles before these patients evolved to a more severe disease course. Specifically, our objective was to show that disease originates within the FLS and that targeting cells directly with advanced therapeutics can have a direct impact on disease progression, particularly in the patients that eventually extend to a more severe disease course. Furthermore, our data suggests that attenuating inflammation within the FLS can lower the overall inflammatory response typically promoted by FLS.
In rheumatoid arthritis (RA), studies have shown that FLS express tumor necrosis factor alpha (TNFα) and secretion of TNFα contributes to the increased inflammatory response associated with active disease [
2,
5‐
8]. Furthermore, studies have demonstrated links between transforming growth factor beta (TGFβ) signaling and TNFα expression in various chronic inflammatory diseases and that anti-TNFα biologic medications cause changes in the TGFβ signaling pathway [
9‐
11]. On the basis of these studies, we examined the impact of adalimumab, an anti-TNFα biologic disease-modifying anti-rheumatic drug (bDMARD), on cytokine expression in FLS. Specifically, we measured global inflammatory profiles in FLS isolated from synovial fluid of patients prior to them extending to a polyarticular disease course and observed differences in expression of TNFα and TNFα-related proteins. We treated cells with adalimumab to determine if this advanced therapy had a direct impact on FLS. The emergence of biologics that target specific cytokines has led to dramatic improvements in the treatment of JIA without widespread risks to the patients [
12]. Distinguishing between persistent and ETB is critical to the study of JIA because patients with these subtypes have the same clinical presentation at diagnosis, making it difficult for clinicians to determine which patients will extend to a more severe disease course. Understanding how ETB JIA differs from persistent JIA could allow for intervention with advanced therapies earlier in the disease course and possibly prevent extension to a more severe disease course. Specifically, treating the FLS directly with advanced medications could prohibit promotion of the inflammatory response in more severe subtypes and potentially prevent patients from extending to a polyarticular disease course.
Methods
Selection of Samples
Samples were selected using the same criteria from our previous work [
13]. Briefly, synovial fluid samples were selected from a repository approved by our Nemours Foundation BioMedical Research Institutional Review Board (IRB#84709-30). Written informed consent was obtained from patients who had clinically indicated arthrocentesis and were offered participation in the repository. This study was performed in accordance with the Helsinki Declaration of 1964, and its later amendments. In our database, 320 patients with oligoarticular JIA were identified. Patients with or diagnoses of psoriatic, inflammatory bowel disease-related, Lyme, syndromic arthritis, and polyarticular JIA were excluded, leaving 83 individuals who underwent one or more arthrocentesis. Out of the 83 patients identified, 41 patients with oligoarticular JIA who underwent an intra-articular arthrocentesis while the patient was on no medications or nonsteroidal anti-inflammatory drugs (NSAIDs) only, and before having received any disease-modifying anti-rheumatic drugs (DMARDs) or biologic disease-modifying anti-rheumatic drugs (bDMARDs), were selected. All patients were followed between 1990 and 2021.
Of the 41 patients who met inclusion criteria and adhere to International League for Associations in Rheumatology (ILAR) classification criteria, 23 patients were in the persistent oligoarticular JIA group and 18 in the extended oligoarticular group. Fifteen of the 23 patients in the persistent group received no medication or received NSAIDs only during their disease course, whereas eight eventually received DMARDs/bDMARDs. From this group, for the cell culture experiments, we selected three persistent oligoarticular synovial fluid samples from which to isolate FLS. All samples had no prior steroid injections, were only on NSAIDs or no medications at time fluid was taken, never received DMARDs/bDMARDs, and were followed for at least 2 years to ensure they remained persistent and did not extend to a more severe disease course. In the extended oligoarticular JIA group, 17 of the 18 patients eventually received DMARDs/bDMARDs, while one did not and remained on no medications. From this group, for the cell culture experiments, we selected three ETB synovial fluid samples from which to isolate FLS. All samples had no prior steroid injections, were only on NSAIDs or no medications at time fluid was taken, had not received DMARDs/bDMARDs at time of injection, and extended after the first 6 months of disease. All selected synovial fluid samples analyzed in this study were obtained before patients received advanced medications, including DMARDs or bDMARDs. At the time of the arthrocentesis, patients were taking either NSAIDs or no medications. If the initial synovial fluid sample could not be located, a sample obtained at least 3 months after the initial arthrocentesis was selected. Paired samples were two samples collected at least 3 months apart from the same patient. For paired synovial fluid samples, all patients were either on NSAIDs or no medications at time of injection, were collected at least 3 months apart, and had received no DMARDs or bDMARDs. All synovial fluid samples were taken from knee joints.
