Regulation of inflammatory arthritis by the upstream kinase mitogen activated protein kinase kinase 7 in the c-Jun N-Terminal kinase pathway

Introduction The c-Jun N-terminal kinase (JNK) is a key regulator of matrix metalloproteinase (MMP) and cytokine production in rheumatoid arthritis (RA) and JNK deficiency markedly protects mice in animal models of arthritis. Cytokine-induced JNK activation is strictly dependent on the mitogen-activated protein kinase kinase 7 (MKK7) in fibroblast-like synoviocytes (FLS). Therefore, we evaluated whether targeting MKK7 using anti-sense oligonucleotides (ASO) would decrease JNK activation and severity in K/BxN serum transfer arthritis. Methods Three 2'-O-methoxyethyl chimeric ASOs for MKK7 and control ASO were injected intravenously in normal C57BL/6 mice. PBS, control ASO or MKK7 ASO was injected from Day -8 to Day 10 in the passive K/BxN model. Ankle histology was evaluated using a semi-quantitative scoring system. Expression of MKK7 and JNK pathways was evaluated by quantitative PCR and Western blot analysis. Results MKK7 ASO decreased MKK7 mRNA and protein levels in ankles by about 40% in normal mice within three days. There was no effect of control ASO on MKK7 expression and MKK7 ASO did not affect MKK3, MKK4 or MKK6. Mice injected with MKK7 ASO had significantly less severe arthritis compared with control ASO (P < 0.01). Histologic evidence of synovial inflammation, bone erosion and cartilage damage was reduced in MKK7 ASO-treated mice (P < 0.01). MKK7 deficiency decreased phospho-JNK and phospho-c-Jun in ankle extracts (P < 0.05), but not phospho-MKK4. Interleukin-1beta (IL-1β), MMP3 and MMP13 gene expression in ankle joints were decreased by MKK7 ASO (P < 0.01). Conclusions MKK7 plays a critical regulatory role in the JNK pathway in a murine model of arthritis. Targeting MKK7 rather than JNK could provide site and event specificity when treating synovitis.


Introduction
Rheumatoid arthritis (RA) is one of the most common immune-mediated diseases and is characterized by synovial inflammation and joint destruction [1]. Mitogenactivated protein kinases (MAPKs) are highly activated in rheumatoid synovium and potentially contribute to inflammatory and destructive mechanisms [2,3]. The c-Jun N-terminal kinases (JNKs), which belong to the MAPK family, play important roles in cytokine production and extracellular matrix degradation by regulating matrix metalloproteinase (MMP) in fibroblastlike synoviocytes (FLS) and animal models of RA [4,5]. Of the three JNK isoforms, JNK1 has been implicated as a pivotal regulator of synovial inflammation in murine arthritis due to its role in mast cell degranulation and macrophage migration [6,7].
JNK is activated via dual phosphorylation by two upstream MAPK kinases (MKKs), MKK4 and MKK7 [8][9][10]. The mice lacking MKK4 or MKK7 are embryonic lethal suggesting the two kinases are non-redundant and serve distinct functions [11]. Some studies suggest that these differences might be due to selective regulation by extracellular stimuli, distinct tissue distribution and different biochemical properties [10]. Thus, an alternative approach targeting the MKKs instead of JNK could suppress signaling responses that contribute to inflammatory arthritis but spare a subset of host defense or homoeostasis pathways.
Our previous studies showed that MKK4 and MKK7 are expressed and phosphorylated in RA synovium and both are activated by cytokines in RA FLS [12]. Surprisingly, cytokine-induced JNK activation and MMP production are strictly dependent on MKK7 in cytokinestimulated FLS and do not require MKK4 [13]. Therefore, we evaluated whether selective targeting of MKK7 using anti-sense oligonucleotides (ASOs) would block arthritis-associated JNK activation and decreased arthritis severity in K/BxN serum transfer arthritis. The data indicate that blockade MKK7 mimics the effect of JNK deficiency and suppresses inflammatory arthritis.

ASO treatment in normal mice
All animal protocols received prior approval by the institutional review board. Pathogen-free male C57BL/6 mice were purchased from The Jackson Laboratory (Bar Harbor, ME, USA) and MKK7 or control ASOs were administered to mice based upon body weight by intravenous injection (25 and 50 mg/kg). Three days after injection of ASOs, mice were sacrificed and various tissues evaluated for MKK7 gene expression.

K/BxN serum transfer arthritis and ASO treatment
To induce K/BxN serum transfer arthritis [15], serum samples were pooled from arthritic adult K/BxN mice and injected intraperitoneally (IP) as previously described [16]. C57BL/6 mice received PBS, MKK7 ASOs or control ASO (50 mg/kg) i.v. twice a week beginning on Day -8 and then administered 100 μl of K/ BxN serum on Day 0. Clinical arthritis scores were evaluated using a scale of 0 to 4 for each paw for a total score of 16. Ankle thickness was measured with a caliper placed across the ankle joint at the widest point.
Histopathologic assessment was performed using a semi-quantitative scoring system as previously described [6], including synovial inflammation, bone erosion and cartilage damage.

