Inhibitory Effects of IL-6-Mediated Matrix Metalloproteinase-3 and -13 by Achyranthes japonica Nakai Root in Osteoarthritis and Rheumatoid Arthritis Mice Models

Achyranthes japonica Nakai root (AJNR) is used to treat osteoarthritis (OA) and rheumatoid arthritis (RA) owing to its anti-inflammatory and antioxidant effects. This study investigated the inhibitory effects of AJNR on arthritis. AJNR was extracted using supercritical carbon dioxide (CO2), and its main compounds, pimaric and kaurenoic acid, were identified. ANJR’s inhibitory effects against arthritis were evaluated using primary cultures of articular chondrocytes and two in vivo arthritis models: destabilization of the medial meniscus (DMM) as an OA model, and collagenase-induced arthritis (CIA) as an RA model. AJNR did not affect pro-inflammatory cytokine (IL-1β, TNF-α, IL-6)-mediated cytotoxicity, but attenuated pro-inflammatory cytokine-mediated increases in catabolic factors, and recovered pro-inflammatory cytokine-mediated decreases in related anabolic factors related to in vitro. The effect of AJNR is particularly specific to IL-6-mediated catabolic or anabolic alteration. In a DMM model, AJNR decreased cartilage erosion, subchondral plate thickness, osteophyte size, and osteophyte maturity. In a CIA model, AJNR effectively inhibited cartilage degeneration and synovium inflammation in either the ankle or knee and reduced pannus formation in both the knee and ankle. Immunohistochemistry analysis revealed that AJNR mainly acted via the inhibitory effects of IL-6-mediated matrix metalloproteinase-3 and -13 in both arthritis models. Therefore, AJNR is a potential therapeutic agent for relieving arthritis symptoms.


Introduction
Osteoarthritis (OA) is a severe chronic degenerative disease of the joints that is common in middle aged and older people [1]. The main clinical manifestations of OA are degeneration of the articular cartilage and changes in the subchondral bone structure [2]. When joint cartilage is completely lost following disruption of cartilage homeostasis through the induction of catabolic factors as well as downregulation of anabolic factors, the bones and soft tissue structures around the joint are altered, resulting in joint pain, swelling, deformity, and disability [3,4]. Although several risk factors associated with OA have

Extraction and Identification of AJNR Components
The AJNR components were extracted using a supercritical CO 2 extraction method ( Figure 1a). The extraction pressure and temperature were set to 40 MPa and 60 • C, respectively. Because we used ethanol as a co-solvent during supercritical CO 2 extraction and an ethanol soluble phase of AJNR extract, in our experiment, ethanol was used as a solvent for gas chromatography-mass spectrometry (GC-MS) analysis. The results showed that the main compounds of AJNR were pimaric acid (74%) and kaurenoic acid (26%) (Figure 1b).

Extraction and Identification of AJNR Components
The AJNR components were extracted using a supercritical CO2 extraction method (Figure 1a). The extraction pressure and temperature were set to 40 MPa and 60 °C, respectively. Because we used ethanol as a co-solvent during supercritical CO2 extraction and an ethanol soluble phase of AJNR extract, in our experiment, ethanol was used as a solvent for gas chromatography-mass spectrometry (GC-MS) analysis. The results showed that the main compounds of AJNR were pimaric acid (74%) and kaurenoic acid (26%) (Figure 1b).  (2), a CO2 storage tank (3), a CO2 pump (4), an extractor (5), and three separators (6). (b) Identification of the main compounds of AJNR extracted using the supercritical CO2 method.

Effects of AJNR on Cell Viability
The effects of AJNR on chondrocyte viability were evaluated by MTT assay. Treatment with AJNR (0-100 µg/mL) for 48 h did not result in a significant cytotoxic effect ( Figure 2a). Moreover, following treatment with AJNR (10, 20, 50 µg/mL) in the absence or presence of the pro-inflammatory cytokines IL-1β (1 ng/mL), IL-6 (100 ng/mL), and TNF-α (10 ng/mL), the cell viability of primary chondrocytes was not recovered ( Figure  2b-d). These results suggest that AJNR does not affect cell viability, either as a single treatment or in the presence of pro-inflammatory cytokines.  (2), a CO 2 storage tank (3), a CO 2 pump (4), an extractor (5), and three separators (6). (b) Identification of the main compounds of AJNR extracted using the supercritical CO 2 method.

