• Roles of histamine in the production of vascular endothelial growth factor (VEGF) in the carrageenin-induced granulation tissue in rats were analysed in vitro and in vivo.
• Incubation of the minced granulation tissue in the presence of histamine (1 and 10μM) increased the content of VEGF protein in the conditioned medium in a time- and concentration-dependent manner. The levels of VEGF mRNA in the minced granulation tissue were also increased by histamine in a concentration-dependent manner.
• The increase in the content of VEGF protein in the conditioned medium by histamine (10μM) was suppressed by the H₂ receptor antagonist cimetidine (IC₅₀ 0.37μM), but not by the H₁ receptor antagonist pyrilamine maleate, the H₃ receptor antagonist thioperamide or the cyclo-oxygenase inhibitor indomethacin.
• The histamine-induced increase in the content of VEGF protein in the conditioned medium was inhibited by the cyclic AMP antagonist Rp-cAMP (IC₅₀ 6.8μM), and the protein kinase A inhibitor H-89 (IC₅₀ 12.5μM), but not by the protein kinase C inhibitors Ro 31-8425 and calphostin C or the tyrosine kinase inhibitor genistein.
• Simultaneous injection of cimetidine (400μg) and indomethacin (100μg) into the air pouch of rats additively reduced the carrageenin-induced increase in VEGF protein levels and angiogenesis in the granulation tissue as assessed by using carmine dye.
• These findings indicate that histamine has an activity to induce VEGF production in the granulation tissue via the H₂ receptor-cyclicAMP-protein kinase A pathway and augments angiogenesis in the granulation tissue.

The rats were treated in accordance with procedures approved by the Animal Ethics Committee of the Graduate School of Pharmaceutical Sciences, Tohoku University, Japan. Air pouch-type inflammation was induced by carrageenin in male Sprague-Dawley rats, specific pathogen-free, weighing 160–170g (Charles River Japan, Inc., Kanagawa, Japan) according to the procedure described by Tsurufuji et al. (1978). Eight milliliters of air was injected subcutaneously into the back to make an air pouch oval in shape. Twenty-four hours later, 4ml of a 2% (w v⁻¹) solution of carrageenin (Seakem No. 202, Marine Colloids Inc., Springfield, NJ, U.S.A.) in saline was injected into the air pouch. The carrageenin solution had been sterilized by autoclaving at 121°C for 15min and supplemented with antibiotics (0.1mg of penicillin G potassium (Meiji Seika, Tokyo, Japan) and 0.1mg of dihydrostreptomycin sulphate (Meiji Seika) per ml of the solution) after cooling to 40–45°C.

Indomethacin (Sigma Chemical Co., St. Louis, MO, U.S.A.) was dissolved in 99.5% of ethanol and diluted with saline. Five-hundred microliters of the solution containing 0.1mg of indomethacin was injected into the pouch of each rat just after and 2 days after the carrageenin injection to minimize the effect of endogeous PGE₂. Three days after injection of the carrageenin solution, the rats were sacrificed by cutting the carotid artery and the granulation tissue was excised and minced into 1–2-mm pieces with a pair of small scissors. The minced granulation tissue was washed twice with 5 volumes of ice-cold phosphate-buffer saline (PBS) (mM): NaCl 137, Na₂HPO₄ 8.1, KCl 2.68, KH₂PO₄ 1.47, (pH 7.4). Four-hundred milligrams of the minced tissue was placed in a 60×15-mm tissue culture dish (Corning Costar, Cambridge, MA, U.S.A.) and incubated in 4ml of Eagle's minimal essential medium (EMEM) (Nissui Pharmaceutical Co., Ltd., Tokyo, Japan) supplemented with 10% (v v⁻¹) calf serum (Dainippon Pharmaceutical, Osaka, Japan), penicillin G potassium (18μgml⁻¹) and streptomycin sulphate (50μgml⁻¹) for 3h at 37°C under an atmosphere of 95% air and 5% CO₂. The tissues were then washed three times with calf serum-free medium and incubated at 37°C for the periods indicated in 4ml of the medium containing various concentrations of histamine (0.1–10μM, Sigma Chemical Co.) in the presence or absence of drugs. After incubation, the minced granulation tissue and the conditioned medium were collected, and centrifuged at 220×g and 4°C for 5min. The supernatant fraction of the conditioned medium was stored at −40°C. The minced granulation tissue was homogenized in 0.5mM sodium hydroxide using the Vir-Tis 45 homogenizer (Virtis Company, Gardiner, NY, U.S.A.) for 4min at scale 40 on an ice bed. The homogenates were centrifuged at 14,000×g and 4°C for 30min and the supernatant fractions were stored at −40°C.

