HK1069316B - Hmgb1 protein inhibitors and/or antagonists for the treatment of vascular diseases - Google Patents

Description

HMGB1 protein inhibitors and/or antagonists for the treatment of vascular diseases

The present invention relates to the field of molecular biology, in particular the treatment of vascular diseases, including vascular diseases due to angioplasty, using inhibitors of the HMGB1 protein and antagonists of HMGB 1.

The HMGB1 protein (HMG; Bustin, 2001, trends biochem. Sci., 26, 152-153, since 2001) is a profilin of the HMG box family, characterized by the presence of a DNA binding domain called HMG box. HMG1 is a small 25KD protein of 215 amino acids with a highly conserved sequence in mammals. The HMGB1 molecule is organized into three regions: two DNA binding regions, HMG box A and box B, which are followed by COOH acidic termini consisting of 30 residues of glutamic and aspartic acids. Box A and box B are fragments of 80 amino Acids (29% identical, 65% similar) with an L-shaped three-dimensional structure (Hardman et al, 1995, biochemistry, 34: 16596-.

HMGB1 was originally thought to be a ubiquitous and abundant nuclear protein. It has more than 1 million copies per nucleus and binds double-stranded DNA non-sequence-specifically. However, HMGB1 binds with strong affinity to specific DNA structures, such as kinked or bent DNA and four-way linkers. However, HMGB1 can be recruited to double-stranded DNA by interacting with several different DNA binding proteins. When bound to double-stranded DNA, it induces structural deformation allowing the formation of a nuclear protein complex, in which case several DNA binding proteins come into contact with each other while binding to respective DNA-related sites (Muller et al, 2001, EMBO J., 16: 4337-. The phenotype of the Hmgbl-/-mouse is consistent with this model (Calogero et al, 1999, nat. Genet., 22: 276-.

Another role of HMGB1 outside the nucleus has recently attracted attention: HMGB1 acts as a late mediator in endotoxin-induced lethality and acute pneumonia in mice; also elevated levels of HMGB1 in the serum of septic patients is a predictive marker (International patent application No. WO 00/47104). HMGB1 is secreted by cultured macrophages and pituitary cells under stress to cytokines and bacterial endotoxins (Abraham et al 2000, J.Immunol., 165: 2950-. If added exogenously, HMGB1 modulates neurite outgrowth, laminin-dependent migration of neuroblastoma and glioma is inhibited by antibodies to HMGB1 (Fages et al, 2000, J.CellSci., 113: 611-620; Merenmies et al, 1991, J.biol.Chem., 266: 16722-16729; Parkkinen et al, 1993, J.biol.Chem., 268: 19726: 19738; Rauvala et al, 1988, J.cell biol., 107: 2293-2305). The interaction of HMGB1 with the plasminogen activation system, in particular with t-PA (tissue plasminogen activator), enhances plasmin formation (Parkkien and Rauvalla, 1991, J.biol.chem., 266: 16730-. The degradation of extracellular matrix proteins is an important step in the cell migration process, which is enabled by the increased extracellular protease activity promoted by HMGB 1.

HMGB1 was identified as a ligand that binds to RAGE receptors (receptors for advanced glycation endproducts) (Hori et al, 1995, J.biol.chem., 270: 25752-25761). RAGE is a multi-ligand receptor of the immunoglobulin superfamily that can be expressed in a wide variety of cell types, including endothelial cells, smooth muscle cells, mononuclear phagocytes, and neurons (Brett et al, 1993, am. J. Phathol., 143: 1699-K1712; Neeper et al, 1992, J. biol. chem., 267: 14998-K15004). It is involved in several different pathological processes, such as: diabetes, amyloidosis, and atherosclerosis (Schmidt et al, 1999, Circ. Res., 84: 489-197). The interaction of HMGB1 and RAGE induces neurite outgrowth, a protein that is co-localized to the leading edge of an extended neurite during embryonic development (Huttunen et al, 1999, J.biol.chem., 274: 19919-. It was observed that the growth and migration of tumor mass prevented the interaction of HMGB1 and RAGE; inhibition of this interaction can inhibit activation of mitogen-activated protein (MAP) kinase and expression of matrix metalloproteinases, molecules that are strongly associated with tumor proliferation and invasion.

The inventors of the present invention demonstrated that HMGB1 has a potent biological effect on Smooth Muscle Cells (SMC), one of the cell types in which RAGE is expressed on the surface. Vascular SMC cells are the most abundant cells in large blood vessels; they are located in the tunica media, which are embedded in the extracellular matrix. In intact blood vessels, SMC cells are in a contracted state, exhibiting a phenotype, i.e., no cell division and migration responsible for the maintenance of the rigidity and elasticity of the vessel wall and blood pressure control.

