Regulation of the JNK pathway by TGF-beta activated kinase 1 in rheumatoid arthritis synoviocytes

c-Jun N-terminal kinase (JNK) contributes to metalloproteinase (MMP) gene expression and joint destruction in inflammatory arthritis. It is phosphorylated by at least two upstream kinases, the mitogen-activated protein kinase kinases (MEK) MKK4 and MKK7, which are, in turn, phosphorylated by MEK kinases (MEKKs). However, the MEKKs that are most relevant to JNK activation in synoviocytes have not been determined. These studies were designed to assess the hierarchy of upstream MEKKs, MEKK1, MEKK2, MEKK3, and transforming growth factor-β activated kinase (TAK)1, in rheumatoid arthritis (RA). Using either small interfering RNA (siRNA) knockdown or knockout fibroblast-like synoviocytes (FLSs), MEKK1, MEKK2, or MEKK3 deficiency (either alone or in combination) had no effect on IL-1β-stimulated phospho-JNK (P-JNK) induction or MMP expression. However, TAK1 deficiency significantly decreased P-JNK, P-MKK4 and P-MKK7 induction compared with scrambled control. TAK1 knockdown did not affect p38 activation. Kinase assays showed that TAK1 siRNA significantly suppressed JNK kinase function. In addition, MKK4 and MKK7 kinase activity were significantly decreased in TAK1 deficient FLSs. Electrophoretic mobility shift assays demonstrated a significant decrease in IL-1β induced AP-1 activation due to TAK1 knockdown. Quantitative PCR showed that TAK1 deficiency significantly decreased IL-1β-induced MMP3 gene expression and IL-6 protein expression. These results show that TAK1 is a critical pathway for IL-1β-induced activation of JNK and JNK-regulated gene expression in FLSs. In contrast to other cell lineages, MEKK1, MEKK2, and MEKK3 did not contribute to JNK phosphorylation in FLSs. The data identify TAK1 as a pivotal upstream kinase and potential therapeutic target to modulate synoviocyte activation in RA.


Introduction
Rheumatoid arthritis (RA) is a chronic inflammatory disease characterized by synovial lining hyperplasia and sublining infiltration of inflammatory cells [1]. Fibroblast-like synoviocytes (FLSs) play a crucial role in joint damage as well as the propagation of inflammation [2]. In response to potent pro-inflammatory cytokines such as IL-1β, FLSs produce large amounts of matrix metalloproteinases (MMP), which are key drivers of matrix destruction [3][4][5]. MMP production is, in turn, regulated by several signal transduction pathways, including the mitogen-activated protein kinases (MAPKs) [6,7].
Multiple upstream MAPKK kinases (MAP3Ks) that activate the MAPKKs and the JNK cascade have been identified in RA. For instance, MEK kinase (MEKK)1, MEKK2, and transforming growth factor-β activated kinase (TAK)1 are the most abundant in inflamed synovium as well as cultured FLSs [20]. Of these MAP3Ks, MEKK2 initially appeared to be the most important in RA because it forms a functional complex with JNK. In the present study, TAK1 functioned as the dominant MAP3K for JNK activation in IL-1-stimulated FLSs. These results were unexpected because several groups have shown that MEKK1, MEKK2 and MEKK3 are indispensable for JNK activation. For instance, MEKK1 is the predominant kinase required for JNK activation in corneal epithelia [20] and murine embryonic fibroblasts (MEFs) [20]. In other culture conditions, JNK activation is inhibited in MEKK3-/-MEFs stimulated with IL-1 [21]. Similarly, fibroblast growth factor (FGF)-2-induced JNK activation and JNK phosphorylation-induced T cell receptor ligation require MEKK2 [22]. Based on our studies using MAP3K deficient cells, these MAP3Ks appear to be redundant in JNK activation in cultured FLSs. Therefore, the diverse and complex functions of MAP3Ks vary depending on the cell type as well as the stimulus. It is precisely this signaling diversity that offers an opportunity to target upstream kinases in the JNK cascade that regulate pathogenic responses in arthritis while potentially sparing other functions that are critical to host responses. This study suggests that TAK1 is a crucial activator of the JNK pathway in FLSs and is a potential target for arthritis therapy.

