Published online before print October 3, 2007
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,1
,1
,2
* Departments of Woman and Child Health, Pediatric Rheumatology Research Unit, and
Medicine, Rheumatology Unit, Karolinska Institutet/Karolinska University Hospital, Stockholm, Sweden;
Division of Rheumatology and Immunology, Department of Medicine, Duke University, Durham, North Carolina, USA; and
Surgery and Bioengineering, DAMP Laboratory, University of Pittsburgh, Pittsburgh, Pennsylvania, USA
2 Correspondence: Department of Woman and Child Health, Astrid Lindgren Children’s Hospital, Q1:02, S-17176 Stockholm, Sweden. E-mail: ulf.andersson{at}ki.se
ABSTRACT
Gold compounds such as gold sodium thiomalate (GST) can reduce the symptoms of rheumatoid arthritis (RA), although their mechanism of action is not well defined. As the proinflammatory mediator high mobility group box chromosomal protein 1 (HMGB1) may play a role in the pathogenesis of RA, we have performed in vitro studies to investigate whether GST inhibits HMGB1 release as the basis of its mode of action. Murine RAW 264.7 or human THP-1 macrophage cells were stimulated in culture with agents causing extracellular HMGB1 release, including LPS, IFN-
, polyinosinic:polycytidylic acid, IFN-β, or NO in the presence of GST, ranging from 0 µM to 250 µM. Secretion and intracellular location of HMGB1 were assessed by Western blotting, HMGB1-specific ELISPOT assay, and immunofluorescent staining. In parallel, TNF and IFN-β levels were analyzed by ELISPOT and/or ELISA. Supernatant NO production was analyzed by the Griess method. At pharmacologically relevant doses, GST inhibited the extracellular release of HMGB1 from activated macrophages and caused the nuclear retention of this protein; in contrast, no effects were observed on the secretion or production of TNF. Release of the key endogenous mediators of HMGB1 translocation, IFN-β and NO, was inhibited by GST. This inhibition required gold, as sodium thiomalate did not affect the responses measured. Furthermore, gold chloride also inhibited release of HMGB1. Together, these results suggest a new mechanism for the anti-rheumatic effects of gold salts in RA and the potential of drugs, which interfere with intracellular HMGB1 transport mechanisms, as novel agents to treat RA.
Key Words: gold salts • arthritis • inflammation • therapy • cytokines • immunomodulation • TNF • IFN-β • nitric oxide
INTRODUCTION
Gold salts occupy an important place in the history of rheumatology and represent the first widely used disease-modifying anti-rheumatic drugs (DMARDs) to treat rheumatoid arthritis (RA). These agents, which include gold sodium thiomalate (GST), aurothioglucose, and auranofin, consist of a gold atom bound to a sulfur-containing molecule. As shown in large clinical trials, gold salts can reduce the signs and symptoms of RA and slow radiographic progression; in some instances, gold salts may induce sustained remission [1 ]. Toxicity limits gold therapy, however, and its use has waned with the introduction of methotrexate and the biological agents. The use of gold salts, however, established an important paradigm for RA therapy and indicated that DMARDs can interrupt synovitis and erosion.
Although the clinical efficacy of gold salts is well established, their mechanism of action in RA still remains poorly understood. As shown in in vitro models, GST, also known as Myochrysine® or Myocrisin®, the most commonly used gold-containing compound, displays diverse anti-inflammatory and immunosuppressive effects on macrophages and monocytes. These effects include suppression of IL-1β and production, increased production of IL-10, modulation of monocyte differentiation, inhibition of PGE2, and NO production [1
]. The effects of GST on TNF-
and IL-6 are less certain [2
3
4
5
]. In the pharmacologic effects of gold, inhibition of protein kinase C (PKC) [6
, 7
], NF-
B activation [8
], and induction of heme oxygenase [9
] have been considered as possible targets of action.
