Susceptibility of cyclin-dependent kinase inhibitor 1-deficient mice to rheumatoid arthritis arising from interleukin-1β-induced inflammation

We recently reported that cyclin-dependent kinase inhibitor 1 (p21) deficiency induces osteoarthritis susceptibility. Here, we determined the mechanism underlying the effect of p21 in synovial and cartilage tissues in RA. The knee joints of p21-knockout (p21−/−) (n = 16) and wild type C57BL/6 (p21+/+) mice (n = 16) served as in vivo models of collagen antibody-induced arthritis (CAIA). Arthritis severity was evaluated by immunological and histological analyses. The response of p21 small-interfering RNA (siRNA)-treated human RA FLSs (n = 5 per group) to interleukin (IL)-1β stimulation was determined in vitro. Arthritis scores were higher in p21−/− mice than in p21+/+ mice. More severe synovitis, earlier loss of Safranin-O staining, and cartilage destruction were observed in p21−/− mice compared to p21+/+ mice. p21−/− mice expressed higher levels of IL-1β, TNF-α, F4/80, CD86, p-IKKα/β, and matrix metalloproteinases (MMPs) in cartilage and synovial tissues via IL-1β-induced NF-kB signaling. IL-1β stimulation significantly increased IL-6, IL-8, and MMP expression, and enhanced IKKα/β and IκBα phosphorylation in human FLSs. p21-deficient CAIA mice are susceptible to RA phenotype alterations, including joint cartilage destruction and severe synovitis. Therefore, p21 may have a regulatory role in inflammatory cytokine production including IL-1β, IL-6, and TNF-α.


Methods
Generation of homozygous p21 −/− mice. Homozygous B6.129S6 (Cg)-Cdkn1atm1 Led/J mice were obtained from the Jackson Laboratory (Bar Harbor, ME, USA). We backcrossed these mice for 10 generations against a C57BL/6 background, obtained from CREA Japan, Inc (Tokyo, Japan), and studied 10-week-old male mice (n = 16). p21 +/+ littermates were used as wild type (WT) controls (n = 16). Genotyping was performed using polymerase chain reaction (PCR)-based amplification of mouse-tail DNA with allele-specific probes. Both the p21 +/+ and p21 −/− groups contained four mice. All animals were bred in mouse houses with automatically controlled lighting (12 h light/dark cycle) and a stable temperature of 23 °C and were allowed ad libitum access to food and water throughout the study. This study was performed in strict accordance with the recommendations of the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health (Bethesda, MD, USA). All procedures were approved by the Animal Studies Committee of Kobe University, Japan (permit number: P180404). Confirms that the authors complied with the ARRIVE guidelines.
Establishment of a CAIA mouse model. A cocktail of five monoclonal antibodies recognizing conserved epitopes on various species of type II collagen (Chondrex Inc., Redmond, WA, USA) was prepared as previously described and used according to the manufacturer's instructions. Mice were injected with the cocktail of antibodies intraperitoneally (i.p.; 5 mg). Three days later, they were injected with 50-µg lipopolysaccharide (LPS) from Escherichia coli 0111: B4 (Chondrex Inc.) i.p. to induce arthritis. On days 7, 14, and 28 (counting from day 0 as the day of cocktail injection), at least four mice each from the p21 −/− and WT groups were euthanized using CO 2 . We defined the mice without injection of monoclonal antibodies as the control mice.
Evaluation of arthritis. The mice were blindly evaluated for disease progression on days 0 (n = 32), 3 (n = 24), 7 (n = 24), 10 (n = 16), 14 (n = 16), and 28 (n = 8). The severity of arthritis in each joint was graded macroscopically on a scale of 0-4, as follows: 0, normal; 1, mild swelling; 2, moderate swelling; 3, severe swelling; 4, pronounced edema of the entire paw by triple-blinded observers. The cumulative score from all four paws (maximum score of 16 per mouse) was used as the overall disease score 34 . Histological evaluation for cartilage degeneration and synovitis. Mouse knee joints were fixed using 4% paraformaldehyde (Wako, Osaka, Japan) for 24 h, decalcified with 14% ethylenediaminetetraacetic acid (EDTA; Dojindo, Kumamoto, Japan) for 7 days, and embedded in paraffin. Histological coronal sections Immunohistochemistry (IHC). Deparaffinized sections were digested with proteinase (Dako, Glostrup, Denmark) for 10 min and treated with 3% hydrogen peroxide (Wako, Osaka, Japan) to block endogenous peroxidase activity. We assessed F4/80 expression-using a previously reported scoring system for immunohistochemistry-as an immune and inflammatory cell marker because it is a well-known macrophage marker 37 , and to investigate the M1/M2 ratio, we assessed CD86 and CD206 expression as M1 and M2 macrophage markers, respectively 38 .