Materials and methods
Fibroblast-like synoviocytes FLSs were isolated from synovial tissues obtained from RA patients at the time of joint replacement as described previously [3]. The diagnosis of RA conformed to the American College of Rheumatology 1987 revised criteria [23]. The protocol was approved by the UCSD Human Subjects Research Protection Program. Synovial tissues were minced and incubated with 0.5 mg/ml collagenase VIII (Sigma, St. Louis, MO, USA) in serum-free RPMI (Mediatech, Herndon, VA, USA) for 1.5 h at 37°C, filtered through a 0.22 μm cell strainer, extensively washed, and cultured in DMEM supplemented with 10% FCS (endotoxin content <0.006 ng/ml; Gemini Biosciences, Calabasas, CA, USA), penicillin, streptomycin, gentamicin and Lglutamine in a humidified 5% CO 2 incubator. After overnight culture, nonadherent cells were removed, and adherent cells were trypsinized, split at a 1:3 ratio, and cultured. Synoviocytes were used from passage 4 through 9, when FLSs were a homogeneous population with <1% CD11b, <1% phagocytic, and <1% FcRγII positive cells.
Mice knee and ankle synovial tissues were isolated, minced and incubated with 0.5 mg/ml collagenase VIII (Sigma) in serum-free RPMI (Mediatech) for 1.5 h at 37°C, extensively washed, and cultured in DMEM supplemented with 10% FCS (endotoxin content <0.006 ng/ml; Gemini Biosciences), penicillin, streptomycin, gentamicin and L-glutamine in a humidified 5% CO 2 incubator. After three days of culture, non-adherent cells were removed, and adherent cells were trypsinized, split at a 1:3 ratio, and cultured. Synoviocytes were then used from passage 4 through 9.

Western blot analysis
After transfection, FLSs were cultured in DMEM with 10% FCS in six-well plates for appropriate times and synchronized in DMEM with 0.1% FCS. FLSs were then treated with medium or rhIL-1β (2 ng/ml; R&D Systems) for 15 minutes. Cell lysates were obtained as described previously [19]. Whole cell lysates (50 μg) were fractionated on Tris-glycinebuffered 10% SDS-PAGE and transferred to nitrocellulose membrane (Biorad, Hercules, CA, USA). The membranes were blocked with 5% nonfat milk in 0.05% Tween 20/Trisbuffered saline(TBS) for 1 h at room temperature, followed by incubation with primary antibody (1:1000) overnight at 4°C. The blots were then incubated in the secondary antibody for 2 h at room temperature. Immunoreactive protein was detected with enhanced chemiluminescence (Perkin Elmer, Waltham, MA, USA) and autoradiography, which was analyzed using NIH Image (version 1.63) and normalized to actin or GAPDH expression.

Electrophoretic mobility shift assay
Following transfection, FLSs were seeded in 10 cm dishes and cultured in DMEM with 10% FCS at 37°C for 24 h. The cells were incubated in fresh media for 48 h and subsequently serum starved (0.1% FCS/DMEM) for 48 h. FLSs were then treated with either medium or IL-1β (2 ng/ml) for 60 minutes. The cells were rinsed twice with phosphate-buffered saline and nuclear extracts were isolated using a nuclear protein extraction kit (Chemicon, Temecula, CA, USA) and protein estimation was performed using the micro-BCA kit (Pierce, Rockford, IL, USA). Nuclear extracts (10 μg) were incubated with [γ-32 P]-ATP labeled or unlabeled AP-1 (5'-CGCTTGAT-GAGTCAGCCGGAA-3'), nuclear factor (NF)-κB (5'-AGTT-GAGGGGACTTTCCCAGGC-3') oligonucleotides (Promega, Madison, WI, USA) for 20 minutes at room temperature and run on a 5% acrylamide/Tris-base EDTA (TBE) gel for 25 minutes at 200 V. The gel was dried and exposed to film. The autoradiograph was analyzed using NIH Image (version 1.63).

IL-6 ELISA
After transfection, FLSs were seeded in 12-well plates and cultured in DMEM with 10% FCS at 37°C for 24 h. The supernatants were aspirated and replaced with fresh medium for 24 h. FLSs were then treated with medium or rhIL-1β (2 ng/ml) for 24 h and the supernatants were harvested. Samples were assayed for IL-6 by ELISA (R&D Systems).

