1. Top of page
  2. Abstract
  6. Acknowledgements


Extracellular high mobility group box chromosomal protein 1 (HMGB-1) is a recently identified, endogenous, potent tumor necrosis factor– and interleukin-1 (IL-1)–inducing protein detectable in inflamed synovia in both human and experimental disease. In the present study, we examined clinical effects in collagen-induced arthritis (CIA) using therapeutic administration of neutralizing HMGB-1 antibodies or truncated HMGB-1–derived A-box protein, a specific, competitive antagonist of HMGB-1.


CIA was induced in DBA/1j mice or dark agouti rats, and animals were examined daily for signs of arthritis. Treatment with polyclonal anti–HMGB-1 antibodies or the A-box protein was initiated at the onset of disease and was administered intraperitoneally twice daily for 7 days. Animals were killed 8 days after initiation of therapy, and immunohistochemical analysis of synovial tissue specimens was performed.


Systemic administration of anti–HMGB-1 antibodies or A-box protein significantly reduced the mean arthritis score, the disease-induced weight loss, and the histologic severity of arthritis. Beneficial effects were observed in both mice and rats. Immunohistochemical analysis revealed pronounced synovial IL-1β expression and articular cartilage destruction in vehicle-treated mice. Both these features were significantly less manifested in animals treated with anti–HMGB-1 antibodies or A-box protein.


Counteracting extracellular HMGB-1 with either neutralizing antibodies or a specific HMGB-1 antagonist may offer a new method for the successful treatment of arthritis. Inflammation and tissue destruction were suppressed in CIA after HMGB-1 blockade.

Rheumatoid arthritis (RA) is a chronic joint disease characterized by synovial inflammation and articular tissue destruction. Excessive local production of proinflammatory cytokines such as tumor necrosis factor (TNF) and interleukin-1 (IL-1) has been demonstrated to play a central role in the pathogenesis of RA (1). Therapeutic intervention through blockade of these molecules has conferred good clinical protection in many, but not all, patients (1). It is plausible that therapeutic efficacy in rheumatic diseases might be achieved by antagonizing additional proinflammatory cytokines.

High mobility group box chromosomal protein 1 (HMGB-1; previously called HMG-1 or amphoterin) is a widely studied, ubiquitous nuclear protein that is present in eukaryotic cells (2–4). As a nuclear protein, HMGB-1 stabilizes nucleosomes and enables nicking of DNA, which facilitates gene transcription (4, 5). HMGB-1 is a 215–amino acid protein with a uniquely conserved sequence among species. Mouse and rat HMGB-1 are identical; they differ from human HMGB-1 by substitution of only 2 amino acid residues. HMGB-1 has 2 separate and characteristic DNA-binding domains, referred to as HMG A box and B box, respectively, each containing ∼80 amino acid residues arranged in 3 α helices (5).

Recent studies quite unexpectedly revealed that extracellularly released HMGB-1 exerts distinctly different actions from its intranuclear functions (6–9). When released as a cytokine, HMGB-1 is an extremely potent macrophage-stimulating factor (7) and a proinflammatory mediator of inflammation (9, 10). It may signal in part through the receptor for advanced glycation end products (RAGE) (11, 12). The interaction of HMGB-1 with RAGE activates several intracellular signal transduction pathways, including mitogen-activated protein kinases, Cdc42, Rac, and a nuclear translocation of nuclear factor κB (13), the transcription factor classically linked to inflammatory processes. Stimulated monocyte/macrophages actively secrete HMGB-1 through lysosomal exocytosis (14). The molecule is released as a late mediator during acute inflammation and participates in an important way in the pathogenesis of systemic inflammation in sepsis, after the early mediator response has resolved (9). Therapeutic administration of neutralizing antibodies against HMGB-1 confers protection against lethality in experimental sepsis and endotoxemia (6, 15). Stimulated macrophages actively secrete HMGB-1, which in turn, stimulates the production of multiple proinflammatory cytokines. We recently demonstrated that extracellular HMGB-1 has a potent ability to induce TNF and IL-1 synthesis and secretion in monocytes and macrophages (7). HMGB-1 caused a delayed and biphasic release of TNF in comparison with that detected after lipopolysaccharide (LPS) activation.

