Histone deacetylase inhibitor suppression of autoantibody-mediated arthritis in mice via regulation of p16INK4a and p21WAF1/Cip1 expression




To examine whether depsipeptide (FK228), a histone deacetylase (HDA) inhibitor, has inhibitory effects on the proliferation of synovial fibroblasts from rheumatoid arthritis (RA) patients, and to examine the effects of systemic administration of FK228 in an animal model of arthritis.


Autoantibody-mediated arthritis (AMA) was induced in 19 male DBA/1 mice (6–7 weeks old); 10 of them were treated by intravenous administration of FK228 (2.5 mg/kg), and 9 were used as controls. The effects of FK228 were examined by radiographic, histologic, and immunohistochemical analyses and arthritis scores. RA synovial fibroblasts (RASFs) were obtained at the time of joint replacement surgery. In vitro effects of FK228 on cell proliferation were assessed by MTT assay. Cell morphology was examined by light and transmission electron microscopy. The effects on the expression of the cell cycle regulators p16INK4a and p21WAF1/Cip1 were examined by real-time polymerase chain reaction and Western blot analysis. The acetylation status of the promoter regions of p16INK4a and p21WAF1/Cip1 were determined by chromatin immunoprecipitation assay.


A single intravenous injection of FK228 (2.5 mg/ml) successfully inhibited joint swelling, synovial inflammation, and subsequent bone and cartilage destruction in mice with AMA. FK228 treatment induced histone hyperacetylation in the synovial cells and decreased the levels of tumor necrosis factor α and interleukin-1β in the synovial tissues of mice with AMA. FK228 inhibited the in vitro proliferation of RASFs in a dose-dependent manner. Treatment of cells with FK228 induced the expression of p16INK4a and up-regulated the expression of p21WAF1/Cip1. These effects of FK228 on p16INK4a and p21WAF1/Cip1 were related to the acetylation of the promoter region of the genes.


Our findings strongly suggest that systemic administration of HDA inhibitors may represent a novel therapeutic target in RA by means of cell cycle arrest in RASFs via induction of p16INK4a expression and increase in p21WAF1/Cip1 expression.

Rheumatoid arthritis (RA) is a chronic inflammatory disease characterized by synovial hyperplasia, with excessive inflammatory cell infiltration in the joints, leading to erosion of the articular cartilage and marginal bone, with subsequent joint destruction (1). Despite an explosion of information over the last 2 decades, a detailed understanding of the mechanisms of synovial hyperplasia and inflammation is lacking. Recent reports have implicated rapid proliferation of synovial cells, overexpression of inflammatory genes, and impairment of apoptosis, which can allow the persistence of abnormal cells (2–4), in the disease process.

Cyclin-dependent kinases (CDKs) are essential for the progression of the cell cycle. The CDK inhibitor genes p21WAF1/Cip1 and p16INK4a have specific, and probably complementary, effects in cell cycle regulation. The p21WAF1/Cip1 gene has multiple functions, such as inhibition of a wide range of CDKs, including CDK-1, CDK-2, CDK-4, and CDK-6 (5), promotion of active CDK complex formation (6), and induction of apoptosis (7–9). The p16INK4a gene specifically binds to cyclin D and prevents it from forming a complex with CDK-4 and CDK-6, which function as regulators of cell cycle progression in the G1 phase by contributing to phosphorylation of the retinoblastoma protein (5, 10). The important physiologic role of p16INK4a lies in the implementation of replicative senescence, the irreversible arrest of cell growth (11). More recently, it was demonstrated that adenoviral gene transfer of p16INK4a inhibited the growth of human RA synovial fibroblasts (RASFs) without inducing apoptosis (12). Forced expression of p21WAF1/Cip1 or p16INK4a genes within the joint by intraarticular administration of the same virus ameliorated adjuvant-induced arthritis in rats (13) as well as collagen-induced arthritis in mice (14).

Transcription factors are central players in cellular events, including proliferation, cell cycle progression, differentiation, cellular stress responses, inflammation, and cell death. It has been proposed that histone acetylation plays an important role in transcriptional regulation. Acetylation loosens histone–DNA bonds and reduces the interaction of histones with ATP-dependent chromatin-remodeling complexes, which facilitates the binding of transcription factors (15). The activity of transcription factors and the expression of various genes are integrated by histone acetyltransferase (HAT) and histone deacetylase (HDA) (16–18), and the balance of the activities of HAT and HDA govern the acetylation status of the histones (19). At present, the roles of transcription factors in arthritis are largely unexplored and may offer multiple therapeutic targets for investigation.

HDA catalyzes the removal of acetyl groups on the amino-terminal lysine residues of core nucleosomal histones. Various natural and synthetic compounds that inhibit HDA activity that have been identified to date include: trichostatin A (TSA) (20), apicidin (21), butylate (22), depsipeptide (FK228) (23), depudecin (24), suberoylanilide hydroxamic acid (SAHA) (25), oxamflatin (26), and MS-27-275 (27). HDA inhibitors have emerged as a potentially promising new class of anticancer drugs based on their ability to activate a variety of genes implicated in the regulation of cell survival, proliferation, differentiation, and apoptosis (28). HDA inhibitor–induced inhibition of growth in tumor cell lines has been shown to accompany cell cycle arrest and induction of the cell cycle inhibitor p21WAF1/Cip1 (29–32).

Chung et al (33) demonstrated that the beneficial effects of ointment therapy in ameliorating symptoms in their rat model of adjuvant-induced arthritis by HDA inhibitors occurred via induction of p16INK4a and p21WAF1/Cip1. Questions arise whether HDA inhibitors have an inhibitory effect on the in vitro proliferation of RASFs and whether systemic administration of HDA inhibitors has therapeutic value in the treatment of arthritis with multiple joint involvement that is seen in RA in humans. To test the question, mice with autoantibody-mediated arthritis (AMA) were treated by systemic administration of FK228. Human RASFs were used for the investigation of the in vitro effects of FK228 and their molecular mechanisms. The results of this study indicate some important roles of the cell cycle regulators and their histone acetylation status in the mechanism of synovial inflammation. The findings also indicate that HDA inhibitors may be worthy of further investigation as novel treatment targets in RA.