Quantitative real time PCR
Ankle joints were collected at study termination, dissected to remove extra-articular tissue and snap-frozen in liquid nitrogen. The specimens were pulverized and total RNA was isolated using Rneasy ® Lipid Tissue kit per manufacturer's protocol (Qiagen, Valencia, CA, USA). MKK7, IL-1β, MMP3 and MMP13 expressions were measured by quantitative real-time PCR as previously described [17]. The threshold cycle (Ct) values were normalized to hypoxanthine-guanine phosphoribosyl transferase (HPRT) or glyceraldehyde-3-phosphate dehydrogenase (GAPDH) expression.

Western blot analysis
Snap-frozen joints were pulverized and homogenized at 100 mg of tissue per 0.5 ml of lysis buffer. Western blot analysis was then performed as described previously [18]. Anti-MKK3, anti-MKK6 and anti-GAPDH antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Anti-MKK4, anti-MKK7, anti-phospho-MKK4, anti-JNK, anti-phospho-JNK, anti-c-Jun and anti-phosphoc-Jun antibodies were purchased from Cell Signaling Technology (Danvers, MA, USA). Immunoreactive protein was detected with Immun-Star Western C kit (Bio-Rad, Hercules, CA, USA) using VersaDoc MP4000 imaging system (Bio-Rad). Densitometry analysis was carried out with Quantity One 1-D analysis software (Bio-Rad).

Statistical analysis
Data are expressed as mean ± SE. Arthritis scores and change of ankle thickness among PBS, control ASO or MKK7 ASO-injected groups were analyzed by one-way ANOVA and Tukey's post-hoc test. Comparisons between control ASO and MKK7 ASO-injected groups were analyzed by two-tailed Student's t-test. In all tests, P-value < 0.05 was considered statistically significant.

MKK7 knockdown by ASO in normal C57BL/6 mice
Three different 2'-MOE chimeric ASOs (MKK7 ASO-1, -2, -3) targeting distinct regions of the MKK7 gene or control ASO were injected i.v. into normal C57BL/6 mice. Three days later ankle joints were harvested and assayed for MKK7 mRNA. As shown in Figure 1A, MKK7 mRNA levels were reduced in a dose-dependent manner ( Figure  1A), with greatest inhibition at a dose of 50 mg/kg of MKK7 ASO-2 compared with control ASO (P < 0.05, n = 3 mice for each group). Control ASO had no effect. MKK7 protein levels in the ankle were also analyzed by Western blot analysis. Figure 1B shows that MKK7 protein levels were also decreased by MKK7 ASO, but MKK3, MKK4 and MKK6 levels were not affected. MKK7 mRNA was also decreased in liver and spleen by up to 37% at a dose of 25 mg/kg of MKK7 ASO-1 ( Figure 1C; P < 0.05, n = 3 mice for each group) and maximally decreased in liver by up to 45% at a dose of 50 mg/kg of MKK7 ASO-1 (data not shown; P < 0.05).
Effect of MKK7 ASO on K/BxN serum transfer arthritis

Effect of MKK7 ASO on histopathology
Histopathologic analysis was performed on ankle joints obtained on Day 10 after K/BxN serum administration. Consistent with the decreased clinical arthritis, MKK7 ASO suppressed synovial inflammation, bone erosions and cartilage destruction compared with control ASO (see Figure 3).

Effect of MKK7 ASO on MKK4, JNK, and c-Jun phosphorylation
Cytokine-induced JNK activation is dependent on MKK7 in cultured FLS and does not require MKK4. To determine the effect of selective MKK7 deficiency on JNK signaling in vivo, the ankle joints were evaluated by Western blot analysis to determine the phosphorylation state of MKK4, JNK and c-Jun. Consistent with the reduction of MKK7 protein level (42.2%, P < 0.05), MKK7 deficiency decreased GAPDH-normalized phospho-JNK by 67% and phospho-c-Jun by 62% compared with control ASO injected mice (P < 0.05, n = 3 for each group) ( Figure 4). However, there was no significant difference of phosphorylation status of MKK4 between MKK7 ASO and control ASO injected groups. Similar results were obtained if the phospho-MKK4 and phospho-JNK were normalized to MKK4 and JNK, respectively. The c-Jun protein levels were higher in the control ASO-treated mice compared with MKK7 treatment due to increased local cytokine production, such as IL-1β (see below). Thus, normalization to GAPDH provides a more reliable assessment of total phospho-c-Jun in the tissue.