Although the documented effects of gold salts could explain their action as DMARDs, recent studies about the pathogenesis of RA suggest new possibilities by which gold salts and other immunomodulatory agents could be therapeutic. These possibilities focus on the role in inflammation of extracellular high mobility group box chromosomal protein 1 (HMGB1). As shown in studies in vivo and in vitro, the nonhistone nuclear protein HMGB1 is a prototype dual-function molecule, which has intracellular and extracellular activity (reviewed in ref. [10 ]). Inside the cell, HMGB1 is a structural DNA-binding nuclear protein regulating transcription. The location of HMGB1 is not fixed, however, and HMGB1 can translocate outside the cell to promote inflammation [10 , 11 ]. In its extracellular action as an alarmin, HMGB1 can elicit a broad range of responses, which resemble those of the Toll ligands as well as cytokines such as TNF and IFNs [12 , 13 ]
To reach the extracellular milieu, HMGB1 undergoes active secretion from myeloid cells or passive release from dead and dying cells [10
, 14
15
16
]. Active secretion involves the acetylation [17
] and phosphorylation [18
] of HMGB1 to alter its charge and binding to chromatin. With these modifications, binding to chromatin diminishes, and HMGB1 can translocate from the nucleus to the cytoplasm to enter endolysosomes for secretion [16
, 17
]. HMGB1 release by activated cells can be stimulated by ligands of the TLRs by cytokines (including TNF- and Types I and II IFNs) and by NO [11
, 19
]. Once outside the cell, HMGB1 can interact with receptor for advanced glycation endproducts and possibly other receptors to trigger a wide range of downstream effects to amplify inflammation [20
].
As shown in our studies and those of others, HMGB1 may play a key role in the pathogenesis of arthritis in humans as well as in animal models [21 22 23 ]. HMGB1 is expressed abundantly in synovial tissue and synovial fluid of patients with RA [21 , 22 ]. Furthermore, in animal models, intra-articular injections of recombinant (r)HMGB1 can cause prolonged, destructive arthritis [24 ], and experimentally induced arthritis can be abrogated by HMGB1 antagonists [25 , 26 ]. In view of these observations, we have questioned whether drugs effective in RA reduce arthritis by inhibiting the release of HMGB1, thereby curtailing synovial inflammation.
In the current studies, we have therefore explored whether gold salts, a well-established class of DMARDs, can affect HMGB1 release as a mechanism of action. We have focused on GST and tested its effects on the translocation of HMGB1 from activated macrophages, using as models the murine macrophage cell line RAW 264.7 or the human macrophage cell line THP-1. In findings present herein, we demonstrate that GST can inhibit the extracellular release of HMGB1 by activated macrophages and cause its nuclear retention. Furthermore, we show that although GST can reduce the production of NO and IFN-β, which can mediate HMGB1 secretion [19 ], exogenous NO or IFN-β do not restore HMGB1 release in cells treated with GST. Together, these findings indicate that GST acts at more than one step to block HMGB1 release and suggest a novel mechanism for its DMARD action.
MATERIALS AND METHODS
Cells
These studies were performed concurrently in two laboratories. Although the overall design of the experiments was the same, there were some differences in the systems and reagents used. The results of the experiments were fully confirmatory, however, and they are presented together. The differences in methods or sources of relevant reagents are indicated.
As a model for macrophages, the RAW 264.7 mouse cell line and the human THP-1 cell line were used. The RAW 264.7 cells were cultured in DMEM supplemented with 5% heat-inactivated FCS (PAA Laboratories, Linz, Austria) and 50 µM 2-ME, 10 mM HEPES, 100 U/ml penicillin, and 100 µg/ml streptomycin (Life Technologies, Paisely, Scotland) and were split 2 days prior to each experiment. In some experiments, RAW 264.7 cells were cultured in RPMI 1640 supplemented with 10% FBS (Hyclone, Logan, UT, USA) and 200 µg/ml gentamicin (Invitrogen, Carlsbad, CA, USA). The THP-1 cells were cultured in RPMI 1640 supplemented with 5% FBS.