Effects of AJNR on Cell Viability
The effects of AJNR on chondrocyte viability were evaluated by MTT assay. Treatment with AJNR (0-100 µg/mL) for 48 h did not result in a significant cytotoxic effect ( Figure 2a). Moreover, following treatment with AJNR (10, 20, 50 µg/mL) in the absence or presence of the pro-inflammatory cytokines IL-1β (1 ng/mL), IL-6 (100 ng/mL), and TNF-α (10 ng/mL), the cell viability of primary chondrocytes was not recovered (Figure 2b-d). These results suggest that AJNR does not affect cell viability, either as a single treatment or in the presence of pro-inflammatory cytokines.  Figure 2. Effects of AJNR on the viability of primary cultured articular chondrocytes. (a) Primary cultured mouse articular chondrocytes were exposed to increasing concentration (0-100 µg/mL) of AJNR for 24 h and then subjected to MTT assay. The other groups were treated with AJNR (10, 20, 50 µg/mL) in the absence or presence of IL-1β (1 ng/mL) (b), IL-6 (100 ng/mL) (c), or TNF-α (10 ng/mL) (d), and then the cell viability of primary chondrocytes was evaluated by MTT assay. The results are representative of three independent experiments. Values are presented as the mean ± SEM and were evaluated using the two-tailed t-test.

Figure 2.
Effects of AJNR on the viability of primary cultured articular chondrocytes. (a) Primary cultured mouse articular chondrocytes were exposed to increasing concentration (0-100 µg/mL) of AJNR for 24 h and then subjected to MTT assay. The other groups were treated with AJNR (10, 20, 50 µg/mL) in the absence or presence of IL-1β (1 ng/mL) (b), IL-6 (100 ng/mL) (c), or TNF-α (10 ng/mL) (d), and then the cell viability of primary chondrocytes was evaluated by MTT assay. The results are representative of three independent experiments. Values are presented as the mean ± SEM and were evaluated using the two-tailed t-test.

AJNR Attenuated DMM-Induced OA Pathogenesis
We further investigated the effect of supercritical CO 2 extract of AJNR on OA pathogenesis in a DMM model. IP injection of AJNR was administered twice a week for 8 weeks after DMM surgery ( Figure 4a). The structural integrity of the articular cartilage was evaluated via staining with Safranin-O/Fast Green (Figure 4b). The results showed that AJNR attenuated DMM-induced cartilage degeneration and erosion. The other OA parameters, such as OARSI (p < 0.0001), sclerosis (p = 0.005), and osteophytes (p < 0.0001), were completely blocked by AJNR in the DMM model (Figure 4c-f). In addition, DMM-induced synovial inflammation (as indicated by increased articular cavity inflammatory cells, thickened synovial membrane, and synovial tissue edema) was attenuated by AJNR (p < 0.0001) (Figure 4g,h). In both sham groups, the articular cartilage surface was smooth, and the synovial tissue was not hyperplastic.

AJNR Attenuated CIA-Induced RA
We determined whether AJNR has inhibitory effects on RA in a CIA model. IP injection of AJNR (2 mg/kg) was performed during CIA induction (Figure 5a). Histological assessments were conducted on mouse ankle, knee, and toe joints. There were no pathological symptoms of arthritis in the non-immunized (NI) group. However, severe synovitis with synovial hyperplasia, erosion of the bone and cartilage, and infiltration of inflammatory cells were observed in the CIA group (Figure 5b,e). CIA-induced cartilage degeneration and synovitis were dramatically reduced in the AJNR-treated group (Figure 5b,e). The histological scores of synovitis and OARSI were also significantly attenuated in the ankle, knee, and toe of mice treated with AJNR (Figure 5c,d,f-i). These results suggest that AJNR has beneficial effects against inflammation and cartilage degeneration in RA. We determined whether AJNR has inhibitory effects on RA in a CIA model. IP injection of AJNR (2 mg/kg) was performed during CIA induction (Figure 5a). Histological assessments were conducted on mouse ankle, knee, and toe joints. There were no pathological symptoms of arthritis in the non-immunized (NI) group. However, severe synovitis with synovial hyperplasia, erosion of the bone and cartilage, and infiltration of inflammatory cells were observed in the CIA group (Figure 5b,e). CIA-induced cartilage degeneration and synovitis were dramatically reduced in the AJNR-treated group (Figure 5b,e). The histological scores of synovitis and OARSI were also significantly attenuated in the ankle, knee, and toe of mice treated with AJNR (Figure 5c,d,f-i). These results suggest that AJNR has beneficial effects against inflammation and cartilage degeneration in RA.

