An in vivo examination of the differences between rapid cardiovascular collapse and prolonged hypotension induced by snake venom

We investigated the cardiovascular effects of venoms from seven medically important species of snakes: Australian Eastern Brown snake (Pseudonaja textilis), Sri Lankan Russell’s viper (Daboia russelii), Javanese Russell’s viper (D. siamensis), Gaboon viper (Bitis gabonica), Uracoan rattlesnake (Crotalus vegrandis), Carpet viper (Echis ocellatus) and Puff adder (Bitis arietans), and identified two distinct patterns of effects: i.e. rapid cardiovascular collapse and prolonged hypotension. P. textilis (5 µg/kg, i.v.) and E. ocellatus (50 µg/kg, i.v.) venoms induced rapid (i.e. within 2 min) cardiovascular collapse in anaesthetised rats. P. textilis (20 mg/kg, i.m.) caused collapse within 10 min. D. russelii (100 µg/kg, i.v.) and D. siamensis (100 µg/kg, i.v.) venoms caused ‘prolonged hypotension’, characterised by a persistent decrease in blood pressure with recovery. D. russelii venom (50 mg/kg and 100 mg/kg, i.m.) also caused prolonged hypotension. A priming dose of P. textilis venom (2 µg/kg, i.v.) prevented collapse by E. ocellatus venom (50 µg/kg, i.v.), but had no significant effect on subsequent addition of D. russelii venom (1 mg/kg, i.v). Two priming doses (1 µg/kg, i.v.) of E. ocellatus venom prevented collapse by E. ocellatus venom (50 µg/kg, i.v.). B. gabonica, C. vegrandis and B. arietans (all at 200 µg/kg, i.v.) induced mild transient hypotension. Artificial respiration prevented D. russelii venom induced prolonged hypotension but not rapid cardiovascular collapse from E. ocellatus venom. D. russelii venom (0.001–1 μg/ml) caused concentration-dependent relaxation (EC50 = 82.2 ± 15.3 ng/ml, Rmax = 91 ± 1%) in pre-contracted mesenteric arteries. In contrast, E. ocellatus venom (1 µg/ml) only produced a maximum relaxant effect of 27 ± 14%, suggesting that rapid cardiovascular collapse is unlikely to be due to peripheral vasodilation. The prevention of rapid cardiovascular collapse, by ‘priming’ doses of venom, supports a role for depletable endogenous mediators in this phenomenon.

years. We have previously defined 'cardiovascular collapse' as a sudden drop in recordable blood pressure 14 following the administration of venom, to a laboratory animal or after human envenoming. The most common snakes responsible for this phenomenon are the brown snakes (Pseudonaja spp.) 15 and, less commonly, taipans (Oxyuranus spp.) 14 and tiger snakes (Notechis spp.) 16 . In some cases, patients spontaneously recover after collapse or respond well to basic and advanced life support 17,18 . In some cases of envenoming, particularly by brown snakes (Pseudonaja spp.), the collapse can be fatal 16,17 . Indeed, in Australia, cardiovascular collapse is the leading cause of death due to snake envenoming 19 .
A number of hypotheses have been proposed to explain the cause of the cardiovascular collapse associated with snake envenoming. Previous studies have postulated that cardiovascular collapse may be due to prothrombin activators or pro-coagulant toxins present in snake venoms 20,21 . We have recently demonstrated that in vivo cardiovascular collapse can be caused by death adder (Acanthophis rugosus) venom, despite a lack of pro-coagulants in this venom. This suggests that pro-coagulant toxins are not required to induce collapse 15 . Furthermore, administering small 'priming' doses of A. rugosus venom, prior to P. textilis venom, prevented subsequent cardiovascular collapse. This indicated that the release of depletable endogenous mediators most likely contribute to cardiovascular collapse. We also showed that the protective effect of priming doses of venom is transient (i.e. lasting up to approximately 1 hour), indicating replenishment of mediators 15 . This suggests that clotting factors are not directly involved in cardiovascular collapse, given the longer time period required for their resynthesis. Commercial polyvalent antivenom demonstrated a protective effect on cardiovascular collapse in vivo, supporting a role for antigenic venom components in cardiovascular collapse 15 .
To further investigate this phenomenon, in the present study we examined the cardiovascular activity of seven medically important snake venoms: Australian Eastern Brown snake (Pseudonaja textilis), Sri Lankan Russell's viper (Daboia russelii), Javanese Russell's viper (D. siamensis), Gaboon viper (Bitis gabonica), Uracoan rattlesnake (Crotalus vegrandis), Carpet viper (Echis ocellatus) and Puff adder (Bitis arietans). We identified the species which caused cardiovascular collapse in vivo to further investigate the possible mechanisms for this phenomenon.