HMGB1 ELISPOT
The HMGB1 ELISPOT was performed as described in ref. [27
]. Briefly, multiscreen, 96-well high-throughput screening plates were pretreated with 70% ethanol, washed with sterile water, and coated with 20 µg/ml mouse mAb 2G7 (Critical Therapeutics Inc., Boston, MA, USA; available via American Type Culture Collection, Manassas, VA, USA) at 4°C overnight in a moist chamber. The plates were blocked with cell medium for 2 h. The RAW 264.7 cells (1000–3000/well) were stimulated with 10 µg/ml LPS (L-6529, Sigma Chemical Co., St Louis, MO, USA) and 100 U rIFN- (Sigma Chemical Co.) for 23 h in the presence of Myocrisin (Aventis Pharma, Sanofi-Aventis, Paris, France), 0, 10, 50, or 250 µM at 37°C with 5% CO2. Cell viability was assessed after 4, 8, 24, and 72 h at a Myocrisin concentration of 250 µM using trypan blue (Merck, Darmstadt, Germany) exclusion analysis and was determined to be 95–100% at all time-points. Signs of cytotoxicity were examined further using the TACS Annexin V-biotin apoptosis detection kit (R&D Systems, Minneapolis, MN, USA) or by assessment of extracellular lactate dehydrogenase (LDH) release with a commercial kit (CytoTox96® nonradioactive cytotoxicity assay, Promega, Madison, WI, USA) or caspase 3 activity measurement (EnzCheck® caspase 3 assay kit, Invitrogen). The human THP-1 cells were stimulated by 100 ng/ml LPS and 10 ng/ml human rIFN- (Biosite Inc., San Diego, CA, USA). After washing the plates, polyclonal antigen affinity-purified rabbit anti-HMGB1 antibodies (#556528, BD Biosciences PharMingen, San Diego, CA, USA), 0.5 µg/ml, were added, and the plates were incubated overnight at 4°C in a moist chamber. The plates were washed again and incubated with biotinylated donkey anti-rabbit antibody (Jackson ImmunoResearch Lab, West Grove, PA, USA), 0.8 µg/ml, at room temperature for 3 h, washed again, and incubated for 2 h at room temperature with Streptavidin-alkaline phosphatase (Mabtech AB, Stockholm, Sweden), diluted 1/1000. Wells were developed for 8–15 min with 5-bromo-4-chloro-3-indolyl phosphate p-toluidine salt/nitroblue tetrazolium chloride chromogen liquide substrate (Sigma-Aldrich Chemie, Schnelldorf, Germany). The reaction was stopped by water, and the plates were dried in room temperature overnight. Analyzing was performed using an AID EliSpot reader system (Germany). High-intensity spots [27
] were analyzed.
TNF- ELISPOT and ELISA
The TNF- ELISPOT was performed according to the manufacturer’s protocol (R&D Systems). Briefly, after pretreatment, 96-well plates were coated overnight at 4°C with the capture antibody diluted 1/60. RAW 264.7 cells (750–1500/well) were added and stimulated for 7 h with 10 ug/ml LPS and 100 U/ml IFN- in the presence of Myocrisin in concentrations, as for the HMGB1 ELISPOT. After washing, the biotinylated anti-TNF- detection antibody, diluted 1/60, was added and incubated at 4°C overnight. The subsequent steps of the procedure were identical with the HMGB1 ELISPOT protocol.
To determine the TNF- level in RAW 264.7 cells, supernatants were collected 20 h after stimulation with LPS or polyinosinic-polycytidylic acid [poly(I:C)] in the presence of 0–50 µM GST. TNF- concentrations in culture media were determined by ELISA. Capture anti-TNF- antibody (BD Biosciences PharMingen) was coated overnight at 4°C on Immunlon® 96-well plates. After three washes, samples were added in duplicate and incubated at room temperature for 2 h. Murine rTNF- (R&D Systems) was included as standard. Biotinylated anti-TNF- antibody (BD Biosciences PharMingen) was added and incubated for an additional 2 h at room temperature. Avidin-conjugated HRP was added with 30 min incubation at room temperature followed by 0.015% 3,3',5,5'-tetramethylbenzidine, 0.01% H2O2 in 0.1 M citrate buffer, pH 4.0, substrate until color development. Three PBS washes were performed between steps, and plates were read at OD 650 using an automated microtiter plate reader.