After mechanical or inflammatory injury, when the endothelium is destroyed, SMC cells switch to a synthetic phenotype, undergoing cell division and cell migration. SMC cells migrate from the tunica media to the intima of the blood vessel, causing intimal thickening, which plays an important role in the pathophysiology of many vascular diseases, such as atherosclerosis and restenosis following coronary angioplasty. In the synthetic state, SMCs also produce large amounts of extracellular proteases, growth factors and cytokines, secreting a fibrous extracellular matrix. Following injury to the vessel wall, several growth factors and/or chemoattractants can induce the SMC cells to switch from a contracting to a synthetic phenotype by circulating monocytes, macrophages and platelets or by release of injured endothelial cells, which can direct SMC cell migration toward the intima of the vessel. Among these factors, bFGF is one of the most important factors, but SMC cells can also start to migrate under stress from angiogenic stimulators. (Schwartz, 1997, J.Clin. Invest., 99: 2814. 2816; Van Leeuwen, 1996, Fibrinolysis, 10: 59-74).

In an attempt to define the role and mechanism of HMGB1 in inducing RSMC cell migration, the inventors demonstrated that HMGB1 is a strong chemoattractant that is able to induce their cell shape change and cytoskeletal reorganization. These events were suppressed by the addition of an antibody against RAGE and by pertussis toxin, suggesting that RAGE and Gi/o proteins may both be involved in this pathway. Furthermore, evidence that HMGB1 promotes migration of phosphorylated ERK1 and 2 proteins into the nucleus suggests involvement of the MAP kinase pathway. Next, HMGB1 was shown to be released by many injured or necrotic cell types, including endothelial cells.

Therefore, HMGB1 has all the features of a molecule that promotes atherosclerosis and restenosis after vascular injury.

The inventors have also demonstrated that the fragment of HMGB1 (corresponding to HMG box) has a higher potency than the entire full-length molecule, and that even the HMG box domain of other proteins of the HMG box family can induce the same effect.

As a result, every molecule that blocks the interaction between HMGB1 and its RAGE receptor (i.e., all molecules belonging to the inhibitor class: antibodies or antibody fragments, four-way DNA; all molecules belonging to the HMG box antagonist class: molecules of the HMGB1 fragment containing the HMG box domain) can be effectively used in the production of pharmacological agents to avoid, delay or inhibit atherosclerosis and restenosis caused by vascular epithelial injury, even due to angioplasty.

The HMGB1 binding molecule or HMGB1 inhibitor may be injected or released from the instrument used in angioplasty procedures, or the molecule can be bound to the surface of the instrument.

The object of the present invention is to provide the use of molecules capable of blocking the interaction of HMGB1 and RAGE for the preparation of a therapeutic agent for the treatment of vascular diseases.

In a preferred embodiment of the invention, said molecule is released from a catheter, surgical instrument or angioplasty stent during or after said surgery.

Further features and advantages of the present invention will become apparent from the following detailed description, taken in conjunction with the accompanying drawings.

In the drawings:

FIG. 1 shows the chemotactic activity of HMGB1 on RSMC in a chemotaxis assay performed using a modified Boyden Chamber. The 100% value corresponds to the number of cells migrating in the absence of any stimulus (random cell migration). Values represent mean ± SD (n ═ 3). FIG. 1-A shows the concentration-dependent migratory response of RSMC to purified HMGB1 from calf thymus. FIG. 1-B shows a comparison of chemotaxis of HMGB1 protein purified from calf thymus or expressed in yeast with chemoattractants fMLP and bFGF. FIG. 1-C shows the effect of antibodies against HMGB1 on fMLP-and HMGB 1-induced migration. The asterisk (—) indicates treatments in which the statistical difference in the student's t-test from the control for the migration response was greater than p ═ 0.0001. FIG. 1-D shows the concentration-dependent migration response of RSMC to HMGB1 expressed in yeast (Pichia pastoris).

Figure 2 shows the effect of HMGB1 on RSMC morphology and cytoskeletal organization. FIG. 2-A shows the effect of HMGB1 purified from calf thymus or expressed in yeast or E.coli on sub-confluent (subonfluent) cultures of RSMC. Actin filaments were visible using TRIC-clitocybin. Figure 2-B shows how anti-HMGB 1 rabbit antibodies inhibit HMGB1 stimulated cytoskeletal reorganization. Resting cells (state 1) showed many stress fibers. Non-resting cells (state 2) showed reorganization of the actin cytoskeleton.

