WO2003020301A1

WO2003020301A1 - Chromogranin a and fragments thereof for the treatment of diseases involving increased vascular permeability

Info

Publication number: WO2003020301A1
Application number: PCT/EP2002/008749
Authority: WO
Inventors: Angelo Corti; Elisabetta Ferrero
Original assignee: Ospedale San Raffaele SRL
Current assignee: Ospedale San Raffaele SRL
Priority date: 2001-08-31
Filing date: 2002-08-06
Publication date: 2003-03-13
Anticipated expiration: 2004-02-29
Also published as: ITMI20011833A1; ITMI20011833A0

Abstract

The use of chromogranin A or fragments thereof for the treatment of diseases related to vascular leakage, such as shock, acute ancute and chronic inflammatory diseases, edema, particulary pulmonary edema, rheumatiod arthritis, congestive heart failure, left ventricular dysfunction and related tissue damages, is herein described.

Description

C_HROM_OG_RANIN A AND FRAGMENTS THEREOF FOR THE TREATMENT OF DISEASES INVOLVING INCREASED VASCULAR PERMEABILITY

The present invention concerns the use of chromogranin A or fragments thereof for the treatment of diseases related to vascular leakage.

Chromogranin A (CgA) is a 439-aminoacids, low isoelectric point acidic protein, characterised by various post-translational modifications, 5 such as glycosylation, sulphatation, phosphorylation, stored in secretory granules of several endocrine and neuroendocrine tissues (Winkler, H. et al., Neuroscience 49, 497-528 (1992) e Rosa, P. et al., J. Endocr. Invest 17, 207-225 (1994). Elevated levels of circulating CgA have been detected in the blood of patients with endocrine and neuroendocrine tumours, kidney 0 failure and heart failure.

The biological role of CgA is still unclear. CgA has been suggested to be involved in the storage of hormones within secretory granules and to be a precursor of biologically active peptides with endocrine, paracrine, and autocrine functions. It has been observed that N-terminal fragments 5 corresponding to aminoacids 1-76 and 1-1 13, respectively named vasostatin I and II, can suppress vasoconstriction in isolated blood vessels. Moreover, the fragment corresponding to residues 1-78 can induce adhesion of fibroblasts and smooth muscle cells on solid-phases, suggesting a role for CgA in modulating cell adhesion. 0 During investigations on the effects of CgA on vascular endothelium, it has surprisingly been found that CgA protein and peptides derived therefrom can significantly reduce vascular leakage induced by agents that alter endothelial barrier function and are involved in various diseases related to increased vascular leakage and edema. It is well known that TNF-α, 5 thrombin, oxidants, bradykinin, histamine and similar substances induce cytoskeletal rearrangements and modifications of adhesion mechanisms in endothelial cells. It has now been found that CgA and fragments thereof can reduce endothelial vascular leakage induced by said active agents, in particular TNF-α, VEGF and thrombin.

The study was carried out in in vitro and in vivo models. As in vivo model the labelled albumin entry within the extra vascular compartment of the liver was measured in mice after administration of the agent that increase vascular leakage; a study on TNF-induced lethal inflammatory response was also carried out in mice. As in vitro model a human umbilical vein endothelial cell monolayer (HUVECs) in a permeability assay in double chamber was used.

A remarkable reduction of induced vascular leakage was observed both with full-length CgA and with fragments, in particular with the fragments corresponding to residues 1-115, 1-78 and 7-57. Structure- activity relationships studies show that region 7-57 contains the active site. The disulfide bridge (Cys-Cys) present in this region is apparently not necessary for the interaction with endothelial cells, since the synthetic peptide CgA₇.₅₇. _SEtM, wherein the two Cys residues are reduced and alkylated, show the same activity of the unmodified product. Previous works demonstrate that the region 47-57 contains the active site for the adhesion of fibroblasts to solid surfaces (Ratti et al J. Biol. Chem. 275 (2000), 29257-29263). The importance of the region 47-57 for the biological activity of CgA is also reflected by the high degree of homology among different animal species, compared with other regions.

Various evidences suggest that the mechanism determining the leakage inhibition observed with CgA or fragments thereof consists in their interaction with a component of the endothelial cellular membrane and in the subsequent inhibition of cytoskeletal rearrangement. Previous observations show that agents that can inhibit actin depolimerisation can also inhibit leakage induced by other agonists (Lum, H. et al., Am. J. Physiol. 267 (1994), 223-241). While the invention should not be bound to any particular mechanism of action, it is likely that CgA and fragments thereof protect vessels by preventing shape modifications and contraction of endothelial cells, irrespectively of the particular stimulus that increases vascular leakage.