AJNR Attenuated CIA-Induced Pannus Formation in Ankle and Knee
Pannus formation is an essential indicator of severe arthritis. Therefore, we evaluated CIA-induced pannus formation in mouse ankles and knees. The results showed that pannus formation in the ankle and knee was not observed in the NI group; however, CIA-induced pannus formation was significantly reduced by AJNR in both the ankles (p = 0.0007) and knee (p = 0.0293) ( Figure 6).
Scores of synovial inflammation (c) and Osteoarthritis Research Society International (OARSI) (d) in the ankle are presented (n ≥ 4 mice per group). (e) Representative Safranin-O and H&E staining images of knee cartilage degeneration and synovial inflammation, respectively, in NI, CIA, and AJNR-treated CIA mice. Scores of synovial inflammation (f) and OARSI (g) in the knee are presented (n ≥ 4 mice per group). Scores of toe synovitis (h) and OARSI (i) are also presented (n ≥ 6 mice per group). The results are representative of three independent experiments. Values are presented as the mean ± SEM and were evaluated using the Mann-Whitney U test. Scale bar: 50 µm.

AJNR Attenuated CIA-Induced Pannus Formation in Ankle and Knee
Pannus formation is an essential indicator of severe arthritis. Therefore, we evaluated CIA-induced pannus formation in mouse ankles and knees. The results showed that pannus formation in the ankle and knee was not observed in the NI group; however, CIAinduced pannus formation was significantly reduced by AJNR in both the ankles (p = 0.0007) and knee (p = 0.0293) ( Figure 6).

AJNR Inhibited IL-6 Mediated Mmp3 and Mmp13 Expression in DMM-Induced OA Model or CIA-Induced RA Model
Our in vitro mechanistic studies revealed that the inhibitory effects of AJNR were specific against IL-6. To further verify the inhibitory effects of AJNR on IL-6-mediated OA pathogenesis, we performed immunohistochemical staining of samples from the DMM model. As shown in Figure 7a,b, IL-6, Mmp3, and Mmp13 expression was elevated in both the knee cartilage and knee synovium of the DMM model, compared with that in the sham group; however, the elevated expression of these proteins was remarkably attenuated in the AJNR-treated group (Figure 7a,b). To further verify the mechanism of IL-6-mediated Mmp3 and Mmp13 expression, we examined the expression of IL-6, Mmp3, and Mmp13 in the CIA model. The expression of IL-6, Mmp3, and Mmp13 was dramatically enhanced in the ankle cartilage and synovium of CIA-induced RA mice (Figure 7c,d). Similar to that in the DMM model, the elevated expression of IL-6, Mmp3, and Mmp13 was reduced in the AJNR-treated CIA model (Figure 7c,d). These results indicate that the inhibitory effects of AJNR are specific for IL-6-mediated Mmp3 and Mmp13 expression.

AJNR Inhibited IL-6 Mediated Mmp3 and Mmp13 Expression in DMM-Induced OA Model or CIA-Induced RA Model
Our in vitro mechanistic studies revealed that the inhibitory effects of AJNR were specific against IL-6. To further verify the inhibitory effects of AJNR on IL-6-mediated OA pathogenesis, we performed immunohistochemical staining of samples from the DMM model. As shown in Figure 7a,b, IL-6, Mmp3, and Mmp13 expression was elevated in both the knee cartilage and knee synovium of the DMM model, compared with that in the sham group; however, the elevated expression of these proteins was remarkably attenuated in the AJNR-treated group (Figure 7a,b). To further verify the mechanism of IL-6-mediated Mmp3 and Mmp13 expression, we examined the expression of IL-6, Mmp3, and Mmp13 in the CIA model. The expression of IL-6, Mmp3, and Mmp13 was dramatically enhanced in the ankle cartilage and synovium of CIA-induced RA mice (Figure 7c,d). Similar to that in the DMM model, the elevated expression of IL-6, Mmp3, and Mmp13 was reduced in the AJNR-treated CIA model (Figure 7c,d). These results indicate that the inhibitory effects of AJNR are specific for IL-6-mediated Mmp3 and Mmp13 expression.

