FLS Isolated from Synovial Fluid of Patients Prior to Disease Extension to a More Severe Disease Course Overexpress Pro-inflammatory Cytokines Related to TNFα Signaling

FLS are thought to play a critical role in disease pathogenesis in JIA [2, 15]. Using protein antibody arrays to measure relative changes in protein expression of 40 different cytokines, we tested FLS isolated from synovial fluid of six patients. FLS from patients that remained persistent throughout disease course were compared to FLS from patients that eventually extended to a more severe disease course (ETB). Of the 40 cytokines, six proteins were elevated in ETB FLS compared to persistent FLS (Fig. 1). Transforming growth factor beta 1 (TGFβ1) had a 1.1-fold increase in ETB FLS compared to persistent FLS (p = 0.022) (Fig. 1). TNFα family members, lymphotoxin alpha (LTα) (fold change [FC] = 3.5 p = 0.014), soluble TNF receptor 1 (sTNFRI) (FC = 4.4 p = 0.014), and soluble TNF receptor II (sTNFRII) (FC = 3.4 p = 0.036) were significantly elevated in ETB FLS compared to persistent FLS (Fig. 1). Platelet-derived growth factor (PDGF), which can stimulate TNFα synthesis and has been shown to be elevated in RA FLS when TGFβ1 expression is increased [16, 17], had a 3.5-fold increase in ETB FLS compared to persistent FLS (p = 0.035) (Fig. 1). Lastly, interleukin-3 (IL-3), a promoter of inflammatory cell proliferation, was elevated in ETB FLS compared to persistent FLS (FC = 1.1 p = 0.008) (Fig. 1). This data suggests that FLS from ETB, prior to extension, secrete higher levels of cytokines that induce TNFα signaling when compared to those patients who remain persistent throughout disease course.

Fig. 1

Protein antibody array on FLS isolated from synovial fluid from patients with persistent and ETB JIA revealed elevation of proteins related to TNFα and TGFβ. We measured 40 cytokines in 3 samples from persistent JIA and 3 samples from ETB JIA. All patients were steroid naïve and on either no medication or only NSAIDs at the time of the arthrocentesis. All synovial fluid samples from the ETB group were taken prior to their disease extending to a more severe disease course. TGFβ1 had a 1.1-fold increase in ETB FLS compared to persistent FLS (p = 0.022). LTα (FC = 3.5 p = 0.014), sTNFRI (FC = 4.4 p = 0.014), sTNFRII (FC = 3.4 p = 0.036) were significantly elevated in ETB FLS compared to persistent FLS. PDGF had a 3.5-fold increase in ETB FLS compared to persistent FLS (p = 0.035). Interleukin-3 (IL-3) was elevated in ETB FLS compared to persistent FLS (FC = 1.1 p = 0.008). FLS fibroblast-like synoviocytes, JIA juvenile idiopathic arthritis, ETB extended-to-be, TNFα tumor necrosis factor alpha, TGFβ1 transforming growth factor beta 1, NSAIDs nonsteroidal anti-inflammatory drugs, FC fold change, LTα lymphotoxin alpha, sTNFRI soluble TNF receptor 1, sTNFRII soluble TNF receptor 2, PDGF platelet-derived growth factor, IL interleukin

There is an Elevated Inflammatory Response in Synovial Fluid from Patients with ETB JIA When Compared to Synovial Fluid from Patients with Persistent JIA

