Defective Angiogenesis in the Inflammatory Granulation Tissue in Histidine Decarboxylase–deficient Mice but not in Mast Cell–deficient Mice

The progress of chronic and proliferative inflammation depends on angiogenesis (1). It is required not only for the maintenance of tissue perfusion, but also to allow an increase in the cellular traffic required for chronicity (2). Therefore, the inhibition of angiogenesis would be a valid target for drug development in chronic inflammatory diseases such as rheumatoid arthritis (3), atherosclerosis (4), diabetic retinopathy (5), psoriasis (6), wound healing (7), and chronic airway inflammation (8), as well as for antineoplastic therapy and Kaposi's sarcoma (9).

Vascular endothelial growth factor (VEGF)^* is a potent inducer of angiogenesis (10). Hypoxia conditions induce VEGF production, in which the transcription factor hypoxia–inducible factor-1 is activated (11). In addition, PGE2 induces VEGF production via the cAMP-protein kinase A pathway (12) and enhances angiogenesis in carrageenin-induced proliferative granulation tissue (13) and in cultures of human umbilical vein endothelial cells treated with the conditioned medium of colon cancer cells (14). However, we have suggested that mediators other than PGE₂ participate in VEGF production by the inflammatory granulation tissue, because the cyclooxygenase inhibitors indomethacin and NS-398 only partially suppressed the VEGF production (13). Zauberman et al. (15) first showed that histamine is an angiogenic in rabbit cornea. In addition, histamine was found to be produced in rapidly growing granulation tissue (16, 17) and was thought to promote neoplastic growth and angiogenesis (18–21). It is also reported that endogenous histamine has dual roles in angiogenesis in rats (22). Although these findings indicate that histamine might have the ability to induce angiogenesis, convincing proof that endogenous histamine upregulates angiogenesis in the granulation tissue is still lacking. Recently, we found that in a carrageenin-induced air pouch-type inflammation model in rats, the angiogenesis that occurs during the formation of granulation tissue was inhibited by treatment with cimetidine through the suppression of VEGF protein induction (23). In addition, in vitro experiments revealed that histamine induces VEGF production in the granulation tissue (23). However, the specific role of histamine in angiogenesis and the source of histamine in the proliferative inflammatory tissue remain to be elucidated.

In this study, we analyzed the role of endogenous histamine in the angiogenesis of the cotton thread-induced granulation tissue using histidine decarboxylase-deficient mice (HDC^−/−), mast cell–deficient mice (WBB6F1-W/W^V), and their corresponding wild-type mice (HDC^+/+ and WBB6F₁^+/+).

HDC-deficient 129 inbred mice (25–28 g) were generated by a gene targeting method as described previously (24). WBB6F1^+/+ and WBB6F1-W/W^V mice (25–28 g) were purchased from Japan SLC, Inc. Drugs used were pyrilamine maleate, cimetidine (Sigma-Aldrich), thioperamide (a gift from J.C. Schwartz, Unite de Neurobiologie et Pharmacologie Moleculaire [U.109] de l'INSERM, Paris, France), dimaprit (Research Biochemicals International), histamine (Sigma-Aldrich), goat anti-VEGF IgG, and control goat IgG (R&D Systems).

A cotton thread (model no. 8; Araiwa Co.) was washed overnight with ethyl acetate, and dried at room temperature before being cut into 1 cm in length (7 mg weight) and sterilized by dry heat at 160°C for 2 h. Male 129 inbred mice (25–28 g) were anesthetized and a cotton thread (1 cm, 7 mg) was implanted subcutaneously in the dorsum using a 13-G implant needle (Natsume Co.) according to the procedure described elsewhere (25) with a slight modification. The mice were treated in accordance with procedures approved by the Animal Ethics Committee of the Graduate School of Pharmaceutical Sciences, Tohoku University, Sendai, Japan.