FK228 was provided by Fujisawa Pharmaceutical (Osaka, Japan). For the in vitro studies, FK228 was dissolved in ethanol and DMSO and then diluted with experimental medium. For the in vivo study, FK228 was dissolved in and diluted with 10% polyoxythylene (60)–hydrogenated castor oil in saline (HCO60 saline).

Animals, arthritis induction, and FK228 treatment.

Nineteen male DBA/1 mice (Charles River Japan, Yokohama, Japan) ages 6–7 weeks old were used to evaluate the disease-modifying activity of FK228 in vivo. Mice were housed at the Laboratory Animal Center for Biochemical Research, Okayama University Graduate School of Medicine and Dentistry, under standard diurnal conditions and were fed a standard commercial diet and given tap water ad libitum. Arthritis was induced by an arthritogenic cocktail of 4 monoclonal antibodies (mAb) to type II collagen (Chondrex, Redmond, WA) combined with lipopolysaccharide simulation according to Terato's method, as previously described (34, 35). Mice were injected intravenously with 2 mg of mAb on day 0 and day 1 (4 mg total) followed by intraperitoneal injection of 50 μg of lipopolysaccharide on day 2. After the onset of clinically distinct arthritis, the treatment group (n = 10) was given a single intravenous dose of FK228 (2.5 mg/kg of body weight) on day 4. Control mice (n = 9) were injected with 10% HCO60 saline alone on day 4 (Figure 1A).

Figure 1.

Effect of intravenous administration of FK228 (2.5 mg/kg of body weight) on the development of arthritis in mice with autoantibody-mediated arthritis (AMA). A, Induction of arthritis and treatment protocol. MAbs = monoclonal antibodies; i.p. = intraperitoneally; LPS = lipopolysaccharide; i.v. = intravenously. B, Representative photographs of the fore paws (top) and hind paws (bottom) of FK228-treated (n = 8) and untreated (AMA; n = 9) mice on day 15. C, Severity of clinical signs in untreated (AMA) and FK228-treated mice (see Materials and Methods for scoring system). D, Incidence of arthritis (average percentage of paws affected) in the 2 groups of mice. E, Changes in body weight in the 2 groups of mice. ∗ = P < 0.01 by Fisher's least significant difference test.

Clinical evaluation of arthritis.

The mice were monitored for the development of arthritis every day after the first mAb injection. Arthritis was scored as 4 grades according to the method of Terato et al (34). Each limb was graded individually on a scale of 0–4 (maximum cumulative clinical arthritis score 16 per mouse), where 0 = normal, 1 = mild but definite redness and swelling of the ankle or wrist or redness and swelling of any degree in any single digit, 2 = moderate to severe redness and swelling of the ankle and wrist, 3 = redness and swelling of the entire foot including the digits, and 4 = maximally inflamed limb, with involvement of multiple joints.

Histologic analysis of hind paws.

Mice were euthanized by systemic perfusion of 4% paraformaldehyde under general anesthesia on day 15. Radiographs were obtained, and limbs were dissected and fixed in the same solution for 24 hours. The samples were decalcified in 0.3M EDTA (pH 7.5) for 7–10 days, divided into 2 blocks along the sagittal plane, dehydrated by a graded series of ethanol, and embedded in paraffin. Standard sagittal sections measuring 4.5 μm were prepared and stained with hematoxylin and eosin.

Histologic examinations for synovial inflammation and bone and cartilage damage were performed independently by 2 of the authors (KN and SM). Sections were graded according to the system described by Sancho et al (36), where 0 = no inflammation, 1 = slight thickening of the synovial cell layer and/or some inflammatory cells in the sublining, 2 = thickening of the synovial lining, infiltration of the sublining, and localized cartilage erosions, and 3 = infiltration in the synovial space, pannus formation, cartilage destruction, and bone erosion.


Immunohistochemistry was performed as previously described (35). Two days after intravenous administration of FK228, formalin-fixed paraffin-embedded hind limbs from a second group of mice with AMA were prepared for analysis. For primary antibodies, polyclonal goat anti-mouse antibody against tumor necrosis factor α (TNFα; 20 μg/ml) and polyclonal rabbit anti-mouse primary antibodies against interleukin-1β (IL-1β; 10 μg/ml) (both from R&D Systems, Minneapolis, MN), p16INK4a (sc-1207; 4 μg/ml) and p21WAF1/Cip1 (sc-756; 4 μg/ml) (both from Santa Cruz Biotechnology, Santa Cruz, CA), and acetylated histone H3 (Lys9; 1 μg/ml) and acetylated histone H4 (Lys12; 1 μg/ml) (both from Cell Signaling Technology, Beverly, MA) were used.

Briefly, sections were treated with primary antibodies diluted in phosphate buffered saline (PBS) containing 0.1% NaN3 and 0.1% bovine serum albumin (Sigma-Aldrich, St. Louis, MO). After incubation for 30 minutes at room temperature with 7.5 μg/ml of biotinylated goat anti-goat or anti-rabbit IgG (Vector, Burlingame, CA), tissues were then treated with avidin–biotin–peroxidase complex for 30 minutes, visualized with diaminobenzidine (Nichirei, Tokyo, Japan) in 0.01% H2O2 for 5 minutes, and counterstained with methyl green or hematoxylin. The negative control was prepared in the same manner, except that the primary antibody was omitted. The sections were then examined under a light microscope.

Isolation and culture of human RASFs.