Since FLS isolated from synovial fluid of patients with ETB JIA secrete factors that induce TNFα signaling, we examined the overall effects on inflammatory cytokines in the milieu of synovial fluid. Using protein antibody arrays to measure relative changes in protein expression of 40 different cytokines, we tested 41 samples of synovial fluid from patients with JIA. We divided these synovial fluid samples into two groups: 23 samples from those that remained persistent throughout their disease course and 18 ETB samples. Of the 40 cytokines measured, 22 were significantly elevated in the ETB group when compared to persistent, suggesting that those patients who eventually extend to a more severe disease course have an elevated inflammatory response and dysregulation of their inflammatory network when compared to persistent samples (Fig. 2). TGFβ1, an activator of T helper cells, was increased 1.3-fold in ETB compared to persistent (p = 0.008) (Fig. 2). Similarly, TNFα, a major pro-inflammatory protein, was increased 1.3-fold in ETB compared to persistent (p = 0.026) (Fig. 2). LTα, a member of the TNF family, was also elevated 1.3-fold in ETB compared to persistent (p = 0.037) (Fig. 1). Interleukin-6 (IL-6), a notable marker of disease activity in JIA and RA, was elevated 1.2-fold in ETB compared to persistent (p = 0.024) (Fig. 2). Additionally, there were 14 pro-inflammatory cytokines that were significantly elevated between one- and threefold in ETB compared to persistent (p < 0.05) (Fig. 2). In response to pro-inflammatory protein levels, anti-inflammatory proteins interleukin-10 (IL-10) (FC = 2.3 p = 0.000) and interleukin-11 (IL-11) (FC = 1.7 p = 0.002) were elevated in ETB compared to persistent (Fig. 2). Notably, tissue inhibitor of metalloproteinases-2 (TIMP-2), was significantly less in ETB compared to persistent (p = 0.001), suggesting a lack of inhibition of metalloproteinases (Fig. 2). Taken together with findings in FLS, this data suggests that ETB JIA have a more robust inflammatory response than persistent JIA and that this could possibly be driven by FLS.

Fig. 2

Protein antibody array on synovial fluid from patients with persistent and ETB JIA reveal an increase in inflammatory proteins in ETB samples compared to persistent samples. We compared 23 samples from persistent JIA to 18 samples from ETB JIA. All patients were steroid naïve and on either no medication or only NSAIDs at the time of the arthrocentesis. All synovial fluid samples from the ETB group were taken prior to their disease extending to a more severe disease course. TGFβ1 was increased 1.3-fold in ETB compared to persistent (p = 0.008). TNFα was increased 1.3-fold in ETB compared to persistent (p = 0.026). LTα was also elevated 1.3-fold in ETB compared to persistent (p = 0.037). IL-6 was elevated 1.2-fold in ETB compared to persistent (p = 0.024). There were 14 pro-inflammatory cytokines that were significantly elevated between one- and threefold in ETB compared to persistent (p < 0.05). Anti-inflammatory proteins IL-10 (FC = 2.3 p = 0.000) and IL-11 (FC = 1.7 p = 0.002) were elevated in ETB compared to persistent. TIMP-2 was significantly less in ETB compared to persistent (p = 0.001). ETB extended-to-be, JIA juvenile idiopathic arthritis, NSAIDs nonsteroidal anti-inflammatory drugs, TGFβ1 transforming growth factor beta 1, TNFα tumor necrosis factor alpha, LTα lymphotoxin alpha, IL interleukin, FC fold change, TIMP-2 tissue inhibitor of metalloproteinases-2, MIP-1-alpha macrophage inflammatory protein 1 alpha, MIP-1-beta macrophage inflammatory protein 1 beta, M-CSF macrophage colony-stimulating factor, CCL chemokine ligand, IFN-gamma interferon-γ, GM-CSF granulocyte-macrophage colony-stimulating factor, G-CSF granulocyte colony-stimulating factor, MIG monokine induced by interferon-γ

Of All Patients Who Eventually Required a DMARD/bDMARD, Synovial Fluid Samples from Patients with ETB JIA had an Elevated Inflammatory Response When Compared to Synovial Fluid Samples from Patients with Persistent JIA