In addition, HMGB-1 may accumulate extracellularly through a completely different mechanism, since it is passively released from disintegrating nuclei of necrotic cells, in contrast to apoptotic cells. Necrotic HMGB-1–deficient cells have a greatly reduced ability to promote inflammation (16). HMGB-1 might thus be a desired link between unprogrammed cell death and an inflammatory response. The B box domain contains the proinflammatory cytokine functionality of the molecule, whereas the A box region has an antagonistic, antiinflammatory effect with therapeutic potential (17). It was recently demonstrated that purified, truncated HMGB-1 A-box protein dose-dependently inhibits HMGB-1–induced TNF and IL-1β release from activated macrophages. Furthermore, administration of highly purified A-box protein or neutralizing antibodies against HMGB-1 rescued mice from lethal sepsis, even when initial treatment was delayed for 24 hours after the onset of infection (8, 15). These results establish a clinically relevant therapeutic window that is significantly wider than that for other known cytokines (6).

Our research group has developed immunohistochemical staining methods for the detection of HMGB-1 in sections of biopsy specimens (18). These tools enabled the identification of a striking cytoplasmic expression of HMGB-1 in synovial tissue cells in experimental arthritis and RA. In contrast, synovial specimens from normal rodent or human synovial tissue exhibited a strictly nuclear HMGB-1 localization. Immunoblotting analysis of synovial fluid samples from patients with RA further confirmed an abundant extracellular presence of HMGB-1 in inflamed joints (18).

Activated macrophages and necrotic cell death due to synovial hypoxia and complement activation are important features of chronic synovitis. Each of these processes leads to extracellular HMGB-1 release, which might contribute to inflammatory pathogenesis in the joints. This study was designed to evaluate the therapeutic potential of HMGB-1 blockade in established collagen-induced arthritis (CIA).


  1. Top of page
  2. Abstract
  6. Acknowledgements


Male DBA/1j mice (Charles River Sweden, Uppsala, Sweden), 6–8 weeks old with a mean weight of 22 gm, and female dark agouti rats (own breeding), 8–12 weeks old with a mean weight of 165 gm, were used in our study. The health status of the animals was monitored according to guidelines from the Swedish Veterinary Board and they were reported free from screened pathogens. Rats were maintained under climate-controlled conditions with a 12-hour light/dark cycle and fed standard rodent chow and water ad libitum. The study was approved by the Ethical Committee of Stockholm North, Sweden.

Induction and evaluation of CIA.

Type II collagen (CII) was prepared from bovine nasal cartilage as previously described (19, 20) and dissolved at 2 mg/ml in 0.1M acetic acid. Collagen solution was emulsified with an equal volume of Freund's complete adjuvant, which was prepared by mixing heat-killed and freeze-dried Mycobacterium tuberculosis (strains C, DT, and PN mixed; Central Veterinary Laboratory, Weybridge, UK) in Freund's incomplete adjuvant (IFA; Difco, Detroit, MI) at 3 mg/ml. Each mouse received 100 μg of CII and 300 μg of M tuberculosis in 0.1 ml of emulsion, injected intradermally in the base of the tail. On day 21, the animals were boosted with an intradermal injection of 100 μg of CII in IFA.

Mice were observed daily for erythema and swelling of the joints. The interphalangeal joints of the digits, the metacarpophalangeal joint and wrist in the forepaw, and the metatarsophalangeal joint and ankle joint in the hind paw were each considered as one category of joint. Individual paws were scored on a scale of 0–3, as follows: 0 = no signs of arthritis, 1 = 1 type of joint affected, 2 = 2 types of joints affected, and 3 = the entire paw affected. Thus, the maximum score for each animal was 12. Arthritis was considered apparent if the total score for the mouse was ≥2. The mean day of arthritis onset was day 28.