A B
Mean arthritis score Days after serum transfer

Regulation of IL-1b and MMP expression by MKK7 deficiency
The JNK pathway regulates MMP gene expression. Consistent with the reduction phospho-JNK and phospho-c-Jun in ankle joints, MMP3 and MMP13 expression were significantly decreased in the mice injected with MKK7 ASO compared with control ASO (77.9%, and 72.6%, respectively, P < 0.05) ( Figure 5). Of interest, IL-1β expression was also decreased. These data suggest that MKK7 plays a key role in regulating the JNK pathway, including transcription of inflammatory cytokines and proteases involved in joint damage.

Discussion
Proinflammatory cytokines and MMPs promote synovial inflammation and facilitate cartilage and bone destruction in RA [5]. The MAPKs (ERK, JNK and p38) contribute by phosphorylating key transcription factors, such as activator protein-1 (AP-1), that are required for gene transcription. JNK, in particular, plays a pivotal role in cytokine-mediated AP-1 induction and MMP gene expression in FLS [4,19]. Three isoforms of JNK have been characterized, namely JNK1, 2 and 3. JNK1 and 2 are ubiquitous while JNK3 is primarily restricted to neurologic tissue [20]. JNK2 deficiency has only modest effects in pre-clinical models of arthritis, but JNK1 deficiency attenuates synovitis and joint destruction in murine antigen-induced arthritis and passive K/BxN serum transfer arthritis [6,7,21]. JNK1 also contributes to osteoclast differentiation, since JNK1-deficient osteoclast progenitors do not mature into bone-resorbing osteoclasts [22]. These data suggest that JNK participates in the synovial inflammation and joint destruction of RA and could potentially be targeted in diseases like RA. While JNKs are attractive targets, they regulate in many normal cell functions, especially in matrix remodeling and host defense [23,24]. Thus, blocking all JNK activity, or even all JNK1 activity, could affect host defense or matrix homeostasis. As an alternative strategy, targeting an individual upstream kinase like MKK4 or MKK7 could permit some normal JNK functions while interfering with a subset that is pathogenic in synovitis. MKK4 and MKK7, two JNK upstream kinases, exhibit some different properties although they can synergistically activate JNKs [10]. TNF and IL-1 mainly activate MKK7 in murine embryonic fibroblasts, while ultraviolet radiation, anisomycin, heat and osmotic shock activate both MKK4 and MKK7 [8,9,25]. These data suggest that MKK4 and MKK7 contribute separately to the activation of JNKs in response to environmental stress or inflammatory cytokines.
We previously showed that MKK7, but not MKK4, is required for IL-1-induced JNK phosphorylation and AP-1-driven MMP expression [13]. Nevertheless, MKK4 is a component of the JNK signal complex and is also readily phosphorylated in FLS [12]. Mice lacking Gadd45β, which serves as an endogenous inhibitor of MKK7, have enhanced JNK activity and disease severity in the passive K/BxN model [16]. These data suggest that selective MKK7 blockade could suppress arthritis and potentially decrease adverse effects by permitting non-pathogenic MKK4-mediated JNK activation. However, there is no direct evidence that MKK7 inhibition would be beneficial in synovitis. Our initial plans to focus on Gadd45β were complicated by the recent observation that Gadd45β deficiency unexpectedly exacerbates disease severity in collagen-induced arthritis [26].
We, therefore, focused on genetic approaches that circumvent the embryonic lethality of MKK7 deficiency. Several small interfering RNA (siRNA) methods were tested because others have reported success [27], but we were unable to consistently knockdown endogenous MKK7 expression (data not shown). Chemically modified ASOs were then tested for applications in animal models of RA because of their nuclease-resistant capacity, potency and long half-life [28][29][30]. Free ASOs are considerably smaller than siRNA-delivery agent complexes and enter many cells types via pinocytosis and phagocytosis, whereas larger siRNA complexes primarily enter macrophages and neutrophils by phagocytosis [31,32]. Thus, we used single strand, 2'-O-methoxyethylribose modified chimeric ASOs to investigate the effect of MKK7 deficiency in mice. Selectivity was confirmed with MKK7 ASOs, which decreased MKK7 mRNA and protein expression but not MKK3, MKK4 or MKK6.
The ASO studies showed that selective MKK7 deficiency significantly reduced arthritis severity and joint destruction compared with control ASO-injected group even though MKK7 was only partially depleted. Downstream events were consistent with previous in vitro studies by demonstrating reduced phosphorylation of JNK and c-Jun in the inflamed joints of MKK7 ASO-treated mice. Decreased joint damage in mice treated with MKK7 ASOs is consistent with previous observations that MKK7 is a pivotal signaling molecule that regulates JNK and MMP expression in FLS [13].
Taken together, these results imply that MKK7 plays a pivotal role in inflammatory arthritis and that MKK7 ASO acts through the inhibition of JNK in passive K/ BxN arthritis. Because JNK2 does not contribute to this model, the effect is most likely due to decreased JNK1 activation with resultant decreased mast cell activation [6]. That observation is supported by the fact that JNK activation is abolished in mkk7 -/mast cell lines, suggesting that MKK7 is essential for JNK activation in mast cells [33].