Trimeresurus flavoviridis (habu) venom (Okinawa), Gloydius blomhoffii venom and Bitis arietans venom were purchased from Japan Snake Institute, Gunma. Crotalus atrox venom was purchased from Sigma-Aldrich. Hemorrhagic toxin b (HTb) from C. atrox venom was prepared by the method reported previously [11]. Bilitoxin-2 and Ac₁-proteinase were isolated using our methods reported previously [12,13] for Agkistrodon bilineatus venom and Deinagkistrodon acutus venom, respectively. Hemorrhagic toxin-1 (HT-1) was obtained from B. arietans venom [14]. Human and bovine fibrinogens were supplied by Sigma-Aldrich, Tokyo, Japan. Type IV collagen was purchased from Nitta Gelatin Inc. Cryo-preserved human umbilical vein endothelial cells (HUVEC), its respective cell culture media (HuMedia EB-2), other cell culture supplements, and reagents were obtained from Kurabo (Osaka, Japan). The cell counting kit was purchased from Dojindo (Kumamoto, Japan). Other chemicals were of analytical grade from commercial sources. All experiments involving the use of animals were carried out in compliance with the guidelines for animal experiments of Faculty of Pharmacy, Meijo University.

Rosmarinic acid (RA) was isolated and purified from methanolic extract of Argusia argentea as reported previously [10]. A voucher sample has been deposited in the Herbarium of the Faculty of Pharmacy, Meijo University, Japan.

Anti-hemorrhagic activity was assayed by the method of Bjarnason and Tu [15] using ddY mice of 20 g average weight. Two groups of four mice were used for the experiment. All crude venom solutions of T. flavoviridis venom, C. atrox venom, G. blomhoffii venom and B. arietans venom, were prepared at a concentration of 0.14 mg/mL in saline. Concentrations of purified hemorrhagic toxin solutions were as follows: HTb (0.41 mg/mL), bilitoxin-2 (2.75 μg/mL), HT-1 (0.29 mg/mL), and Ac₁‑proteinase (1.04 mg/mL). A test solution was prepared by mixing the venom solution or the toxin solution (50 µL) and RA (0.5 mg/mL in 10% DMSO-saline, 50 µL) followed by 10 min incubation at 37 °C. These test solutions (100 µL) were injected subcutaneously (s.c.) in the abdomen of mice. Similarly, a group of mice which were injected with a venom solution without RA was used as a control group, and also a group which was only injected with 10% DMSO-saline (50 μL) served as a blank group. Prior to this study, effects of DMSO at several concentrations were investigated, and DMSO at less than 10% was found to cause no significant inactivation of venom. After 24 h, mice were euthanized by inhalation of chloroform, the skin covering the abdomen was removed and hemorrhagic lesions were analyzed as follows. The area of the lesion was estimated by major and minor axes measurements, since the shape of the lesions are always amorphous, like an ellipse.

Fibrinogen hydrolytic activity was assayed by the method of Ouyang and Teng [16]. A solution of 0.1% human fibrinogen in 50 mM Tris-HCl buffer (pH 7.5) (1 mL) and a venom solution (50 µL of 0.21 mg/mL of T. flavoviridis venom or 5.5 µg/mL of bilitoxin-2) were incubated in the presence or absence of RA (0.5 mg/mL) at 37 °C. At various time intervals, aliquots of 100 µL of denaturing solution (10 mM phosphate buffer, pH 7.2, containing 10 M urea, 4% sodium dodecyl sulfate (SDS), and 4% β-mercaptoethanol) were added. This solution was incubated at 37 °C for 6 h and then run on 10% polyacrylamide slab gel electrophoresis. Electrophoresis was carried out for 2 h with a current of 25 mA per slab gel. Bromophenol blue (BPB) solution was used as an indicator.

Collagen hydrolytic activity was assayed as follows. Sodium hydrogen carbonate (60 µL, pH 12) was added to 0.3% type IV collagen (0.9 mL) and adjusted to pH 8. Aliquots of type IV collagen were incubated with T. flavoviridis venom (0.21 μg/mL) in the presence or absence of RA (0.5 mg/mL). At various time intervals, aliquots of 100 µL of denaturing solution (10 mM phosphate buffer, pH 7.2, containing 10 M urea, 4% SDS, and 4% β-mercaptoethanol) were added. This solution was boiled for 3 min and run on SDS-PAGE using a 7.5% polyacrylamide slab gel electrophoresis.

The effects of RA and T. flavoviridis venom on cultured human umbilical vein endothelial cells (HUVEC) were investigated [17,18,19]. Frozen HUVEC were cultured and maintained in commercially available media, HuMedia-EB2, supplemented with fetal calf serum (2% v/v), hEGF (10 ng/mL), hFGF-B (5 ng/mL), hydrocortisone (1 µg/mL), heparin (10 µg/mL), gentamicin (50 µg/mL), and amphotericin B (50 ng/mL). At confluency, cells were trypsinized, washed with the same medium and then resuspended in growth media. These cells were seeded in 96-multiwell plates (5 × 10³ cells per well in 100 µL medium) and were allowed to attach and reach log phase of growth. Aliquots of venom and RA to be assayed were diluted in saline and were sterilized by filtration with cellulose acetate 0.22 µm membrane filters. Various concentrations of RA (0.5, 0.25, 0.125, 0.06, 0.03 mg/mL in 10% DMSO-saline) in the presence or absence of T. flavoviridis venom (0.14 mg/mL) were added to each well in 100 µL medium. The plate was incubated at 37 °C under 5% CO₂ atmosphere for 17 h. Ten microliters of cell counting kit-8 was added to each well, and the microplate was incubated for 1 h, after which cell densities were measured at 450 nm using Bio-RAD Model 550 MicroplateR eader.