Accordingly, object of the present invention is the use of CgA or a fragment thereof for the preparation of a medicament useful for the treatment of diseases involving increased vascular leakage. The term "active fragment" means a portion of CgA endowed with protection activity against increased leakage, preferably the portion containing the amino acid sequence 47-57 (sequence numbers according to Koneki et al. J.Biol. Chem. 262, 17026-17030, 1993), more preferably the sequence 7-57. Preferred fragments are contained in the region comprised between residues 1-115, preferably between residues 1-78.

CgA can be prepared by purification from a tissutal or cellular extract, or can be produced by recombinant DNA techniques in eucariotic or procariotic cells (see Gasparri et al J. Biol. Chem. 272, 20835-20843, 1997, and Corti et al. Eur J. Biochem. 248, 692-699, 1997, herein fully incorporated by reference; as far as recombinant DNA techniques are concerned, see also Sambrook et al., Molecular cloning. A Laboratory Manual, Cold Spring Harbor Press, Cold Spring Harbor, New York (1982)). The fragments can be prepared by synthesis (see for instance Stewart and Young, Solid Phase Peptide Synthesis, 2^nd ed., Pierce Chemical Co. (1984)), by recombinant DNA techniques or by fragmentation or enzymatic digestion of the native protein. Moreover, the aminoacidic residues of the native protein or of the fragments can be modified so as to improve the pharmacokinetic properties of the molecules for therapeutical use, without varying the specific biological activity. For instance, single residues can be replaced by other L- or D-aminoacids or can be chemically modified by glycosylation, coupling to groups with different polarity or lipophilicity, amidation or esterification of carboxy groups, conjugation with other molecules or peptides.

Diseases that may benefit from the treatment with CgA or fragments thereof are those whose symptomatology or etiopathogenesis involve endothelial barrier alteration, and increased vascular leakage and/or leucocyte extravasation. Without limiting in any way the invention, the pathologies involving increased vascular leakage include shock, acute and chronic inflammatory diseases, edema, particularly pulmonary edema, rheumatoid arthritis, congestive hearth failure, left ventricular dysfunction, and related tissutal damage.

According to a preferred embodiment, chromogranin or fragments thereof are used for reducing TNFα-induced vascular leakage increase, which is associated to different pathologic conditions such as acute respiratory distress syndrome (SDRA) in adults, systemic inflammatory response syndrome, chronic or congestive hearth failure, rheumatoid arthritis and multiple sclerosis. In addition, CgA and fragments thereof can be used in angiogenic therapy, particularly as coadjuvants of gene therapy to reduce vascular leakage induced by VEGF or FGF.

For therapeutical use, CgA and fragments thereof can be formulated with pharmaceutically acceptable carriers and excipients, such as buffers, stabilizers, dissolution agents, preservatives, dispersants, surfactants, disintegrants, lubrificants, thickening agents, fillers, dyes and the like. Pharmaceutical compositions can be administered through oral, intravenous, subcutaneous, intramuscular, transdermal, nasal, sublingual and topical route. Principles and methods for the preparation of pharmaceutical compositions are known to the expert in the art and are described for instance in Remington: The Science and Practice of Pharmacy, Lippincott, Williams and Wilkins Eds, Dec. 2000. The dosage of CgA or fragments thereof varies according to the kind and severity of the disease and the general conditions of the patient. In general an amount of active principle comprised between 0.03 and 3 mg/kg, preferably 0.1-0.3 mg/kg, can be administered once or twice a day.

The following experimental results illustrate the invention in greater detail.

MATERIALS AND METHODS

Cell cultures and reagents.

HUVECs were isolated from human umbilical veins by collagenase treatment as described (Ferrero, E. et al. FEBS Lett. 374, (1995), 323-326) and cultured in 1% gelatin coated flasks (Falcon, Becton-Dickinson,

Bedford, MA) containing endotoxin-free Medium 199 (Sigma, S. Louis,

MO), 20% heat-inactivated fetal calf serum (Hyclone, Logan, UT), 1% bovine retina derived growth factor, 90 μg/ml heparin, 100 IU/ml penicillin,

100 μg/ml streptomycin (Biochrom, Berlin, Germany) (complete medium, CM). All experiments were carried out with HUVECs at passage 2-4. Mouse anti-CgA mAb 5A8 (IgGl) was described previously (Ratti, S. et al. J. Biol.