Results
In vivo experiments. For  To investigate the effects of artificial respiratory support, a higher dose of D. russelii venom (1 mg/kg, i.v.) was used, which caused a 100% decrease in blood pressure. This hypotensive effect (i.e. 100%) of D. russelii venom (1 mg/kg, i.v.) was significantly attenuated by artificial respiratory support, reducing the hypotensive effect to 42% (Fig. 2a). In contrast, the rapid cardiovascular collapse induced by E. ocellatus venom (50 µg/kg, i.v.) was not attenuated by artificial respiratory support (Figs. 1b and 2b).
To further investigate the above effects of the venoms, a representative venom that caused collapse (i.e. P. textilis) and a representative venom that caused hypotension (i.e. D. russelii) were injected intramuscularly. Venom doses were increased to better mimic a bite scenario. P. textilis venom (20 mg/kg, i.m.; Table 2) caused collapse within 10 min of administration to the left bicep femoris muscle. D. russelii venom (50 mg/kg or 100 mg/kg, i.m.; Table 2) caused hypotension, but not collapse, within 30 min of administration.  (Fig. 4).

Discussion
We have demonstrated two distinct patterns of cardiovascular effects caused by the intravenous administration of different snake venoms. The first group of venoms cause a rapid decrease in blood pressure, often without recovery. We refer to this as 'rapid cardiovascular collapse' and it is the same phenomenon that we have previously described with Australian elapid venom 15 . A defining feature of this hypotensive response is that it is attenuated by sub-toxic 'priming' doses of venom of the same, or different snake species 15 . Snake venoms reported to induce this effect include P. textilis and E. ocellatus in this study, and previously, O. scutellatus (Coastal taipan) 14 . The second group of venoms, which include D. russelii and D. siamensis, caused a slower and prolonged decrease in blood pressure, with recovery occurring in most cases. In contrast to the first group, the drop in blood pressure is not prevented by prior administration of priming doses. We refer to this effect as 'prolonged hypotension' . www.nature.com/scientificreports www.nature.com/scientificreports/ We have previously postulated that the attenuation of the hypotensive effect with prior administration of smaller sub-toxic doses of venom is due to the pre-release, and depletion, of mediators which induce collapse 15 . This phenomenon was observed in the current study in which smaller priming doses of E. ocellatus venom or P. textilis venom prevented cardiovascular collapse caused by a larger dose of E. ocellatus venom. This suggests that these venoms are inducing their cardiovascular effects via a common mechanism.
For a high dose of D. russelii venom (i.e. 1 mg/kg), a response similar to rapid cardiovascular collapse occurred. However, when the rat was placed on a ventilator prior to administration of venom, this so called 'collapse' was prevented. In contrast, when rats administered E. ocellatus venom were placed on the ventilator, rapid cardiovascular collapse still occurred. The reasons for the protective effects of supportive respiration are unclear. We have previously shown that the neurotoxins present in D. russelii venom are relatively weak 22 . However, given that a rat has approximately 64 ml of circulating blood per kg body weight, an intravenous dose of 1 mg/kg of venom leads to a blood concentration of approximately 16 µg/ml. This very high venom concentration may be sufficient to cause paralysis of the diaphragm given that a 30 ng/ml concentration of the same venom caused complete   www.nature.com/scientificreports www.nature.com/scientificreports/ neuromuscular blockade in the chick biventer nerve-muscle preparation 22 . Therefore, it could be argued that artificial respiration is preventing or overcoming the paralytic effects of the neurotoxins on the rat diaphragm. The different effects of supportive respiration on the cardiovascular effects of the venoms also supports the fact that collapse due to E. ocellatus venom occurs via a different mechanism. These studies were conducted in vivo using ketamine/xylazine as anesthesia, which may have affected the blood pressure, although ketamine is more likely to cause a slight increase in blood pressure.    www.nature.com/scientificreports www.nature.com/scientificreports/ To ensure that these cardiovascular effects seen in the in vivo model occurs in an actual snake bite, the effects of P. textilis venom and D. russelii venom were also tested via intramuscular administration. At 20 mg/kg (i.m.), P. textilis venom caused collapse within 10 min of administration. This delay in response is likely to be due to the time it takes for the venom to be absorbed. In contrast, when D. russelii venom was administered via intramuscular injection prolonged hypotension occurred, similar to that observed when venom was administrated intravenously. Even at 100 mg/kg concentration, collapse did not occur, further highlighting that both collapse and hypotension are not dose-dependent responses but represent two distinct cardiovascular effects.
