Introduction
Rheumatoid arthritis (RA) is a chronic inflammatory disorder with a variable disease outcome and is characterized by inflammation of multiple joints. The prognosis of RA patients has improved significantly in recent years after the introduction of tumor necrosis factor-alpha (TNFα)-based therapy [
1]. Despite the wide use of these biologics, their precise mechanisms of action in RA remain unclear.
Several animal models of RA have been described; however, the therapeutic benefits of TNF antagonists have been confirmed in only a few of these models. Schubert and colleagues [
2] reported that continuous injections of human TNF receptor (TNFR) p75-IgG-Fc fusion protein (Etanercept) from days 0 to 9 completely protected against the development of arthritis in glucose-6-phosphate isomerase (GPI)-induced arthritis. In this regard, we recently demonstrated a clear therapeutic effect of anti-TNF monoclonal antibody (mAb) in mice with GPI-induced arthritis, and the therapeutic response correlated with the
in vitro regulation of TNF production [
3]. We also identified that anti-interleukin-6 (IL-6) receptor mAb blocks the development of GPI-induced arthritis [
3,
4]. These results indicate that the GPI-induced arthritis model is suitable for studying the mechanisms of action of TNFα antagonists as well as IL-6 antagonists in RA patients.
Using such a TNFα-dependent arthritis model, we investigated TNFα-related molecules by GeneChip analysis. The expression of TNFα-induced adipose-related protein (TIARP) was the highest in GeneChip study. TIARP was identified as a transmembrane protein that is highly regulated by TNFα in adipocytes [
5]. Not only TNFα but also IL-6 regulated the expression of TIARP [
6], suggesting the involvement of the inflammatory cascade in RA. To our knowledge, however, no information on its role in arthritis or its localization in joints has been published.
To explore the role of TIARP in arthritis, we conducted the present study in GPI-induced arthritis. TIARP mRNA and proteins were upregulated in joints and spleens in mice with GPI-induced arthritis. Administration of anti-TNFα mAb reduced TIARP mRNA in splenocytes. In arthritic mice, TIARP mRNA was expressed mainly in CD11b+ cells in the spleen, and TIARP mRNA level was increased in the joints (accompanied by joint swelling), especially in hyperplastic synovium. Overexpression of the human TIARP counterpart, such as six-transmembrane epithelial antigen of the prostate-4 (STEAP4), was noted in the synovia of patients with RA. The results provide the first characterization of the role of TIARP in inflammatory arthritis.
Materials and methods
Glucose-6-phosphate isomerase-induced arthritis
Male DBA/1 mice (6 to 8 weeks old) were obtained from Charles River Laboratories (Yokohama, Japan). Recombinant human GPI was prepared as described previously [
7]. Mice were immunized by intradermal injection of 300 μg of recombinant human GPI-GST (glutathione S-transferase) (hGPI) in emulsified complete Freund's adjuvant (CFA) (Difco Laboratories Inc., now part of Becton Dickinson and Company, Franklin Lakes, NJ, USA). Control mice were immunized with 100 μg of GST in CFA. Arthritic animals were assessed visually, and changes in each paw were scored on a scale of 0 to 3. A score of 0 indicates no evidence of inflammation, 1 indicates subtle inflammation or localized edema, 2 indicates swelling that is easily identified but localized to the dorsal or ventral surface of paws, and 3 indicates swelling on all aspects of paws, and the maximum possible score was 12 per mouse. The experimental protocol was approved by the Ethics Review Committee for Animal Experimentation of the University of Tsukuba (Japan).
GeneChip analysis of splenocytes from glucose-6-phosphate isomerase-induced arthritis
The spleens of three GPI-GST (molecular weight [MW] = 89 kDa) (300 μg)-immunized DBA/1 mice were harvested on day 10. As a control, the spleens of three GST (MW = 26 kDa) (100 μg)-immunized DBA/1 mice were used. Total RNA was extracted from the splenocytes using ISOGEN (Nippon Gene Co., Ltd., Toyama, Japan), and then 15 μg of RNA was used for cDNA synthesis by reverse transcription followed by synthesis of biotinylated cRNA through in vitro transcription. After cRNA fragmentation, hybridization with mouse 430A2.0 GeneChip (Affymetrix, Santa Clara, CA, USA) with probes for 43,000 mouse gene ESTs (expressed sequence tags) was performed in accordance with the protocol provided by the manufacturer. Analysis was performed by gene expression software.