One coronal section from the center of the most severe lesion in each tibial plateau was scored. The numbers of stained cells were counted in three areas of high-magnification fields at both superficial and deep zones of the cartilage tissue by triple-blinded observers. The average percentages of MMP-3-, MMP-13-, p-IKKα/β-, TNF-α-, IL-1β-positive cells relative to total hematoxylin-stained cells were calculated. The positive cells were included superior of the tidemark. CD86 and CD206 expression levels in the synovium were determined semiquantitatively using the National Institutes of Health ImageJ software (http:// imagej. nih. gov/ ij/) and digitally captured images. Expression levels were determined as the average of the gray values normalized by the number of nuclei, similar to a previously published method 39 . In brief, for color deconvolution of IHC images, DAB and hematoxylin staining were digitally separated using ImageJ software with a color deconvolution plugin. Deconvoluted images with DAB staining were subjected to measurement of mean gray values, with the lower and upper thresholds set at 0 and 120 for CD86 and CD206, respectively. Coronal sections from the 16 mice were evaluated for each group.
Preparation of human synovium. Synovial tissues were obtained during a total knee joint replacement surgery from five patients with RA. All RA patients fulfilled the American College of Rheumatology 1987 revised criteria for RA 40 . OA synovial tissues were also obtained during total knee joint replacement surgery from five patients as controls. Diagnoses of OA were based on clinical, laboratory, and radiographic evaluations. All samples were obtained following the World Medical Association Declaration of Helsinki Ethical Principles for Medical Research Involving Human Subjects. The study protocol was approved by the Kobe University Graduate School of Medicine Ethics Committee, and all participants provided informed consent.
Preparation of cell culture. Primary synoviocytes were isolated and cultured from the RA and OA synovial tissues. The tissues were minced and incubated with trypsin (0.5 mg/ml; Sigma-Aldrich, St. Louis, MO, USA) for 15 min at 37 °C, and then the synovium was treated with Dulbecco's modified Eagle's medium (DMEM; Gibco/Life Technologies, Grand Island, NY, USA) containing 0.2% collagenase (Sigma-Aldrich, St. Louis, MO, USA) at 37 °C for 15 h. Dissociated cells were cultured overnight in DMEM supplemented with 10% fetal bovine serum (BioWhittaker FBS; Lonza, Walkersville, MD, USA) and 100 U/ml penicillin-streptomycin. The non-adherent cells were removed, and the adherent cells were further incubated on a 6-well plate with fresh medium (3 × 10 5 cells/well). All experiments were conducted using 3-5 passage cells.
Transfection of small-interfering RNA (siRNA). Lipofectamine RNAiMax transfection reagent (Invitrogen, Carlsbad, CA, USA) was used to transfect p21 siRNA and nonspecific siRNA control into the RA and OA human knee synoviocytes, respectively, according to the manufacturer's recommendations. Briefly, a day before transfection, the cells (3 × 10 5 cells/well) were seeded on a 6-well plate without antibiotics to achieve 30 www.nature.com/scientificreports/ were prepared and added to each well. After transfection for 24 h, the complexes were removed, and fresh medium containing 10% FBS was added.
Quantitative reverse transcriptase-polymerase chain reaction (RT-PCR). Cultured RA and OA synoviocytes were transfected with the p21 siRNA or nonspecific siRNA control. FLSs without siRNA transfection were used as controls. After transfection for 24 h, the cells were incubated for another 24 h with or without stimulation with 10-ng/ml recombinant human IL-1β (R&D systems, McKinnley, MN, USA), followed by RNA extraction using a QIA shredder and RNeasy Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer's protocol. Briefly, 1 μg of total RNA was reverse-transcribed to first-strand cDNA using 1.25-μM oligo-dT primer in 40-μl PCR buffer II containing 2.5-mM MgC1 2 , 0.5-mM dNTP mix, 0.5 U of RNase inhibitor, and 1.25 U of murine leukemia virus reverse transcriptase (PerkinElmer/Applied Biosystems, Foster City, CA, USA), at 42 °C for 1 h. The relative expression levels of mRNA of human p21, IL-6, IL-8, MMP-3, and MMP-9, were analyzed using SYBR Green RT-PCR on an ABI Prism 7500 sequence detection system (Applied Biosystems, Foster City, CA, USA). Relative gene expression was normalized against the Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) housekeeping gene using the comparative cycle threshold (Ct) method. The difference between the mean Ct values of the gene of interest and those of the housekeeping gene is denoted as ΔCt, whereas the difference between the ΔCt and the Ct value of the calibrator sample is denoted as ΔΔCt. The log 2 (ΔΔCt) value gives the relative level of gene expression. The primer sequences used to detect human p21, IL-6, IL-8, MMP-3, and MMP-9, are listed in Supplementary Table S1.