Rat CII was prepared from a rat chondrosarcoma and dissolved in 0.01M acetic acid, as previously described (19, 20). Equal volumes of collagen solution and IFA were emulsified at 4°C so that 300 μl of emulsion contained 225 μg of rat CII. Rats were immunized intradermally at the base of the tail with a volume of 300 μl per animal. Rats were observed daily for clinical signs of arthritis, and each paw was scored on a scale of 0–4 as follows: 0 = unaffected, 1 = 1 type of joint affected, 2 = 2 types of joints affected, 3 = 3 types of joints affected, 4 = 3 types of joints affected and maximal erythema and swelling. The joints were categorized similarly as for CIA in mice. The total score for each rat was calculated as an arthritis index, with a maximum value of 16 per animal. All arthritis evaluation was performed by 2 independent observers (RK and HEH) blinded to the identity of the animals.

Production of anti–HMGB-1 antibodies.

Polyclonal antibodies against the B-box domain of HMGB-1 were raised in rabbits (Cocalico Biologicals, Reamstown, PA). Antibody titers were determined by immunoblotting. Antibodies were affinity purified by a protein A–based binding/elution buffer system according to the manufacturer's instructions (ImmunoPure; Pierce, Rockford, IL). The neutralizing activity of anti–HMGB-1 antibodies was confirmed in macrophage cultures exposed to recombinant HMGB-1 and assayed for TNF release.

Production and purification of A-box protein of HMGB-1.

The primers for amplification of complementary DNA (cDNA) encoding the A-box protein of HMGB-1 (261 bp) were as follows: 5′GATGGGCAAAGGAGATCCTAAG3′ and 5′TCACTTTTTTGTCTCCCCTTTGGG3′. A stop codon was added to ensure the accuracy of protein size. Polymerase chain reaction (PCR) products were subcloned into pCRII-TOPO vector Eco RI sites using the TA cloning method (Invitrogen, Carlsbad, CA) according to the manufacturer's instructions. After amplification, the PCR product was digested with Eco RI and subcloned into an expression vector (pGEX) with a glutathione S-transferase (GST) tag (Pharmacia, Piscataway, NJ); correct orientation in positive clones was confirmed by DNA sequencing of both strands. The recombinant plasmids were transformed into protease-deficient Escherichia coli strains BL21 (Novagen, Madison, WI) and incubated in 2× YT medium containing ampicillin (50 μg/ml) for 5–7 hours at 37°C, with shaking, until the optical density at 600 nm reached 1–1.5. Fusion protein expression was then induced by the addition of 1 mM isopropyl thiogalactose for 3 hours.

Recombinant proteins were purified using a glutathione–Sepharose resin affinity column (Pharmacia). A GST vector lacking the cDNA insert was expressed and purified as a control for experiments using the A-box protein. DNase I was added (80 units/ml) to glutathione–Sepharose beads and incubated for 20 minutes at room temperature in buffer containing 100 mM sodium acetate (pH 5) and 5 mM magnesium chloride to remove any contaminating DNA. To remove any contaminating LPS, the A-box protein was purified using a polymyxin B column (Pierce). The purity and integrity of the protein was verified by Coomassie blue staining after sodium dodecyl sulfate–polyacrylamide gel electrophoresis.

Treatment of CIA with neutralizing anti–HMGB-1 antibodies.

Neutralizing anti–HMGB-1 antibodies, as defined by the activity of antibodies to inhibit HMGB-1–mediated stimulation of macrophage TNF release, were administered for 7 days in mice with established CIA with a minimum arthritis index of 2. A total dose of 0.2 mg/day was given intraperitoneally, divided between 2 administrations. An irrelevant rabbit IgG antibody (I-5006; Sigma, St. Louis, MO) served as a control.

Treatment of CIA with the A-box protein.

A-box protein was administered for 7 days in mice and rats with established CIA with a minimum arthritis index of 2. A total dose of 1.2 mg (mice) or 10 mg (rats) per day was given intraperitoneally, divided between 2 administrations. GST-tag protein purified similarly to the A-box protein served as a control.

Immunohistochemical analysis.