Mice were anesthetized and killed by dislocation of the cervical spine at the indicated days after cotton thread implantation. The granulation tissue formed was dissected together with the cotton thread and weighed. Angiogenesis was assessed by taking photographs of the vascular network formation around the cotton thread and the subcutaneous tissue beneath the cotton thread. The dissected granulation tissue together with the cotton thread was washed in PBS, pH 7.4, and cut into small pieces with scissors before being homogenized in 20 vol of 0.5 mM sodium hydroxide using a Vir-Tis 45 homogenizer (Virtis Company) for 4 min at the scale 40 on an ice bed. The tissue homogenate was centrifuged at 10,000 g and 4°C for 30 min. An aliquot (200 μl) of the supernatant was centrifuged again at 14,000 g and 4°C for 30 min. The hemoglobin concentration in the supernatant was then determined spectrophotometrically by measuring the absorbance at 540 nm using a hemoglobin assay kit (Hemoglobin B Test Wako; Wako Pure Chemical Ind.). The amount of hemoglobin in the granulation tissue was expressed as μg hemoglobin/mg wet tissue.

After protein levels in the supernatant of the homogenate of the granulation tissue were determined (26), aliquots containing 0.56 μg of protein were separated by electrophoresis on a 12% (wt/vol) SDS-polyacrylamide gel and then transferred onto a nitrocellulose membrane (Schleicher and Schuell, Inc.). The membrane was first incubated with mouse monoclonal anti-VEGF (1:200; Santa Cruz Biotechnology, Inc.) at 4°C for 12 h, followed by biotinylated anti–mouse IgG (1:2,000; Vector Laboratories) at 4°C for 3 h, and finally in the avidin-biotin-peroxidase complex (Vector Laboratories) at room temperature for 30 min. The reaction product was visualized with an ECL kit (ECL System; Amersham Pharmacia Biotech).

Concentrations of VEGF proteins in the supernatant of the homogenate of the granulation tissue were determined using a commercially available ELISA kit (R&D Systems). In brief, 50 μl of the assay diluent and 50 μl of the diluted sample solution were incubated at room temperature in the dark for 2 h in each well of 96-well microplates precoated with polyclonal antibody specific to mouse VEGF. After washing, 100 μl solution of an antibody against mouse VEGF conjugated to horseradish peroxidase was added to each well and incubated at room temperature in the dark for 2 h.

After incubation, the plates were washed again and 100 μl of the substrate solution containing stabilized hydrogen peroxide and tetramethylbenzidine was added to each well and incubated for 30 min at room temperature in the dark. Thereafter, 100 μl of the stop solution was added to each well and the absorbance at 450 nm was determined. The intensity of the colored product is proportional to the concentration of mouse VEGF present in the original samples.

Coefficient of variations in intraassay and interassay precision of this assay kit is 4.3–8.2% and 5.7–8.4%, respectively. Total protein contents in the supernatant of the homogenate of the granulation tissue were determined (26), and the contents of VEGF proteins were expressed as pg VEGF per μg total protein.

A piece of the tissue, 2 cm in diameter surrounding the implanted cotton thread including the skin, cutaneous muscle layer, subcutaneous tissues, and the granulation tissue, was dissected together with the cotton thread, 0, 1, 3, and 5 d after cotton thread implantation. The dissected tissue including the cotton thread was homogenized in 20 vol of an HDC reaction buffer (0.1 M potassium-phosphate buffer, pH 6.8, 0.2 mM dithiothreitol, 0.01 mM pyridoxal 5′-phosphate, 2 μg/ml leupeptin, 2 μg/ml pepstatin, and 1% [vol/vol] polyethylene glycol, MW 300) for 4 min at the scale 40 on an ice bed using a Vir-Tis 45 homogenizer. The tissue homogenate was centrifuged at 10,000 g and 4°C for 30 min. The supernatant of the homogenate was then dialyzed overnight against the same buffer to remove endogenous histamine. Next, 1 ml of the dialyzed supernatant was preincubated at 37°C for 10 min. The sample was then incubated with or without l-histidine at 37°C for 3 h. The reaction was stopped by adding perchloric acid. Final concentrations of l-histidine and perchloric acid in the reaction mixture were 0.25 mM and 0.4 N, respectively. After centrifugation at 220 g and 4°C for 3 min, the supernatant was collected, and the amount of histamine in the supernatant determined fluorometrically (27). Protein levels in the dialyzed supernatant were determined (26), and HDC activity was expressed as the amount of histamine formed per min per mg protein.