Discussion
In this study, we investigated the effect of supercritical CO 2 extract of AJNR against arthritis pathogenesis in vitro and in vivo. We used AJNR at a concentration of 50 µg/mL in the in vitro study on the basis of the results of the articular chondrocyte viability test (Figure 2). AJNR particularly inhibits catabolic alteration mediated by IL-6, among proinflammatory cytokines (i.e., IL-1β, TNF-α, and IL-6) (Figure 3). Based on our preliminary test of AJNR in animals, IP injection of AJNR was administered at a concentration of 2 mg/mL. IP injection of AJNR effectively protected and slowed down the pathogenesis of post-traumatic OA in mouse DMM and CIA models of RA. Compared with those in the arthritis group, cartilage erosion and proteoglycan loss, synovitis, and subchondral plate thickness were reduced in the AJNR treatment group. These phenomena are explicitly related to IL-6-mediated Mmp3 and Mmp13 expression. Taken together, our findings showed that the potential of AJNR for joint protection and arthritis treatment was significant.
In accordance with the OA concept put forward by Garrod in 1890, the pathological characteristics of OA in inflammatory lesions of the joints were evaluated [37]. The main pathological changes in OA are articular cartilage loss and collagen fiber degeneration [38], which can lead to bone hyperplasia and osteophyte formation when the original mechanical balance of the knee joint is disrupted [39]. Therefore, these pathological changes can reflect the clinical characteristics of this degenerative joint disease [40,41]. Mmps participate in the degradation of many matrix components, and the activities of Mmps are regulated by hormones and cytokines in vivo [20,42]. Mmps play an essential role in synthesizing and decomposing the matrix and in the intervention of many physiological and pathological processes, such as arthritis and tissue remodeling [42]. Our results showed that AJNR effectively inhibited IL-6-induced Mmp3, Mmp10, and Mmp13 expression, but this effect was not specific to IL-1β and TNF-α (Figure 3). ADAMTS4 and ADAMTS5 are the main proteases responsible for the degradation of proteoglycans in the articular cartilage of OA [43][44][45][46]. AJNR was only effective against IL-1β-or TNF-α-induced ADAMTS5 expression ( Figure 3). In addition, AJNR recovered the reduction of SOX9 and Aggrecan mediated by all three pro-inflammatory cytokines (Figure 3).
There has been no evidence of the complete curation of OA. The main clinical treatments for this disease are oral drugs, such as NSAIDs, which can cause many adverse side effects [47][48][49]. Nevertheless, it is important to protect the joints of patients with the disease. The inhibitory effect of AJNR against OA symptoms indicates its potential as a therapeutic agent against OA; however, further validation in other animal models and clinical trials is required [50].
The saponins and sterones in AJN, which have antitumor and anti-inflammatory effects, are believed to be the main medicinal substances of AJN [51,52]. The flavonoids, alkaloids, allantoin, succinic acid, and β-sitosterol of AJN have also been evaluated for pharmacological activities [52]. So far, researchers have focused only on the extract ingredients of AJN, which possess inhibitory effects against arthritis [53,54]. Moreover, the animal model they used was not a post-traumatic model such as monosodium iodoacetate-induced osteoarthritis animal model [55] or rabbit CIA model [56]. Furthermore, there is a lack of available information on the active compounds of AJN. In this study, we identified two novel main components from AJNR, namely, pimaric acid and kaurenoic acid, extracted using CO 2 supercritical fluid (Figure 1). Because there has been no report of these compounds from AJNR extraction via normal organic solvent extraction methods nor functional research, our current study guaranteed functional studies of these novel compounds in arthritis pathogenesis in the future. As a traditional plant medicine, AJNR has been used in the clinical treatment of OA for many years [52,57,58]. Our data indicated the inhibitory effects of AJNR on all OA parameters, e.g., Safranin-O-stained tissue histology, OARSI grade, subchondral bone plate thickness, osteophyte size, osteophyte maturity, and synovitis ( Figure 4). Moreover, AJNR displayed an inhibitory effect against CIA-induced arthritis in the knee, ankle, and toe ( Figures 5 and 6). Finally, our data suggested that AJNR exerted a specific effect on IL-6-mediated alterations in Mmp3 and Mmp13 expression in OA and RA mouse models, as confirmed by immunohistochemistry analysis.

Preparation of AJNR Extracts by Supercritical CO 2
The extracts for the experiment were prepared using supercritical CO 2 extraction, which is widely applied to extract bioactive substances from natural materials [31,59]. The extraction was performed using a laboratory-scale supercritical CO 2 extraction device (RZSCF130-65-01L SUPERCRITICAL CO2 EXTRACTION EQUIPMENT) manufactured by Nantong Wisdom Supercritical Science & Technology Development Co., Ltd., Haian, China, which essentially consisted of a CO 2 source, a condenser, a CO 2 storage tank, a CO 2 pump, an extractor, and three separators, as shown in the schematic drawing in Figure 1a. During extraction, 500 g of dried and powdered AJNR (approximately 40 meshes) mixed with 250 mL of absolute ethanol were loaded into the extractor, in which the air was replaced with CO 2 three times. The CO 2 was pumped into the extractor, and the extracts dissolved in CO 2 were transported out from the extractor and separated from CO 2 at the separators. The extraction pressure and temperature were set to 40 MPa and 60 • C, respectively. The pressure and temperature of the separators were set as 8 MPa, 50 • C (separators 6-I), and 5 MPa, 40 • C (separators 6-II and 6-III). The CO 2 flow rate was controlled at 20 kg/h, and the extraction time was 2 h. After being separated from the extract, the CO 2 would return to the condenser and repeat the cycle. After the extraction was complete, CO 2 was released to the atmosphere via an empty valve, the extracts were collected from the separators and further evaporated using a rotavapor to remove the ethanol. The final extract (approximately 10 g) was stored in a sealed airtight tube for subsequent tests.