A solution containing soluble starch (Wako Pure Chemicals, Osaka, Japan) and Bacto peptone (Difco Laboratories, Detroit, MI, U.S.A.) (5% each) was injected i.p. into male Sprague-Dawley rats (400–600g, specific pathogen-free, Charles River Japan, Kanagawa, Japan) at a dose of 5ml per 100g body weight. Four days after the injection, the rats were sacrificed and peritoneal cells were harvested according to the procedure described by Ohuchi et al. (1985). The peritoneal cells were suspended in EMEM (Nissui Pharmaceutical Co.) supplemented with 10% (v v⁻¹) calf serum (Dainippon Pharmaceutical), penicillin G potassium (18μgml⁻¹) and streptomycin sulphate (50μgml⁻¹). The peritoneal macrophages (1.5×10⁷ cells) were plated in a 100-mm tissue culture dish (Corning Costar) and incubated at 37°C for 2h in 10ml medium. The dishes were washed three times to remove non-adherent cells and the adherent cells were further incubated for 20h at 37°C, after which the cells were washed and incubated at 37°C for the periods indicated in 10ml of EMEM supplemented with 10% (v v⁻¹) calf serum containing various concentrations of histamine (0.1–10μM).

For isolation of fibroblasts, the granulation tissue dissected 3 days after the injection of a 2% (w v⁻¹) solution of carrageenin was minced into 1–2-mm pieces, and 1g of the minced granulation tissue was incubated in a 100-mm tissue culture dish (Corning Costar) in 10ml of Dulbecco's modified Eagle's medium (DMEM) (Nissui Pharmaceutical Co.) supplemented with 10% (v v⁻¹) foetal bovine serum (Dainippon Pharmaceutical), penicillin G potassium (18μgml⁻¹) and streptomycin sulphate (50μgml⁻¹) at 37°C for 8 days under an atmosphere of 95% air and 5% CO₂. The medium was changed each 48h interval and the minced granulation tissues were removed from the dish. The adherant fibroblasts were washed three times with PBS and detached from the dish by the addition of 0.05% trypsin and 0.02mM EDTA in PBS. After washing, fibroblasts (1.5×10⁷ cells) were incubated in a 100-mm tissue culture dish (Corning Costar) at 37°C for 2h in 10ml of DMEM containing 10% (v v⁻¹) foetal bovine serum. The cells were then washed three times with foetal bovine serum-free DMEM and incubated at 37°C for the periods indicated in 10ml of DMEM supplemented with 10% (v v⁻¹) foetal bovine serum containing various concentrations of histamine (0.1 to 10μM).

Drugs used were pyrilamine maleate (Sigma Chemical Co.), cimetidine (Sigma Chemical Co.), thioperamide (a gift from Dr J.C. Schwartz at Unite de Neurobiologie et Pharmacologie Moleculaire (U.109) de l'INSERM, Paris, France), indomethacin (Sigma Chemical Co.), Rp-cAMPs triethylamine salt (Research Biochemicals International, Natick, MA, U.S.A.), Ro 31-8425 (Amersham Pharmacia Biotech Co., Piscataway, NJ, U.S.A.), calphostin C (Kyowa Medex, Tokyo, Japan), H-89 (Biomol Research Lab., Polymouth Meeting, PA, U.S.A.), genistein (Wako Pure Chemical Ind.), brefeldin A (Sigma Chemical Co.), forskolin (Seikagaku Co., Tokyo, Japan), histamine (Sigma Chemical Co.), and PGE₂ (Sigma Chemical Co.). Histamine, pyrilamine maleate, cimetidine, thioperamide, and Rp-cAMPs triethylamine salt were dissolved in EMEM. PGE₂, brefeldin A, forskolin and indomethacin were dissolved in 99.5% of ethanol. Ro 31-8425, calphostin C, H-89, and genistein were dissolved in dimethyl sulphoxide. Final concentration of ethanol and dimethyl sulphoxide in medium was adjusted to 0.1% (v v⁻¹). The control medium contained the same amount of ethanol and dimethyl sulphoxide.