After dissection, the granulation tissue together with the cotton thread were fixed in PBS containing 10% (vol/vol) formalin for 72 h at 4°C. They were then dehydrated through three changes of 70, 80, 90, and 95% (vol/vol) ethanol, two changes of absolute ethanol and two changes of pure chloroform each 12 h. The samples were then embedded in paraffin and cut into sections. The sections (5 μm) were then mounted on glass slides and spread by warming at 60°C for 30 min. The tissue sections were then deparaffined with xylene and ethanol, treated with 0.3% (vol/vol) hydrogen peroxide in methanol to inactivate endogenous peroxidase activity, and finally incubated in PBS containing 3% (wt/vol) BSA (Sigma-Aldrich) at room temperature for 1 h to block nonspecific staining. The slides were then incubated with rabbit polyclonal anti-HDC (1:750; Euro-diagnostica) in 3% (wt/vol) BSA-PBS at 4°C for 1 h. After being washed with PBS, they were incubated with biotinylated anti-rabbit IgG (1:2,000; Vector Laboratories) in 3% (wt/vol) BSA-PBS at 4°C for 1 h and then with the avidin-biotin peroxidase complex (1:500; Vector Laboratories) in 3% (wt/vol) BSA-PBS at room temperature for 30 min. The reaction product was detected with 50 mM Tris-HCl buffer (pH 7.4) containing 0.05% (wt/vol) 3,3′-diaminobenzidine tetrahydrochloride (Dojin Chemical Laboratories), 0.04% (wt/vol) nickel chloride, and 0.033% (vol/vol) hydrogen peroxide and HDC-expressing cells were observed with a light microscope (original magnification: ×400).

After dissection, the granulation tissue together with the cotton thread was fixed, dehydrated, and embedded in paraffin as described above. The sections (5 μm) were then mounted on glass slides and spread by warming at 60°C for 30 min. The tissue sections were then deparaffined with xylene and ethanol, treated with 0.3% (vol/vol) hydrogen peroxide in methanol to inactivate endogenous peroxidase activity, and incubated in PBS containing 1.5% (wt/vol) blocking serum (Santa Cruz Biotechnology, Inc.) at room temperature for 1 h to block nonspecific staining. The slides were then incubated with goat polyclonal anti-CD31 (2 μg/ml; Santa Cruz Biotechnology, Inc.) in PBS containing 1.5% (wt/vol) blocking serum at room temperature for 1 h. After being washed with PBS, they were incubated with biotinylated anti–goat IgG (1 μg/ml; Santa Cruz Biotechnology, Inc.) in PBS containing 1.5% (wt/vol) blocking serum at room temperature for 1 h and then with the avidin-biotin peroxidase complex (1:500; Santa Cruz Biotechnology, Inc.) in PBS at room temperature for 30 min. The reaction product was detected with freshly prepared peroxidase substrate mixture containing 1.6 ml distilled water, five drops 10× substrate buffer, 1 drop 50 × 3,3′-diaminobenzidine tetrahydrochloride chromogen, and 1 drop 50× peroxidase substrate (Santa Cruz Biotechnology, Inc.). CD31-expressing cells were observed with a light microscope (original magnification: × 200).

Statistical significance of the results was analyzed by Dunnett's test for multiple comparisons and Student's t test for unpaired observations.

Subcutaneous implantation of a cotton thread in the dorsum of HDC^+/+ mice induced gradual increases in the weight and hemoglobin levels of the granulation tissue from 3 to 7 d after cotton thread implantation (Fig. 1 A and B). In addition, the vascular network formation of the granulation tissue and subcutaneous tissue beneath the cotton thread also increased over the same time points (Figs. 2 A and 3 A). The levels of VEGF protein in the granulation tissue were also increased (Fig. 1 C). To clarify the role of VEGF in the angiogenesis and granulation tissue formation, goat anti-VEGF IgG or control goat IgG was subcutaneously injected at the site of cotton thread implantation just after the implantation and once a day on consecutive days. The injection of goat anti-VEGF IgG, but not control goat IgG, suppressed the vascular network formation, as well as the increases in hemoglobin levels and granulation tissue weight 5–7 d after cotton thread implantation (Fig. 2). These findings suggest that the cotton thread-induced angiogenesis is highly dependent on VEGF and that the inhibition of angiogenesis results in the inhibition of granulation tissue formation.