Animals were perfused in vivo with formaldehyde solution. Paws were dissected and decalcified, as previously described (21). Sections (8 μm thick) were cut using a Leica cryostat and mounted on glass slides that had been coated with chromium potassium sulfate and gelatin (Novakemi, Stockholm, Sweden). The sections were stained for intracellular production of IL-1β, as previously described (22). The cell membrane and the Golgi organelle were permeabilized by the use of Earle's balanced salt solution (EBSS; Gibco, Paisley, UK) supplemented with 0.1% saponin (Riedel de Haen, Seelze, Germany) in all subsequent washes and incubation steps. Endogenous peroxidase activity was blocked with 1% hydrogen peroxide and 2% sodium azide dissolved in EBSS–saponin for 1 hour at room temperature in the dark. Sections were then washed twice in EBSS–saponin and thereafter blocked with 2% normal human AB serum for 30 minutes. Endogenous biotin was blocked with avidin for 30 minutes and with biotin for 15 minutes (avidin/biotin blocking kit; Vector, Burlingame, CA), both completed with 0.1% saponin.

Sections were thereafter incubated overnight in a humidified chamber with 50 μl of primary antibody: a polyclonal ligand affinity-purified goat anti-rat IL-1β (AF-501-NA; R&D Systems, Minneapolis, MN), used at a final concentration of 2 μg/ml. Slides were then incubated for 30 minutes with the appropriate biotin-labeled, F(ab′)2-fragmented antibody: donkey anti-goat (catalog 705-066-147; Jackson ImmunoResearch, West Grove, PA), final concentration 1:1,000. Thereafter, 50 μl of a solution of avidin–biotin–horseradish peroxidase (Vectastain Elite ABC kit; Vector) was applied for 30 minutes. After final washes in EBSS, the substrate diaminobenzidine (peroxidase substrate kit; Vector) was added and sections were incubated for 5 minutes. Finally, sections were counterstained with Mayer's hematoxylin and mounted with buffered glycerol. Slides were evaluated using a Polyvar II microscope (Reichert-Jung, Vienna, Austria) connected to a charge-coupled device color camera (DXC-750P; Sony, Tokyo, Japan).

Statistical analysis.

The differences between the mean arthritis scores in different groups were compared using the Mann-Whitney U test for independent groups. Then, t-tests were used to evaluate differences in numbers of affected paws.


  1. Top of page
  2. Abstract
  6. Acknowledgements

Therapeutic effects of A-box protein and control therapy in established murine CIA.

DBA/1j mice with early CIA were treated with purified GST-tagged A-box protein or control GST tag alone. Three separate experiments based on a similar treatment protocol were performed, including a total of 23 animals treated with A-box protein and 23 animals treated with GST tag. All 3 experimental trials gave consistent results, demonstrating that repeated A-box injections ameliorated the clinical course of arthritis, in contrast to control treatment. Therapy with twice daily intraperitoneal injections (0.6 mg/dose) of A-box protein or control was initiated when animals expressed a clinical arthritis score of 2 and was continued for 7 days. The animals were killed 8 days after the start of the therapy for immunohistochemical analysis of synovial tissue. The results of the experiment (Figure 1) indicate that A-box treatment led to a final clinical outcome of a mean (±SEM) arthritis index of 3.2 ± 1.8 on day 8 after the start of therapy, which was statistically significantly better than the mean final score of 5.4 ± 1.7 in the control group (Figure 1A).

thumbnail image

Figure 1. Effect of anti–high mobility group box chromosomal protein 1 (anti–HMGB-1) antibodies and A-box protein on severity of established collagen-induced arthritis in DBA/1j mice. Mice were injected intraperitoneally twice daily with either anti–HMGB-1 antibodies (0.1 mg/dose) or A-box protein (0.6 mg/dose). Irrelevant IgG antibody and glutathione S-transferase (GST)–tag protein served as controls, respectively. A, Mean arthritis indices and B, number of affected paws in the different treatment groups. Treatment was initiated when the animals had a minimum arthritis index of 2 (arrow in B) and was administered for 7 days. Significant differences were evident between the anti–HMGB-1 antibody–treated group and the group treated with the irrelevant control antibody (P < 0.05 from day 2 to day 8) and between the A-box protein–treated group and the GST tag–treated group (P < 0.05 from day 2 to day 8). The numbers of mice per group were A-box n = 5, GST tag n = 5, anti–HMGB-1 antibody n = 12 and control antibody n = 11.