Cell Culture and Adalimumab Treatment
We selected primary cells from three subjects with persistent oligoarticular JIA and three subjects who eventually extended to a polyarticular disease course (termed extended-to-be, ETB). All samples had no prior steroid injections and were on NSAIDs or no medications at time synovial fluid was obtained. Synovial fluid was spun at 500 rpm for 5 min, and cell pellets were plated in T-25 culture flasks with 15% fetal bovine serum/Dulbecco’s modified Eagle’s medium. Initial primary cultures were passaged into T-75 flasks and, at confluence, passaged three more times, to ensure FLS were the predominant cell type in culture. FLS from persistent oligoarticular and ETB oligoarticular JIA were grown in medium containing 15% FBS. At sub-confluence, media were removed, and cells were incubated overnight in growth medium containing 0.2% serum. The serum-starved cells were then seeded on six-well plates for experiments. Cells were stimulated with either 10 µg/ml adalimumab [
14] or normal media for 24 h. After 24 h, cell media supernatants were obtained from each condition. Power analysis was performed on the basis of synovial fluid data to determine the number of FLS samples needed to meet 0.80 power with an error rate of 5%. On the basis of this data, it was determined that for inflammatory proteins measured, two to four samples were needed to reach a power of 0.80 with a 5% error rate. As a result of the number of proteins analyzed, we chose the median of three samples per group.
Antibody Arrays and Proteomic Analysis
Data analysis was performed using the same techniques from our previous work [
13]. Briefly, protein concentrations were calculated using a Bradford assay. RayBiotech performed Human Inflammation Array C3 (AAH-INF-3) according to the manufacturer’s protocols. The array membranes were processed similarly to a western blot (chemiluminescent readout). Densitometry and background correction was performed. Averages were calculated between technical duplicates for each sample. Semiquantitative changes in protein expression were determined and compared between JIA subtypes. Scaling was performed by converting intensity measurements or relative protein amounts, to the log
10 value. Statistical analysis was performed using Excel. For full data calculations please reference our previous work [
13]. Briefly, we calculated average log
10 values and standard deviation for the combined biological replicates in each group. Using a Welsh’s
t test, we determined significant differences between two comparisons as having a
p value of less than 0.05.
Discussion
Our data suggests that when well-characterized markers of inflammation were measured, FLS isolated from synovial fluid of patients with ETB JIA had increased expression of inflammatory markers, namely TNFα-related cytokines, when compared to FLS isolated from synovial fluid of patients with persistent JIA, suggesting that FLS may contribute to the overall inflammation associated with disease pathogenesis. When these cells were treated with TNFα inhibitor, adalimumab, expression of pro-inflammatory proteins downstream of TNFα signaling were significantly reduced in ETB FLS treated with adalimumab.
FLS are the major component of the synovial membrane and contribute to inflammation by producing pro-inflammatory cytokines and participating in cell-to-cell communication to recruit inflammatory cells to affected joints [
18,
19]. FLS are thought to play a key role in disease pathogenesis in JIA [
2] which prompted us to examine the inflammatory profiles of FLS isolated from synovial fluid of patients with persistent and ETB JIA and study the effects of adalimumab on this cell type. ETB FLS have increased expression of inflammatory promoters TGFβ and LTα and both TNFα receptors, suggesting that targeting the TNFα signaling pathway earlier in the disease course could be effective at delaying or preventing disease extension.
It has been shown that cell death can induce inflammation in autoimmune diseases and chronic inflammatory diseases [
20]. Specifically, binding of TNFα to TNFR1 exacerbates inflammation by inducing cell death [
21]. This is thought to be achieved by overcoming cell death checkpoints that allow for the activation of downstream inflammatory pathways MAPK and NF-κB [
22,
23]. TGFβ signaling acts as an intermediary activator of both MAPK and NF-κB activation, downstream of TNFα binding to TNFRI [
22,
23]. Specifically, TNFα induces TGFβ expression in lung fibroblasts and, conversely, TGFβ1 caused increased expression of TNFα in FLS isolated from patients with RA, and subsequent activation of NF-κB signaling, suggesting TGFβ is downstream of TNFα and can contribute directly to the inflammatory response of FLS [
10,
24].