Analysis of TIARP and tumor necrosis factor-alpha gene expression
Spleens and lymph nodes were isolated, cut into small pieces, and passed through cell strainers (BD Biosciences, Erembodegem, Belgium) to obtain single-cell suspensions. The remaining cells were washed twice with phosphate-buffered saline (PBS). Synovial tissues from the ankle joints were isolated and minced by scissors. Total RNA was extracted with ISOGEN in accordance with the instructions provided by the manufacturer. cDNA was obtained by reverse transcription with a commercially available kit (Fermentas, Glen Burnie, MD, USA). Primers sequenced were as follows: TIARP sense 5'-AGCCCACGTGGTCAAAGCAT-3' and antisense 5'-CCTTGGTCCAGTGGGGTGA-3' and glyceraldehydes-3-phosphate dehydrogenase (GAPDH) sense 5'-CGTCCCGTAGACAAAATGGT-3' and antisense 5'-GAATTTGCCGTGAGTGGAGT-3'.
All polymerase chain reactions (PCRs) were performed in a Takara PCR Thermal Cycler (Takara Bio Inc., Shiga, Japan). After denaturation at 95°C for 5 minutes, cycles were set at 10 seconds at 94°C, 10 seconds at 60°C, and 30 seconds at 72°C. Cycling was followed by 10 minutes of elongation at 72°C. PCR products were subjected to electrophoresis in 1% agarose gels in Tris-borate-EDTA (ethylenediaminetetraacetic acid) electrophoresis buffer, stained with ethidium bromide, and detected by ultraviolet transillumination. cDNA samples were normalized for the housekeeping gene GAPDH.
For real-time PCR, we used a TaqMan Assay-on-Demand gene expression product (Applied Biosystems, Foster City, CA, USA). The expression levels of TIARP, TNFα, and GAPDH (assay ID Mm00475402_m1, Mm00443258_m1, and Mm99999915_g1, respectively; Applied Biosystems) were normalized relative to the expression of GAPDH. Analysis was performed with an ABI Prism 7500 apparatus (Applied Biosystems) under the following conditions: inactivation of possible contaminating amplicons with AmpErase UNG for 2 minutes at 50°C, initial denaturation for 10 minutes at 95°C, followed by 45 thermal cycles of 15 seconds at 95°C and 60 seconds at 60°C. The serum TNFα level was measured by an enzyme-linked immunosorbent assay (ELISA) kit (eBioscience, Inc., San Diego, CA, USA). After conditioning, the detection limit of TNFα concentration was 2 μg/mL.
Preparation of anti-TIARP and anti-STEAP4 antibodies
One rabbit was immunized subcutaneously by TIARP peptide
5–19 (HADEFPLTTDSSEKQ, amino-terminal peptide coupled to keyhole limpet hemocyanin) or human ortholog STEAP4 peptide
3–15 (KTCIDALPLTMNS) [
8] with CFA four times, on days 0, 14, 28, and 42. The rabbit was sacrificed on day 52, and serum was collected. Serum was first purified by protein A column and then affinity-purified by TIARP-peptide
5–19 or STEAP4 peptide
3–15 column. The purified fraction was confirmed by TIARP peptide
5–19 or STEAP4 peptide
3–15 ELISA.
Western blotting
The cells were washed with PBS and incubated with lysis buffer (pH 7.4, 50 mM Tris-HCl, 5 mM MgCl2, 2 mM phenylmethylsulfonyl fluoride [PMSF], and 0.5% NP-40). Where indicated, protein concentrations were quantified using the bicinchoninic acid reagent (Pierce, Rockford, IL, USA). Samples (10 μg of total protein) were separated by SDS-PAGE (4/20% acrylamide; Daiichi Pure Chemicals Co., Ltd., Tokyo, Japan) and transferred to polyvinylidene fluoride membranes (Bio-Rad Laboratories, Inc., Hercules, CA, USA). All subsequent wash buffers contained 0.05% Tween-20 in PBS. Four percent Block Ace (Dainippon Pharmaceutical, Osaka, Japan) was used to block the membranes and to dilute antibodies. Rabbit polyclonal anti-TIARP antibodies and rabbit anti-actin antibodies (Sigma-Aldrich, Munich, Germany) were used at 1:3,000 dilution. Horseradish peroxidase (HRP)-conjugated anti-rabbit secondary antibodies (1:6,000 dilution; Bio-Rad Laboratories, Inc.) were used to visualize bound anti-TIARP antibodies or anti-actin antibodies with the ECL [enhanced chemiluminescence] Western blot detection kit (Amersham, now part of GE Healthcare, Little Chalfont, Buckinghamshire, UK).