Western blot analysis. First, the cultured RA synoviocytes were treated with or without 10-ng/ml recombinant human IL-1β (R&D Systems, Minneapolis, Minnesota, USA) for 5, 10, 15, 30, 60 min, and 24 h; stimulation time for IL-1β was determined as previously reported 41 . The synoviocytes were washed with Tris-buffered saline with Tween-20 (TBST) and lysed in a buffer containing 25-mM Tris, 1% Nonidet P-40, 150 mM NaCl, 1.5 mM ethylene glycol tetraacetic acid, and a protease/phosphatase inhibitor mix (Roche Diagnostics, Basel, Switzerland). The lysates were centrifuged at 4 °C at 15,000×g for 10 min to remove cellular debris. Next, the cellular debris-free lysates were collected and mixed with 4× electrophoresis sample buffer; 15 μl of cell lysates (1.0 × 10 7 cells/ml) were electrophoresed on a 7.5-15% SDS-polyacrylamide gradient gel (Biocraft, Tokyo, Japan) and electrically transferred onto a polyvinylidene difluoride blotting membrane (GE Healthcare Life Sciences, Little Chalfont, UK). The membrane was blocked with 5% skimmed milk in TBST at 25 °C for 30 min, incubated with antibodies against anti-p-IKKα/β (Cell Signaling Technology, Danvers, MA, USA), anti-phosphor-inhibitor of κB (IκB) α (Abcam, Cambridge, UK) and anti-IκBα (Abcam, Cambridge, UK) at 4 °C for 12 h, and further incubated with horseradish peroxidase-conjugated goat anti-rabbit IgG secondary antibody at 25 °C for 1 h. The proteins were subsequently visualized using ECL Plus reagent (GE Healthcare Life Sciences, Little Chalfont, UK) in a chemilumino analyzer (LAS-3000 mini; Fujifilm, Tokyo, Japan). The cultured RA and OA synoviocytes were then transfected with p21 siRNA or nonspecific siRNA control. After 24 h of transfection, the cells were incubated with or without IL-1β stimulation for the period with the highest level of p-IKKα/β, p-IκBα, and IκBα in the western blot. Western blots of the synoviocytes were subsequently subjected to the same procedure as described above. Expression of the alpha-tubulin protein was detected using rabbit anti-alpha-tubulin polyclonal antibody (Abcam, Cambridge, UK) as a primary antibody. Protein expression was determined semi-quantitatively with the National Institutes of Health ImageJ using digitally captured images. Five different samples were analyzed for each experiment.
Statistical analysis. Statistical analysis was performed using one-way (Figs. 5b, 6a,c) or two-way (Figs. 1a,c,e,f, 2b,e, 3b,e, 4c,f,h, 6b,d) analysis of variance and Tukey's post hoc test for multiple comparisons of paired samples. The Mann-Whitney U test was used to compare two groups in vitro the relative expression of p21, ILs, and MMPs mRNA (Fig. 5a). Results are presented as means with 95% confidence intervals and were considered statistically significant at P < 0.05.
Ethics approval and consent to participate. This study was performed in strict accordance with the recommendations of the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health (Bethesda, MD, USA). All procedures were approved by the Animal Studies Committee of Kobe University, Japan (permit number: P180404). Synovial tissues were obtained during a total knee joint replacement surgery from five patients with RA and OA. All samples were obtained in accordance with the World Medical Association Declaration of Helsinki Ethical Principles for Medical Research Involving Human Subjects. The study protocol was approved by the Kobe University Graduate School of Medicine Ethics Committee, and all participants provided informed consent.

