Venom is a functional trait, used by venomous species of snake to subdue prey animals or deter potential predators (Jackson and Fry, 2016). The venom itself is often described as a “cocktail” (Casewell et al., 2013) because it contains numerous toxin species, each of which may interact with a different component of the target organism’s physiology. Venom achieves its goal of subjugation or deterrence by virtue of these toxins interfering with the normal functioning of various regulatory (homeostatic) networks. These networks operate in a state of “criticality” (marginal stability—Daniels et al., 2018) and thus their subversion may lead to a catastrophic failure of homeostasis, either locally (i.e., in the vicinity of the bite site) or systemically. On the other hand, organisms are buffered systems and the effects of interference with a single gene product are often minimal at the level of the integrated phenotype (Noble, 2011). The complexity of the venom cocktail, therefore, is in service of the overall function of the trait–venoms typically attack multiple regulatory networks simultaneously and may attack each network on multiple fronts. This is also highlighted by the complexity of clinical envenoming syndromes in human bite victims (Gutiérrez et al., 2017).

According to the World Health Organisation, an estimated 2 million people are envenomed following a snakebite every year and approximately 100,000 of these cases are fatal (The Lancet, 2017). The majority of these cases occur in developing countries in sub-Saharan Africa, South Asia and South East Asia (Gutiérrez et al., 2017; Longbottom et al., 2018). Despite Australia’s reputation for deadly snakes, snakebite envenoming is far less frequent on that continent. However, deaths do occur there every year because of snakebite and brown snakes (genus Pseudonaja) are the most “medically significant” snake taxa, being responsible for 41% of all deaths from snakebite envenoming between 2005 and 2015 (Johnston et al., 2017).

Patients who are envenomed by Pseudonaja spp. may exhibit a range of clinical symptoms, including venom-induced consumption coagulopathy (VICC), hypotension or cardiovascular collapse. Despite the presence of both pre- and post-synaptic neurotoxins in the venom, paralysis is only a rare consequence of brown snake envenoming, a fact known as the “brown snake paradox” (Barber et al., 2012). VICC is the most common severe symptom observed in cases of human envenoming by Pseudonaja spp., though hypotension and subsequent cardiovascular collapse are also potentially life-threatening (Allen et al., 2012). Unlike VICC, which is caused by the prothrombinase toxin pseutarin C interfering with the coagulation cascade (Rao and Kini, 2002), the mechanisms underlying the induction of hypotension and collapse following snakebite remain contentious, although pseutarin C has also been implicated in these pathologies (Chaisakul et al., 2015). Hypotensive effects of snake venoms have also been attributed to the release of dilator autacoids (Chaisakul et al., 2014), blockade of L-type voltage-gated Ca²⁺ channels (Yagami et al., 2013), presence of bradykinin-potentiating (Ferreira et al., 1970; Ianzer et al., 2004; Camargo et al., 2012) and natriuretic (Fry et al., 2005; St Pierre et al., 2005) peptides, or as a side effect of disseminated intravascular coagulation (Tibballs et al., 1992).

The venoms of Pseudonaja spp. are notably diverse in toxin content and highly variable–intra-as well as interspecific variations have been documented (Flight et al., 2006; Jackson et al., 2016; Reeks et al., 2016). Diversity exists both in terms of the variety of toxin classes present within a given venom, as well as within toxin classes, such as post-synaptically neurotoxic “3-finger toxins” (3FTx), which may exist in more than 40 variants within a single venom (Jackson et al., 2013). As well as the prothrombinase homolog pseutarin C (Rao and Kini, 2002), the venoms of Pseudonaja spp. contain phospholipases A₂ (PLA₂), including the presynaptic neurotoxin textilotoxin (Su et al., 1983); long chain and short chain 3FTx including the post-synaptic neurotoxins pseudonajatoxin a (Barnett et al., 1980) and b (Tyler et al., 1987); serine protease inhibitors including textilinin-1 and -2 (Masci et al., 2000); cysteine-rich secretory peptides; and vasoactive C-type natriuretic peptides (Judge et al., 2002; Birrell et al., 2006; Viala et al., 2015; Jackson et al., 2016; Reeks et al., 2016).