FIG. 3 shows the chemotactic response of RSMC to the HMG box domain of HMGB 1. FIG. 3-A shows concentration-dependent responses to cassette B and (e) cassette A, both of which were expressed in E.coli. Random cell migration was set as 100% migration. Data represent mean ± SD (n ═ 3). Statistical significance P of the results in the ANOVA model was <0.0001 for both box a and box B. FIG. 3-B shows the effect of full-length HMGB1, Box A + B, Box A or Box B expressed in E.coli on actin cytoskeletal organization. Actin filaments were visualized using TRIC-clitocybin.

Fig. 4 shows the effect of HMGB1 and its HMG box on migration of RSMC to the wound. The value 100% corresponds to the amount of cell migration (basal migration) when no stimulus is present. Data represent mean ± SD (N ═ 5). Statistical significance of bFGF and full length HMGB1 treatment made by bacteria was 0.05< p <0.01, 0.01< p <0.001 for box a and box B treatment, 0.001< p <0.0001 for calf thymus HMGB1 treatment.

Figure 5 shows how HMGB1 binds to the surface of RSMC and stimulates the motility of cells by RAGE. Fig. 5-a shows that a large amount of HMGB1 binds to the surface of the RMSC. RSMC expressing RAGE is shown in FIG. 5-B. FIG. 5-C shows how antibodies against RAGE inhibit HMGB 1-induced migration of RSMC. Statistical significance of treatment with HMGB1 and HMGB1 plus nonspecific antibodies 0.001< p < 0.0001.

Figure 6 shows how pertussis toxin inhibits HMGB 1-induced RSMC migration and actin cytoskeleton reorganization. Chemotaxis assays performed using a modified Boyden chamber are shown in FIG. 6-A. A value of 100% corresponds to the migration of the basal cells in the absence of any stimulus; data represent mean ± SD. FIG. 6-B shows significant cytoskeletal reorganization, and actin filaments were visible using conjugated TRITC-clitocide.

Fig. 7 demonstrates that the MAP kinase pathway is involved in HMGB1 signaling. Cells were stained with antibodies specific for phosphorylated ERK1/2 and DAPI, and individual cell samples were stained with TRITC-curculin to visualize reorganization of the cytoskeleton.

Fig. 8 shows HMGB1 released by necrotic and damaged cells. FIG. 8-A shows the results of Western blot analysis of proteins released by necrotic or permeabilized Hela cells; the presence of HMGB1 is evident in lanes 1 and 3. FIG. 8-B shows the results of immunofluorescence analysis of necrotic and viable Hela cells.

Figure 9 shows that HMGB1 is present in the nucleus of endothelial cells, but not in the nucleus of vascular SMC. HMGB1 was shown to be present in the nuclei of endothelial cells in fig. 9-a and 9-B, but was not found in vascular smooth muscle cell nuclei of human pancreatic arterial sections stained with antibodies against HMGB1 and counterstained with ematoxylin, as indicated by low (a) and high (B) magnification. The red box indicates the location of the region shown in the graph B, with the arrow pointing to the core of the SMC. Western blot analysis in fig. 9-C shows the expression level of HMGB1 in RSMC (in contrast to Hela cells).

FIG. 10 shows the chemotactic effect of HMGB1 on mouse embryonic fibroblasts in a chemotaxis assay performed using a modified Boyden chamber in the presence or absence of anti-RAGE antibody (1000 ng/ml). The value of 100% corresponds to the number of cell migrations without any stimulus (random cell migration). Data represent mean ± SD (n ═ 3).

Expression and purification of HMGB1 and derivatives thereof

In the first step, HMGB1 and its derivatives must be expressed and purified.

Expression of full-length HMGB1 was accomplished in E.coli transformed with pT7-7-rHMGB1cm plasmid (gif. J. O. Thomas, Cambridge University) and purification was accomplished following a well-known procedure (Muller et al, 2001, Biochemistry, 40: 10254-10261).

Expression and purification of full-length HMGB1 in Pichia pastoris was accomplished following a well-known procedure (Mitry et al, 1997, Biotechniques, 22: 718-.

The well-known plasmids pRNHMG1/M1-V176, pT7HMG1bA and pT7HMG1bB were used for expression and purification of BoxA + BoxB, BoxA and BoxB, respectively, following well-known single-and two-cassette purification procedures (Bianchi et al, 1992, EMBO J., 11: 1055-.

To demonstrate the chemotactic effect of HMGB1, three independent cell migration assays were performed: chemotaxis assays, and in vitro wound assays. The functional relationship between HMGB 1-induced cell migration and morphological changes (i.e. actin fiber reorganization, cell elongation and cell shape polarization) of non-resting cells was investigated.