Immunostaining
Intracellular HMGB1 expression was studied using an immunofluorescent staining technique. RAW 264.7 cells were cultured in chamber slides (BD Biosciences, Discovery Labware, Bedford, MA, USA) in a concentration of 400,000 cells/ml in the presence of 250 µM Myocrisin. Cells were stimulated with 10 ug/ml LPS and 100 U/ml IFN- for a total time of 4, 8, or 24 h. Slides were washed in PBS, fixed in 4% formaldehyde for 12 min, and stored in PBS at 4°C until staining. Cells were permeabilized with 20 mM HEPES, pH 7.4, 300 mM sucrose, 50 mM NaCl, 3 mM MgCl2, and 0.5% (vol/vol) Triton X-100 for 3 min, followed by rinsing in PBS containing 0.1% saponin. The slides were blocked with 5% normal horse serum followed by avidin and biotin-blocking (Vector Laboratories, Burlingame, CA, USA) for 15 min, respectively.
Staining was performed with 2 µg/ml mouse anti-HMGB1 mAb 2G7 (Critical Therapeutics Inc., Boston, MA, USA) or with a HMGB1-specific polyclonal peptide affinity-purified rabbit antibody (#556528, BD Biosciences PharMingen; 0.5ug/ml) for 1 h, followed by incubation with biotinylated horse anti-mouse IgG (H+L) (Vector Laboratories; 2 µg/ml) or with biotinylated goat anti-rabbit IgG (BA-1000, Vector Laboratories), diluted at a concentration of 1:800 for 30 min, followed by a third step with Avidin-Oregon Green 1:500 (Molecular Probes, Eugene, OR, USA) for 30 min and nuclear staining with 4',6-diamidino-2-phenylindole (DAPI; Sigma-Aldrich Chemie), diluted 1:2000 in PBS. All washes were performed in 0.1% saponin (Riedel de Haen AG, Seelze, Germany) in PBS and all incubations in room temperature. Coverslips were mounted with PBS/glycerol, and microscopic analysis was conducted using a Polyvar 2 UV microscope (Reichert-Jung, Vienna, Austria).
Immunostaining for TNF- was performed according to a slightly different protocol. Briefly, permeabilization of cell membranes was performed by the use of Earle’s balanced salt solution (Gibco Ltd., Paisley, UK), supplemented with saponin, 0.1%, in all washes and incubations during the procedure. The slides were blocked with 5% FCS followed by avidin and biotin-blocking. Primary antibody used was a polyclonal peptide antigen affinity-purified rabbit anti-rat/mouse TNF- (Lot #8-14, Dr. Peter van der Meide, Biomedical Primate Research Center, Rijswijk, The Netherlands) at a concentration of 5 µg/ml for 30 min in room temperature. Secondary antibody used was a biotinylated donkey anti-rabbit IgG (Jackson ImmunoResearch Lab; diluted 1:1000) for 30 min. The rest of the procedure was identical as for HMGB1.
HMGB1 Western blotting
The HMGB1 Western blotting was performed as described previously [19
]. Briefly, RAW 264.7 cells were plated in six-well culture plates in Opti-MEM (Invitrogen) for 3–4 h and then stimulated with 0.5 µg/ml LPS (Escherichia coli 0111:B4, Sigma Chemical Co.) or 0.25 µg/ml poly(I:C) (Invivogen, San Diego, CA, USA) in the presence of sodium aurothiomalate hydrate (Sigma Chemical Co.) or mercaptosuccinic acid (thiomalate, Sigma Chemical Co.), 0–50 µM, or gold chloride (AuCl3; Sigma Chemical Co.), 0–50 µM, for 20 h at 37°C. Culture supernatants were collected and assayed for nitrite levels by the Griess method [22
], IFN-β levels by an IFN-β ELISA kit (R&D Systems), and HMGB1 levels by Western blotting. For Western blotting of HMGB1, supernatants were concentrated by Centricon YM-10 (Millipore, Billerica, MA, USA). The volume of the concentrated supernatants was adjusted to 70 µl for equal loading, and samples were resolved on 4–12% NuPAGE® Tris-Bis SDS polyacrylamide gel (Invitrogen). Protein was transferred to polyvinylidene difluoride membranes (Invitrogen), blocked with 5% dry milk in TBS-Tween, and blotted with a mouse anti-HMGB1 mAb (R&D Systems). The membrane was then incubated with HRP-conjugated anti-rabbit IgG, followed by Super Signal® West Femto substrate (Pierce, Rockford, IL, USA). Images were captured by exposing the membrane to a charged-coupled device camera (FluorChem8900, Alpha Innotech, San Leandro, CA, USA). Each experiment was repeated at least twice, and one representative result was shown.