There are many factors that could lead to venom-induced hypotension 11 , as distinct from cardiovascular collapse. Some snake venoms have highly evolved toxins such as calciseptine, FS2 toxins, C10S2C2 and S4C8 which block L-type Ca 2+ currents 23,24 . Increasing capillary permeability protein (ICPP), isolated from Blunt-nosed viper (V. lebtina) venom is similar in potency and structure to vascular endothelial growth factor (VEGF) and is responsible for increasing vascular permeability 25 . Natriuretic peptides found in Green Mamba (D. angusticeps) venom 26 and bradykinin potentiating peptides found in Bothrops spp. are also potent vaso-relaxants 10,27,28 . In the current study, D. russelii venom caused concentration-dependent relaxation of rat small mesenteric arteries suggesting peripheral vasodilation contributes to the prolonged hypotension observed in vivo. D. siamensis venom was also an efficacious dilator of rat small mesnteric arteries, though less potent than D. russelii venom. In contrast, the venoms which had a modest hypotensive effect in vivo (B. arientans, C. vegrandis and B. gabonica) were poor vasorelaxants of isolated mesenteric arteries. Although vasorelaxant responses can exhibit heterogeneity throughout the vasculature, the mesenteric vascular bed was chosen for this study given it makes a significant contribution to overall total peripheral resistance, receiving 25% of total cardiac output. As such, characterising vasorelaxation responses in these small mesenteric arteries (approx. 300microns in diameter), is of physiological relevance to blood pressure control. Gaboon viper (B. gabonica) venom has been shown to induce vasodilation resulting in a drop in peripheral resistance, leading to reduction in stroke volume due to cardiotoxins 29 . In another study using isolated heart preparations, Rhinoceros viper (B. nasicornis) venom produced an increase in left ventricular pressure, pacemaker activity and heart rate, indicating that the venom contains toxins that disrupt [Ca 2+ ] and ion conductance 30 .
PLA 2 toxins are ubiquitous components of snake venoms and display an array of activities including neurotoxicity, myotoxicity, cardiotoxicity, anti-coagulation, haemolytic, hypotensive and local tissue necrotic activity 31 . Interestingly, P. textilis and E. ocellatus venom, which induced rapid cardiovascular collapse, had the lowest PLA 2 activity, whereas Daboia spp. had the highest amount of PLA 2 activity. Therefore, there does not seem to be a link between PLA 2 activity and cardiovascular collapse, although these toxins may contribute, directly or indirectly, to prolonged hypotension via vasodilation.
Daboia spp. and Pseudonaja spp. contain pro-coagulants factors in their venom 32 . While both P. textilis and E. ocellatus venoms caused rapid cardiovascular collapse in vivo, P. textilis venom was most potent in causing coagulation while E. ocellatus venom possessed comparatively less pro-coagulant activity. D. russelii and D. siamensis venom also caused coagulation. D. russelii and D. siamensis venoms are known to contain Factor X 33,34 while both E. ocellatus 35 and P. textilis venom contain prothrombin activators 21,36,37 . Therefore, pro-coagulant activity is unlikely to be directly related to the cardiovascular collapse induced by the venoms.
In conclusion, we have shown that the in vivo cardiovascular effects of venom include, at least, two distinct phenomena i.e. rapid cardiovascular collapse and prolonged hypotension and that both effects involve different mechanisms. Rapid cardiovascular collapse has a sudden onset and appears to be mediated by depletable endogenous mediators. In contrast, prolonged hypotension has a slower onset and appears to be due mainly to vasodilation. Anaesthetised rats. Male Sprague-Dawley rats (280-350 g) were anaesthetised with a mixture of ketamine (100 mg/kg, i.p.) and xylazine (10 mg/kg, i.p.). Ketamine/xylazine cocktail was used as it provides sedation and muscle relaxation as well as deep analgesia and anesthesia without compromising blood pressure. A midline incision was made and a cannula inserted into the trachea for mechanical ventilation (~1 ml/100 g of body weight at 55 strokes/min) if required. Cannulae were inserted into the left jugular vein for administration of venom and the right carotid artery to record arterial blood pressure. The arterial cannula was connected to a pressure transducer. Blood pressure was then allowed to stabilise for approximately 10-15 min. Body temperature was maintained at 37 °C using an overhead lamp and heated dissection table. Venom was administered via the jugular vein followed by flushing with saline or via a bolus administration into the left bicep femoralis muscle. Responses to venom were measured as percentage change in mean arterial pressure (MAP). www.nature.com/scientificreports www.nature.com/scientificreports/ mesenteric artery) were isolated, cut into 2 mm lengths, and mounted in isometric