Treatment with anti-tumor necrosis factor-alpha monoclonalantibody
We used commercially available anti-TNFα mAb (eBioscience, Inc.). For a control antibody, we used similar amounts of rat IgG1 isotype control (R&D Systems, Inc., Minneapolis, MN, USA). Just after the onset of arthritis (on day 8), a single dose of 100 μg of anti-TNFα mAb or control antibody was injected. Spleen was harvested at the indicated time points and analyzed for TIARP expression. Three independent experiments were performed.
Identification of TIARP-positive cells in splenocytes of mice with glucose-6-phosphate isomerase-induced arthritis
The spleens were harvested on day 12 after GPI immunization and single-splenocyte cell suspensions were prepared as described above. CD4+, CD19+, CD11b+, and CD11c+ cells from splenocytes were isolated by magnetic beads using the MACS™ [magnetic-activated cell sorting] system (Miltenyi Biotec, Bergisch Gladbach, Germany). The cells contained more than 97% CD4+, CD19+, CD11b+, and CD11c+ cells as confirmed by fluorescence-activated cell sorting analysis. The cells were dispensed at 1 × 106 cells to analyze the expression of TIARP mRNA.
Immunohistochemical staining for TIARP/STEAP4
At the indicated time points, the ankles of the mice were removed, fixed, decalcified, and paraffin-embedded. Sections (5-μm thick) were stained with hematoxylin and eosin and were evaluated for histological changes. For immunohistochemical study, endogenous peroxidase activity was inhibited using 3% hydrogen peroxidase in methanol. Sections were blocked by 5% bovine serum albumin in PBS for 10 minutes and then incubated with rabbit anti-TIARP antibody (1:100 dilution) or normal rabbit Ig (1:100 dilution; Dako, Tokyo, Japan). Isotype-matched HRP-conjugated anti-rabbit IgG antibody (Bio-Rad Laboratories, Inc.) was added for 30 minutes. HRP activity was detected using 3,3-diaminobendine (DAB) (Nichirei Corporation, Tokyo, Japan) as a substrate. The stained sections were counterstained with Mayer's hematoxylin for 10 seconds and mounted with aqueous mounting medium.
For human STEAP4 staining, synovial tissues were obtained after informed consent was given by RA patients at the time of joint replacement. All RA patients satisfied the classification criteria of the American College of Rheumatology (1987) [
9]. The synovium was embedded in optimal cutting temperature compound and frozen in dry ice isopentane, and 5-μm-thick sections were mounted at -25°C. Anti-human STEAP4 polyclonal antibody conjugated with fluorescein isothiocyanate (FITC protein labeling kit; Pierce) and purified anti-human CD68 (BD Pharmingen, San Diego, CA, USA) conjugated with rhodamine (1:100 dilution, Rhodamine protein labeling kit; Pierce) were used. Nuclei were counterstained with 4'-6'-diamidine-2-phenylindole dihydrochloride (DAPI) (Molecular Probes, Inc., now part of Invitrogen Corporation, Carlsbad, CA, USA). The stained sections were examined under a fluorescent microscope (model FW4000; Leica Microsystems, Tokyo, Japan).
Patients and analysis of human peripheral blood mononuclear cells and synovium for STEAP proteins
Peripheral blood mononuclear cells (PBMCs) from three female patients with RA and three healthy control subjects were obtained. All RA patients satisfied the classification criteria of the American College of Rheumatology (1987) [
9]. Synovial tissues from 36 RA and 19 osteoarthritis (OA) patients were obtained at the time of total knee replacement. Written informed consent was obtained from all subjects, and the study was approved by the ethics review committee. Total RNA was extracted with ISOGEN in accordance with the protocol provided by the manufacturer. cDNA was obtained by reverse transcription with a commercially available kit. The following primers were used: STEAP2 sense 5'-CCTACAGCCTCTGCTTACCG-3' and antisense 5'-GAGGGCAAAACAAGAGCAAG-3', STEAP3 sense 5'-GCCAGAAGAGATGGACAAGC-3' and antisense 5'-GGTGCTCTTGCTCTGTAGGG-3', STEAP4 sense 5'-GCTCTCCAGTCAGGAGCACT-3' and antisense 5'-CACACAGCACAGCAGACAAA-3', and GAPDH sense 5'-GAAGGTGAAGGTCGGAGTC-3' and antisense 5'-GAAGATGGTGATGGGATTTC-3'. For real-time PCR, we used a TaqMan Assay-on-Demand gene expression product (Applied Biosystems). The expression level of STEAP4 was normalized relative to the expression of GAPDH. Methods were described above.