On the basis of our data that FLS isolated from synovial fluid of patients with ETB JIA have increased secretion of TNFα-related cytokines and may contribute to an elevated inflammatory response in synovial fluid from patients with ETB, we tested a subset of 25 samples of synovial fluid from patients with persistent and ETB JIA who eventually required either a DMARD or bDMARD. Using protein antibody arrays, we measured relative changes in protein expression of 40 different cytokines between eight samples from those that remained persistent throughout their disease course and 17 ETB samples. TNF family member LTα (p = 0.040) was elevated 1.5-fold in ETB who eventually required a DMARD/bDMARD compared to persistent who eventually required a DMARD/bDMARD (Fig. 3a). TGFβ1 was elevated 1.4-fold in ETB who eventually required a DMARD/bDMARD compared to persistent who eventually required a DMARD/bDMARD (p = 0.003) (Fig. 3a). In addition to these TNFα-related cytokines, there were 18 pro-inflammatory cytokines (FC between 1.5-fold and 9.0-fold) that were significantly elevated in ETB who eventually required a DMARD/bDMARD compared to persistent who eventually required a DMARD/bDMARD (p < 0.05) (Fig. 3a). In response to this increase in inflammatory cytokines, IL-10 (FC = 8.5 p = 0.000) and IL-11 (FC = 4.2 p = 0.000), anti-inflammatory proteins were significantly elevated in ETB who eventually required a DMARD/bDMARD compared to persistent who eventually required a DMARD/bDMARD (Fig. 3a). The elevation of LTα and TGFβ suggests that patients with ETB JIA who eventually require a DMARD/bDMARD may benefit from TNFα inhibitors earlier in the disease course.

Fig. 3

Protein antibody array on synovial fluid from patients with persistent and ETB JIA who eventually required a DMARD or bDMARD reveal that prior to receiving advanced therapies, patients who eventually extend to a more severe disease course have elevated inflammatory protein expression. We measured protein expression of 40 cytokines between 8 samples from those that remained persistent throughout their disease course and 17 ETB samples (a). All patients had not yet received a DMARD or bDMARD, were steroid naïve, and on either no medication or NSAIDs only at the time of arthrocentesis and all synovial fluid from the ETB group were taken prior to their disease extending to a more severe disease course (a). LTα (p = 0.040) was elevated 1.5-fold in ETB who eventually required a DMARD/bDMARD compared to persistent who eventually required a DMARD/bDMARD. TGFβ1 was elevated 1.4-fold in ETB who eventually required a DMARD/bDMARD compared to persistent who eventually required a DMARD/bDMARD (p = 0.003). There were 18 pro-inflammatory cytokines (FC between 1.5-fold and 9.0-fold) that were significantly elevated in ETB who eventually required a DMARD/bDMARD compared to persistent who eventually required a DMARD/bDMARD (p < 0.05). IL-10 (FC = 8.5 p = 0.000) and IL-11 (FC = 4.2 p = 0.000) were significantly elevated in ETB who eventually required a DMARD/bDMARD compared to persistent who eventually required a DMARD/bDMARD. We tested serial synovial fluid samples from the 25 persistent and ETB samples (b). At the time of arthrocentesis, none of the patients had received a DMARD or bDMARD, had only one previous steroid injection, and were on no medication or NSAIDs only (b). TNFα (p = 0.003) and LTα (p = 0.007) were both elevated 1.8-fold in ETB who eventually required a DMARD/bDMARD compared to persistent who eventually required a DMARD/bDMARD. TGFβ1 was significantly elevated in ETB who eventually required a DMARD/bDMARD compared to persistent who eventually required a DMARD/bDMARD (FC = 2.1-fold p = 0.001). In addition to these promoters of inflammation, 20 other cytokines were elevated in ETB who eventually required a DMARD/bDMARD compared to persistent who eventually required a DMARD/bDMARD after a single intra-articular corticosteroid injection (FC range 1.1–5.1 p < 0.050) (c). ETB extended-to-be, JIA juvenile idiopathic arthritis, DMARD disease-modifying anti-rheumatic drug, bDMARD biologic disease-modifying anti-rheumatic drug, LTα lymphotoxin alpha, NSAIDs nonsteroidal anti-inflammatory drugs, IL interleukin, TGFβ1 transforming growth factor beta 1, FC fold change, M-CSF macrophage colony-stimulating factor, MIP-1 macrophage inflammatory protein 1, TIMP-2 tissue inhibitor of metalloproteinases-2, G-CSF granulocyte colony-stimulating factor, CCL chemokine ligand, IFN-gamma interferon-γ, GM-CSF granulocyte-macrophage colony-stimulating factor, MIG monokine induced by interferon-γ