Download figure to PowerPoint

A significant reduction in the clinical score in the group treated with A-box protein was not only apparent at the end of the trial, but was evident throughout the whole treatment period (Figure 1A). The mean (±SEM) number of affected paws was also beneficially influenced in the A-box protein–treated group (1.6 ± 0.5) (Figure 1B), which was significantly lower than the number of affected paws in the GST-treated group (2.8 ± 0.4). Consistent results were observed in a separate experiment in which a group of mice with CIA that received no therapy was included (Table 1). The clinical outcome in animals with no treatment or with control GST treatment was not statistically different, while A-box protein–treated animals had a lower arthritis score and fewer affected paws (Table 1). Mice in the A-box protein–treated group also appeared to be generally more healthy and active than the control animals in all trials. Weight loss was significantly lower in the group receiving A-box protein (Table 1).

Table 1. Effect of HMGB-1 targeted therapy on established CIA in mice*
A-boxGST tagUntreatedAnti–HMGB-1 antibodyControl antibody
  • *

    Values are from day 8 after the initiation of therapy. HMGB-1 = high mobility group box chromosomal protein 1; CIA = collagen-induced arthritis; GST tag = glutathione S-transferase tag.

No. of animals6761211
Maximum arthritis index, mean ± SEM3.3 ± 2.36.0 ± 0.77.2 ± 1.42.4 ± 1.85.7 ± 1.8
No. of mice with maximum arthritis score in any paw24647
No. of paws with maximum arthritis score3/246/2810/245/4811/44
No. of paws affected11/2421/2819/2418/4833/44
Weight loss, %

Clinical effects based on therapeutic injections of neutralizing anti–HMGB-1 antibodies or control antibodies in established murine CIA.

DBA/1j mice with early CIA were treated with neutralizing polyclonal anti–HMGB-1 rabbit antibodies or control rabbit antibodies for 7 days (0.1 mg/dose injected intraperitoneally twice daily). Anti–HMGB-1 antibody therapy had effects on established arthritis that were similar to those observed after A-box protein therapy (Figures 1A and B). Mice injected with anti–HMGB-1 antibodies developed a milder form of arthritis throughout the treatment period and reached a final mean arthritis index of 2.4 ± 1.8 on day 8 of arthritis. The corresponding arthritis score in the control antibody–treated group was 5.7 ± 1.8 (Figure 1A). Significantly improved outcomes with regard to the number of affected paws were also documented during the treatment period in the anti–HMGB-1 antibody–treated group (mean ± SEM 1.5 ± 0.8, compared with 3.0 ± 0.7 in the control group) (Figure 1B and Table 1). Mice treated with anti–HMGB-1 antibodies were also generally healthier than mice treated with control antibodies and lost less weight than the controls (Table 1). The results from anti–HMGB1 antibody–treated animals were not statistically different from those obtained with A-box therapy.

Histopathologic assessment of synovial tissue in murine CIA after HMGB-1 blockade.

Immunohistochemical staining of synovial tissue sections was used to study local effects in the diseased target organ. Cryopreserved sections from mouse paws that had exhibited clinical signs of inflammation at the start of the study were obtained and studied 8 days after the start of therapy with anti–HMGB-1 antibodies, A-box protein, or control therapy, respectively. For immunohistochemical analysis, arthritic paws with good response to A-box protein or anti–HMGB-1 antibody treatment (paws with an initial score ≥1 and an end score of 0) were chosen, together with paws from control animals with similar initial scores.

Microscopic examination after hematoxylin staining revealed that mice treated with HMGB-1 blockade had significantly less cellular synovial proliferation than did control-treated animals. Severe destruction of cartilage and bone was visible in articular specimens from mice treated with GST tag or control antibody. In contrast, much less destruction of cartilage or bone was observed in mice treated with A-box protein or anti–HMGB-1 antibody.