Statistical analysis
All data were expressed as mean ± standard error of the mean. Differences between groups were examined for statistical significance using the Mann-Whitney U test. A P value of less than 0.05 denoted the presence of a statistically significant difference.
Discussion
Although the therapeutic effect of TNF antagonists is confirmed in RA [
1], only a few animal models of arthritis have been used to confirm the beneficial effects of TNF antagonists. For example, a recent study reported the therapeutic effect of anti-TNF mAb in DNaseII, type I interferon receptor (IFN-IR) double-knockout mice [
11], although this was not a genetically unaltered mouse. Furthermore, Schubert and colleagues reported the protective effect of TNF antagonist in GPI-induced arthritis [
2] and arthritis was clearly B cell-dependent [
12]. We recently demonstrated the therapeutic effect of TNF antagonist in GPI-induced mice. Thus, it is important to explore TNF-regulated genes in the latter model to understand the mechanisms of action of TNFα antagonists in RA patients. When the GeneChip analysis was used, the present results showed upregulation of TIARP mRNA in the spleen of arthritic mice. TIARP was first identified as TNFα-induced cell surface protein in adipose tissues and is also known to be localized in the liver, kidney, heart, and skeletal muscle [
5]. This protein was detected in the course of adipocyte differentiation and conversion and is also induced by IL-6 [
6]. In this study, we confirmed its induction in CD11b
+ splenocytes in arthritis and we confirmed that it is upregulated in the arthritic synovium of murine GPI-induced arthritis. These findings suggest the involvement of TIARP in the process of proliferation or differentiation state induced by inflammation. In fact, previous studies indicated that TIARP is induced by TNFα and IL-6 in adipocytes [
5,
6]. TNFα and IL-6 are pleiotropic cytokines known to play crucial roles in human RA, and significant therapeutic effects of their antagonists have been confirmed in recent years [
1,
13]. In GPI-induced arthritis, both TNFα and IL-6 antagonists have protective effects [
3,
4], and these cytokines play important roles in the induction of arthritis in collaboration with autoantibodies (anti-GPI antibodies) [
14]. However, there is no clear scenario of balance between IL-6 and TNFα in arthritis. In TIARP knockdown animals, exposure to TNFα induced a greater amount of IL-6, suggesting a crucial role of TIARP in the balance between TNFα and IL-6 [
15]. It is possible that TIARP expression plays a downregulatory role in the inflammatory cascade.
At this stage, there is no information on whether TIARP act in an antagonistic or agonistic manner with arthritis. However, one report on STAMP2 (a homolog of TIARP protein) [
15] confirmed (a) upregulation of inflammatory cytokines such as TNFα and IL-6 in STAMP2-deficient mice, (b) upregulation of macrophage-specific antigens such as CD68 and CD11b, (c) infiltration of CD68
+ cells in adipose tissues, and (d) STAMP2-induced suppression of IL-6 expression upon stimulation by TNFα. These findings suggest that STAMP2 (TIARP) suppresses inflammatory cytokines such as TNFα and IL-6 and also blocks the activation of macrophages/monocytes.
Is this scenario applicable to patients with RA? In humans, the STEAP protein family was identified in prostate tumors [
16,
17] and is also known to be involved in cell apoptosis [
18]. Among this family of genes, STEAP4 is highly expressed in the bone marrow, followed by placenta and fetal liver [
19]. The STEAP4 expression was induced by TNFα in human adipose tissue [
20] and also by TNFα in human synovial cells (our preliminary result). However, there is no report regarding the expression of this molecule in articular joints. The present study identified the expression of human ortholog STEAP4 in the synovium, especially in CD68
+ macrophages of patients with RA. In addition, our preliminary data using human synovial cell lines provide evidence that TNFα stimulation enhances the expression of STEAP4 protein and that a stably expressed form of STEAP4 is partially co-localized with endosomes (Tanaka and colleagues, manuscript in preparation). Further large-scale studies are required to assess the expression of STEAP4 in the joints and PBMCs of RA patients before and after treatment with TNF antagonists.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
AI helped to write the manuscript, conceive of the study, perform all experiments, and coordinate statistical study. IM wrote the manuscript and conceived of the study. YT helped to perform all experiments and coordinate statistical study. KI participated in the clinical assessment. AK and NO collected the synovial samples. DG and SI participated in discussion. TS participated in the full design and coordination of the study. All authors read and approved the final manuscript.