Chemotaxis assay

Chemotaxis assays were performed using well-known procedures (Degryse et al, 1999, Blood, 94: 649-. A modified Boyden chamber was used with a 0.5um pore size filter (Coring, Acton, MA) and treated with fibronectin (10ug/ml) (Roche) and collagen I (100ug/ml in 0.5M acetic acid). RSMC cells (dr. marcobertuli, Bayer Research Laboratories, Milan) were cultured in DMEM without serum, and 20.000-40.000 cell samples were added to the upper well of the Boyden chamber. The molecules to be tested were diluted in the same serum-free medium and added to the lower well.

Different HMGB1 formulations were used: purified HMGB1 from calf thymus (as given by J.Bernous, C.S.I.C., Barcelona, Spain), E.coli-expressed recombinant HMGB1, and a slightly modified HMGB1 form produced on Pichia pastoris (containing EAEAEAYVEF amino acids bound to the N-terminus) (Mistry et al, 1997, Biotechniques, 22: 718-729).

If desired, polyclonal rabbit anti-HMGB 1(Pharmingen BD, TorreyPines, CA), Pertussis Toxin (PT) from Bordetella pertussis (a gift from dr.m.g. pizza, i.r.i.s., Siena), or an inhibitor was added to both wells.

Cell migration was maintained at 37 ℃ throughout the night, followed by scraping off cells remaining on the upper surface of the filter, which was fixed in methanol and stained with a 20% methanol solution containing 10% gentian violet. All experiments were done at least twice, in triplicate.

The results shown in FIGS. 1-A, 1-B, 1-C, 1-D are the mean. + -. SD of the number of cells in the high density region (high power fields) per filter 10, expressed as a multiple of the control. Random cell migration (i.e., migration in the absence of chemoattractant) gave an artificial value of 100%.

Treatments were either compared in pairs using student's-t test for statistical analysis or treatment with increasing amounts of reagents were evaluated using the ANOVA model.

HMGB1 from calf thymus stimulated migration of RSMC in a concentration-dependent manner, beginning at a low dose of 0.1ng/ml with a 2.5-fold maximal response at 100ng/ml (fig. 1-a). The effect of HMGB1 was comparable in magnitude to the effects of the identified attractants fMLP and bFGF (fig. 1-B). Polyclonal HMGB1 antibody, but not the non-specific control antibody, completely blocked the migration response (fig. 1C), indicating that this is specifically attributed to HMGB 1. These antibodies were unable to alter the effect of the chemoattractant fMLP peptide as a positive control. Similar results were obtained with recombinant HMGB1 produced in Pichia pastoris (FIG. 1-D).

Immunofluorescence assay

15.000-20.000RSMC (20-40% confluency) samples at 2cm²Glass coverslips in wells were seeded, incubated in DMEM supplemented with 10% FCS for 24 hours, washed with PBS and incubated in DMEM without FCS for a further 24 hours. RSMC was stimulated with 100ng/ml HMGB1 at 37 ℃ for increasing time intervals of 5 to 120 minutes. After stimulation, the RSMC was fixed with 3% paraformaldehyde and 2% sucrose in PBS (pH 7.5) for 20 minutes at room temperature, followed by three washes with 0.2% PBS-BSA. Cells were permeabilized with 20mM hepes, pH7.4, 300mM sucrose, 50mM sodium chloride, 3mM magnesium chloride, 0.5% (V/V) Triton X-100 for three minutes at 4 ℃ and washed three times with 0.2% PBS-BSA. The RSMC was then incubated with 2% PBS-BSA at 37 ℃ for 15 minutes, with primary antibody at 37 ℃ for 30 minutes, washed three times with 0.2% PBS-BSA and incubated with 2% PBS-BSA for 15 minutes. Finally, thinCells were stained with secondary antibodies and/or rhodamine (rodamin) -bound curculin to visualize filamentous actin; in some cases, DAPI (4', 6-diamidino-2-phenandiolo, Roche) is used to label the nucleus.

After all subsequent incubations were completed, the coverslips were washed three times with 0.2% PBS-BSA, washed twice with distilled water, mounted in 20% Mowiol PBS and analyzed under an Axiophot microscope (Carl Zeiss). Fluorescent photographs were taken on T-Max400 or EPH P1600X film (Eastman kodak) with Zeiss40 or 100 neoflurar lenses.

Figure 2-a low magnification photograph shows that stress fiber content, cell shape and size, cytoskeletal tissue changed within 30 minutes, but recovered after 120 minutes. High magnification photographs (FIG. 2-B) show many clearly visible stress fibers before stimulation and the cell shape is non-polarized. Within 15-30 minutes, morphology and cytoskeletal organization were completely altered: RSMC show an extended, polarized morphology, which reflects spatial reorganization of the actin cytoskeleton. The effect of HMGB1 slowly fades: after 1-2 hours, the stress fiber content increased back to the original level and cell morphology restored to a similar extent to the unstimulated control cells.