Only very few IL-1β–producing cells were seen in synovial tissue sections from mice treated with A-box protein or with anti–HMGB-1 antibodies (Figures 2A and C). In contrast, specimens from mice treated with GST tag or irrelevant antibody displayed a strong production of IL-1β, mostly localized in areas infiltrated with inflammatory cells (Figures 2B and D).

thumbnail image

Figure 2. Immunostaining of hind paw ankle joints in mice with collagen-induced arthritis (CIA) treated with A-box protein and anti–high mobility group box chromosomal protein 1 (anti–HMGB-1) antibody. Immunohistochemical staining for interleukin-1β (IL-1β) in synovial tissue from mice treated with A, anti–HMGB-1 antibody, B, irrelevant control antibody, C, A-box protein or D, glutathione S-transferase tag on day 8 after the onset of CIA. A mild synovitis is evident in A and C, while marked cellular infiltration and erosion of cartilage and underlying bone are evident in B and D from control-treated animals. IL-1β–producing cells were particularly visualized in the pannus tissue (arrow in B). c = cartilage; p = pannus; sc = synovial cavity; b = bone. Original magnification × 250.

Download figure to PowerPoint

Clinical effects of A-box therapy in established CIA in dark agouti rats.

Dark agouti rats with early CIA were also treated in an analogous manner to the protocol used for murine CIA, but received a higher daily dose of A-box protein or GST tag (10 mg/day). Rats treated with A-box protein had a significant reduction of the clinical severity of arthritis compared with the control group (Table 2), consistent with results obtained in the murine studies. Eight days after the start of A-box therapy, the maximum mean arthritis index was 5.6 ± 2.2, which was statistically significantly lower than in the control group, with a mean score of 9.5 ± 1.6 (Table 2). No significant differences with regard to weight loss and the number of affected paws were observed during the study period. Outside the study protocol, we followed the clinical severity of arthritis for another 7 days after completion of A-box protein treatment in a subgroup of animals. The arthritis score increased rapidly in these animals after A-box protein therapy was withheld, with the animals developing severe clinical disease equivalent to that in control-treated animals (data not included).

Table 2. Effect of A-box protein therapy on established CIA in rats*
A-boxGST tag
  • *

    Values are from day 8 after the initiation of therapy. CIA = collagen-induced arthritis; GST tag = glutathione S-transferase tag.

No. of animals1111
Maximum arthritis index, mean ± SEM5.6 ± 2.29.5 ± 1.6
No. of rats with maximum arthritis score in any paw27
No. of paws with maximum arthritis score2/4410/44
No. of paws affected35/4443/44
Weight loss, %9.810.6


  1. Top of page
  2. Abstract
  6. Acknowledgements

We recently reported that the pattern of HMGB-1 expression is altered in synovial biopsy tissues obtained from RA patients and from animals with experimental arthritis, as compared with nonarthritic controls (18), with the proinflammatory mediator being overexpressed cytoplasmically in macrophages. It is also expressed extracellularly in the synovial membrane and is detectable in synovial fluid from patients with RA. Synovial tissue from uninflamed joints mainly displays a cellular HMGB-1 expression that is restricted to the nuclear compartment. When considered with the results described here, it now appears that HMGB-1 may be targeted to therapeutic advantage in a standard preclinical model of arthritis.

The outcome of the present study demonstrates that HMGB-1 blockade therapy with antibodies or antagonizing A-box protein ameliorates CIA in both mice and rats and thus strongly implicates HMGB-1 as a pathogenetic factor in the arthritic process. The molecule has also been shown to be a potent proinflammatory mediator in a number of other conditions, including experimental septicemia (23), endotoxemia (6), and acute lung inflammation (10). Therapies based on neutralizing anti–HMGB-1 antibodies and truncated A-box protein conferred significant clinical protection in all these other disease models, with systemic in vivo administration of A-box protein even significantly increasing the survival rate in a standardized murine sepsis model when treatment was delayed 24 hours from disease onset (15). Additionally, A-box protein dose-dependently inhibited HMGB-1–induced TNF and IL-1 release in cultured macrophages.

We evaluated the effects of HMGB-1 blockade therapy on established disease in experimental arthritis, since this may be more relevant to the design of clinical trials than a pretreatment approach. We did not observe toxicity in the animals treated with either anti–HMGB-1 antibodies or A-box protein. The results of the present protocol using neutralizing anti–HMGB-1 antibodies or purified A-box protein were clearly beneficial, because CIA was ameliorated to a similar extent as that found after anti-TNF antibody treatment of established murine CIA in previous studies (27). HMGB-1 and TNF blockade regimens both led to clinical improvement within 1 or 2 days. Future studies are required to determine whether therapeutic targeting of HMGB-1 or TNF may provide selective opportunities or drawbacks in the treatment of arthritis. Currently, it is not fully understood to what extent HMGB-1 depends on TNF to mediate its proinflammatory function.