p21-deficient mice exhibited enhanced local expression of inflammatory cytokine in knee
joints. p21 −/− mice showed higher IL-1β and TNF-α expression in cartilage tissues than p21 +/+ mice at each time point except for the control (Fig. 2a,d). IL-1β and TNF-α expression in synovial tissues were also increased in p21 −/− mice compared with that in the p21 +/+ mice at each time point except for the control (Fig. 2c,f). The ratios of IL-1β-positive cells and TNF-α-positive cells in cartilage were significantly higher in p21 −/− mice than in p21 +/+ mice at each time point except for the control (Fig. 2b,e). Of note, those of IL-1β and TNF-α were significantly higher on day 7 and thereafter than those of the control (Fig. 2b,e). There was no significant difference in the positive-cell ratio of p21 −/− mice between days 7 and 14 ( Fig. 2a,b,d,e). p21 deficient mice exhibited enhanced M1 macrophage infiltration in synovial tissues. F4/80expression levels were elevated in the synovial tissue of p21 −/− mice at each time point except for the control, indicating increased macrophage infiltration (Fig. 3a). The F4/80 score was significantly higher in p21 −/− mice than in p21 +/+ mice at each time point except for the control (Fig. 3a,b). CD 86 expression in synovial tissues increased in p21 −/− mice compared with p21 +/+ mice at each time point except for the control (Fig. 3c). CD86/ CD206 expression ratio (M1/M2 ratio) was significantly higher in p21 −/− mice than in p21 +/+ mice at each time point except for the control, indicating increased M1 macrophage infiltration (Fig. 3c-e). On day 7 and after that, the F4/80 score and M1/M2 ratio were significantly higher in p21 −/− mice than in controls (Fig. 3b,e). There was no significant difference in the M1/M2 ratio of p21 −/− mice between days 7 and 14 (Fig. 3e).

p21-deficient mice increased expression of inflammatory transcription factors and exhibited
joint destruction through elevated MMPs expression. p21 −/− mice showed higher p-IKKα/β expression in the cartilage and synovial tissues than p21 +/+ mice at each time point except for the control (Fig. 4a,b). The positive-cell ratio of p-IKKα/β in the cartilage was significantly higher in p21 −/− mice than in p21 +/+ mice at each time point except for control (Fig. 4a,c), and the ratios of p-IKKα/β in p21 −/− mice on day 7 and thereafter were significantly higher than those in the control (Fig. 4c). There was no significant difference in the positivecell ratio of p21 −/− mice between days 7 and 14 (Fig. 4c). MMP-3 and MMP-9 expression levels were elevated in the synovial tissues of p21 −/− mice at each time point except for the control (Fig. 4e, Supplementary Fig.S1a).
The positive-cell ratios of MMP-3 and MMP-13 in the cartilage were significantly higher in p21 −/− mice than in p21 +/+ mice at each time point except for the control (Fig. 4d,f-h), and positive-cell ratios of MMP-3 and MMP-13 were significantly higher on day 7 and thereafter than those of the control (Fig. 4d,f-h).

Knockdown of p21 gene expression enhanced phosphorylation of IKKα/β and IκBα, and degradation of IκBα in RA synovial tissues.
Western blot results demonstrated that the expression of p-IKKα/β and p-IκBα markedly increased after 15 min treatment with 10 ng/ml IL-1β (Fig. 6a,c). Therefore, the RA FLS and OA FLS were treated with IL-1β for 15 min in subsequent experiments. The knockdown of p21 gene expression significantly enhanced IKKα/β phosphorylation in RA FLS compared with OA FLS (Fig. 6b). Western blot also demonstrated that IκBα expression markedly decreased following 15 min treatment with IL-1β (Fig. 6c). The knockdown of p21 gene expression significantly enhanced the IκBα phosphorylation and IκBα degradation in RA FLS compared with OA FLS (Fig. 6d).