PLA₂ toxins exhibit diverse activities, including neurotoxicity (Su et al., 1983; Lambeau et al., 1989); myotoxicity (Harris and MacDonell, 1981; Dixon and Harris, 1996); procoagulant (Judge et al., 2002); synergistic/auxillary interactions with other toxins (Bougis et al., 1987; Mukherjee, 2010; Mora-Obando et al., 2014; Laustsen, 2016); and others, including both enzymatic and non-enzymatic activities (Kini, 2003). Inhibition of PLA₂ toxins, therefore, may have considerable therapeutic benefit for patients suffering from snakebite envenoming, including those envenomed by Australian brown snakes. Lewin et al. (2016) found that varespladib–a secretory phospholipase A₂ (sPLA₂) inhibitor–inhibited the phospholipase activities of 28 snake venoms. Further, rats treated with varespladib before or after envenoming had higher survival rates. Its pro-drug methyl-varespladib, when administered in conjunction with antivenom, also rescued mice administered otherwise lethal doses of Oxyuranus scutellatus venom (Lewin et al., 2018). These studies have revealed the potential of varespladib as an adjunct therapy, to be administered along with antivenom and/or additional small molecule therapeutics, for snakebite envenoming.

In the present study, we examined the cardiac, vascular, and sympathetic effects induced by five Pseudonaja spp. venoms in rat isolated atria and small mesenteric arteries. Where the brown snake venoms caused significant effects, the role of sPLA₂s in mediating these responses was assessed by treatment with varespladib in vitro. Further, the role of nitric oxide, prostaglandins or CGRP in venom-induced vascular relaxation was investigated.

The results of the present study highlight the variability of brown snake venoms, both in terms of their baseline toxicity and their relative inhibition with the sPLA₂ inhibitor varespladib. They also contribute additional evidence in support of the potential utilisation of varespladib as an adjunct therapy for snakebite envenoming. Perhaps their most important contribution is to the investigation of the cardiovascular effects of these venoms, which have largely been neglected, despite cardiac arrest being the most common cause of death following snakebite envenoming in Australia (Allen et al., 2012; Johnston et al., 2017). Few studies have concentrated on deconvoluting the causes of cardiovascular collapse following brown snake envenoming, with the notable exceptions of in vitro (Chaisakul et al., 2015) and in vivo (Chaisakul et al., 2013) studies of the effects of P. textilis venom. Interestingly, in the present study we found little (P. inframacula, P. nuchalis and P. textilis) to no (P. affinis) positive inotropic effects, and no effects on chronotropy, induced by the four brown snake venoms tested. These results corroborate those of the in vivo study of P. textilis venom (Chaisakul et al., 2013), as well as one conducted by the same team with isolated sPLA₂s from Oxyuranus scutellatus (taipan) venom (Chaisakul et al., 2014). Collectively, these results suggest that venom-induced cardiovascular collapse following envenoming by Australian snakes does not result from direct cardiac toxicity but is more likely to be mediated by effects in the vasculature.

Chaisakul et al. (2013) examined the pharmacology of whole P. textilis venom and found that it caused vasorelaxation. Our results agree with this finding for P. textilis and demonstrate that vasorelaxant activity is a marked effect of the venoms of four additional species of Pseudonaja (i.e., P. affinis, aspidorhyncha, inframacula and nuchalis), suggesting it is a widespread property of brown snake venoms. Inhibition with varespladib significantly attenuated the vasorelaxant effects of three of the venoms (P. nuchalis, textilis and aspidorhyncha), thus our results indicate that sPLA₂s of these species may functionally antagonise vasocontraction (vascular tone) in mammalian arteries. However, varespladib (1 µM) pretreatment did not significantly inhibit vasorelaxant effects of P. affinis or P. inframacula venoms, suggesting that these effects may be mediated by a combination of toxins which are present in differing ratios in each of the venoms. One of the limitations of our study was the protocol of using only a single venom concentration, rather than the construction of a concentration-response curve for each Pseudonaja venom. However, it has been previously reported for both P. textilis and Oxyuranus scutellatus (Papuan taipan) venoms that rapid desensitization occurs with repeated exposure of isolated arteries in vitro (Chaisakul et al., 2012; Chaisakul et al., 2013; Chaisakul et al., 2014). The use of only a single venom concentration per tissue was therefore considered the most rigorous pharmacological approach.

Although our study did not examine the cardiovascular effects of isolated toxin fractions, the sPLA₂-mediated vasorelaxant activity of Oxyuranus scutellatus (coastal taipan) venom has previously been investigated in a similar bioassay (Chaisakul et al., 2014). Oxyuranus is the sister genus to Pseudonaja and the venoms of both genera contain closely related toxins, a number of which are unique to the Oxyuranus/Pseudonaja clade (Jackson et al., 2016).