Histopathological study for RA was performed by intramuscular (i.m.) injection of T. flavoviridis venom solution into the medial aspect of the thigh muscle of ddY strain white mice. Histopathological study of muscle was conducted in three groups. Group A was injected with the venom (0.21 mg/mL, 100 µL), while group B was injected with RA (0.5 mg/mL, 100 µL). Group C was injected with a mixture of the venom (0.41 mg/mL, 50 µL) and RA (0.25 mg/mL, 50 µL). Test solutions were preincubated at 37 °C for 10 min before injection. The mice were killed by chloroform inhalation 24 h after injection. Tissue samples were immediately fixed in buffered formate fixative for 24 h at room temperature. The tissue was then washed for 4 h in running water, dehydrated in an autotechnicon, and stained with hematoxylin and eosin for observation under light microscope.

RA (0.5 mg/mL) in 10% DMSO-saline was heated at 37 °C, 50 °C and 100 °C for 10 min, respectively. Fifty microliters of each heat-treated RA solution was mixed with T. flavoviridis venom (0.21 mg/mL, 50 µL) and incubated at 37 °C for 10 min. An aliquot of 0.1% human fibrinogen in 50 mM Tris-HCl buffer, pH 7.5 was added to each test tube. The reactions were stopped by adding 100 µL of denaturing solution (10 mM phosphate buffer, pH 7.2, containing 10 M urea, 4% SDS, and 4% β-mercaptoethanol). These solutions were incubated at 37 °C for 6 h and then run on 10% polyacrylamide slab gel electrophoresis. The venom solution without a sample was also subjected to SDS-PAGE for comparison. Moreover, by using RA (0.5 mg/mL), which was treated at 100 °C, antihemorrhage activity was also examined as mentioned above.

Hind-paw edema activity was assayed by the method of Ho et al. [20]. Four ddY strain white mice (20–23 g) were individually injected in the right foot pad with Trimeresurus elegans venom (12.5 µg in 50 µL of 10% DMSO-saline). An equal volume of 10% DMSO saline was injected into the left paws as control. Inhibition assays were performed by preincubated rosmarinic acid (0.5 mg/mL in 10% DMSO saline) with toxin for 10 min at 37 °C. The volume of each paw was measured with a slide caliper. The degree of paw swelling was expressed as % increase of the initial paw volume.

When crude venom (T. flavoviridis (habu) venom, C. atrox venom, G. blomhoffii venom, and B. arietans venom) or purified toxin, (HTb, bilitoxin-2, HT-1 and Ac₁-proteinase) was injected s.c. in the abdomen of mice, a distinct hemorrhagic lesion was observed (Figure 1c) [10]. No hemorrhagic spots were produced after s.c. injection of crude venom or purified toxin with RA (Figure 1b). RA effectively inhibited the hemorrhagic activities of crude venoms as well as purified hemorrhagic toxins.

When human fibrinogen was incubated with T. flavoviridis venom, the Aα band of the fibrinogen disappeared on SDS-PAGE, whereas the Bβ chain and γ chain were essentially unaffected (Figure 2a). The venom with RA did not reveal any apparent degradation of human fibrinogen (Figure 2b). RA also inhibited Aα hydrolysis by bilitoxin-2 (Figure 2c and 2d).

RA alone had no effect on the viability of HUVEC, but it markedly protected HUVEC from the toxic effects of T. flavoviridis venom (0.14 mg/mL) at all concentrations of RA tested (0.50, 0.25, 0.125, 0.06, and 0.03 mg/mL) (Figure 3). The maximum (84.2%) protective effect of RA was exhibited with RA at 0.5 mg/mL.

Type IV collagen was incubated with T. flavoviridis venom for different periods of time. The venom completely degraded type IV collagen (104 kDa), especially over 1 h, and degradates with smaller molecular weights (43 and 35 kDa) appeared, as shown in Figure 4a. In the presence of RA (0.5 mg/mL), type IV collagen was not digested by incubation with the venom (Figure 4b).

Both hemorrhage and neutrophil infiltrations were observed in a wide area (of the circle) after injection of T. flavoviridis venom (0.21 mg/mL) (Figure 5a). The result showed normal musculature devoid of hemorrhage and neutrophils in the muscle fibers after injection of RA (0.5 mg/mL) (Figure 5b). There was no hemorrhage or neutrolphil infiltration in the muscle fibers after injection of a mixture of the venom (0.41 mg/mL) and RA (0.5 mg/mL, or 0.25 mg/mL; Figure 5c).

RA, which was heated at various temperatures, was incubated with human fibrinogen and T. flavoviridis venom. RA inhibited the digestion of Aα chain of human fibrinogen after treatment at each temperature (Figure 6). RA (0.5 mg/mL), which was heated at 100 °C for 10 min, also showed complete inhibition against T. flavoviridis venom.

The edema-forming activity was assayed using four mice. T. elegans venom induced an edema of 30% in the mouse footpad, at a dose of 12.5 µg. When T. elegans venom was preincubated with RA (0.5 mg/mL), the edema-forming was significantly reduced.