Chem. 275, (2000), 29257-29263). Human TNF (5.45 x 10⁷) + 3.1 units/mg and murine TNF (1.2 ± 0.14) x 10⁸ units/mg were prepared by recombinant

DNA techniques as described (Curnis, F. et al. Nat. Biotechnol. 18, (2000), 1185- 1190).

Production of natural CgA and recombinant CgA N-terminal fragments.

Natural human CgA was purified from pheochromocytoma tissue extracts (heat stable fraction) by immunoaffinity chromatography, essentially as previously described (Corti, A. et al. Eur. J. Biochem. 235, (1996), 275-280), using a column bearing the anti-CgA mAb 5A8. SDS- PAGE of the final product revealed two bands of 70 and 60 kDa. To investigate the structure-activity relationships of CgA various recombinant and synthetic fragments were prepared. Each fragment is herein indicated with its sequence numbers (Konecki, D.S. et al. J. Biol. Chem. 262, (1987), 17026-17030). Recombinant

were obtained by expression in E.coli (Corti, A. et al. Eur. J. Biochem. 248, (1997), 692-699) and purified by reverse-phase HPLC using a SOURCE 15 RPC column (Pharmacia-Upjohn, Uppsala, Sweden) followed by gelfiltration chromatography on a Sephacryl S-200 HR column, as previously described (Ratti, S. et al. J. Biol. Chem. 275, (2000), 29257-29263) SDS-PAGE of and

showed that both products were homogeneous under reducing and non-reducing conditions. The molecular mass of CgA,.₇₈ and CgA,._U5, as measured by electrospray mass spectrometry, were 9069.7 and 13247.4 Da, respectively (expected 9069.3 and 13247.5 Da). Endotoxin content was 0.008-0.016 units/μg by the Limulus Amoebocyte Lysate (LAL) Pyrotest (Difco Laboratories, Detroit, MI). Peptide 7-57 with cysteines reduced and alkylated with N-ethylmaleimide (CgA₇.₅₇._SEtM) was prepared as described (Ratti, S. et al. J. Biol. Chem. 275, (2000), 29257-29263). The molecular mass of CgA_7-57._SEtM was 6038 Da (expected 6039 Da). In vitro permeability assay. The assay was carried out by measuring the flux of radiolabeled albumin through HUVEC monolayers, cultured on gelatin-coated membranes of Transwell cell culture chambers (0.4 μm filters, Costar). HUVECs (5xl0⁴ cells/well, in CM) were cultured for 5 days in the upper compartment of the Transwell device. The culture medium was then replaced with human TNF (4 ng/ml) or thrombin (2 U/ml) solutions, either in the absence or in the presence of CgA or N-terminal fragments, in medium 199 containing 2 mM glutamine, 100 IU/ml penicillin, and 100 μg/ml streptomycin (200 μl/well). After 1-2 h a mixture of 4 μCi/ml ¹²⁵I- labeled bovine serum albumin (NEN-Boston, MA) and 1.5 mg/ml human serum albumin (Farma-Biagini SpA, Lucca, Italy), was added to the upper compartment (50 μl/well). The lower compartment was then filled with medium 199 containing 2 mM glutamine, 100 IU/ml penicillin, and 100 μg/ml streptomycin (600μl/well) and incubated for 1 hour with gentle agitation. At various times, samples (30 μl) were taken from the lower compartment and their radioactivity was measured using a γ-counter (Packard, Sterling, VA). Each experiment was carried out in triplicate and the results were expressed as mean + SD. In vivo permeability assay. The effect of murine TNF on the leakage of trypan blue-albumin complex from liver vessels was evaluated in anaesthetized male BALB/c mice (Charles River Italia, Calco, Italy) as described (Ferrero, E. et al. Cancer Res. 56, (1996), 3211-3215). Murine TNF (2 ng) in 0.2 ml 0.9% sodium chloride was injected intraperitoneally. Other animals were treated with a mixture of TNF (2 ng) and CgA (1 μg), with or without mAb 5A8 (20 μg). After 30 min, the liver of each animal was perfused for 5 min with 4 ml of a solution containing 0.4% bovine serum albumin, 0.5% trypan blue, 0.85% sodium chloride through the inferior vena cava. The perfusate was drained from the portal vein, as described (Branster, M.V. & Morton, R.K. Nature (London) 180, (1957), 1283-1284). Then, 5 ml 0.9% sodium chloride were injected trough the inferior vena cava, to remove the dye present in the lumen of vessels. The livers were dissected, homogenized with 3 ml PBS, and centrifuged. The absorbance at 540 nm of the supernatants was then measured using a Pye Unicam SP-550 spectrophotometer (Cambridge, United Kingdom).