In some experiments, cells were pretreated overnight with antibody or PT or inhibitor. As shown in fig. 2-B, HMGB1 antibody completely inhibited cytoskeletal reorganization and RSMC morphological changes induced by HMGB 1. The control antibody was unable to inhibit the effect of HMGB 1.

Finally, to determine whether the observed effect of HMGB1 on RSMC actually reflects a dynamic transition from a resting state to a moving state, the proportion of cells in different states was quantified. Low magnification photographs were taken and the cells were divided into two states:

state 1, the cells show the appearance of typical unstimulated cells, characterized by a large number of stress fibers and a non-polarized cell shape;

state 2, RSMC shows low stress fiber content, membrane ruffling, actin semiring or an elongated shape.

FIG. 2-C clearly shows that 60% of the cells in unstimulated culture are in State 1 and 40% are in State 2; the proportion of state 2 cells increased to 60% within 5 minutes after stimulation and to 80% after 15-30 minutes. One hour after HMGB1 stimulation, these ratios returned to the values of the unstimulated culture state, with 60% RSMC in state 1 and 40% in state 2. These data demonstrate that the effect of HMGB1 is transient, representing a change from a quiescent state to a transitional state, and confirm the chemotaxis results: HMGB1 is a chemoattractant for RSMC.

In vitro wound analysis

Confluent culture of RSMC at 2-cm²Glass coverslips were grown in wells, washed once with PBS, and starved for 24 hours with FCS in serum-free DMEM. The wound was then stimulated and a single line drawn in the central area of the monolayer with the tip of a pipette. The treated monolayers were washed once with PBS and recovered for 48 hours in serum-free medium with or without HMGB1(100 ng/ml). The cells were then fixed and stained with TRITC-umbrella toxin. Migration was quantified by taking a low magnification photograph and counting cells that had migrated into the cell-free space. Data represent mean ± SD, values of 100% correspond to the number of cells migrating without any stimulus (basal migration).

Figure 4 shows that HMGB1 stimulation increased the number of cells migrating 5-2 fold. Cassette A and cassette B (10ng/ml) were also tested, both stimulating 1.8 fold cell migration. Finally, comparison with bFGF indicates that the molecules described above are more potent. It is postulated that wound healing and chemochemotaxis, chemotaxis may be based on the same signaling pathways.

Signal path

Then, a signal path is detected.

As a migratory signal, HMGB1 must reach the responding cell and bind to the receptor. To test whether HMGB1 bound to the surface of RSMC, 1 million cells were trypsinized and incubated for 20 minutes at 4 ℃ in PBS containing 800ng of cassette a + B peptide and 5ug of bsa. The BoxA + BoxB polypeptide is slightly smaller than the endogenous full-length HMGB1 and is easily identified in SDS-PAGE gels. Then the cells are agglomerated, and the supernatant is reserved; cells were rinsed twice in 500ul cold PBS, resuspended in SDS-PAGE sample buffer, heated at 100 ℃ for 5 minutes, loaded onto 12% tricine-SDS gel (lineP), and 20ul of supernatant (lineS) was adjacent. Gels were then blotted onto Immobilon filters and stained with printing ink.

Figure 5-a shows an SDS-PAGE gel from which the number of cassette a + B recovered from the cell pellet and supernatant can be calculated and it can be estimated that more than 500000 cassette a + B molecules bind on a single RSMC. This result demonstrates that extracellular HMGB1 can bind RSMC, but most likely does not reflect the actual number of receptors. Indeed HMGB1 has been shown to bind to heparin and proteoglycans (Bianchi, 1988, EMBO J., 7: 843-; thus, HMGB1 may also be associated with the extracellular matrix produced by RSMC, as demonstrated by the inventors in Hela cells, where only a small amount of HMGB1 binds the cells, since these cells produce little extracellular matrix.

