Introduction

The New World coral snakes consist of three genera: Leptomicrurus (five species), Micruroides (one species), and Micrurus (approx. 70 species) (Roze 1996). The two less speciose genera are restricted in their geographic range; Leptomicrurus is only found in close proximity to the Amazon basin (Campbell and Lamar 1989) while Micruroides is restricted in the United States to the southern half of Arizona and extreme southeastern New Mexico, south into the northwestern Mexican states of Sonora, extreme southwestern Chihuahua, and Sinaloa, with an insular population occurring on Isla Tiburón (Campbell and Lamar 2004). In contrast, snakes of the genus Micrurus distribution ranges from the southeastern USA to Argentina (Jorge da Silva and Sites 2001).

Bites from most of these snakes, with the exception of the very small Micruroides, are considered potentially life-threatening due to deaths reported from species across their geographical range (Bucaretchi et al. 2016; Manock et al. 2008; Morgan et al. 2007). The venom is considered to be primarily neurotoxic, and the most common cause of death is paralysis-induced hypoxia (Vital-Brazil 1987) highlighting the importance for each country that contains coral snakes in having access to antivenoms capable of neutralizing the neurotoxic effects of coral snake envenomation. Currently, the most effective treatment for coral snake envenomations is the administration of specific antivenom coupled with direct treatment of symptoms such as pre-synaptic neurotoxicity not neutralized by antivenom (Peterson 2006). As antivenoms are produced by inoculating livestock (generally sheep or horses) against venom from a particular species or selection of species and purifying the resulting antibodies (Angulo et al. 1997; Rojas et al. 1994), antivenoms can be partially or wholly ineffective against the venom from species not included in production (Tanaka et al. 2010; Boyer et al. 2015). Nonetheless, experimental (Bolanos et al. 1975; Silva et al. 2016) as well as clinical evidence of cross-neutralization has been reported even for non-related snake species (Isbister et al. 2014).

The reason antivenoms sometimes fail when treating bites from coral snake species, whose venom was not included in the production of antivenom, has recently become more salient as studies have revealed that the variation in coral snake venoms is deeper and more widespread than previously presumed. In particular, Aird and Jorge da Silva (1991) provided evidence based on enzymatic activities that some coral snake venoms more closely resemble the venoms of more distantly related snakes such as Bungarus, Bothrops, or Naja than other coral snakes. Proteomic and transcriptomic analyses of coral snakes from across the Americas indicate that, broadly speaking, some coral snake venoms contain three-finger toxins (3FTx) as their most abundant components while others are composed primarily of phospholipase A2 (PLA2) (Fernandez et al. 2015; Rey-Suarez et al. 2016, Vergara et al. 2014; Lomonte et al. 2016a).

There are currently four antivenoms primarily used in the treatment of New World coral snake bites (Table 1). The production of North American coral snake antivenom (NACSA) was halted in 2008, and remaining stocks have diminished while repeatedly having their effectiveness tested and expiry dates extended (Wood et al. 2013); a single lot was released in 2016 for limited distribution in the USA (Supplementary File 1). The general shortage of coral snake antivenom in the USA has led to the proposal that the Mexican Coralmyn® coral snake antivenom (Bioclon, Mexico) be used in its place. Due to the diversity of coral snakes in each country where they are found being greater than the diversity used in production, testing the cross reactivity of antivenom on species that were not part of its manufacture remains medically important.

Table 1 Currently available coral snake antivenoms
Full size table

The use of commercially available coral snake antivenoms is effective in neutralizing the lethal doses (LD50) of Micrurus fulvius and Micrurus tener venoms (Sánchez et al. 2008). Sánchez et al. (2008) tested the efficacy of two types of commercially available coral snake antivenom, North American coral snake antivenom (NACSA; Wyeth, United States) and Coralmyn® coral snake antivenom (Bioclon, Mexico), in order to ascertain their ability to neutralize the lethal components in both M. fulvius and M. tener, two American coral snake species. While the use of Coralmyn® was effective in neutralizing the LD50 doses of both M. fulvius and M. tener venoms, the use of NACSA was ineffective in neutralizing the lethality of three varying doses of venom from M. tener. In South America, a study by Camargo et al. (2011) investigated the efficacy of coral snake antivenom against the venom of Micrurus pyrrhocryptus. It was shown that both the commercial coral snake antivenom (Instituto Butantan, Brazil) and the specific antivenom raised against the venom of M. pyrrhocryptus were effective in neutralizing its neurotoxic effects. However, it remained to be seen whether Coralmyn® coral snake antivenom, which was raised against Micrurus nigrocinctus nigrocinctus (black-banded coral snake) venom, would be effective against M. pyrrhocryptus venom. Moreover, as the use of Coralmyn® was effective against the venoms of M. fulvius and M. tener in a murine model, it was hypothesized that the antivenom would be equally as effective in an in vitro preparation. The current study is of clinical significance as the production of NACSA was discontinued by Wyeth in 2008 (Wood et al. 2013) and, thus, it is imperative that the efficacy of other coral snake antivenom is determined.