The clinical results of the HMGB-1 blockade therapies in CIA are quite promising and are of great conceptual interest. From a practical point of view, it is conceivable that refinement of HMGB-1 antagonists will further improve the results. In the future, neutralizing monoclonal antibodies might lend themselves to new therapeutic trials. The recent recognition of the B-box motif as the cytokine-inducing part of the molecule has been instrumental in generating effective anti–HMGB-1 IgG monoclonal antibodies for future therapeutic studies. The truncated A-box protein used in our studies is a low molecular construct of 80 amino acid residues. The fact that the cytokine-inducing part of the HMGB-1 protein is identical in humans, mice, and rats provides an unusual therapeutic opportunity. HMGB-1–neutralizing antibodies or fusion proteins identified in experimental studies may confer good clinical protection in human disease without any need for modification.

The precise mode of action through which HMGB-1 blockade therapy reduced joint disease cannot be elucidated from the present study and needs further investigation. Immunohistochemical staining assessments indicate inhibited synovial proliferation and pannus formation in animals treated with anti–HMGB-1 antibodies or A-box protein. The joints from these animals also expressed a reduced number of IL-1β–producing cells compared with controls. Efficient neutralization of extracellular HMGB-1 activity would be expected to cause a down-regulation of IL-1β and TNF formation, since HMGB-1 is a potent stimulus of the production of these cytokines (7). Another consequence of HMGB-1 interaction with RAGE concerns its facilitation of tissue penetration. HMGB-1 activates plasminogen (25) and stimulates additional proteolytic enzymes, including several metalloproteinases. The reduced tissue destruction seen in anti–HMGB-1–treated animals may thus be a direct consequence of HMGB-1 neutralization. Taken together, these results depict HMGB-1 as a therapeutic target molecule of considerable interest in human disease.


  1. Top of page
  2. Abstract
  6. Acknowledgements

The authors thank Associate Professor R. A. Harris for linguistic advice and critical reading of the manuscript.