HMGB1 has been reported to bind to RAGE expressed by many cell types. To demonstrate the presence of RAGE on RSMC membranes, one million RSMCs were lysed on plates containing SDS-PAGE sample buffer (50mM pH6.8 Tris, 2% 2-mercaptoethanol, 4% SDS, 12% glycerol (glicerol), 0.05% bromophenol blue), denatured at 100 ℃ for 5 min, and separated on 12% acrylamide. The isolated proteins were blotted onto Immobilon (Millipore) membranes using a tankbit system containing 25mM PH7.5Tris, 0.192M glycine, 20% methanol. Blots were blocked in 5% skim milk/TBST for one hour at room temperature (20mM Tris, pH7.5, 137mM NaCl, 0.1% Tween20), TBST rinsed three times, and incubated with anti-HMGB 1 antibody in TBST-0.01% BSA. After rinsing with TBST-0.01% BSA, incubation with secondary antibody was performed. Proteins were detected using the ECL system (Amersham). The presence of RAGE was detected using an anti-RAGE antibody (as gifted by dr.a.m.schmidt, Columbia University, NY). The results of FIG. 5-B demonstrate the presence of RAGE in RSMC. Furthermore, HMGB 1-induced chemical chemotaxis could be inhibited not only by anti-HMGB 1 antibody, but also by anti-RAGE antibody, as shown in fig. 5-C. anti-RAGE antibodies block RSMC cytoskeletal reorganization and morphological changes in response to HMGB1 migration signals; unrelated antibodies were not able to block cytoskeletal reorganization.

These data indicate that RAGE receptors are required for HMGB 1-induced RSMC responses.

Many chemoattractants are known to act through membrane receptors associated with heterotrimeric GTP-binding proteins (G proteins), determining whether the G proteins are involved in the HMGB1 signaling pathway. Pertussis toxin was used because it inhibits a specific subclass of G proteins (Gi/o proteins), revealing their involvement in the signaling pathway (Baggiolini et al, 1994, adv. Immunol.55: 97-179; Haribabu et al, 1999, J.biol. chem., 274: 37087-. mPT (inactivated mutant of PT) was used as a control. RSMC was pretreated with PT or mPT (50ng/ml) for 6 hours and stimulated with HMGB1(100ng/ml), BoxA or BoxB (10ng/ml) for 30 minutes. Chemical chemotaxis assays were performed as described previously. Data represent mean ± SD, values 100% represent basal migration without any stimulus present. Fig. 6-a shows the inhibitory effect of PT on HMGB 1-induced chemical chemotaxis. These data show that the Gi/o protein is involved in the signaling pathway controlled by HMGB 1. FIG. 6-B shows cytoskeletal reorganization, with actin filaments seen as described previously. Then, it was investigated whether the signals induced by HMGB1 are involved in the MAP kinase pathway; indeed, these proteins are known to be activated by RAGE and have a direct role in the regulation of intracellular motor mechanisms. RSMC were pretreated with PD98059(50mM) for one hour or without pretreatment, stimulated with HMGB1(100ng/ml) from calf thymus for 30 minutes, stained with specific ERK 1/2-phosphorylating antibodies (New England Biolabs, Beverly, Mass.) and DAPI. To visualize cytoskeletal reorganization, individual cell samples were stained with TRITC-umbrella toxin. FIG. 7 shows how stimulation by HMGB1 induces activation of ERK1/2 proteins of RSMC and their nuclear transport within 30 minutes; in contrast, phosphorylated ERK protein is hardly detectable in unstimulated RSMC and is located in the cytoplasm. Furthermore PD98059 (a selective inhibitor of MEK, which is an upstream regulator of ERK) inhibits HMGB 1-induced ERK phosphorylation and nuclear transport, as well as RSMC migration and cytoskeletal reorganization. As a result, these data indicate that MAP kinase plays an important role in HMGB 1-induced cell migration.

Cell damage induction

In view of the state of the art, it was examined whether damaged cells or cells undergoing necrosis release HMGB1 in the extracellular medium.

HeLa cells and HUVEC induced necrosis with 5um ionomycin (Sigma) and 20um CCCP, or mM deoxyglucose and 10mM sodium azide. After 16 hours at 37 ℃, the number of cells undergoing necrosis was noted according to morphology, and when it reached 50%, the supernatant was collected.

For Western blot analysis, the culture medium of treated and untreated cells was collected and concentrated 50-fold using Amicon Ultrafree-MC filters; cells were lysed with SDS-PAGE sample buffer.

In immunofluorescence assays, cells were fixed with 4% PFA, incubated with anti-HMGB 1 antibody, stained with secondary antibody and DAPI, and permeabilized using 0.1% NP-40 in PBS.

FIG. 8-A shows Western blot analysis of supernatant (S) and cell pellet (P). HMGB1 was recovered from the supernatant of necrotic and injured cells. Figure 8-a shows immunofluorescence analysis performed on single live and necrotic Hela cells, with HMGB1 not bound to residual necrotic cells.

Fig. 9 shows the results of immunohistochemical analysis, and these data indicate that HMGB1 is contained in the nuclei of endothelial cells along human arteries, but not in the nuclei of RSMC (fig. 9-a is low magnification, fig. 9-B is high magnification), and in fact, most nuclei of smooth muscle cells contain an undetectable amount of HMGB1 (box of fig. 9-B). FIG. 9-C Western blot showing the expression level of HMGB1 in RSMC in comparison to Hela cells, indicating that RSMC in vitro culture contains a lower amount of HMGB1 than Hela cells.