Quantification of MMP mRNA in FLS
mRNA from cultured FLSs was isolated using RNA-STAT (Tel-Stat, Friendswood, TX, USA) and cDNA was prepared, according to the manufacturer's instructions using GeneAmp 2400 (Applied Biosystems, Foster City, CA, USA). Quantitative real-time PCR was performed using Assays On Demand (Applied Biosystems) to determine relative mRNA levels using the GeneAmp 5700 Sequence Detection System (Applied Biosystems) as described previously [24]. Sample threshold cycle (Ct) values were used to calculate the number of cell equivalents in the test samples. The data were then normalized to GAPDH expression to obtain relative cell equivalents.

Statistical analysis
Data are expressed as mean ± standard error of the mean. Comparisons between two groups were performed using Student's t-test. A comparison was considered statistically significant if p < 0.05.

MEKK1, MEKK2 and MEKK3 knockdown do not alter IL-1-induced MAPK activation
To determine the relative contributions of MEKK1, MEKK2, and MEKK3 to IL-1β-induced MAPK activation, FLSs were transfected with the corresponding siRNA, individually or in combination. On day 3, the serum starved FLSs were stimulated with IL-1β (2 ng/ml) and the lysates were evaluated by western blot analysis using anti-P-JNK, anti-P-p38, and anti-P-ERK antibodies. Figure 2a shows that MEKK1, MEKK2, and MEKK3 deficiency alone or in combination had no effect on IL-1-stimulated JNK, p38 or ERK activation. We then repeated the experiment using MEKK1-/-mouse FLS. Western blot analysis of IL-1β or medium treated MEKK1-/-FLS lysates confirmed data obtained from human FLSs transfected with MEKK1 siRNA (Figure 2b). Next, we transfected MEKK1-/mouse FLS with MEKK2 and MEKK3 siRNA. IL-1β-induced JNK, p38, and ERK activation was then determined ( Figure  2c). The results indicate that MEKK1, MEKK2 and MEKK3 deficiency do not alter the activation of JNK, p38 or ERK in IL-1β-stimulated FLSs.