With the patients' written permission, fresh synovial tissues were obtained from the knee joints of 7 RA patients who were undergoing total knee arthroplasty. Tissues were minced and digested immediately with collagenase (Wako, Osaka, Japan) and DNase (Sigma-Aldrich) at 37°C, as previously described (37). Tissue debris was removed with a cell strainer, and cells were washed twice with medium consisting of Dulbecco's modified Eagle's medium (DMEM; Sigma-Aldrich) supplemented with 10% HEPES (Life Technologies, Tokyo, Japan), 100 IU/ml of penicillin, and 100 mg/ml of streptomycin. The resultant single cells were dispensed into the wells of a 24-well microtiter plate (Costar, Cambridge, MA) at a density of 2 × 106 cells/ml in 2 ml of DMEM supplemented with 10% HEPES, 100 IU/ml of penicillin, and 100 mg/ml of streptomycin. The plates were incubated at 37°C in a humidified atmosphere containing 5% CO2. Synovial tissue cell cultures were divided once weekly until the primary cultures had reached confluence. After the third passage, the cells appeared to be morphologically homogeneous fibroblast-like cells.

Cell proliferation assay.

The cells were seeded at a density of 1 × 103/well into a 96-well culture plate (Costar) containing 0.1 ml of DMEM and 10% FCS and were allowed to adhere overnight. Cells were then treated with both recombinant human TNFα (1 ng/ml) and recombinant human IL-1β (10 ng/ml) for 1 hour to activate synovial fibroblasts, followed by treatment with FK228 at various concentrations. Cell viability in the 96-well culture plates was evaluated at 24, 48, and 72 hours after FK228 treatment, using the colorimetric MTT assay (Chemicon, Temecula, CA) according to the manufacturer's instructions. The experiments were repeated 4 times on cells from 2 different RA patients.

Assessment of cell morphology.

Since HDA inhibitors have been shown to induce morphologic changes in tumor cells (20, 24, 26), a morphologic examination of the FK228-treated RASFs was performed by phase-contrast microscopy, light microscopy, and transmission electron microscopy. Cells were stimulated with both recombinant human TNFα (1 ng/ml) and recombinant human IL-1β (10 ng/ml) and then incubated with or without FK228 (10 nM) for 24 or 48 hours. After examination by phase contrast microscopy, cells were pelletized, fixed, and embedded in hydrophilic resin (LR-White). Semithin sections were stained with toluidine blue and examined by light microscopy. Ultrathin sections were contrasted with aquenous uranyl acetate and lead citrate for examination by transmission electron microscopy (model 7100 transmission electron microscope; Hitachi, Tokyo, Japan) (38).

Western blotting for the detection of p16INK4a and p21WAF1/Cip1 proteins.

The data obtained above suggested that the HDA inhibitor may have a therapeutic effect, at least in part, via regulation of the cell cycle of RASFs. In cancer cells, p21WAF1/Cip1 and p16INK4a are among the genes that are highly responsive to HDA inhibitors. These data prompted us to examine the effects of FK228 on the expression of p21WAF1/Cip1 and p16INK4a in RASFs. Cells were seeded at a density of 2 × 105/well into a 6-well culture dish. Cells were stimulated with both TNFα and IL-1β as described above, and then treated with FK228 for up to 24 hours.

The cells were then washed twice with cold PBS, lysed in lysis buffer (10 mM Tris HCl, pH 7.4, 150 mM NaCl, 1% Nonidet P40 [NP40], 1 mM EDTA, 0.1% sodium dodecyl sulfate [SDS], 0.1% deoxycholic acid, and protease inhibitor mixture) for 10 minutes, and scraped. The extracts were centrifuged at 14,000 revolutions per minute for 15 minutes at 4°C. Protein concentrations were measured and equalized using a Bio-Rad protein assay kit (Bio-Rad, Hercules, CA) according to the manufacturer's instructions. A total of 30 μg of protein per lane was run on 12% SDS–polyacrylamide electrophoresis gels and then transferred onto polyvinylidene difluoride membranes (Immobilon; Millipore, Bedford, MA).

Membranes were blocked overnight at 4°C in blocking buffer (5% nonfat dry milk). The blots were then probed for 2 hours with mouse monoclonal anti-p16INK4a antibody or anti-p21WAF1/Cip1 antibody and mouse monoclonal anti–β-actin antibody, respectively. After washing, blots were exposed to horseradish peroxidase–conjugated anti-mouse secondary antibody (Amersham, Arlington Heights, IL) and visualized by the enhanced chemiluminescence detection system (Amersham) according to the manufacturer's instructions. We used β-actin as an internal control to confirm that the amounts of protein were equal.

Real-time polymerase chain reaction (PCR) for the quantitative detection of p16INK4a and p21WAF1/Cip1 messenger RNA (mRNA).

Cells were seeded at a density of 1 × 106/well into 6-well dishes, stimulated with both TNFα and IL-1β for 1 hour, and incubated with or without 10 ng/ml of FK228 under an atmosphere of 5% CO2 for the time periods indicated below. Total RNA was isolated from cultured cells with Isogen reagent (Nippon Gene, Toyama, Japan). The RNA was reverse-transcribed using Rever Tra Ace (Toyobo, Tokyo, Japan). Primers for p21WAF1/Cip1 were from Search-LC (Heidelberg, Germany). The primer sequences of p16INK4a (39) and GAPDH used were as follows: for p16INK4a, 5′-AAG-CCA-TTG-CGA-GAA-CTT-3′ (forward) and 5′-CAG-AGG-GCA-GAA-AGA-AAA-3′ (reverse); and for GAPDH, 5′-CAT-TGG-CAA-TGA-GCG-GTT-C-3′ (forward) and 5′-GGT-AGT-TTC-GTG-GAT-GCC-ACA-3′ (reverse).