Protein levels in the supernatant fractions were determined according to the method described by Bradford (1976). Proteins, at 100μg aliquot for the conditioned medium, 4.6μg aliquot for the granulation tissue and 1.4μg aliquot for the pouch fluid, were separated by electrophoresis on a 12% (w v⁻¹) sodium dodecylsulphate-polyacrylamide gel and transferred onto a nitrocellulose membrane (Schleicher and Schuell Inc., Dassel, Germany). The membrane was incubated at 4°C for 12h with mouse monoclonal anti-VEGF (1:200, Santa Cruz Biotechnology, Santa Cruz, CA, U.S.A.). It was then incubated at 4°C for 3h with biotinylated anti-mouse IgG (1:2000, Vector Laboratories Inc., CA, U.S.A.) and in avidin-biotin-peroxidase complex (Vector Laboratories Inc.) for 30min at room temperature. The reaction product was visualized with an ECL kit (ECL system; Amersham, Arlington Heights, IL, U.S.A.).

After incubation, the minced granulation tissue was collected and homogenized in D-solution (47.3% (w v⁻¹) guanidine thiocyanate, 0.5% (w v⁻¹) sarcosyl, 2.5% (v v⁻¹) 1M sodium citrate and 0.1% (v v⁻¹) 2-mercaptoethanol) using the Vir-Tis 45 homogenizer for 3min at scale 40 on an ice bed. Total RNA was prepared from each sample by guanidinium-phenol-chloroform extraction (Chomczynski & Sacchi, 1987) and the yield of RNA extracted was determined by spectrophotometry. The primers for VEGF mRNA were designed as described by Asano et al. (1997); forward (5′-CGCGAATTCCATGAACTTTCTGCTCTCT-3′) and reverse (5′-TGAGAATTCTAGTTCCCGAAACCCTGA-3′). These primers amplify three isoforms of VEGF mRNA, 711bp (VEGF 188), 636bp (VEGF 164) and 504bp (VEGF 120). The rat glyceraldehyde 3-phosphate dehydrogenase (GAPDH) gene (a housekeeping gene) was used as an internal standard gene. The primers were designed as described by Robbins & McKinney (1992); forward (5′-TGATGACAAGAAGGTGGTGAAG-3′) and reverse (5′-TCCTTGGAGGCCATGTAGGCCAT-3′). Two micrograms of RNA from each sample and 20pmol of reverse primers of GAPDH and VEGF were incubated at 80°C for 5min, cooled immediately and reverse-transcribed in 20μl of a buffer (mM): Tris-HCl (pH 8.3) 50, KCl 75, MgCl₂ 3, 200 units of the reverse-transcriptase from moloney murine leukaemia virus (Life Technologies, Gaithersburg, MD, U.S.A.), deoxyribonucleotide triphosphates (Amersham Pharmacia Biotech Co.) 0.5 and dithiothreitol 10, at 37°C for 1h. For VEGF mRNA, PCR amplification was performed in 50μl of a PCR buffer (Tris-HCl (pH 8.3) 2.5mM, KCl 50mM, MgCl₂ 1.5mM, 2μl of the reverse transcribed RNA solution, 20pmol of each primer, 170μM dNTPs and 1.25 units of Taq polymerase (Takara Shuzo, Shiga, Japan)) with a thermal cycler (GeneAmp PCR system 2400, Perkin Elmer Cetus, Norwalk, CT, U.S.A.). PCR was performed at 94°C for 3min for denaturation, 57°C for 1min for annealing and 72°C for 3min for extension in the first round, 94°C for 1min, 57°C for 1min and 72°C for 3min in the next 38 rounds, and 94°C for 1min, 57°C for 1min and 72°C for 10min in the last round. For GAPDH mRNA, PCR amplification was performed for 27 cycles at 30s denaturation at 94°C, 1min annealing at 57°C and 2min extension at 72°C. The other conditions were the same as those used for VEGF mRNA. After PCR, 10μl of the PCR reaction mixture was loaded onto a 2% (w v⁻¹) agarose minigel and the PCR products were visualized by ethidium bromide staining after electrophoresis at 100V for 40min.