Of All Patients Who Eventually Required a DMARD/bDMARD, TNFα Remains Elevated in ETB After Intra-articular Corticosteroid Injection Compared to Persistent After Intra-articular Corticosteroid Injection

To determine if intra-articular corticosteroids change the expression of inflammatory promoter TNFα, we tested serial synovial fluid samples from the 25 persistent and ETB samples analyzed above using protein antibody arrays. At the time of arthrocentesis, none of the patients had received a DMARD or bDMARD, had only one previous steroid injection, and were on no medication or NSAIDs only. In the synovial fluid samples of the subsequent arthrocentesis, TNF family members TNFα (p = 0.003) and LTα (p = 0.007) were both elevated 1.8-fold in ETB who eventually required a DMARD/bDMARD compared to persistent who eventually required a DMARD/bDMARD (Fig. 3b). TGFβ1 was significantly elevated in ETB who eventually required a DMARD/bDMARD compared to persistent who eventually required a DMARD/bDMARD (FC = 2.1-fold p = 0.001) (Fig. 3b). In addition to these TNFα-related promoters of inflammation, after a single intra-articular corticosteroid injection, 20 other cytokines were elevated in ETB who eventually required a DMARD/bDMARD compared to persistent who eventually required a DMARD/bDMARD (FC range 1.1–5.1 p < 0.050) (Fig. 3c). This data suggests that while intra-articular corticosteroids may initially lower inflammatory response in those patients who eventually extend to a more severe disease course, intra-articular corticosteroids did not alter overall inflammatory response in the long term.

Adalimumab is Effective at Targeting Downstream TNFα Signaling and Overall Inflammatory Response in FLS Isolated from Synovial Fluid of Patients with ETB JIA

We showed that FLS isolated from synovial fluid of patients with ETB JIA have increased TNFα expression and that this may contribute to the elevated inflammatory response seen in synovial fluid from patients with ETB JIA that eventually required a DMARD/bDMARD. Therefore, using protein antibody arrays we examined the effects of a short course (24 h) of adalimumab, a TNFα inhibitor, on inflammatory cytokines in FLS isolated from synovial fluid of six patients. In the persistent FLS treated with adalimumab for 24 h, 17 of the 40 inflammatory proteins remained significantly elevated when compared to untreated cells (Fig. 4). Specifically, TNFα (FC = 1.2 p = 0.001) and TGFβ (FC = 1.5 p = 0.001) were significantly elevated in persistent FLS treated with adalimumab compared to untreated persistent FLS (Fig. 4). LTα (FC = 4.3 p = 0.015), sTNFRI (FC = 5.1 p = 0.008), and sTNFRII (FC = 3.8 p = 0.025) were significantly overexpressed in persistent FLS treated with adalimumab compared to untreated persistent FLS (Fig. 4). The remaining 12 markers of inflammation fold change ranged from 1.1 to 4.4 increase in persistent FLS treated with adalimumab compared to untreated FLS (p < 0.05) (Fig. 4). Interestingly, this data suggests that one dose of adalimumab for 24 h is not effective at inhibiting an inflammatory response in persistent FLS, the least severe disease course of JIA.