myograph baths. Vessels were maintained in physiological salt solution, composed of (in mM): 119 NaCl, 4.7 KCl, 1.17 MgSO 4 , 25 NaHCO 3 , 1.8 KH 2 PO 4 , 2.5 CaCl 2 , 11 glucose, and 0.026 EDTA, at 37 °C and supplied with carbogen (95% O 2 ; 5% CO 2 ). The mesenteric arteries were allowed to equilibrate for 30 min under zero force and then a 5 mN resting tension was applied. Changes in isometric tension were recorded using Myography Interface Model 610 M version 2.2 (ADInstruments, Pty Ltd, USA) and a chart recorder (Yokogawa, Japan). Following a 15 min equilibration period at 5 mN, the mesenteric arteries were contracted maximally (F max ) using a K + depolarizing solution [K + -containing physiological salt solution (KPSS); composed of (in mM) 123 KCl, 1.17 MgSO 4 , 1.18 KH 2 PO 4 , 2.5 CaCl 2 , 25 NaHCO 3 , and 11 glucose]. The integrity of the endothelium was confirmed by relaxation to acetylcholine (ACh, 10 µM) in tissues pre-contracted with the thromboxane A 2 mimetic, U46619 (1 µM), then washed with physiological salt solution and the tension allowed to return to baseline. Relaxation of >80% to ACh was used to indicate vessels with an intact endothelium. There were no significant differences in response to ACh between the groups studied. If endothelial damage was evident (ACh relaxation <80%) then the vessel was not used for experimentation. Cumulative concentration-response curves to venom (1 ng/ml-1µg/ml) were constructed in vessels pre-contracted with titrated concentrations of U46619 (~50% F max ). Sodium nitroprusside (SNP, 10 µM) was added at the end of each concentration-response curve to ensure maximum relaxation. Only one concentration-response curve to venom was obtained in each vessel segment 38,39 . Pro-coagulation assay. Aliquots (10 ml) of fresh frozen plasma were thawed at 37 °C, then spun at 2500 rpm for 10 min. Venom solutions (100 µL) were placed in the wells of a 96 well microtitre plate at room temperature or at 37 °C in a BioTek ELx808 plate reader. Plasma (100 µL) and calcium (0.2 M/ml) were then added simultaneously to each well using a multichannel pipette. After a 5 s shake step for mixing, the optical density at 340 nm was monitored every 30 s over 20 min 40 . PLA 2 assay. PLA 2 activity of the venoms was determined using a secretory PLA 2 colourmetric assay kit (Cayman Chemical; MI, USA) according to manufacturer's instructions. This assay used 1, 2-dithio-analogue of diheptanoyl phosphatidylcholine, which serves as a substrate for PLA 2 enzymes. Free thiols generated following the hydrolysis of the thioester bond at the sn-2 position by PLA 2 are detected using DTNB (5, 5′-dithio-bis-[2-nitrobenzoic acid]). Colour changes were monitored at 405 nm in a fusion α microplate reader (PerkinElmer; MA, USA), sampling every minute for a 5 min period. PLA 2 activity was expressed as micromoles of phosphatidylcholine hydrolysed per minute per milligram of enzyme 2 .

Myograph experiments. Male
Statistical analysis. For the anaesthetized rat experiments, pulse pressure was defined as the difference between systolic and diastolic blood pressures. Mean arterial pressure (MAP) was calculated as diastolic blood pressure plus one-third of pulse pressure. These data were tested using a D' Agostino-Pearson normality test and found to be normally distributed. Therefore, differences in MAP between treatment groups were analysed using a one-way ANOVA with Dunnett's multiple comparison test. Sample sizes are based on the number of animals required to provide >85% power to detect an effect size of 35% with a confidence level (α) of 5% for the in vivo endpoint measure of blood pressure (standard deviation (SD) <15%). This ensured that experimental design was sufficiently powered.
For the myography experiments, blood vessel relaxation was expressed as a percentage reversal of the U46619 pre-contraction. Individual relaxation curves to D. russelii venom were fitted to a sigmoidal logistic equation and EC 50 values (concentration of agonist resulting in a 50% relaxation) calculated 41 . Where EC 50 values could not be obtained, concentration-response curves to venoms were compared by means of a two-way repeated measures ANOVA (n = number of artery segments from separate animals). Data represent the mean ± SEM (error bars on graph). Statistical significance was defined as P < 0.05. All data analysis was performed using GraphPad Prism version 5.02 (GraphPad Software, San Diego, CA, USA).
For the coagulation assay, responses were plotted as 30 s/[clotting time(s)] against the logarithm of the venom concentration. This provided a normalised measure of the clotting effect and produced normalised concentration-clotting curves, which were fitted with a standard sigmoidal curve (Hill slope = 1) to calculate the effective concentration 50 (EC 50 ). The EC 50 is the concentration of venom that resulted in a pro-coagulant effect halfway between no clotting effect and maximal clotting effect 42 .