Real-time quantitative PCR reactions were performed on a LightCycler instrument (Roche Diagnostics, Mannheim, Germany) using a LightCycler FastStart DNA Master SYBR Green I kit (Roche Molecular Biochemicals, Mannheim, Germany) as recommended by the manufacturer. The final expression value was calculated by dividing the level of p21WAF1/Cip1 or p16INK4a mRNA expression by the level of GAPDH mRNA expression, and each value at time 0 was set as 1.


Cells were cultured in 2-well slide chambers, fixed with 2% paraformaldehyde in PBS, washed with PBS, permeabilized with 0.1% Tween 20 (Sigma-Aldrich) for 10 minutes at room temperature, and nonspecific staining was blocked by 1% bovine serum albumin. For immunocytochemistry, cells were incubated overnight at 4°C with rabbit polyclonal anti-p16INK4a antibody (5 μg/ml), then for 1 hour at room temperature with Alexa Fluor 488 goat anti-rabbit IgG (Molecular Probes, Eugene, OR), then overnight at 4°C with mouse monoclonal anti–Ki-67 antibody (5 μg/ml; Zymed, Burlingame, CA), and then for 1 hour at room temperature with Alexa Fluor 594 mouse anti-rabbit IgG (Molecular Probes). After labeling of the cell nuclei with Hoechst 33342 (ICN Biomedicals, Aurora, OH), they were examined by fluorescence microscopy (model CW-4000 fluorescence microscope; Leica, Wetzlar, Germany).

Chromatin immunoprecipitation assays for acetylation of p16INK4a and p21WAF1/Cip1 promoter regions.

Chromatin immunoprecipitation assays were performed using 8 × 105 cells on 6-cm dishes for each sample. The cells were stimulated with both TNFα (10 ng/ml) and IL-1β (1 ng/ml) for 1 hour, FK228 (50 nM) was added, and the cells were incubated for 24 hours. Cells were fixed with 1% formaldehyde for 20 minutes at room temperature, and then incubated with 0.125M glycine for 5 minutes. The precipitated cells were suspended in cell lysis buffer (5 mM PIPES, pH 8.0, 85 mM KCl, 0.5% NP40, and protease inhibitors) and incubated on ice for 10 minutes. Cells were then pelletized, resuspended in 100 μl of nuclear lysis buffer (50 mM Tris HCl, pH 8.1, 10 mM EDTA, 1% SDS, and protease inhibitors) for 10 minutes on ice. Lysates were sonicated 3 times with 10-second bursts. Supernatants were diluted in equal amounts of immunoprecipitation buffer (50 mM Tris HCl, pH 7.5, 100 mM NaCl, 0.05% NP40, 1 mM EDTA, 2.5 mM EGTA, 10% glycerol, and protease inhibitors).

A 10% volume of each sample was stored as an input sample. One-half of the remaining sample was rotated with protein A/G beads (Sigma-Aldrich) containing 1 μl of anti–acetyl-lysine antibodies (Upstate Biotechnology, Lake Placid, NY) or 1 μg of rabbit IgG (Sigma-Aldrich) for 3 hours at 4°C. Pelletized beads were washed with immunoprecipitation buffer and suspended in elution buffer 1 (10 mM Tris HCl, pH 8.1, 1 mM EDTA, and 1% SDS) for 15 minutes at 65°C. Supernatants were removed and stored. Beads were resuspended in elution buffer 2 (10 mM Tris HCl, pH 8.1, 1 mM EDTA, 0.67% SDS) for 15 minutes at 65°C. The supernatants were removed and combined with stored supernatants.

For the input fractions, 4 μl of 10% SDS and 40 μl of Tris–EDTA buffer (pH 7.5) were added. Samples were incubated for 4 hours at 65°C, and treated overnight at 37°C with 2 μl of proteinase K (20 mg/ml; Invitrogen, San Diego, CA) and 100 μl of Tris–EDTA buffer. DNA samples were recovered by phenol/chloroform extraction and ethanol precipitation.

The following primers were used to amplify the promoter regions of p16INK4a and p21WAF1/Cip1 in 35 cycles of PCR reactions: for p16INK4a, 5′-CCG-CGA-TAC-AAC-CTT-CCT-AAC-3′ (forward) and 5′-ATC-AAG-GGT-TGA-GGG-GGT-AG-3′ (reverse); and for p21WAF1/Cip1, 5′-ACT-TCC-CTC-CTC-CCC-CAG-T-3′ (forward) and 5′-TCA-GCT-GCA-TTG-GGT-AAA-TC-3′ (reverse).

Statistical analysis.

Statistical analysis was performed by one-way analysis of variance and subsequent Fisher's least significant difference test or Mann-Whitney U test using StatView-J 5.0 software (SAS Institute, Cary, NC) for the Macintosh. P values less than 0.05 were considered statistically significant.


Amelioration of synovial proliferation and joint destruction after FK228 treatment in mice with AMA.

To evaluate the in vivo effects of FK228, AMA was induced in 19 mice. Clinically apparent arthritis developed on day 3, with marked swelling or redness of the joints of the limbs. Arthritis in the control mice with AMA (n = 9) progressed rapidly and markedly. Clinical symptoms of active arthritis reached a peak on day 8, with a mean ± SD score of 15.4 ± 1.0. In contrast, arthritis decreased rapidly in the mice treated on day 4 with a single intravenous injection of FK228 (2.5 mg/kg; n = 10). The mean score in the FK228-treated mice on day 8 was 0.38 ± 0.74. The symptoms of clinical arthritis were barely apparent on day 10 in these mice, and continued to diminish until the end of the observation period, on day 15 (Figures 1B, C, and D).