Taken together, these data indicate that the HMGB1 molecule that signals vascular smooth muscle cells may only result from necrosis or mechanical injury of nearby cells.

In conclusion, the experimental data described above, which are the basis of the present application, demonstrate that HMGB1 nuclear protein is an important mediator of vascular remodeling following mechanical injury and/or inflammation, which can be passively released by injured or necrotic cells.

These data particularly indicate the following:

HMGB1 as a chemoattractant

Like bFGF or fMLP, HMGB1 is a strong chemoattractant in chemotaxis and wound assays, promoting cell shape changes and cytoskeletal tissue changes, similar to those observed with prourokinase; these uses are attributed specifically to HMGB1 and not to potential contaminants. In addition, antibodies to HMGB1 inhibited its effect on cell migration, whereas non-specific control antibodies did not.

Binding of RAGE in RSMC triggers the signaling pathway of HMGB1

The experiments reported above show that RAGE is expressed in RSMC and that anti-RAGE antibodies inhibit the effect of HMGB1 on RSMC.

It was confirmed that MAP kinase is involved in HMGB 1-induced migration of RSMC cells, and MEK inhibitor PD98059 is able to block HMGB 1-induced migration of cells, since ERK1/2 is phosphorylated and is transferred to the nucleus under stimulation by HMGB 1. The data also indicate that the Gi/o protein is involved in the process of activation by HMGB1, since HMGB 1-induced cell migration can be blocked by bordetella pertussis toxin. G proteins are usually associated to seven transmembrane elix receptors (7TM), but direct binding between RAGE and G proteins has not been described to date. To date, it is not known whether HMGB1, in addition to binding RAGE, needs to bind to the 7TM receptor/G protein receptor, or whether the G protein is involved downstream of RAGE, or in the feedback mechanism.

Paracrine function of HMGB1

HMGB1 is released in an unregulated manner, which means that cells are mechanically damaged or undergo necrosis under the stimulation of cytokines or lipopolysaccharides. Thus, HMGB1 can signal injury or destruction of a single cell to adjacent cells in a paracrine manner. Cells that respond to extracellular HMGB1 contain little HMGB1 by themselves and little within the nucleus. RSMC contain little HMGB1 compared to Hela cells or endothelial cells, and they contain little HMGB1 mostly localized in the cytoplasm. Migrating RSMC tend to concentrate HMGB1 on the surface of the cell front. It can be hypothesized that cells responding to HMGB1 contain little HMGB1 in order to reduce the chance of an inappropriate response to their own HMGB 1. HMGB1 was focused on the leading edge of migrating cells to induce HMGB 1-induced responses in neighboring cells: the relocation of molecules involved in cell migration (e.g. integrins, urokinase receptors or c-Src) is characteristic of the motor RSMC. Migration is also involved in the activation of extracellular proteases, and the interaction of HMGB1 with the plasminogen activation system can drive the migration of cells in the extracellular matrix.

Role of HMGB1 in vascular diseases

The responsiveness of smooth muscle cells to HMGB1, the observation that endothelial cells contain a large amount of HMGB1 and vascular SMC contain little, and that cells subjected to mechanical injury release HMGB1, all the above results indicate that HMGB1 may play a role in the tissue remodeling process that occurs in atherosclerosis and restenosis.

The above specific experimental results enable the identification of molecules inhibiting the interaction between HMGB1 and RAGE receptor, which is also the object of the present invention, which, considering their structural and functional characteristics, are classified as follows:

1. HMGB1 antagonist: HMGB1 fragment, HMG box analogs that are more potent than the intact full-length molecule, and proteins containing the HMG box domain, the latter two being able to bind to RAGE receptors.

2.HMGB1 inhibitor: as antibodies or antibody fragments and molecules directed back to DNA, they are able to bind to HMG box domain and avoid HMGB1 binding to RAGE.

These molecules are advantageously applied in pharmaceutical formulations to prevent, delay or minimize atherosclerosis and/or restenosis following vascular epithelial injury, including those arising following angioplasty.

Furthermore, the inventors of the present invention showed that HMGB1 has a strong biological effect on murine embryonic fibroblasts. Fibroblasts are known to be the major cellular component of connective tissue, and they contribute to the synthesis and maintenance of the attached extracellular matrix. More specifically, HMGB1 acts in vitro as a potent chemoattractant for fibroblasts, and anti-RAGE antibodies block this effect.