Statistical analysis
Cytokine ELISA and nitrite results were analyzed by two-tailed, unequal variance Student’s t-test. A P value less than 0.05 was considered significant.
RESULTS
The effects of GST on HMGB1 release after LPS and IFN- activation
In these experiments, we have used murine RAW 264.7 and human THP-1 cell lines as models to assess the effects of GST on macrophages, activating cells by LPS or the combination of LPS and IFN-, which can stimulate HMGB1 secretion [27
]. To measure HMGB1 release, we have used an ELISPOT assay, recently developed in our laboratory. As these experiments indicate (Fig. 1A
), GST (10–250 µM) caused a dose-dependent inhibition of extracellular HMGB1 translocation from unstimulated and activated RAW 264.7 cells (Fig. 1A)
and from activated THP-1 cells (Fig. 1D)
; similar results were obtained in four independent experiments with each cell line, using LPS or LPS plus IFN-. The inhibition of HMGB1 secretion by GST was not a result of cell toxicity, as cell viability was unaffected by GST at concentrations up to 250 µM (95–100% viability in all cultures) when assessed by trypan blue; visual inspection of cultures, furthermore, failed to show evidence of apoptosis. Lack of GST-mediated cytotoxicity was confirmed further using Annexin V immunofluorescent cellular staining and by assessments of extracellular release of LDH and caspase 3 in the culture supernatants (data not shown). The effects of GST on HMGB1 release by activated RAW 264.7 cells were confirmed by Western blotting (Fig. 1C)
.
 View larger version (20K): [in a new window] |
Figure 1. Effects of GST on HMGB1 or TNF release from RAW 264.7 cells (A–C) or from THP-1 cells (D) activated by LPS (10 µg/ml with 264.7 RAW cells and 100 ng/ml with THP-1 cells) and rIFN- (100 U/ml with 264.7 RAW cells and 10 ng/ml with THP-1 cells) or without exogenous stimulus. Secretion of HMGB1 or TNF was detected by ELISPOT assays. One representative experiment out of four performed is shown (A). Results are expressed as means ± SD of three different well counts. For statistical evaluation, raw data each of the four experiments were normalized by denoting the number of HMGB1 spots from LPS + IFN--stimulated cells as 100% and subsequently calculating the suppression achieved by the addition of GST. The normalized values were pooled in a box-plot analysis (B), and P values were calculated by the Kruskal-Wallis rank sum test. **, P < 0.01, when compared with stimulated cells without GST addition. The release of HMGB1 in culture supernatants was also measured by Western blotting (C). (D) HMGB1 ELISPOT-based results from one of three representative experiments performed with human THP-1 cells.
|
release after LPS and IFN- activation secretion in unstimulated or LPS + IFN--activated RAW 264.7 cells treated in a similar way as that used for HMGB1 assessments. As these data indicate, GST produced divergent effects on TNF- secretion as compared with HMGB1 secretion. Thus, GST, in concentrations ranging from 10 µM to 250 µM, did not affect spontaneous TNF- secretion or LPS + IFN--induced TNF- secretion from RAW 264.7 cells (Fig. 1A)
. Cell viability was more than 95%, as determined by the trypan blue exclusion assay at all studied GST concentrations.