The in vivo toxicity of FK228 was assessed by daily monitoring of body weight, food intake, and behavior in the treated and control groups. One mouse in the FK228-treated group died of unknown causes on day 5. None of the control group mice died during the treatment period. At 14 days after the initiation of arthritis, there were no significant differences in the body weights in the 2 groups of mice (Figure 1E). A skin ulcer developed at the site of FK228 injection (on the tail) of 1 of the remaining 9 mice; this mouse was excluded from the analysis. Sclerotic changes of vessels and edematous changes around them made it difficult to pursue a second intravenous injection of reagents in the treatment group.

Radiologic and histologic assessment of the severity of joint destruction.

Radiographs of the hind paws of mice with AMA (control group) showed joint space narrowing, bone erosion, and atrophy (Figure 2B). Sections of paw joint tissues from the control group showed marked pathologic changes, including synovial hyperplasia, with a large number of infiltrated inflammatory cells, extensive pannus formation at the cartilage–bone junction, and severe cartilage destruction (Figures 2E and H). In contrast, radiographs of the hind paws of FK228-treated mice with AMA showed normal-appearing joints, with intact joint spaces and smooth joint surfaces (Figure 2C). Histologic sections of the paw joints from the treatment group showed little or mild inflammatory cell infiltration, no pannus invasion, and intact bone and cartilage structures (Figures 2F and I).

Figure 2.

Effects of FK228 on joint pathology in mice with autoantibody-mediated arthritis (AMA). Representative results of A–C, radiologic and D–I, histologic examinations of hind paws harvested on day 15 are shown. A, D, and G, Normal joints, showing no evidence of arthritis. B, E, and H, Arthritic joints of untreated mice. The radiograph shows joint space narrowing, bone atrophy (arrowhead), and erosions at the base of the proximal phalanx (arrow). Histologic sections show marked synovial hyperplasia, inflammatory cell infiltration, pannus invasion (arrow), and bone and cartilage destruction. C, F, and I, Joints of FK228-treated mice show normal joint surfaces, bone density, and bone and cartilage structure, and there is no sign of synovial inflammation. (Hematoxylin and eosin stained; original magnification × 50 in D, E, and F; × 200 in G, H, and I.)

The numbers of bilateral hind paw samples with histologic grades 0, 1, 2, and 3 were 0, 3, 7, and 8, respectively, in the control group (n = 9) (Table 1). The corresponding values were 10, 6, 0, and 0, respectively, in the treatment group (n = 8). Histologic changes observed in the treatment group were significantly less severe than those observed in the control group (P < 0.0001).

Table 1. Histologic findings in the hind paws of untreated and FK228-treated mice with autoantibody-mediated arthritis*
 Histologic scoreTotal
  • *

    Mice treated with FK228 received 2.5 mg/kg intravenously. Values are the number (%) of animals. Histologic changes in the 2 groups differed significantly (P < 0.0001, by Mann-Whitney U test).

Untreated (n = 9)0 (0)3 (16.7)7 (38.9)8 (44.4)18 (100)
FK228 treated (n = 8)10 (62.5)6 (37.5)0 (0)0 (0)16 (100)

Immunohistochemistry findings.

To evaluate the in vivo effects of FK228, immunohistochemical studies were performed using synovial tissues obtained on day 6 from untreated and treated (48 hours after FK228 administration) mice with AMA. Histologically, samples from both untreated and treated mice showed synovial proliferation and infiltration by inflammatory cells.

We next examined the endogenous levels of acetylated histones H3 and H4 to determine the amount of histone hyperacetylation in the nucleus of the synovial cells, which served as a marker of the effects of HDA inhibition. Histone hyperacetylation was not apparent in the synovial tissue of untreated mice. However, 48 hours after FK228 treatment, numerous synovial cells showed positive reactivity for acetylated histones H3 and H4 in the nucleus (Figures 3A and B). Synovial lining cells and fibroblasts in the sublining cell layer showed strong staining for both TNFα and IL-1β. However, staining for both TNFα and IL-1β was markedly reduced in the synovium of mice treated with FK228 (Figures 3C and D) compared with that in untreated mice. In addition, synovial cells with nuclear expression of p16INK4a and p21WAF1/Cip1 were increased in the samples from FK228-treated mice (Figures 3E and F).

Figure 3.

Representative immunohistology sections of the hind paws obtained on day 6 from mice with autoantibody-mediated arthritis (AMA). Intravenous treatment with FK228 induced the hyperacetylation of A, histone H3 and B, histone H4 in synovial cells of the lining and sublining layers (arrows), as compared with untreated mice (AMA). There is increased expression of C, tumor necrosis factor α (TNFα) and D, interleukin-1β (IL-1β) in the synovial lining cells of untreated mice and a marked reduction in the expression of these cytokines in the synovial tissues of FK228-treated mice. FK228 treatment also induced an up-regulation of the nuclear expression of E, p16INK4a and F, p21WAF1/Cip1 proteins in synovial cells (arrows). (Original magnification × 200.)

Inhibition of cell proliferation and alteration of cell morphology after FK228 treatment.

We examined the effects of the HDA inhibitor FK228 on the proliferation of human RASFs in vitro. Cells were cultured with or without 0.1, 0.5, 1, 10, or 100 nM FK228 for 1–3 days. FK228 concentrations of 10 nM and 100 nM inhibited cell proliferation at each time point examined (Figure 4).

Figure 4.

Effects of FK228 on the viability of rheumatoid arthritis synovial fibroblasts, as determined by MTT assay. Cells were treated with various concentrations of FK228 and were compared with vehicle-treated cells at the indicated times. The results were determined as optical density units and are expressed as a percentage of the control (0.1% ethanol [EtOH]). ∗ = P < 0.001 versus vehicle treatment.