The minced granulation tissue was incubated at 37°C for 6h in the presence of histamine (10μM) and brefeldin A (50μM, Sigma Chemical Co.). Brefeldin A reversibly blocks protein translocation from endoplasmic reticulum to the Golgi apparatus (Strous et al., 1993) resulting in the inhibition of the release of VEGF. After the incubation, the tissue was fixed in 10% (v v⁻¹) formalin in PBS for 72h at 4°C and then dehydrated through three changes of 70% (v v⁻¹), 80% (v v⁻¹), 90% (v v⁻¹), and 95% (v v⁻¹) alcohol, three changes of absolute alcohol and three changes of pure chloroform for 12h each. The tissues were then embedded in paraffin. The sections (5μm) from the paraffin embedded tissues were mounted on glass slides and fixed by warming at 60°C for 30min. The tissue sections were deparaffined with xylene and ethanol, treated with 0.3% (v v⁻¹) hydrogen peroxide in methanol to inactivate endogenous peroxidase activity, and incubated in 3% (w v⁻¹) bovine serum albumin (BSA) (Sigma Chemical Co.) in PBS for 1h at room temperature to block nonspecific staining. The slides were then incubated with mouse monoclonal anti-VEGF (Santa Cruz Biotechnology) in 3% BSA-PBS at a concentration of 1μgml⁻¹ for 1h at 4°C. After being washed with PBS, the slides were treated with biotinylated anti-mouse IgG (1:2000, Vector Laboratories Inc.) in 3% BSA-PBS for 1h at 4°C and then treated with avidin-biotin peroxidase complex for 30min at room temperature. The reaction product was detected with 0.05% (w v⁻¹) 3,3′-diaminobenzidine-tetrahydrochloride (Dojindo Laboratories, Kumamoto, Japan), and 0.04% (w v⁻¹) nickel chloride in 50mM Tris-HCl buffer (pH 7.4) containing 0.033% (v v⁻¹) hydrogen peroxide.

Drugs used were cimetidine (Sigma Chemical Co.) and indomethacin (Sigma Chemical Co.). Cimetidine was dissolved in saline and indomethacin was dissolved in 99.5% of ethanol, and diluted with saline. Concentration of ethanol was adjusted to 0.1% (v v⁻¹). Five-hundred microliters saline containing 400μg cimetidine, 100μg indomethacin, or both was injected into the pouch of each rat just after carrageenin injection and then once a day for five consecutive days. Control rats received the same amount of saline containing 0.1% (v v⁻¹) ethanol. Six days after the injection of the carrageenin solution, total pouch fluid was collected and its volume was measured. The infiltrating leucocytes in the pouch fluid were enumerated using a hemocytometer. The population of infiltrating leucocytes in the pouch fluid were analysed by May-Gruenwald-Giemsa staining. The granulation tissue was dissected, weighed and homogenized as described above. The pouch fluid was centrifuged at 10,000×g and 4°C for 10min. VEGF protein levels in the supernatant fractions of the homogenates and the pouch fluid were determined by Western blot analysis as described above.

The angiogenesis in the granulation tissue was assessed using carmine dye (Natural Red 4, Sigma Chemical Co.) according to the method described by Kimura et al. (1986) and Colville-Nash et al. (1995) with slight modifications (Ghosh et al., 2000). Three milliliters of 5% (w v⁻¹) carmine dye in 5% (w v⁻¹) gelatin (Sankoh Jun-yaku, Tokyo, Japan) in saline at 37°C was injected into the tail vein of each rat. The carcasses were chilled on ice for 3h, and the entire granulation tissue was dissected and weighed. After 3 washes with PBS, the granulation tissue was homogenized in 2 volumes of 0.5mM sodium hydroxide using the Vir-Tis 45 homogenizer for 4min at scale 40 on an ice bed. The tissue homogenate was centrifuged at 10,000×g and 4°C for 30min. Five-hundred microliters of the supernatant was diluted 2 fold with 0.5mM sodium hydroxide and centrifuged again at 14,000×g and 4°C for 30min. The dye content in 200μl of the supernatant was determined spectrophotometrically by measuring absorbance at 490nm. For the standard curve, known amounts of carmine dye were added to the final supernatant of the granulation tissue of control rats which were injected with 3ml of a 5% (w v⁻¹) gelatin solution in saline without carmine dye, and the absorbance determined. The amount of carmine dye in the whole granulation tissue was then calculated.

The statistical significance of the results was analysed by Dunnett's test for multiple comparisons and Student's t-test for unpaired observations.