The effects of GST on intracellular HMGB1 translocation after LPS and IFN- activation
Immunohistochemical staining of intracellular HMGB1 in RAW 264.7 cells cultured with or without GST was next performed to investigate the mechanisms by which GST suppresses HMGB1 secretion. Unstimulated RAW 264.7 cells cocultured for 24 h with 10–250 µM GST or without GST differed in the pattern and intensity of HMGB1 nuclear staining with a stronger nuclear HMGB1 signal in cells cultured with GST in doses ranging from 10 µM to 250 µM (data not shown). Important differences were observed in HMGB1 staining patterns in LPS + IFN--activated RAW 264.7 cells treated with 10–250 µM GST compared with cells, which had not been exposed to GST (Fig. 2
). Thus, the staining pattern of activated cells not treated with GST demonstrated low or absent nuclear HMGB1 signals and strong cytoplasmic HMGB1 accumulation (Fig. 2A)
. Stimulated cells cultured with GST, however, showed the reverse image, with an intense nuclear and faint cytoplasmic HMGB1 signal (Fig. 2C)
.
 View larger version (32K): [in a new window] |
Figure 2. Effects of GST on intracellular HMGB1 localization in RAW 264.7 cells activated by LPS and IFN-. (A and B) RAW cells stimulated for 24 h with LPS (10 ug/ml) and IFN- (100 U/ml) without GST addition and then stained by immunofluorescence for HMGB1 (green) or nucleus (blue DAPI). HMGB1 was mainly observed cytoplasmically (A). (C and D) A much-altered HMGB1 staining pattern after cell activation in the presence of GST when most of the HMGB1 was retained in the nuclei.
|
accumulation after LPS and IFN- activation was examined. UV microscopy demonstrated that TNF- production in unstimulated RAW 264.7 cells was low in cultures, with and without GST (data not shown). The majority of cells stimulated with LPS and IFN- showed strong intracellular TNF- production, and no differences were observed in the numbers of TNF--expressing cells in cultures up to 250 µM GST versus no GST addition (data not shown). The TNF--producing cells could be recognized as a result of the characteristic morphology caused by the accumulation of TNF in the Golgi apparatus [28
]. This appearance differed from that caused by HMGB1, which does not traverse the endoplasmic reticulum-Golgi system.
The effects of GST on HMGB1, NO, or IFN-β release after TLR stimulation
Having shown the effects of GST on HMGB1 release after LPS and LPS + IFN- stimulation, we then tested another TLR ligand and examined the effects of stimulation with poly(I:C), a ligand of TLR3. In these experiments, as a comparison for poly(I:C), LPS, a ligand of TLR4, was used to stimulate cells alone. Western blotting was used to measure the release of HMGB1, as other experiments indicated that results of HMGB1 determination by ELISPOT and Western blotting were comparable. As these three experiments presented in Figure 3
indicate, LPS and poly(I:C) induced extracellular HMGB1 release, which was detected by Western blotting in cultures with activated cells (Fig. 3A)
. In contrast, HMGB1 levels were reduced in cocultures with 50 µM GST.
 View larger version (20K): [in a new window] |
Figure 3. Effects of GST on HMGB1, NO, IFN-β, and TNF- from RAW 264.7 cells activated by LPS, poly(I:C) (pIC), or dsRNA. The release of HMGB1 was measured by Western blotting and quantified by densitometry. Supernatants were collected and assayed for IFN-β and TNF- (ELISA) and NO (Nitrite). As these data from three separate experiments indicate, GST reduced HMGB1 release induced by LPS as well as poly(I:C) (A). Production of NO and IFN-β induced by LPS or poly(I:C) was inhibited by GST, whereas TNF- production was not affected (B). *, P < 0.05, when compared with LPS or poly(I:C) treatment alone.
|
release, we were interested in the possibility that it could affect other mediators contributing to HMGB1 release. In a previous study, we demonstrated that stimulation of HMGB1 release by LPS depends on NO, and release induced by poly(I:C) depends on Type I IFNs [19
]. In addition, IFN-β may mediate LPS- and poly(I:C)-induced HMGB1 release (unpublished data). We therefore measured the effects of GST on the production of these mediators in the cultures. As these studies indicate, LPS as well as poly(I:C) stimulated the RAW 264.7 cells to produce NO and IFN-β, as assessed by measurement of nitrite and IFN-β in culture supernatants. These responses were strongly diminished in cultures treated with 50 µM GST (Fig. 3B)
. These findings demonstrate the inhibitory activities of GST on the production of two mediators of HMGB1 release. Confirming previous ELISPOT results on TNF- release, the addition of GST did not inhibit TNF- production induced by LPS or poly(I:C) (Fig. 3B)
.