RASFs showed a morphologically homogeneous fibroblast-like appearance when cultured in the absence of FK228 (Figure 5A). FK228 treatment (10 nM) dramatically changed the cell morphology to an elongated shape, with filamentous protrusions (Figures 5B and C). Light microscopic examination of pelletized cells from untreated mice showed round cell shapes, with clear nuclear membranes and central bodies (Figures 5D and F). After 48 hours of treatment with FK228, most cells remained viable but had a large or bipolar shape, with an obscure nucleus and increased cytoplasmic granularity (Figures 5E and G). Cells with an appearance consistent with apoptosis, such as shrinkage, nuclear condensation, and membrane blebbing, were rarely seen.

Figure 5.

Morphologic features of human rheumatoid arthritis synovial fibroblasts (RASFs) were examined by A–C, phase-contrast microscopy, D and E, light microscopy, insets, oil-immersion light microscopy, and F and G, electron microscopy. A, D, and F, Untreated RASFs, showing a homogeneous fibroblast-like appearance. B, C, E, and G, RASFs treated with 10 nM FK228 for the indicated times, showing dramatic alterations in morphology. Light and electron microscopy were performed on pelletized RASFs (see Materials and Methods for details). Sections shown in D and E and their insets were stained with toluidine blue. (Original magnification x 100.) Bar = 1 μm.

Induction of p16INK4a and increase in p21WAF1/Cip1 expression at the mRNA and protein levels after FK228 treatment.

The effects of FK228 on p16INK4a and p21WAF1/Cip1 protein levels were determined by Western blot analysis. RASFs were pretreated with a combination of TNFα and IL-1β for 1 hour and then cultured in the presence and absence of FK228 for 2, 6, 15, and 24 hours. Proteins and mRNA were prepared from the cells at each time point. RASFs expressed p21WAF1/Cip1 protein but not p16INK4a protein under the in vitro conditions used in our experiment (results not shown). Stimulation with the proinflammatory cytokines had no apparent effect on the expression of p21WAF1/Cip1 and p16INK4a at the protein level. However, after incubation with FK228, the protein levels of both p16INK4a and p21WAF1/Cip1 were increased within 24 hours (Figure 6A).

Figure 6.

Expression of p16INK4a and p21WAF1/Cip1 at the protein and mRNA levels induced by treatment with FK228 (10 nM). A, Western blot analysis of p16INK4a and p21WAF1/Cip1 protein levels with and without FK228 treatment for up to 24 hours. B and C, Real-time polymerase chain reaction analysis of early changes in the expression of p16INK4a and p21WAF1/Cip1 mRNA, with and without FK228 treatment for up to 48 hours. Values are the mean and SD and are representative of 3 independent experiments. ∗ = P < 0.05 versus time 0.

The effect of FK228 on p16INK4a and p21WAF1/Cip1 mRNA levels in RASFs was determined by real-time PCR. Levels of p16INK4a mRNA were increased 4.5-fold at 2 hours of culture with FK228 and were decreased at 24 hours, but remained elevated above the levels in cultures without FK228. Levels of p21WAF1/Cip1 mRNA also increased by 2-fold at 2 hours, but decreased to the level detected in cultures without FK228 by 24 hours (Figures 6B and C).

Cell cycle arrest and p16INK4a expression.

To demonstrate that cell cycle arrest occurred in synchrony with the up-regulation of the nuclear expression of p16INK4a after FK228 treatment, we performed double immunocytochemistry using p16INK4a and Ki-67, a marker of cell proliferation that recognizes all but the G0 phase of the cell cycle. At 48 hours, RASFs stimulated with both TNFα and IL-1β clearly showed positive reactivity for Ki-67, but positive staining for p16INK4a was not seen in the nucleus. In contrast, FK228 treatment induced p16INK4a protein expression in RASFs, but Ki-67 staining was negative. This finding suggests that FK228 induced cell cycle arrest in synchrony with the induction of p16INK4a expression (Figure 7).

Figure 7.

Double immunocytochemistry for the expression of p16INK4a and Ki-67 protein in rheumatoid arthritis synovial fibroblasts (RASFs) with or without FK228 treatment (10 nM) for 48 hours. Shown are nuclear staining with Hoechst 33342 and immunostaining for p16INK4a and Ki-67. FK228 treatment induced nuclear expression of p16INK4a in RASFs that were negative for Ki-67, which indicates arrest of the cell cycle (original magnification × 200).

Increased acetylated histones in chromatin associated with the p16INK4a promoter regions after FK228 treatment.

To confirm the effects of FK228 on the augmentation of p16INK4a and p21WAF1/Cip1 expression, we performed chromatin immunoprecipitation assays using anti–acetyl-lysine antibodies and the specific primers of their promoter regions. After stimulation with both TNFα and IL-1β, FK228 treatment increased the acetylation of p16INK4a promoter regions (Figure 8A). In contrast, the promoter regions of the p21WAF1/Cip1 gene were acetylated by cytokine stimulation (Figure 8B). In addition, no significant differences were observed in the acetylation status of p21WAF1/Cip1 promoter regions after FK228 treatment (Figure 8B). These findings suggest that under these in vitro conditions, FK228 may modify the progression of the cell cycle, at least in part, by inducing histone acetylation of the p16INK4a promoter regions.

Figure 8.

Induction of the acetylation of p16INK4a promoter regions by treatment with FK228 (50 nM). A, The 184-bp polymerase chain reaction fragment amplified with specific primers for the p16INK4a promoter regions. DNA isolated from the immunoprecipitated chromatin using anti–acetyl-lysine antibodies was amplified in the cells treated with histone deacetylase (HDA) inhibitors. B, The 199-bp fragments of the promoter region of the p21WAF1/Cip1 gene were detected in untreated cells and in HDA inhibitor–treated cells after stimulation with both tumor necrosis factor α (TNFα) and interleukin-1β (IL-1β). Samples were also immunoprecipitated with rabbit IgG and with anti–acetyl-lysine antibodies (acetyl).