Fig. 4

Adalimumab, a TNFα inhibitor, did not decrease expression of pro-inflammatory cytokines in FLS isolated from synovial fluid of patients with persistent JIA after 24 h of treatment. All patients were steroid naïve and on no medication or only NSAIDs at the time of the arthrocentesis. All FLS isolated from synovial fluid from the ETB group were taken prior to their disease extending to a more severe disease course. TNFα (FC = 1.2 p = 0.001), TGFβ (FC = 1.5 p = 0.001), LTα (FC = 4.3 p = 0.015), sTNFRI (FC = 5.1 p = 0.008), and sTNFRII (FC = 3.8 p = 0.025) were significantly elevated in persistent FLS treated with adalimumab compared to untreated persistent FLS. The remaining 12 markers of inflammation fold change ranged from 1.1 to 4.4 increase in persistent FLS treated with adalimumab compared to untreated FLS (p < 0.05). TNFα tumor necrosis factor alpha, FLS fibroblast-like synoviocytes, JIA juvenile idiopathic arthritis, ETB extended-to-be, NSAIDs nonsteroidal anti-inflammatory drugs, TGFβ transforming growth factor beta, LTα lymphotoxin alpha, FC fold change, sTNFRI soluble TNF receptor 1, sTNFRII soluble TNF receptor 2, IL interleukin, MIP-1 macrophage inflammatory protein 1, PDGF platelet-derived growth factor

Adalimumab is Effective at Decreasing Inflammation that Occurs Downstream of Initial TNFα Signaling After 24 h in FLS Isolated from Synovial Fluid of Patients Who Eventually Extend to a More Severe Disease Course

In the ETB FLS treated with adalimumab eight of the 40 inflammatory markers had significant changes in protein expression compared to the untreated ETB FLS. Interestingly, seven proteins decreased after exposure to adalimumab for 24 h and only TNFα remained significantly elevated in treated ETB FLS compared to untreated ETB FLS (FC = 1.04 p = 0.023) (Fig. 5). TGFβ (FC = 1.03 p = 0.037) was significantly decreased in treated ETB FLS compared to untreated ETB FLS (Fig. 5). Additionally, granulocyte colony-stimulating factor (G-CSF), interferon gamma (IFNγ), IL-10, macrophage colony-stimulating factor (M-CSF), macrophage inflammatory protein 1 alpha (MIP-1-alpha), and macrophage inflammatory protein 1 delta (MIP-1-delta) were all significantly decreased in treated ETB FLS compared to untreated ETB FLS (FC range from 1 to 3.6 p < 0.05) (Fig. 5). This data suggests that although adalimumab did not lower TNFα expression in ETB FLS, after 24 h, adalimumab is effective at decreasing inflammation that occurs downstream of initial TNFα signaling after a short time. Differing responses to adalimumab between persistent and ETB FLS are another distinguishable characteristic between these two JIA subtypes.

Fig. 5

Effects of a short course (24 h) of adalimumab, a TNFα inhibitor, on inflammatory cytokines in FLS isolated from synovial fluid of patients with ETB JIA reveal a decrease in pro-inflammatory cytokines downstream of TNFα. In the ETB FLS treated with adalimumab, TNFα remained significantly elevated compared to untreated ETB FLS (FC = 1.04 p = 0.023). TGFβ (FC = 1.03 p = 0.037) was significantly decreased in treated ETB FLS compared to untreated ETB FLS. G-CSF, IFNγ, IL-10, M-CSF, MIP-1-alpha, and MIP-1-delta were all significantly decreased in treated ETB FLS compared to untreated ETB FLS (FC range from 1 to 3.6 p < 0.05). TNFα tumor necrosis factor alpha, FLS fibroblast-like synoviocytes, JIA juvenile idiopathic arthritis, ETB extended-to-be, FC fold change, TGFβ1 transforming growth factor beta 1, G-CSF granulocyte colony-stimulating factor, IFNγ interferon-γ, IL interleukin, M-CSF macrophage colony-stimulating factor, MIP-1 macrophage inflammatory protein 1