The effects of GST on HMGB1 release after NO or IFN-β treatment
To evaluate further the influence of GST on NO- or IFN-β-dependent processes to explain GST inhibition of HMGB1 release, we investigated whether exogenous NO or IFN-β can reverse the blocking effects of GST. For this purpose, RAW 264.7 cells were treated with the NO donor compound NOC-15 or with rIFN-β, both of which can induce HMGB1 release (ref. [19
] and unpublished findings). As these studies indicate, GST can block the response of RAW264.7 cells to these mediators, and stimulated cells are unable to release HMGB1 in the presence of 50 µM GST (Fig. 4
). These findings indicate that GST affects HMGB1 release at more than one step.
 View larger version (25K): [in a new window] |
Figure 4. Effects of GST on NO- or IFN-β-dependent HMGB1 release. RAW 264.7 cells were cultured with the NO donor compound NOC-15 or with rIFN-β, both of which induce HMGB1 release, as assessed by Western blotting. As these data indicate, stimulated cells release less HMGB1 in cocultures with 50 µM GST.
|
 View larger version (25K): [in a new window] |
Figure 5. The effect of thiomalate (TM) in three separate experiments on HMGB1 release induced by LPS or poly(I:C) (A). RAW 264.7 cells were pretreated with 50 µM GST or 50 µM thiomalate for 30 min and then stimulated with LPS or poly(I:C) for 20 h. Supernatant HMGB1 levels were measured by Western blotting (A). Supernatant nitrite levels were also measured (A). The effects of AuCl3 on HMGB1 and nitrite release from RAW 264.7 cells after poly(I:C) or LPS stimulation for 20 h were also determined in three separate experiments (B). *, P < 0.05, when compared with LPS or poly(I:C) treatment alone.
|
DISCUSSION
These studies provide new insights into the mechanism of action of gold salts and provide the first demonstration that a DMARD can decrease the release of the proinflammatory mediator HMGB1 from activated macrophages. Thus, using three independent assays (Western blotting, a HMGB1-specific ELISPOT assay, and intracellular immunostaining), we have shown that GST can inhibit HMGB1 release from activated macrophages. In contrast to these findings, parallel studies demonstrated that GST does not block TNF- release in agreement with a previous report [1
]. As TNF- production was not blocked by GST, these findings argue against GST-mediated cell toxicity or death as the basis for decreased HMGB1 secretion. This conclusion is supported by results using specific methods for assessing cell death. Together, these findings point to an effect of GST on the intracellular trafficking of HMGB1 as the mechanism by which this agent affects the nuclear retention of this proinflammatory mediator.
As shown by a variety of cell biology techniques, the extracellular transport of the HMGB1 occurs by a nonconventional pathway, which differs from that of most other secreted proinflammatory proteins [16 , 17 , 29 ]. Inside the cell, HMGB1 can shuttle between the nucleus and cytoplasm. With activation, however, acetylation and phosphorylation alter the charge of HMGB1 and its interaction with chromatin. These post-translational modifications cause HMGB1 relocation to the cytosol [17 , 18 ]. A specific ATP-binding cassette protein transporter, multidrug resistance-related protein 1 (MRP1), then translocates HMGB1 into secretory lysosomes for extracellular exocytosis. The transport by MRP1 requires covalent linkage of HMGB1 to glutathione [12 ].
Although HMGB1 release was defined originally with LPS, other TLR ligands as well as endogenous mediators such as cytokines and NO can all induce HMGB1 translocation and release. In previous studies, we demonstrated that NO and Type I IFNs are key down-stream mediators for HMGB1 release [19
]. In these responses, LPS, poly(I:C), and TNF- and IFN- are all potent inducers of NO production, and TLR3 and TLR4 ligands are potent inducers of IFN-β. It is thus of interest that GST can inhibit the release the key mediators, NO and IFN-β, following stimulation by different TLRs. In the present study, we observed that AuCl3, like GST, down-regulated NO release, and the effects on IFN-β were not studied. Previous studies have shown that another gold compound, auranofin, can also down-regulate NO synthesis in activated rat macrophages [1
]. Taken together, these observations indicate that this effect may be a general characteristic of therapeutic gold compounds [1
].