  1. Top of page
  2. Abstract
  6. Acknowledgements
  • 1
    Taylor PC, Williams RO, Maini RN. Immunotherapy for rheumatoid arthritis. Curr Opin Immunol 2001; 13: 6116.
  • 2
    Mosevitsky MI, Novitskaya VA, Iogannsen MG, Zabezhinsky MA. Tissue specificity of nucleo-cytoplasmic distribution of HMG1 and HMG2 proteins and their probable functions. Eur J Biochem 1989; 185: 30310.
  • 3
    Bustin M, Lehn DA, Landsman D. Structural features of the HMG chromosomal proteins and their genes. Biochim Biophys Acta 1990; 1049: 23143.
  • 4
    Vaccari T, Beltrame M, Ferrari S, Bianchi ME. Hmg4, a new member of the Hmg1/2 gene family. Genomics 1998; 49: 24752.
  • 5
    Bustin M. Regulation of DNA-dependent activities by the functional motifs of the high-mobility-group chromosomal proteins. Mol Cell Biol 1999; 19: 523746.
  • 6
    Wang H, Bloom O, Zhang M, Vishnubhakat JM, Ombrellino M, Che J, et al. HMG-1 as a late mediator of endotoxin lethality in mice. Science 1999; 285: 24851.
  • 7
    Andersson U, Wang H, Palmblad K, Aveberger AC, Bloom O, Erlandsson-Harris H, et al. High mobility group 1 protein (HMG-1) stimulates proinflammatory cytokine synthesis in human monocytes. J Exp Med 2000; 192: 56570.
  • 8
    Andersson U, Erlandsson-Harris H, Yang H, Tracey KJ. HMGB1 as a DNA-binding cytokine. J Leukoc Biol 2002; 72: 108491.
  • 9
    Wang H, Yang H, Czura CJ, Sama AE, Tracey KJ. HMGB1 as a late mediator of lethal systemic inflammation. Am J Respir Crit Care Med 2001; 164: 176873.
  • 10
    Abraham E, Arcaroli J, Carmody A, Wang H, Tracey KJ. Cutting edge: HMG-1 as a mediator of acute lung inflammation. J Immunol 2000; 165: 29504.
  • 11
    Hori O, Brett J, Slattery T, Cao R, Zhang J, Chen JX, et al. The receptor for advanced glycation end products (RAGE) is a cellular binding site for amphoterin: mediation of neurite outgrowth and co-expression of RAGE and amphoterin in the developing nervous system. J Biol Chem 1995; 270: 2575261.
  • 12
    Schmidt AM, Yan SD, Yan SF, Stern DM. The biology of the receptor for advanced glycation end products and its ligands. Biochim Biophys Acta 2000; 1498: 99111.
  • 13
    Huttunen HJ, Fages C, Rauvala H. Receptor for advanced glycation end products (RAGE)-mediated neurite outgrowth and activation of NF-κB require the cytoplasmic domain of the receptor but different downstream signaling pathways. J Biol Chem 1999; 274: 1991924.
  • 14
    Gardella S, Andrei C, Ferrera D, Lotti LV, Torrisi MR, Bianchi ME, et al. The nuclear protein HMGB1 is secreted by monocytes via a non-classical, vesicle-mediated secretory pathway. EMBO Rep 2002; 3: 9951001.
  • 15
    Yang H, Wang H, Erlandsson-Harris H, Czura CJ, Qiang X, Susarla SM, et al. HMGB1A box antagonizes HMGB1 and rescues mice with established sepsis. Submitted for publication.
  • 16
    Scaffidi P, Misteli T, Bianchi ME. Release of chromatin protein HMGB1 by necrotic cells triggers inflammation. Nature 2002; 418: 1915.
  • 17
    Li J, Kokkola R, Tabibzadeh S, Yang R, Ochani M, Qiang X, et al. Structural basis for the proinflammatory cytokine activity of high mobility group box 1 (HMGB1). Mol Med. In press.
  • 18
    Kokkola R, Sundberg E, Ulfgren AK, Palmblad K, Li J, Wang H, et al. High mobility group box chromosomal protein 1: a novel proinflammatory mediator in synovitis. Arthritis Rheum 2002; 46: 2598603.
  • 19
    Smith BD, Martin GR, Miller EJ, Dorfman A, Swarm R. Nature of the collagen synthesized by a transplanted chondrosarcoma. Arch Biochem Biophys 1975; 166: 1816.
  • 20
    Andersson M, Holmdahl R. Analysis of type II collagen-reactive T cells in the mouse. I. Different regulation of autoreactive vs. non-autoreactive anti-type II collagen T cells in the DBA/1 mouse. Eur J Immunol 1990; 20: 10616.
  • 21
    Ahmed M, Bjurholm A, Schultzberg M, Theodorsson E, Kreicbergs A. Increased levels of substance P and calcitonin gene-related peptide in rat adjuvant arthritis: a combined immunohistochemical and radioimmunoassay analysis. Arthritis Rheum 1995; 38: 699709.
  • 22
    Palmblad K, Andersson U. Identification of rat IL-β, IL-2, IFN-γ and TNF-α in activated splenocytes by intracellular immunostaining. Biotech Histochem 2000; 75: 1019.
  • 23
    Ulloa L, Ochani M, Yang H, Tanovic M, Halperin D, Yang R, et al. Ethyl pyruvate prevents lethality in mice with established lethal sepsis and systemic inflammation. Proc Natl Acad Sci U S A 2002; 99: 123516.
  • 24
    Williams RO, Feldmann M, Maini RN. Anti-tumor necrosis factor ameliorates joint disease in murine collagen-induced arthritis. Proc Natl Acad Sci U S A 1992; 89: 97848.
  • 25
    Parkkinen J, Raulo E, Merenmies J, Nolo R, Kajander EO, Baumann M, et al. Amphoterin, the 30-kDa protein in a family of HMG1-type polypeptides: enhanced expression in transformed cells, leading edge localization, and interactions with plasminogen activation. J Biol Chem 1993; 268: 1972638.