Various experiments examining the enhanced expression of immunosuppressive molecules have been performed in animal models of arthritis. Therapeutic induction of p16INK4a or p21WAF1/Cip1 by means of gene transfer into the joints of rats or mice resulted in a dramatic amelioration of synovial inflammation and cartilage destruction (12–14). The results of in vivo gene transfer were further supported by the recent report that adenovirus-mediated gene transfer of p21WAF1/Cip1 suppressed the spontaneous induction of IL-6 and matrix metalloproteinase 1 (MMP-1) in RASFs (40). However, the method of local injection of viruses into a single joint is not practical for treatment in the clinical setting. In another recent study, Chung et al (33) first demonstrated that HDA inhibitors (10% phenylbutyrate cream and 1% TSA ointment) induced the expression of p21WAF1/Cip1 and p16INK4a in synovial cells and inhibited the expression of TNFα in affected tissues of rats with adjuvant-induced arthritis. Ointment therapy with HDA inhibitors was shown to reduce joint swelling and ameliorate the pathologic features of joint destruction. However, treatment was limited to small joints close to the body surface, and the precise measurement of uptake of the agents was not determined.

Findings of our study also support the idea that modulation of the transcription activity of specific promoters by the local release or perturbation of the chromatin structure by treatment with FK228 might effectively prevent the synovial proliferation and joint destruction seen in RA. Surprisingly, a single intravenous injection of FK228 at a concentration of 2.5 mg/kg of body weight (maximum tolerated dose 3.2 mg/kg in mice) was sufficient to inhibit synovial proliferation, mononuclear cell infiltration, pannus invasion, and cartilage destruction over the 2 weeks of observation. Tissue levels of TNFα and IL-1β in the FK228-treated mice were clearly lower on day 6 than those in the untreated mice. These data indicate that the antiinflammatory properties of FK228 occur via suppression of cytokines and inhibition of the proliferation of RASFs.

Our in vitro experiments with RASFs demonstrated that FK228 inhibited their proliferation. We also demonstrated an epigenetic mechanism for the induction of p16INK4a and the up-regulation of p21WAF1/Cip1 expression induced in vitro in RASFs by FK228 treatment. These results may shed new light on the possible pharmacologic modulation of cell cycle regulators as an attractive treatment strategy for RA. TSA, another HDA inhibitor, also altered the cell morphology and inhibited the proliferation of RASFs in a dose-dependent manner (data not shown). However, an attempt to compare the in vivo effects of FK228 with a low dose of TSA (0.5 and 1 mg/kg of body weight by daily subcutaneous injection) in the same mouse model failed to show adequate suppressive effects for analytical purposes. This was probably due to its instability in vivo. Further experiments using a higher dose of TSA are under way in our laboratory, and the findings will be reported separately.

Previous reports have suggested that HDA inhibitors induce apoptosis in several types of tumor cells through cell cycle arrest that is mediated by CDK inhibitor p21WAF1/Cip1 (23, 29–32, 41, 42). The mitochondrial membrane damage and apoptosis induced by HDA inhibitors such as SAHA, FK228, and oxamflatin were shown to be inhibited by the overexpression of Bcl-2 (41). TSA has been shown to increase the activity of the proapoptotic proteins or Bax (31), Bid (41), and Bad (42) and to decrease the activity of Bcl-2 (31). Another possible apoptotic mechanism might be that normally silenced cell death genes are transcriptionally activated by stimulation with HDA inhibitors. However, our phase-contrast microscopy analyses showed that FK228 induced morphologic changes in human RASFs, from cells that were fibroblast-like to cells that had an elongated shape with filamentous protrusions, suggesting a possible role of HDA inhibitors in the formation of stress fibers and in the control of cell growth as reported previously (20, 24, 26).

In contrast, only small populations of morphologically apoptotic cells were seen when examined by light and electron microscopy, even after FK228 treatment. This finding is consistent with those of previous studies showing that the expression of p16INK4a did not correlate with the detection of apoptosis in human synovial membrane (43), and in vitro, p21WAF1/Cip1 or p16INK4a gene transfer of RASFs inhibited the cell cycle without having any effect on apoptosis (12, 40). It is known that the progressive accumulation of p16INK4a as cells age may be induced by a senescence timer (5). Alcorta et al (44) reported the involvement of p16INK4a in the replicative senescence of normal human fibroblasts, suggesting that senescence requires the expression of both p21WAF1/Cip1 and p16INK4a in a multistep process. Since the senescence of human fibroblasts is closely correlated with the expression of p21WAF1/Cip1 and p16INK4a genes, we speculate that cell cycle arrest and altered cell shape and cytoskeletal architecture induced by FK228 might be related to the senescence-like cell changes in human RASFs, which could contribute to the suppression of synovial proliferation and the subsequent inflammatory processes.

The precise mechanisms that regulate the expression of p16INK4a and p21WAF1/Cip1 in RA synovial cells remain unknown. Since rheumatoid synovial cells vigorously proliferate in affected joints, they may have certain failures in cell cycle regulation. Western blot analysis revealed that cultured human RASFs did not express p16INK4a, even after stimulation with cytokines. Taniguchi et al (12) analyzed the pattern of CDK inhibitor gene expression in RASFs and reported that neither p16INK4a nor p21WAF1/Cip1 was expressed, but both were readily induced when cell growth was inhibited in vitro. Although the induction of p21WAF1/Cip1 was observed in normal fibroblasts and in fibroblasts derived from the synovial tissue of osteoarthritis patients, the induction of p16INK4a was characteristic of RASFs. Those authors speculated that continuous stimulation by high synovial concentrations of proinflammatory cytokines, such as TNFα or IL-1β, in RA joints might suppress p16INK4a expression in synovial fibroblasts in vivo.