Confocal microscopy.

HUVECs (1 x 10⁵ cells in CM) were grown on glass coverslips for five days and incubated for 2 h with 4 ng/ml TNF alone, or in combination with

CgA,_₇₈ or CgA, in CM. The cells were then washed, fixed with 2% paraformaldehyde for 15 min and stained with fluorescein isothiocyanate- conjugated goat anti-mouse immunoglobulins (Zymed Laboratories, Inc.,

South San Francisco, CA). The cells were then permeabilized with 0.01% Triton X-100 in PBS and incubated with fluorescein tetraisothiocyanate phalloidin (Sigma) to stain F-actin. Coverslips were mounted using 50% glycerol in PBS. Microscopic analysis was performed using a Bio-Rad MRC

1000 confocal scanning microscope (Biorad Laboratories, Milan, Italy) and fluorescence images were recorded on Kodak T-Max 100 film using a Focus Image recorder Plus (Focus Graphics, Foster City, CA).

TNF cytotoxicity assay.

The cytotoxic activity of TNF in the presence or absence of CgA was estimated by standard cytolytic assays using L-M mouse fibroblasts (ATCC CCL1.2) as described (Corti, A. et al. J. Immunol. Meth. 177, (1994), 191- 198).

RESULTS

CgA and its N-terminal fragments (CgAj.₇₈ and CgAι__ns) protect vessels against TNF-induced plasma leakage in mice

To assess whether CgA can inhibit TNF-induced vascular leakage, the effect of murine TNF on the leakage of trypan blue-albumin from liver vessels in BALB/c mice was measured. This assay is based on the measurement of trypan blue-albumin entry within the extra vascular compartment of the liver after treatment with TNF, as a measure of vessel permeability to macromolecules (Ferrero, E. et al. Cancer Res. 56, (1996), 321 1-3215). Three separate experiments in different conditions were performed (Fig. 1). As expected, TNF induced marked dye entry into liver parenchyma in the absence of inhibitors, whereas CgA alone did not affect dye entry (Fig. la and c). Administration of 0.4 μg or 1 μg of natural CgA significantly inhibited TNF-induced dye entry (Fig. la and c). This effect was specific, as co-administration of the anti-CgA monoclonal antibody (mAb 5A8), but not of a control antibody (mAb 19E12), partially neutralized the activity of CgA (Fig. la). Furthermore, injection of the recombinant CgA fragment

(3 μg) inhibited the effect of TNF (Fig. lb) indicating that both CgA and its N-terminal fragments can inhibit the TNF-induced leakage of macromolecules from liver vessels.

CgA and its N-terminal fragments inhibit TNF-induced permeability of cultured HUVEC monolayers The mechanism of action was then investigated. To assess whether the effect of CgA on TNF-induced vascular leakage was related to regulation of endothelial function we measured the activity of natural CgA and human TNF on the flux of radiolabeled albumin through confluent HUVECs monolayers, using the double chamber permeability assay. Natural CgA (3 μg/ml) did not affect the transendothelial flux of ¹²⁵I-albumin when added alone (Fig. 2a). In contrast, TNF (4 ng/ml) significantly increased the permeability of HUVECs (Fig. 2b). The effect of TNF was inhibited by CgA in a dose dependent manner (Fig. 2b). To assess the specificity of these effects we measured the activity of CgA in the presence of the anti-CgA mAb 5A8. A mixture of 3 μg/ml CgA and 30 μg/ml 5A8 behaved like 0.3 μg/ml CgA (Fig. 2c), indicating this antibody could block most of the CgA activity. Since mAb 5A8 is directed to an epitope located in the N-terminal domain of CgA (Corti, A. et al. Eur. J. Biochem. 248, (1997), 692-699), these results suggest that the N-terminal region is critical for activity.

To provide further information on the location of the active site of CgA, we measured the effects of recombinant CgA^, ,^ CgA₁.₇₈ and synthetic peptide CgA₇.₅₇._SEtM on the in vitro TNF-induced permeability. Both recombinant and synthetic fragments inhibited the TNF-induced permeability (Fig. 3c and d) suggesting that the region 7-57 contains the active site. However, dose-response experiments showed that natural CgA was at least 10 times more potent (on a molar basis) than recombinant CgA,.₇₈ (Fig. 2b and Fig. 3 ). Taken together, these results indicate that the N-terminal domain of

CgA contains a site that affects endothelial cell function in vitro and in vivo. These effects were observed with HUVECs prepared from several different donors.