A fraction containing pseutarin C, the procoagulant toxin in P. textilis venom, was also found to induce a significant fall in mean arterial pressure without altering cardiac function (Chaisakul et al., 2015). In vitro experiments from the same study demonstrated that pseutarin C did not possess any vasorelaxant activity, further reinforcing the role of sPLA₂s in the induction of vasorelaxation in mesenteric arteries. Taken together, these results indicate that multiple components–in this case, a prothrombin activator and sPLA₂s (perhaps in addition to C-type natriuretic peptides also present in these venoms; St Pierre et al., 2006)–may contribute to the overall cardiovascular symptoms of snake envenomation. Notably, synergistic (potentiating) effects between PLA₂s and other venom components have been described previously in a number of snake venoms, including those of the cobras Naja mossambica and Hemachatus haemachatus (Louw and Visser, 1978), the pit viper Bothrops asper (Mora-Obando et al., 2014), and the taipan O. scutellatus (Herrera et al., 2016).

Another popular hypothesis about snake venom-induced vasorelaxation invokes the release of dilator autacoids, particularly histamine (Feldberg and Kellaway, 1937; Teixeira et al., 2003) and bradykinin (Ferreira et al., 1970; Xu et al., 2015), as the proximal cause. Although this has largely been investigated using the venoms of viperid snakes (family Viperidae), a study using crude O. scutellatus venom showed that pre-treating rats with the cyclooxygenase inhibitor indomethacin, but not atropine (muscarinic receptor antagonist), mepyramine (histamine H₁ receptor antagonist) or the nitric oxide synthase inhibitor L-NAME, could prevent vasodilation and sudden collapse (Chaisakul et al., 2012). PLA₂ enzymes cause the release of lysophospholipids and fatty acids via hydrolysing glycerophospholipids (Kini, 2003). Chaisakul et al. (2014) showed that sPLA₂-mediated hypotensive effects of O. scutellatus venom were inhibited in rats administered with indomethacin prior to venom administration, concluding significant release of dilator autacoids. An earlier study of the PLA₂ fraction from V. russelli venom showed that PLA₂ relaxed rat isolated aortic rings in an endothelium-independent manner, ruling out a role for the release of endothelial autacoids such as histamine and bradykinin (Huang and Lee, 1985). They found that inhibition of soluble guanylate cyclase with the early antagonist methylene blue partly attenuated the PLA₂-induced aortic relaxation, suggesting a role for cyclic GMP in this response. In rat isolated mesenteric arteries, the relaxation caused by Papuan taipan venom was shown by Chaisakul et al. (2012) to involve, in part, protein kinase A and large-conductance Ca²⁺-activated K⁺ channels. In our study, pretreatment with indomethacin, L-NAME or the CGRP receptor antagonist CGRP_8-37 did not abrogate the vasorelaxant effects of whole P. textilis venom in isolated mesenteric arteries, suggesting that the relaxation was not due to local release of dilator prostanoids or nitric oxide, or sensory nerve-released CGRP; the possible involvement of cyclic GMP formation, protein kinase A or large-conductance Ca²⁺-activated K⁺ channels was not tested and warrant further experimentation.

Although studies investigating the structure and activity of the main components of P. textilis venom–particularly pseutarin C and the presynaptically neurotoxic sPLA₂ textilotoxin–have been conducted for more than 50 years, the remaining members of the genus Pseudonaja have received far less attention. PLA₂s appear to be an almost ubiquitous component of Australian elapid snake venoms (Jackson et al., 2016) and have been detected in every Pseudonaja venom system investigated to-date (see e.g., Jackson et al., 2013; Jackson et al., 2016; Reeks et al., 2016). PLA₂s have been identified in P. nuchalis and P. aspidorhyncha venoms as “homologues of textilotoxin PLA₂ subunits” (Reeks et al., 2016); and in P. affinis venom as a “toxic PLA₂ variant” (Judge et al., 2002). The latter study assayed the activity of purified PLA₂s from P. affinis venom and found they induced contraction in rat tracheae–an activity like the venoms of O. microlepidotus and O. scutellatus, which was hypothesised by the authors to result from the downstream release of leukotrienes.