Secretion of TNFα-related cytokines by FLS prompted us to examine inflammatory cytokine profiles in synovial fluid of patients with JIA. Inflammatory cytokine expression is significantly elevated in synovial fluid from patients who eventually extend to a more severe disease course when compared to those patients who remain persistent. In RA, patients with chronic inflammation have elevated IL-1, IL-6, and TNFα [
25] as well as elevated GM-CSF, IL-10, and TGFβ [
26]. Similarly, we observed increased expression of these cytokines in synovial fluid from patients with ETB JIA when compared to patients with a persistent disease course.
When we tested a subset of synovial fluid from patients who eventually required either a DMARD/bDMARD but prior to receiving an intra-articular corticosteroid injection or DMARD/bDMARD treatment in persistent and ETB JIA, we observed significant elevation of inflammatory cytokines related to TNFα signaling in patients with ETB JIA when compared to persistent. When we measured cytokine protein expression after patients received an intra-articular corticosteroid, we still observed significant increases in TNFα and cytokines downstream of TNFα. Elevation of these cytokines in synovial fluid from patients with ETB JIA who eventually require a DMARD/bDMARD could indicate the need for introduction of DMARD/bDMARD earlier in the disease course. Specifically, a TNFα inhibitor introduced earlier in the disease course may effectively delay extension to a polyarticular course in these patients. While there currently exists no clinical indicators for patients who will eventually extend to a more severe disease course, global protein expression of cytokines could be used to indicate the need for intervention with advanced medications earlier in the disease course.
Anti-TNF agent adalimumab is a neutralizing antibody that effectively binds to soluble and membrane-bound TNFα and prevents receptor/ligand binding to inhibit TNFα-driven inflammation [
27,
28]. In patients treated with adalimumab, a response to treatment is usually seen in 131 ± 56 h [
29]. Here, we treated FLS isolated from synovial fluid of patients with a persistent and ETB disease course for 24 h, to determine if a short course of treatment could have a direct effect on FLS in culture. In both persistent FLS and ETB FLS treated with adalimumab, TNFα remained elevated compared to respective untreated cells. In the persistent FLS, TGFβ1 was also still significantly overexpressed in treated cells compared to untreated cells. Berkhout et al. showed that circulating TNF actually increased in patients with RA treated with adalimumab and they saw an increase in healthy volunteers after one dose of adalimumab, implying that TNF ligand in circulation may not be associated with disease activity in patients with RA [
30]. Interestingly, TGFβ expression levels significantly decreased in ETB FLS treated with adalimumab when compared to untreated ETB FLS, suggesting that after 24 h, adalimumab may have an effect downstream of TNFα on TGFβ signaling. In persistent FLS, sTNFRI and sTNFRII were significantly overexpressed in adalimumab-treated cells, suggesting that soluble receptor expression in response to elevated TNFα could bind available ligand and prevent membrane-bound receptor signaling. Interestingly, there were no significant differences in the soluble receptors in ETB FLS treated with adalimumab. Increased expression of sTNFRI could indicate increased cell death in persistent FLS compared to ETB FLS [
20,
31]. Increased cell death has been linked to increased inflammation in chronic inflammatory diseases [
22,
23,
31‐
33].
Adalimumab may induce a faster response in more severe disease in JIA. Our findings show that TGFβ expression is lower in ETB FLS treated with adalimumab compared to untreated cells. Both soluble TNF receptors protein expression were not significantly different and IFNγ expression was significantly lower in ETB FLS treated with adalimumab compared to untreated cells, suggesting that ETB FLS are not secreting these downstream signaling proteins that activate other inflammatory cell types like macrophages and neutrophils [
34].
While our data demonstrates increased inflammation in the ETB subtype of JIA when compared to persistent JIA in both FLS isolated from synovial fluid and synovial fluid samples, we recognize there are some limitations to this study. Cells were treated with adalimumab for 24 h. While in the more severe disease course we saw downregulation of TNFα downstream signaling, with prolonged exposure using a similar timeline to in vivo studies we may see more pronounced effects of adalimumab on FLS directly.