CgA does not block the TNF-TNF receptor interactions and inhibits the permeability of HUVEC monolayers induced with thrombin

It has previously been shown that the interaction of TNF with the p55-

TNF-receptor is necessary and sufficient for triggering an increase in vascular leakage in vivo and HUVEC monolayer permeability in vitro

(Ferrero, E. et al. Dominant role of tumour necrosis factor (TNF) p55- receptor in TNFalpha-induced vascular permeability. Am. J. Physiol. (2001)281 :C1 173-C1179, 2001). Thus, it was investigated whether CgA could bind TNF or TNF-Rs and consequently inhibit the TNF/TNF-Rs interactions. No binding of human TNF to CgA, immobilized onto microtiter plates, was observed by ELISA, using anti-TNF polyclonal IgGs as detecting reagent (not shown). Moreover, neither CgA nor CgA,.^ inhibited the cytolytic activity of human TNF (a selective murine p55-TNF agonist) or murine TNF (a p55- and p75-TNF-receptor agonist) against mouse L-M cells in a standard cytotoxicity assay. Since the cytolytic activity of murine TNF on these cells is dependent on both the p55 and p75 receptors (Pelagi, M. et al. Eur. Cytokine Net. 11, (2000), 580-588), these results imply that CgA does not inhibit the binding of TNF to both membrane receptors. To confirm this hypothesis and to assess whether CgA can also inhibit the activity of other inflammatory agents independent of the TNF/TNF receptor system, the effect of natural CgA on the thrombin-induced permeability of HUVECs was measured. As shown in Fig. 4, 3 μg/ml CgA inhibited the increase of cell permeability caused by thrombin. CgA and CgA,_₇₈ inhibit endothelial cytoskeleton reorganization

TNF-α and thrombin are known to promote changes in the size and shape of endothelial cells by rearranging the F-actin cytoskeleton and converting peripheral actin bundles into stress fibers (Goldblum, S.E. et al. Am. J. Physiol. 264, (1993), C894-905). Since this phenomenon is believed to correlate with an increase in endothelial permeability to macromolecules, we investigated the effect of CgA^g and CgA on TNF-induced cytoskeleton rearrangement of HUVECs by confocal microscopy. As expected, human TNF induced marked actin reorganization (Fig. 5a and b). When 3 μg/ml CgA^-₇₈ was added to the cultures actin reorganization was inhibited, peripheral bundles being observed both in the presence and absence of TNF (Fig. 5c and d). The inhibition was reverted by the anti-CgA mAb 5A8, as indicated by stress fibers formation in the presence of a mixture of antibody, and TNF (Fig. 5e). Similarly, 3 μg/ml CgA inhibited TNF-induced stress fiber formation (not shown). These results indicate that CgA^s and CgA can inhibit the TNF-induced morphological changes associated with disruption of endothelial barrier function.

CgAχ.₇₈ increases the survival of C57BL6 mice at lethal doses of TNF and galactosamine. To verify the ability of CgA].₇₈ to protect the animals from the lethal effect of TNF, some C57BL6 mice were treated with 150 ng of mTNF and 36 mg of galactosamine, with or without pre-treatment with CgA_1-78 (3 μg). In this model TNF can trigger a strong inflammatory response, leukocyte transmigration through the endothelial barrier and fulminating hepatitis within 24 h. The results, reported in Table 1 , show that CgA,.₇₈ increases mice survival.

The animals treated with TNF and galactosamine, but not those pre- treated with CgA,_₇₈, showed suffering signs some hours after treatment (hair erection, difficulty in breathing, lethargy, scarce reaction to stimuli).

^a) Each compound was administered through intraperitoneal route (100 μl/mice). The animals were treated three times with CgA₁.₇₈: 1 h before mTNF and galactosamine administration, 5 min and 1 h after treatment. Galactosamine (Sigma) was diluted with 0.9% sodium chloride; mTNF and CgA₁.₇₈ were diluted with 0.9% sodium chloride containing 100 μg/ml human serum albumin.