Discussion
This study demonstrated enhanced cartilage degradation and more severe synovitis in response to systemic inflammation in p21-deficient mice through NF-κB signaling in cartilage and synovial tissues. Similar to our study, another report has shown increased arthritis scores and observed histological changes, including a marked increase in macrophage infiltration, in the knee synovial membrane of p21 +/+ CAIA mice 42 . Moreover, we found more severe arthritis phenotype changes in p21 −/− mice than in p21 +/+ mice on day 7 and thereafter, suggesting that p21 knockdown may exacerbate CAIA and enhance M1 macrophage infiltration in mice. IL-1β expression in p21 −/− mice reportedly increased up to 2.4-fold compared with that in p21 +/+ mice in an experimental endotoxic shock model in vivo, and that in p21 −/− bone marrow-derived macrophages with LPS stimulation increased 3.4-folds compared with p21 +/+ cells in vitro 43 . We have previously reported that p21 knockout in a murine model of destabilization of the medial meniscus increased IL-1β serum levels and local IL-1β expression in knee joints on days 1 and 56 post-surgery 28 . Macrophages initiate and maintain the inflammation, in the range of antigen presentation and phagocytosis and contribute to immunomodulation via the production of various inflammatory cytokines, including IL-1β and TNF-α [43][44][45] . The relevance between macrophages and p21 in RA has been highlighted by several studies, including the studies by Trakala et al. that reported increased IL-1β and TNF-α expression in p21 −/− macrophages in vitro 46 , and by Mavers et al., who found remarkably increased macrophage infiltration in the ankles of a p21 −/− arthritis mouse model and significantly elevated IL-1α serum levels 31 . Our current study also found that p21 −/− CAIA model mice exhibited severe joint arthritis with macrophage infiltration, especially M1 macrophages, and elevated IL-1β, IL-6, and TNF-α serum levels, as well as local IL-1β and TNF-α expression in vivo. These findings suggest that activated macrophages enhance systemic and local inflammation in this murine model of RA.
Several reports have shown that IL-1β and TNF-α stimulate NF-κB signaling and induces the expression of MMP-3, MMP-9, and MMP-13 13,47 . It is known that the IKK complex and IκBα proteins play a central role in the regulation of NF-κB activity 48,49 . IκBα protein binds tightly to NF-κB dimer and dampen NF-κB activation. Since bound IκBα requires IKK phosphorylation for basal degradation, increased p-IKKα/β and p-IκBα along with IκBα degradation activate NF-κB. In this study, we confirmed that the expression levels of p-IKKα/β, MMP-3, and MMP-13 in chondrocytes and of p-IKKα/β, MMP-3, and MMP-9 in synovial tissues were elevated in p21 −/− mice compared with those in p21 +/+ mice in vivo. These findings suggest that rapid joint destruction was caused by elevated levels of inflammation induced via p-IKKα/β signaling. Perlman et al. reported that synovial fibroblasts from p21-deficient mice enhance IL-6 and MMP-3 mRNA levels, causing a 100-fold increase in IL-6 protein levels 33 . Hence, alterations in p21 expression may enhance pro-inflammatory cytokine and MMP production, thereby promoting the development of autoimmune diseases 33 . These earlier studies support our results that p21-deficient mice exhibit enhanced MMP-3, MMP-9, and MMP-13 expression through IL-1β-and TNF-α-induced NF-κB signaling.
Consistent with our results, RA FLSs have exhibited decreased p21 expression 33 , higher IL-1β levels 6,50 , higher IL-6 and IL-8 levels 51,52 compared with OA FLSs. Moreover, other studies have demonstrated increased IL-6 and MMP levels in RA FLSs through TNF-α stimulation 53,54 . Consequently, the decline in p21 expression in RA FLSs might cause severe inflammation through IL-1β-induced NF-κB signaling. Furthermore, we have demonstrated that the knockdown of p21 in FLSs alters the cellular response to IL-1β stimulation. IL-1β stimulation increased IL-6, IL-8, MMP-3, and MMP-9 expression, and enhanced p-IKKα/β, p-IκBα activation, and IκBα degradation in both RA FLSs and OA FLSs. However, the responses to IL-1β in RA FLSs and OA FLSs were different, indicating that decreased expression of p21 in RA joints may account for the observed joint destruction.
This study has a limitation. The applicability of the results from this study to human arthritis is limited using an experimental animal model of antibody to collagen-induced arthritis. The human scenario is much more complex than can be recreated in the CAIA model.

Conclusions
Overall, we demonstrated that p21-deficient CAIA mice were susceptible to joint cartilage destruction and severe synovitis via IL-1β-and TNF-α-induced inflammation in vivo. We have also shown that knockdown of p21 led to enhanced susceptibility to inflammation through IL-1β stimulation in RA FLSs compared with that in OA FLSs. Therefore, p21 may suppress inflammatory cytokine production, including IL-1β, IL-6, and TNF-α, and represent a potential therapeutic target for novel RA treatment. However, given that p21 is an oncogene involved in cell-cycle regulation, further research is required to verify the viability and safety of p21-based therapies for RA treatment.

Data availability
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request. www.nature.com/scientificreports/ Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.