In the present study, only P. affinis and P. textilis venoms markedly inhibited sympathetic nerve-mediated vascular contractions of isolated mesenteric arteries. These effects were significantly inhibited by pretreatment with varespladib, indicating that PLA₂ toxins play an important role in inducing this sympatholysis. As no studies have been performed which assess the sympatholytic effects of isolated textilotoxin homologues, their involvement in the inhibition of sympathetic transmission is unknown. Furthermore, neither P. nuchalis nor P. aspidorhyncha venoms, which contain textilotoxin isoforms (Reeks et al., 2016), induced significant sympatholytic effects in our assays. These data raise several questions: whether only certain isoforms of textilotoxin can cause these effects in rat mesenteric arteries; if particular concentrations must be achieved (i.e., whether the absence of the effect for some venoms is a consequence of the relative abundance of this toxin type); if synergy with post-synaptic neurotoxins (i.e., 3FTx) is required; or indeed if this presynaptic neurotoxin is involved at all. As varespladib significantly attenuated the sympatholytic effects of P. affinis and P. textilis venoms, it is reasonable to infer that some forms of sPLA₂ act at synaptic terminals in vascular smooth muscle. Alternatively, they may act post-synaptically at the α₁-adrenoceptor, which would explain the lack of effects on sympathetic nerve stimulation-induced inotropy (mediated by β₁-adrenoceptors) in rat left atria–in contrast to the effects in blood vessels–in this study. Whether sPLA₂s affect the function of α₁-adrenoceptors in vascular smooth muscle is not known, however, PLA₂ has been shown to desensitize α₂-adrenoceptors in platelets (Rao and White, 1989), and in isolated membrane preparations (Cohen et al., 1985). This latter effect on α₂-adrenoceptors is not relevant to the current study as these receptors were irreversibly blocked in the mesenteric arteries used in the sympathetic nerve stimulation experiments. The precise reason for the differential effects in atria and arteries is therefore impossible to ascertain based on our data but represents an interesting question for further investigation.

Previous studies indicate that presynaptic neurotoxins in snake venoms can deplete acetylcholine release from motor neurons in skeletal muscle via the destruction of synaptic terminals (Chang et al., 1973; Su and Chang, 1984; Cavalcante et al., 2017) and vesicular membranes (Montecucco et al., 2009). To complicate the matter further, whilst a proteomic study of P. affinis venom did not detect any textilotoxin homologues (Judge et al., 2002), multiple studies (Judge et al., 2002; Jackson et al., 2016) have detected an abundance of proteins in the 12–14 kDa range–the expected size of venom sPLA₂s following decomplexation–in P. affinis venom. Thus, it is possible that the sympatholytic effects induced by this venom are caused by unknown sPLA₂s. More detailed investigation into the composition of Pseudonaja venoms is therefore warranted to ascertain which PLA₂ toxins are responsible for the sympatholytic effects observed in the present study.

Whilst all five brown snake venoms examined in the present study induced vasorelaxant effects, only two venoms–P. textilis and P. affinis–induced marked sympatholytic effects. The activity of the venoms (and their relative inhibition with varespladib) thus varied amongst species at the level of gross activity, as well as more subtly in terms of the potency of each activity. These findings corroborate earlier studies (e.g., Flight et al., 2006; Jackson et al., 2016; Reeks et al., 2016) that have uncovered considerable variability (at both inter- and intraspecific levels) in the composition and activity of brown snake venoms. Thus, our data contribute to the ongoing investigation of these biologically fascinating and medically significant venoms. However, it would be premature at this stage to conjecture about the evolutionary origins (i.e., the function sensu stricto), or possible clinical relevance of the variability we have reported here. Based on their inhibition with the sPLA₂ inhibitor varespladib, we infer that PLA₂ toxins in P. aspidorhyncha, P. nuchalis and P. textilis venoms are involved in brown snake venom-induced vasorelaxation and the sympatholytic effects of P. affinis and P. textilis venoms. Observed vasodilatory effects were likely endothelium-independent, although determination of the exact mechanism of action was outside the scope of this study. Given that little to no direct cardiac effects were observed in our assays, the cause of cardiovascular collapse following brown snake envenoming is likely due to a fall in mean arterial pressure unrelated to cardiac function. Further proteomic and pharmacological studies remain necessary to resolve the mechanistic details of this dangerous venom-induced pathology. Finally, our results are promising regarding the potential role of varespladib as an initial (pre-referral) and/or adjunct (in combination with antivenom) therapeutic agent for brown snake envenoming.