CgA and its fragments inhibit VEGF-induced permeability of cultured HUVEC monolayers The experiment was carried out using HUVEC monolayers as described in "Materials and Methods". The cell monolayers were preincubated with peptide 7-57 (3 μg/ml, 15 min) and peptide 7-57 (3 μg/ml) plus VEGF (lOng/ml) for 2 h before ¹²⁵I-BSA. Control wells were incubated with medium alone. The results are reported in the following Table. Table 2. Effect of peptide CgA 7-57 on VEGF-induced endothelial cell permeability

Treatment ¹²⁵I-BSA flux (cpm)^b) P value

A) None 8198±720

B) VEGF 10480±552 p-0.036 (B vs A) C) VEGF+CgA7-57 (3μg/ml) 8393±336 p=0.012 (B vs C)

b) Mean ± SEM of 5 wells DESCRIPTION OF THE FIGURES Figure 1. Effect of natural CgA (α) and recombinant CgA N-terminal fragments (b) on TNF-induced leakage of trypan blue-albumin from liver vessels of mice. Animals were treated with the reagents indicated above each panel, 30 min before liver perfusion with trypan-blue albumin as described in Materials and Methods. Three separate experiments were carried out. In "Experiment 1" (upper panel, a) and "Experiment 3" (b) each compound was injected sequentially. In Experiment 2 (lower panel, a) the reagents were pre-mixed and incubated 30 min before injection. The following doses of each reagent were injected intraperitoneally to mice: TNF (2 ng); mAb 5A8 (20 μg); mAb 19E12 (20 μg); CgA,_₇₈ (3 μg); CgA_1-U5 (3 μg); CgA (0.4 μg, upper panel; 1 μg, lower panel, a). Each bar represents the mean + SD of three mice. <0.05 (*); PO.005 (**), by unpaired t-test, two-tailed.

Figure 2. Effect of CgA on human TNF-induced permeability of cultured HUVEC monolayers. The assay was carried out with cells untreated (a) or treated (b and c) with 4 ng/ml human TNF. The effect of CgA and TNF was measured also in the presence of the anti-CgA mAb 5A8 (30 μg/ml) (c). The (Y) axis reports the ¹²⁵I-albumin present in the lower chamber of the Transwell systems after 1 h, expressed as % of control (cells treated with TNF alone). Dashed bars represent the basal permeability observed in the absence of TNF; black bars represent the TNF-induced permeability. (*) O.05; (**) P< 0.01 (cells treated with TNF vs CgA/TNF mixtures). Figure 3. Effect of CgA N-terminal fragments on human TNF-induced permeability of cultured HUVEC monolayers. The assay was carried out with cells treated with culture medium alone (a) or medium containing 3 μg/ml of CgA (b), CgA_{M 15} (c), CgA₇.₅₇._SEtM (d) or

(e, j) in the absence (o) or in the presence (•) of 4 ng/ml human TNF. The radioactivity present in the lower chamber of the Transwell systems was measured at the indicated times. Panel /: effect of different doses of CgA₁.₇₈ on cell permeability, as measured after 2 h (see Fig. 2 for bar explanation).

Figure 4. Effect of CgA on thrombin-induced permeability of cultured HUVEC monolayers. Untreated cells (Δ); cells treated with 2 U/ml thrombin (o); cells treated with 2 U/ml thrombin and 3 μg/ml CgA (•).

Claims

1. CgA or fragments thereof for use as a medicament.

2. CgA or fragments thereof according to claim 1 , for use in the treatment of diseases involving increased vascular leakage.

3. CgA or fragments thereof according to claims 1-2, wherein said fragments contain sequence 47-57 of CgA.

4. CgA or fragments thereof according to claim 3, wherein said fragments contain sequence 7-57 of CgA.

5. CgA or fragments thereof according to claims 1-4 wherein the diseases involving an increased vascular leakage are shock, acute and chronic inflammatory diseases, edema, particularly pulmonary edema, rheumatoid arthritis, congestive hearth failure, left ventricular dysfunction, and related tissue damages.

6. CgA or fragments thereof according to claims 1-4, wherein said increased vascular leakage is TNFα-induced.

7. CgA or fragments thereof according to claims 1-4, wherein said increase vascular leakage is VEGF-induced.

8. CgA or fragments thereof according to claims 1-4, for use in angiogenic therapy to reduce the vascular leakage induced by VEGF or

FGF.

9. Pharmaceutical composition containing CgA of fragments thereof as the active ingredient.

10. Pharmaceutical composition according to claim 7, wherein CgA fragments are as defined in claims 3 or 4.