The results of our chromatin immunoprecipitation assay demonstrated that acetylation of promoter regions of p16INK4a was not enhanced by treatment with the combination of TNFα and IL-1β, but was enhanced by treatment with FK228. It is known that the promoter regions of the p16INK4a and p15INK4b genes located at chromosome band 9p21 are frequently silenced by CpG island methylation in hematologic malignancies and solid tumors (45). Densely methylated DNA associates with transcriptionally repressive chromatin, which is characterized by the presence of undeacetylated histones (46, 47). Histone deacetylation and DNA methylation can act as synergistic layers to inhibit the transcription of tumor-suppressor genes in tumor cells (48). Our results suggest that HDA inhibition by FK228 promoted the reexpression of the p16INK4a gene, which could be silenced in RASFs. Whether the severity of synovial proliferation, the type of disease progression, or the response to drugs could affect the methylation status of p16INK4a in RA synovial cells merits further investigation.

The acetylation of p21WAF1/Cip1 promoter regions was in direct contrast to the acetylation of p16INK4a. Interestingly, stimulation with both TNFα and IL-1β, which promotes the fibroblast proliferation, induced the acetylation of promoter regions of p21WAF1/Cip1, which is known to halt the progression of the cell cycle. It seems very likely that the level and location of p21 protein within individual cells is related to the function of this gene. First, Taniguchi et al (12) reported their finding that p21WAF1/Cip1 and p16INK4a were inducible in RASFs in response to growth inhibition induced by irradiation, high-density culturing, or serum starvation. Thus, the in vitro culture conditions might affect the level of expression of p21WAF1/Cip1 in RASFs.

Second, the expression of the p21WAF1/Cip1 gene is induced by activation of wild-type p53 or during cellular senescence. Tak et al (49) reported the overexpression of immunoactive p53 protein in synovial fibroblasts and synovial tissues from patients with early and longstanding RA, which is probably secondary to an increased production of wild-type p53 protein in response to DNA damage as well as to somatic mutations caused by the genotoxic local environment of the inflamed synovial tissue. Thus, p21WAF1/Cip1 protein might be induced by the increased p53 protein in response to treatment with both TNFα and IL-1β, which is known to stimulate cell proliferation under in vitro conditions.

Third, p21WAF1/Cip1 functions as a cell cycle inhibitor, and this activity is closely associated with its nuclear localization in various tissues, such as fibroblasts and epithelial cells. Recently, Asada et al (50) demonstrated by Western blot analysis the expression of cytoplasmic p21WAF1/Cip1 in normal human monocytes, suggesting that p21WAF1/Cip1 is expressed at a low level in normal cells. Those investigators also demonstrated that the function of p21WAF1/Cip1 as an inhibitor of cell cycle progression or of apoptosis is determined by its subcellular localization. Based on these findings and the findings of the present study, it is reasonable to suggest that the induction of p21WAF1/Cip1 in RASFs by the combination of TNFα and IL-1β might be related to the important role of p21WAF1/Cip1 in the protection of cells against cell death/apoptosis promoted by cytotoxic stimulation, such as high concentrations of inflammatory cytokines.

Although intravenous administration of 2.5 mg/kg of FK228 showed a dramatic amelioration of the clinical symptoms of arthritis, possible serious side effects that might be induced by the modulation of the histone acetylation status by HDA inhibitors could not be identified. When FK228 was injected at a dose of 1 mg/kg, no significant inhibitory effect on arthritis was seen (data not shown). No death or local reaction was noted in the group given 1 mg/kg, but 1 mouse given 2.5 mg/kg of FK228 died of unknown causes on day 5. A skin ulcer developed at the injection site on the tail in 1 of the remaining 9 FK228-treated mice; this was probably due to subcutaneous leakage of the agent.

The initial development of FK228 as antitumor therapy was stopped because of significant cardiac toxicity in dogs, specifically, myocardial hemorrhage and ischemia (51). However, subsequent studies conducted at the National Cancer Institute, National Institutes of Health, demonstrated that treatment with FK228 without resultant cardiotoxicity was possible by varying the schedule of administration. FK228 showed therapeutic efficacy in phase I trials of patients with malignant lymphoma (48) and refractory neoplasms (52). TSA treatment reportedly modulated 1–2% of genes in human lymphoid cell lines (53). In a small preliminary complementary DNA (cDNA) microarray experiment, we found that multiple genes were up-regulated or down-regulated by FK228 treatment (50 nM at 24 hours) on cytokine-stimulated RASFs (data not shown). These findings would suggest that in arthritis, there may be multiple therapeutic targets among HDA inhibitor–responsive genes.

The results of our cDNA microarray analysis also revealed a down-regulation of proinflammatory cytokines IL-6, IL-8, IL-11, and type 1 IL-1 receptor. FK228 treatment down-regulated the expression of matrix-degrading proteases MMPs 1 and 14, and up-regulated the expression of tissue inhibitors of metalloproteinases 1 and 3. In the current study, FK228 exerted inhibitory effects not only on cell proliferation, but also on the in vivo expression of TNFα and IL-1β in synovial tissues. The inhibitory effect on cytokine expression might be partly related to the up-regulation of p21WAF1/Cip1, the overexpression of which was recently reported to reduce the amplification of inflammatory pathways by cytokines in RASFs (40, 54). Further studies are required to identify precisely all of the genes that are up-regulated and down-regulated in RASFs by HDA inhibitors. In vivo treatment protocol is sufficiently attractive in terms of the regulation of specific genes involved in the pathogenesis of chronic arthritis, as well as the specific histologic effects on joint destruction to warrant further investigation of HDA inhibitors in the treatment of RA.


The authors thank Mrs. C. McCown for editing the manuscript. We also thank Dr. Toshihiro Ito, Department of Orthopaedic Surgery, Okayama University Graduate School of Medicine and Dentistry, for useful discussions.