Reverse Transcription-Polymerase Chain Reaction (RT-PCR)
Total RNA was extracted from primary cultured chondrocytes using TRIzol reagent (Molecular Research Center, Inc., Cincinnati, OH, USA). The quality and concentration of RNA were assessed using a NanoDrop™ 2000 spectrophotometer (Thermo Scientific, Waltham, MA, USA). The RNA was reverse transcribed, and the resulting cDNA was amplified by PCR or BIO-RAD Real-Time PCR system (CFX96™ Real-Time System, Bio-Health Materials Core-Facility, Jeju National University) using SYBR premixed Extaq reagent (Takara Bio, Mountain View, CA, USA). Target bands were quantified using the ImageJ densitometry software (NIH, Bethesda, MD). The PCR primers and experimental conditions are summarized in Supplementary Table S1. Glyceraldehyde-3-phosphate dehydrogenase (Gapdh) was used as an internal control.

Animal Model and AJNR Treatment
All experiments were approved by the Jeju National University Animal Care and Use Committee. The DMM-mouse model was established using 12-week-old C57BL/6J male mice (9 mice/group), as previously described [13,14,19]. The medial anterior meniscotibial ligament of the right knee was cut using a surgical scissor in the patellar tendon in the middle and tendon of the tibial plateau. In the sham operation group, only arthrotomy was performed without resection of the tibial ligament of the medial meniscus. The mice were administered intraperitoneal (IP) injection of AJNR (2 mg/kg) in 200 µL polyethylene glycol 400 (PEG-400) twice a week. The control group was injected with the same volume of PEG-400 reagent on the same schedule. The CIA mouse model was established using 7-week-old DBA/1J male mice (8 mice/group). RA was induced by type II collagen soluble in 0.05 M acetic acid at 4 • C. Emulsification was performed with an equal volume of Freund's complete adjuvant and boosted with Freund's incomplete adjuvant on day 21. Mice were administered AJNR (2 mg/kg) via IP injection with a newly formulated PEG-400. Control mice were injected with the same volume of PEG-400 reagent on the same schedule. The mice were examined every 3 days to measure the visual appearance of arthritis around the knee and ankle and to determine the corresponding severity and arthritis score [60,61].

Histological Analysis
The CIA (knee and ankle) or DMM (knee) samples were fixed with 4% paraffin formaldehyde for 24 h and then decalcified in 0.5 M EDTA solution at 4 • C for 4 weeks. Next, the samples were dehydrated using gradient ethanol and embedded in paraffin. Finally, the sample was sliced at a thickness of 5 mm and stained with Safranin-O/Fast Green for evaluation. Cartilage destruction and severity of synovitis were evaluated by experienced histological researchers who were blinded to the study groups. The Osteoarthritis Research Society International (OARSI) scoring system was used to evaluate cartilage degeneration, as previously described [38,[62][63][64]. Subchondral plate thickness was measured using the Aperio Image Scope V12 software (Leica Biosystems, Buffalo Grove, IL, USA) [13,14,16].

Statistical Analysis
All statistical analyses were performed using IBM SPSS Statistics 24 (IBM Corp., Armonk, NY, USA). Experimental data were analyzed using the non-parametric Mann-Whitney U test and two-tailed Student's t-tests with unequal sample sizes. Significance was accepted at a level of probability of 0.05 (p < 0.05). The results are shown as the mean ± standard error of the mean (SEM).

Conclusions
The present study showed that AJNR inhibited arthritis pathogenesis by reducing the expression of catabolic factors and enhancing the expression of anabolic factors without altering cell viability. Interestingly, AJNR exerted therapeutic effects in both OA and RA animal models. In particular, AJNR showed specific inhibitory effects against IL-6mediated anabolic and catabolic imbalances. These findings indicate that AJNR has vast potential for the treatment of arthritis pathogenesis, which can lead to drug development.