To clarify the role of histamine in angiogenesis, we implanted a cotton thread in the dorsum of HDC^−/− mice. Vascular network formation in the granulation tissue and the subcutaneous tissue beneath the cotton thread in HDC^−/− mice 3 and 5 d after cotton thread implantation was notably less than those in HDC^+/+ mice (Fig. 3 A). In addition, the granulation tissue formation (Fig. 3 B), as well as the levels of hemoglobin (Fig. 3 C) and VEGF protein in the granulation tissue (Fig. 3 D) of HDC^−/− mice were significantly less than those in HDC^+/+ mice 3 and 5 d after cotton thread implantation. Concentrations of VEGF proteins in the supernatant of the homogenate of the cotton thread induced-granulation tissue determined by ELISA also revealed that the levels of VEGF proteins in the granulation tissue 3 to 7 d after cotton thread implantation were lower in HDC^−/− mice than in HDC^+/+ mice (Table I). In addition, immunohistochemical analysis of the granulation tissue dissected 5 d after the cotton thread implantation using an antibody to CD31, a marker for endothelial cells, indicated that the neovascular density in the granulation tissue in HDC^−/− mice was obviously less than that in HDC^+/+ mice (Fig. 4). These findings indicate that histamine plays an important role in the induction of VEGF and angiogenesis during the early stage of cotton thread-induced granulation tissue.

Subcutaneous injection of histamine (0.1 and 1 μg) or the H₂ receptor agonist dimaprit (0.01, 0.1, and 1 μg) at the site of implantation just after the implantation and then once a day on consecutive days significantly enhanced angiogenesis in the granulation tissue as well as the granulation tissue formation itself compared with those in mice treated with saline (Fig. 5 A–C). The VEGF level in the granulation tissue was also increased by the injection of histamine or dimaprit in a dose-dependent manner (Fig. 5 D). In addition, topical injections of the H2 receptor antagonist cimetidine inhibited the histamine-induced angiogenesis in the granulation tissue and subcutaneous tissue beneath the cotton thread (Fig. 5 E). In contrast, the H₁ receptor antagonist pyrilamine maleate and the H₃ receptor antagonist thioperamide had no effect. Cimetidine also inhibited the histamine-induced increases in granulation tissue weight, and levels of hemoglobin and VEGF in the granulation tissue (data not shown). These findings indicate that histamine augments VEGF production via H₂ receptors.

To clarify whether the histamine responsible for angiogenesis was derived from mast cells, we compared the cotton thread-induced granulation tissue formation and angiogenesis in mast cell–deficient mice (WBB6F₁-W/W^V) with those in wild-type mice (WBB6F₁^+/+). We found that there was no significant difference in the granulation tissue formation, angiogenesis and VEGF levels in the granulation tissue between the two groups (Fig. 6). In HDC^+/+ mice, the implantation of a cotton thread increased the HDC activity in the tissue surrounding the implanted cotton thread including the skin, cutaneous muscle layer, subcutaneous tissues and the granulation tissue, from 1 to 5 d after implantation (Fig. 7 A). In WBB6F₁-W/W^V mice, although the HDC activity in the tissue was very low due to the mast cell deficiency, the cotton thread implantation induced a marked increase in HDC activity in the tissue to a similar extent as that in WBB6F₁-^+/+ mice (Fig. 7 B). In addition, histochemical analysis of the granulation tissue dissected 5 d after the cotton thread implantation indicated the absence of mast cells in the tissue (data not shown). However, the HDC-producing cells were shown to be infiltrating macrophages (Fig. 7 C). These findings indicate that HDC expression is induced in the nonmast cell, macrophages.

Histamine plays a variety of roles as an autacoid which regulates allergic inflammatory reactions (18, 28), differentiation of leukocytes precursors (29), neutrophil infiltration into the inflammatory sites (17), and gastric acid secretion (30). It also acts as a neurotransmitter in the central nervous system (31). In this study, we demonstrated a significant role of histamine in angiogenesis, i.e., histamine enhances angiogenesis in the inflammatory granulation tissue via the induction of VEGF production.