In the present study, an interesting observation concerns the effect of GST on HMGB1 release induced by NO or IFN-β. Thus, neither NO nor IFN-β was able to induce HMGB1 release from RAW 264.7 cells treated with GST. These results thus indicate that in addition to blocking the production of the pivotal mediators NO and IFN-β, GST can counteract a biological response to these mediators. As such, GST can interdict HMGB1 release at more than one step. Our studies also indicate that the inhibitory effects of GST relate to the gold component of this compound rather than the thiomalate moiety. This conclusion is supported by the fact that AuCl3 also prevented HMGB1 release from activated 264.7 RAW macrophages. Studies investigating other effects of gold support the role of the heavy metal in the anti-inflammatory activities, perhaps by interacting with thiol-containing molecules involved in processes such as activation of NF-B [8
].
The present study demonstrates that GST inhibits cytoplasmic and extracellular HMGB1 translocation in cultured myeloid cells at pharmacologically relevant concentrations. Assuming that similar effects occur in vivo, this mechanism could explain the important anti-rheumatic effects of gold therapy in RA, as decreased levels of HMGB1 should ameliorate synovial inflammation and tissue destruction [25 , 26 ]. Of note, we have demonstrated that maximal HMGB1 expression in untreated collagen type 2-induced synovitis occurs in the pannus tissue [30 ]. Together, these findings suggest that gold, although exerting systemic, anti-inflammatory effects, may act locally. As gold affects a number of intracellular signaling systems (e.g., PKC and heme oxygenase), studies are in progress to identify the key system responsible for the effects on HMGB1 release.
ACKNOWLEDGEMENTS
Financial support was provided through the regional agreement on medical training and clinical research (ALF) between the Stockholm County Council and the Karolinska Institutet, King Gustaf V 80-year-foundation, the Freemason Lodge Barnhuset in Stockholm, Foundation for Technical Support to Disabled, The Swedish Research Council, nos. K2005-74X-09082 and K2005-73X-14642, the Swedish Rheumatism Association, The Lupus Research Institute, and VA Medical Research Service. W. J. was supported by National Institutes of Health Training Grant AI007217.
FOOTNOTES
1 These authors contributed equally to this work. 
Received May 24, 2007; revised September 7, 2007; accepted September 8, 2007.
REFERENCES
levels J. Rheumatol. 29,21-28 B activation by blocking I B kinase J. Immunol. 164,5981-5989 and nitric oxide in the release of HMGB1 by RAW 264.7 cells stimulated with polyinosinic-polycytidylic acid or lipopolysaccharide J. Immunol. 177,3337-3343This article has been cited by other articles:
 |
|
 |
|
  I. Garcia-Arnandis, M. I. Guillen, M. A. Castejon, F. Gomar, and M. J. Alcaraz Haem oxygenase-1 down-regulates high mobility group box 1 and matrix metalloproteinases in osteoarthritic synoviocytes Rheumatology, January 27, 2010; (2010): kep463v1 - kep463. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Pan, J. Cardinal, R. Dhupar, M. R. Rosengart, M. T. Lotze, D. A. Geller, T. R. Billiar, and A. Tsung Low-dose cisplatin administration in murine cecal ligation and puncture prevents the systemic release of HMGB1 and attenuates lethality J. Leukoc. Biol., September 1, 2009; 86(3): 625 - 632. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Di Comite, C. M. Rossi, A. Marinosci, K. Lolmede, E. Baldissera, P. Aiello, R. B. Mueller, M. Herrmann, R. E. Voll, P. Rovere-Querini, et al. Circulating chromogranin A reveals extra-articular involvement in patients with rheumatoid arthritis and curbs TNF-{alpha}-elicited endothelial activation J. Leukoc. Biol., January 1, 2009; 85(1): 81 - 87. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