As shown in Figures 1b and 2a, histamine increased the VEGF protein content in the conditioned medium in a time- and concentration-dependent manner. mRNA (504, 636 and 711bp) levels for the three isoforms of VEGF in the minced granulation tissue were also increased by histamine with a maximum at 1h (Figure 1c), and in a concentration-dependent manner (Figure 2b). VEGF protein levels in the minced granulation tissue reached a maximum at 3h and then declined (Figure 1a). VEGF protein levels in the conditioned medium of the minced granulation tissue were also increased by incubating the granulation tissue in the presence of PGE₂ (1μM). The effect of PGE₂ (1μM) was almost the same as that of histamine at 1–10μM (Figure 2a).

As shown in Figure 3, the histamine-induced increase in VEGF protein levels in the conditioned medium of the minced granulation tissue was suppressed by the H₂ receptor antagonist cimetidine, but not by the H₁ receptor antagonist pyrilamine maleate (100μM) or the H₃ receptor antagonist thioperamide (100μM). The IC₅₀ value of cimetidine was calculated to be 0.37μM (Figure 4). The possibility that histamine induces VEGF production by increasing the production of PGE₂, which is one of the inducers of VEGF protein, was excluded because indomethacin at 1μM which inhibits PGE₂ production almost completely did not inhibit the histamine-induced VEGF production in the minced granulation tissues (Figure 3a).

The histamine-induced VEGF production in the minced granulation tissue was markedly inhibited by the protein kinase A (PKA) inhibitor H-89 (15–50μM) in a concentration-dependent manner, with IC₅₀ 12.5μM, and the cyclic AMP antagonist Rp-cAMP (10–100μM) in a concentration-dependent manner, with IC₅₀ 6.8μM (Figure 4). In contrast, the protein kinase C (PKC) inhibitors Ro 31-8425 (10μM) and calphostin C (10μM) did not inhibit the histamine-induced VEGF production (Figure 3b). The protein tyrosine kinase inhibitor genistein (30μM) also showed no effect (Figure 3b). To confirm the role of the cyclic AMP-PKA pathway in the histamine-induced VEGF production, effects of the adenylate cyclase activator forskolin by itself was examined. In the absence of histamine, forskolin (10μM) increased VEGF protein expression at 6h in the minced granulation tissue (Fold Increase: Control 1.00±0.12, Forskolin (10μM) 1.86±0.24*, the means from four samples with s.e.mean, *P<0.05 versus the Control). These findings indicated that the cyclic AMP-PKA pathway participates in VEGF protein expression.

The minced granulation tissue was incubated at 37°C for 6h in medium containing brefeldin A (50μM) in the presence or absence of histamine (10μM) and the expression of VEGF protein was analysed by immunostaining. Brefeldin A was added to block protein translocation from endoplasmic reticulum to the Golgi apparatus resulting in the inhibition of the release of VEGF into the conditioned medium. Histochemical analysis demonstrated that histamine increased the number of cells stained by anti-VEGF (Figure 5) and the VEGF-positive cells were mainly macrophages, endothelial cells and fibroblasts (Figure 5).

In rat peritoneal macrophages, 1h incubation with histamine increased the level of VEGF mRNA at 1 and 10μM in a concentration-dependent manner (Figure 6a). In fibroblasts obtained from the granulation tissue 3 days after the carrageenin injection, the levels of VEGF mRNA were also increased by histamine at 1 and 10μM (Figure 6b). In both cases, the levels of three isoforms of VEGF mRNA (504, 636 and 711bp) were increased.

Carrageenin-induced pouch fluid accumulation, leucocyte infiltration, and formation of granulation tissue 6 days after carrageenin injection were significantly inhibited by the administration of cimetidine (400μg) into the air-pouch (Figure 7). Effects of indomethacin (100μg) were more potent than those of cimetidine (400μg), and the additive effects were observed by the combined treatment with cimetidine (400μg) and indomethacin (100μg) (Figure 7). The angiogenesis in the granulation tissue was evaluated by the carmine dye method. Cimetidine (400μg) inhibited the angiogenesis in the whole granulation tissue at day 6 and the inhibition by indomethacin (100μg) was more potent than that by cimetidine (400μg) (Figure 8). The combined treatment with indomethacin (100μg) and cimetidine (400μg) further suppressed the angiogenesis (Figure 8).

Administration of cimetidine (400μg) and indomethacin (100μg) into the air pouch lowered the VEGF protein levels in the granulation tissue (Figure 9a) and in the pouch fluid (Figure 9b) at day 6. The effect of cimetidine was less potent than that of indomethacin (Figure 9a,b), and the combined treatment with cimetidine and indomethacin further decreased the VEGF production.