Discussion
RA is a chronic inflammatory disease of unknown etiology that targets the synovium. Intimal lining macrophages and fibroblast-like synoviocytes produce pro-inflammatory cytokines that contribute to synovial inflammation and production of destructive enzymes like MMPs [25]. The MMPs can then degrade components of the extracellular matrix, especially native interstitial collagen [26], initiating a series of events leading to irreversible joint damage [27]. MMP gene expression in FLSs is regulated by many signaling pathways, although MAPKs play a prominent role [9].
MKK4 and MKK7 are, in turn, regulated by a large family of serine/threonine kinases known as the MAP3Ks that integrate extracellular stimuli and activate transcription factors in a celltype and stimulus-specific manner [29]. Little is known about how MAP3Ks regulate the JNK pathway or MMP gene expression in RA. To address this issue, we previously examined MAP3K gene and protein expression in RA synovial tissue and FLSs [20]. These studies showed that four of the MAP3Ks, namely MEKK1, MEKK2, MEKK3 and TAK1, are abundant in RA FLSs. Kinase assays suggested that MEKK2 and TAK1 are especially important activators of the JNK pathway in FLSs.
In the present study, we examined the hierarchy of these proteins in IL-1β-mediated JNK activation in RA FLSs using siRNA. The results showed that MEKK1, MEKK2 and MEKK3 are not necessary for IL-1β-mediated JNK phosphorylation, either individually or in combination. The data also demonstrate that the pathways utilized by stress kinases can be cell and stimulus specific. For instance, MEKK1 is required for JNK and c-Jun activation in the corneal epithelia of MEKK1-/-mice [30]. MEKK1 is also critical for JNK activation in response to pro-inflammatory stimuli and cell migration in MEKK1-/-MEFs [31]. Unlike MEKK1, MEKK3 knockout is embryonic lethal [32] and JNK and p38 activation are defective in MEKK3-/-MEFs stimulated with IL-1β [21].
Several additional studies indicate that MEKK2 can play a celllineage specific role in JNK activation, and our previous studies showed that it could form a functional signaling complex in FLSs [20]. MEKK2 also associates with MKK7 and JNK1 in Cos cells [33]. Its role in JNK function was suggested by studies showing that MEKK2 gene disruption inhibits JNK activation in mast cells in response to c-Kit and Fcε RI stimulation [34]. Kesavan and colleagues showed that FGF-2induced JNK activation also required MEKK2 in knockout MEFs [22]. Furthermore, MEKK2 is required for JNK activation in T cell receptor signaling and IL-2 gene expression [35]. Despite these data, siRNA and MEKK2-/-studies indicate that MEKK2 is not required for IL-1β-induced JNK activation in FLSs.
In contrast, TAK1 is a critical upstream kinase regulating JNK in FLSs. TAK1 is an evolutionarily conserved MAP3K that is essential for some innate and adaptive immune responses [36]. Signaling through the IL-1 receptor leads to ubiquination and activation of the tumor necrosis factor receptor-associated factor 6 (TRAF6)/TAB1/TAB2/TAB3 (TAB, TAK1-binding protein) complex through IL-1 receptor-associated kinase (IRAK) [37][38][39][40][41]. TAK1 is then activated by autophosphorylation of serine/threonine residues within its activation loop [42]. It can then engage I-kappaB kinase and MAPK pathways leading to the activation of NF-κB, p38, and JNK [43]. TAK1, unlike MEKK1, MEKK2, and MEKK3, is intimately involved in IL-1β-induced JNK activation in FLSs. TAK1 knockdown significantly inhibited the kinase activity of MKK4, MKK7, and JNK. However, TAK1 deficiency did not affect the p38 pathway or interferon gene expression (IP-10 and IFNβ). This result in FLSs differs from studies using 293 cells where p38 and JNK activation by IL-1β required TAK1 [44]. Once JNK is phosphorylated, its effect on downstream gene expression typically involves AP-1 activation. Transcription factor studies in FLSs confirmed that TAK1 deficiency not only decreased JNK activation but also suppressed AP-1 binding. The effect on NF-κB was less prominent and is consistent with previous studies in RANK ligand-stimulated 293 cells expressing dominant-negative TAK1 [45]. The variability of MAP3K function in different cell types is also underscored by studies implicating TAK1 in the NF-κB pathway in HeLa cells [46] and NIH3T3 cells [47]. Therefore, it is important to evaluate signaling mechanisms in tissue-specific cell lineages when considering their potential role in inflammatory diseases.
The functional consequences of activating the TAK1-JNK-AP1 pathway can be evaluated by determining expression of key AP-1-driven genes implicated in RA. The AP-1 consensus Regulation of AP-1 binding and transcriptional activity is TAK1-depend-ent Regulation of AP-1 binding and transcriptional activity is TAK1-dependent. The effect of TAK1 deficiency on IL-1β-induced AP-1 activation was evaluated by electrophoretic mobility shift assay. Five days after small interfering RNA transfection, cultured fibroblast-like synoviocytes were stimulated with IL-1β (2 ng/ml) for 60 minutes. Nuclear extracts were obtained and evaluated for AP-1 and NF-κB binding activity. A representative experiment is shown (n = 3). IL-1β-induced AP-1 activity was significantly decreased in the absence of TAK1 (84 ± 8% inhibition compared to control, p = 0.03). NF-κB binding, on the other hand, was not significantly decreased with TAK1 deficiency (34 ± 7% inhibition, p = 0.21). Cold oligo, non-radioactive oligo; Sc, scrambled control.
(page number not for citation purposes) sequence is located at -70 base-pairs in the promoter region of the gene encoding MMP3, making it a useful biomarker for TAK1 in cells such as osteocytes [48] and chondrocytes [49]. siRNA studies showed that TAK1 inhibition significantly decreased MMP3 gene expression in cultured FLSs. Of interest, IL-1β-induced IL-6 production was also decreased by TAK1 deficiency, which could reflect an effect on AP-1 because this transcription factor also binds to the IL-6 promoter. The modest effect of TAK1 on NF-κB could also contribute to MMP3 and IL-6 expression.

Conclusion
These data suggest that TAK1 is a key element in JNK activation, IL-6 production, and MMP expression by FLSs. Surprisingly, other MAP3Ks implicated in JNK activation, such as MEKK1, MEKK2, and MEKK3, do not have a major contribution to this pathway in FLSs. Therefore, targeting TAK1 might represent an alternative way to regulate JNK activation and matrix degradation in inflammatory arthritis.