As a result, each molecule homologous to HMGB1 can be used as a full-length protein for the preparation of a pharmaceutical agent that can be upregulated to drive and/or induce cell migration of fibroblasts. Likewise, every molecule that blocks the interaction of HMGB1 with its RAGE receptor (i.e.all molecules belonging to the inhibitor group: antibodies or antibody fragments, four-way DNA; and all molecules belonging to the HMG box antagonist group: HMGB1 fragments, molecules containing the HMG box domain) can be effectively used in the production of pharmaceutical preparations in order to avoid, delay or reduce the regeneration of connective tissues.

Another object of the invention is the use of HMGB1, HMGB1 fragment corresponding to HMG box, HMG box domain of other proteins belonging to the HMG box family and other proteins of the HMG box family for the preparation of a therapeutic drug for promoting and/or inducing fibroblast migration to up-regulate connective tissue regeneration.

Part of the present invention is the use of all molecules, antagonists and/or inhibitors capable of inhibiting the interaction of HMGB1 and RAGE receptor for the preparation of a therapeutic agent for reducing, delaying and avoiding connective tissue regeneration, as indicated in the following experiments.

Chemotaxis assay of fibroblasts

Chemotaxis assays were performed using well-known protocols (Degryse et al, 1999, Blood, 94: 649-. A modified Boyden chamber (Corning, Acton, Mass.) with a 0.5um pore size filter was used, treated with collagen I (100ug/ml in 0.5M acetic acid) and fibronectin (10ug/ml) (Roche). Murine embryonic fibroblasts were cultured using well-known protocols (Calogero et al, 1999, nat. Genet., 22: 276-. Recombinant HMGB1 expressed in e.coli was diluted in the same serum-free medium and added to the lower well.

anti-RAGE antibody (1000ng/ml) (as given by Dr. A. M. Schimdt, Columbia university, NY) was added to both wells.

The cells migrated overnight at 37 ℃, then the cells stored on the upper surface of the filter were scraped off, the filter was fixed with methanol, and stained with 10% gentian violet in 20% methanol. All experiments were performed at least twice, in triplicate.

The results shown in figure 10 are the mean ± SD of the number of cells calculated at 10 high density zones per filter, expressed as a fold of the untreated control. An artificial value of 100% was given to random cell migration (i.e., migration in the absence of chemoattractant). Statistical analysis was performed using an ANOVA model to evaluate treatments with increasing doses of agents.

Coli expressed recombinant HMGB1 stimulated fibroblast migration in a concentration-dependent manner, starting at a dose of 0.1ng/ml, with a maximal response at 100ng/ml and a lower response at higher doses (1000ng/ml) than the control. The complete blocking of the migration response by the anti-RAGE antibody (1000ng/ml) (right side of the diagram in fig. 10) shows that the above phenomenon is specifically attributed to HMGB 1.

Role of HMGB1 in the regulation of connective tissue regeneration

The responsiveness of fibroblasts to HMGB1 indicates that HMGB1 may play a role in the remodeling of connective tissue that occurs following injury due to trauma or surgery. Furthermore, the fact that anti-RAGE antibodies block the response suggests that the interaction of HMGB1 and RAGE receptor at the cell surface is the basis for the sensitivity of fibroblasts to HMGB 1.

And (4) conclusion:

HMGB1 and/or fragments of HMGB1 corresponding to HMG box, the HMG box domain of other proteins belonging to the HMG box family and other proteins of the HMG box family can be advantageously used in pharmaceutical formulations which are positively regulating (i.e. promoting and/or inducing connective tissue regeneration).

Each molecule capable of inhibiting the interaction between HMGB1 and RAGE and belonging to the group of antagonists (capable of binding to RAGE receptor) and to the group of inhibitors (capable of binding to HMG box domain thereby blocking HMGB1 binding to RAGE receptor) may be advantageously used in pharmaceutical formulations for negative regulation (i.e. blocking, delaying or reducing connective tissue regeneration).

Claims (4)

1. Use of an HMG box-binding molecule in the manufacture of a therapeutic agent for the treatment of a vascular disease, excluding diabetes-induced consequences, wherein said molecule is an antibody to HMGB 1.
2. Use of HMGB1 and/or the fragment of HMGB1 corresponding to HMG box, the HMG box domain of other proteins belonging to the HMG box family and of other proteins of the HMG box family for the preparation of a therapeutic agent capable of promoting the restructuring of connective tissue.
3. 3. Use according to claim 2, in the preparation of a therapeutic agent capable of inducing the healing of wounds and/or burns and/or bedsores.
4. Use of an HMG box-binding molecule in the manufacture of a therapeutic agent capable of blocking, delaying or reducing connective tissue regeneration, wherein said molecule is an antibody to HMGB 1.