Subcutaneous implantation of a cotton thread induced rapid formation of granulation tissue together with angiogenesis (Figs. 1 and 3). The angiogenesis was highly dependent on VEGF production, as it was strongly inhibited by goat anti-VEGF IgG (Fig. 2). In addition, we also demonstrated that a reduction in the granulation tissue formation was accompanied by the inhibition of angiogenesis (Fig. 1). Thus, the mechanism of angiogenesis in the proliferative inflammatory tissue can be studied in vivo by using the mouse model of inflammation induced by cotton thread implantation. Using HDC^−/− mice, we found that the development of angiogenesis especially in the early phase (3–5 d after cotton thread implantation) depends on endogenous histamine (Figs. 3 and 4). The possibility that the functioning of hemangioblasts and endothelial cells were defective due to the destruction of the HDC gene was ruled out because the injection of histamine or dimaprit rescued the defective angiogenesis in HDC^−/− mice (Fig. 5).

The production of VEGF in HDC^−/− mice was significantly less than HDC^+/+ mice, and the injection of dimaprit or histamine increased the VEGF levels in the granulation tissue (Fig. 5 D). In addition, topical injection of cimetidine inhibited the histamine-induced angiogenesis in HDC^−/− mice (Fig. 5 E). Therefore, we concluded that endogenous histamine enhances VEGF production via H₂ receptors. This conclusion corresponds with our previous in vitro findings that exogenous histamine induces VEGF production by macrophages, endothelial cells, and fibroblasts present in the granulation tissue via the H₂ receptor-cAMP-protein kinase A pathway (23).

Although it is known that there is increased histamine production in rapidly growing tissues (16, 17), the histamine-producing cells have not yet been identified. It has been reported that the number of mast cells in the rapidly growing tissues increases (32) and that mast cell–derived histamine is angiogenic (33). However, we found that HDC activity in the tissue surrounding the implanted cotton thread including the skin, cutaneous muscle layer, subcutaneous tissues, and the granulation tissue increased even in WBB6F₁-W/W^V mice (Fig. 7 A and B). In addition, histochemical analysis of the granulation tissue dissected 5 d after the cotton thread implantation indicated the absence of mast cells in the granulation tissue (data not shown). In contrast, HDC-producing cells in the granulation tissue were identified as infiltrating macrophages (Fig. 7 C). Histamine production by nonmast cells was observed in mouse skin treated with PMA (34), in various tissues such as liver and lung in IL-1–treated mice (35), and in the infiltrating leukocytes in allergic inflammation in rats (36). Our findings did not exclude the possibility that mast cells in the granulation tissue and the surrounding tissues release histamine.

Because the cotton thread implantation induced an apparent angiogenesis in WBB6F₁-W/W^V mice as well as in WBB6F₁^+/+ mice, we concluded that histamine from infiltrating macrophages play a significant role in angiogenesis of the granulation tissue. It is reported that cimetidine, an H₂ antagonist, delays wound healing especially in peptic ulcer (37, 38), but its mechanism of action has not been clarified. We found that cimetidine reduces VEGF production in the carrageenin-induced inflammation model in rats (23). In the cotton thread–induced inflammation model in mice, treatment with cimetidine showed partial but significant inhibition of angiogenesis and VEGF production (data not shown). Therefore, we hypothesize that cimetidine delays wound healing by inhibiting histamine-mediated VEGF production and angiogenesis. Our findings also suggest that H₂ receptor antagonists would be useful for suppressing the angiogenesis-dependent formation of inflammatory granulation tissue, and conversely, that H2 receptor agonists may enhance angiogenesis in wound healing and several ischemic diseases. In addition, because cimetidine inhibits tumor growth in vivo (39, 40), there is a possibility that the growth of some tumors may also be regulated by histamine-mediated angiogenesis.

These findings demonstrate that the defective angiogenesis in the inflammatory granulation tissue is induced in HDC^−/− mice but not in mast cell–deficient mice, and histamine derived from nonmast cells plays a significant role in the angiogenesis of the inflammatory granulation tissue.

We are grateful to Dr. J.C. Schwartz (Unite de Neurobiologie et Pharmacologie Moleculaire (U.109) de l'INSERM, Paris, France) for providing the H3 receptor antagonist thioperamide. We thank Dr. T. Terui and Prof. H. Tagami (Department of Dermatology, Tohoku University, Graduate School of Medicine, Sendai, Japan) for use of the microscopy facility.

This study was supported in part by Scientific Research on Priority Areas (12139202) from the Ministry of Education, Culture, Sports, Science, and Technology (MEXT) of Japan.