Historically, murine LD50 studies have been performed in order to determine relative neurotoxicity of venoms (WHO 2010). In some circumstances, however, these can be supplemented by in vitro studies such as that of the chick biventer cervicis nerve-muscle preparation (Hodgson and Wickramaratna 2002). Using this preparation, the neurotoxicity of venoms can be examined by observing the time in which the venom/toxins produces 90% inhibition of nerve-mediated twitches (t 90 values). However, there may be extreme variation between the synapsid (mouse) and diapsid (bird) results if the venoms are evolutionarily shaped to be taxon-specific (Fry et al. 2003; Heyborne and Mackessy 2013; Lumsden et al. 2004a, b, Lumsden et al. 2005a; Utkin et al. 2015). Such specificity has been described in vivo for the venoms of M. pyrrhocryptus and M. spixii (Jorge da Silva and Aird 2001). There, they determined LD50 values of whole venoms on prey and non-prey models and showed that they are, respectively, six and two times more lethal to the snake Lyophis typhlus than to mice. The case of M. tener, on the other hand, appears to be somewhat different because, even though a few apparently snake specific toxins have been isolated from it, the whole venom has no proven specificity towards snakes (Bénard-Valle et al. 2014).

Chick biventer assays are sometimes utilized in predicting clinical efficacy of antivenoms; however, chick biventer assays are not an antivenom challenge as the antivenom is pre-incubated with the venom prior to the administration. Thus, efficacy under this protocol does not indicate clinical efficacy. However, inability of the antivenom to neutralize under such an ideal scenario is highly likely to be predictive of clinical failure.

In the present study, we examined the neurotoxicity and antivenom cross-reactivity of venoms from Micruroides euryxanthus, M. fulvius, M. pyrrhocryptus, Micrurus spixii, and M. tener representing a broad phylogenetic diversity of coral snake species (Fig. 1). Of these species, the latter three are found in the USA. We used the Mexican coral snake antivenom, Coralmyn®, in our experiments as it is the most readily available substitute should the remaining supplies of NACSA prove insufficient. Given the phylogenetic distribution of PLA2 and 3FTx dominated venoms, we would expect Coralmyn® to be more effective against M. fulvius and M. tener venoms than the other venoms. This is due to M. fulvius and M. tener belonging, along with M. nigrocinctus (the species used in the production of Coralmyn®), to the PLA2-abundant group (Fernandez et al. 2015).

Fig. 1
figure 1

Phylogenetic relationships of species studied and millions of years of separation (following Lee et al. 2016)

Full size image

Methods

Venom Preparation and Storage

M. euryxanthus (AZ, USA collected by CC), M. fulvius (FL, USA supplied by Miami Serpentarium), M. pyrrhocryptus (IW Surinam captive specimen), M. spixii (Brazil, supplied by Miami Serpentarium), and M. tener (TX, USA, supplied by Miami Serpentarium) freeze-dried venoms were prepared in MilliQ-H2O and stored at −80 °C until required.

Drugs and Chemicals

The following drugs and chemicals were used: acetylcholine chloride (Sigma), carbamylcholine chloride (Sigma), potassium chloride (Ajax Finechem), (+)-tubocurarine chloride (Sigma), and Coralmyn® coral snake antivenom (Bioclon, Mexico; Batch #B-2D-06/2004). Stock solutions of drugs were made up MilliQ-H2O unless otherwise specified.

Neurotoxicity and Antivenom Studies

Male chicks (4–10 days) were sacrificed by CO2 and exsanguination. Both chick biventer cervicis nerve-muscle preparations were isolated and mounted on wire tissue holders under 1-g resting tension in 5-ml organ baths containing physiological salt solution (NaCl, 118.4 mM; KCl, 4.7 mM; MgSO4, 1.2 mM; KH2PO4, 1.2 mM; CaCl2, 2.5 mM; NaHCO3, 25 mM; and glucose, 11.1 mM), maintained at 34 °C, and bubbled with 95% O2/5% CO2. Indirect twitches were evoked by electrical stimulation of the motor nerve (supramaximal voltage, 0.2 ms, 0.1 Hz) using a Grass S88 stimulator (Grass Instruments, Quincy, MA). d-Tubocurarine (10 μM) was added, and the subsequent abolition of twitches confirmed selective stimulation of the motor nerve, after which thorough washing with physiological salt solution was applied to re-establish twitches. In the absence of electrical stimulation, contractile responses to acetylcholine (ACh; 1 mM for 30 s), carbachol (CCh; 20 μM for 60 s), and potassium (KCl; 40 mM for 30 s) were obtained prior to the addition of the venoms and at the conclusion of the experiment. Venoms were left in contact with the preparation for a maximum of 3 h to test for slow developing effects. Efficacy of Coralmyn® coral snake antivenom (10 units/mL) was assessed via a 10-min pre-incubation in the organ bath.

Data Analysis and Statistics

Twitch tension was measured from the baseline in 2-min intervals. Responses were expressed as a percentage of twitch tension prior to the addition of the venom. Contractile responses to agonists obtained at the conclusion of the experiment were measured and expressed as a percentage of the response obtained prior to the addition of venom. The time taken to inhibit 90% of twitch contractions (i.e., t 90) was measured as a quantitative means of measuring neurotoxicity. Values for t 90 were measured by the time elapsed to reach 10% twitch tension amplitude following addition of venom. Where indicated, a two-way analysis of variance (ANOVA) or a paired t test was used to determine statistical significance of responses. Statistical analysis was performed using the Prism 5 (GraphPad Software, San Diego, CA, USA) software package. Unless otherwise indicated, data are expressed as mean ± S.E.

Animal Ethics

All animal experiments used in this study were approved by the SOBS-B Monash University Animal Ethics Committee.

Phylogenetic Comparative Analyses

A phylogeny was assembled using Lee et al. 2016, as this is currently the most comprehensive phylogeny available for Micrurus and was used for all further analyses conducted in R v3.2.5 (R-Core-Team 2011) using the ape package (Paradis et al., 2004) for general handling of phylogenetic and trait data. Ancestral states were estimated and reconstructed over the tree in order to visualize the changes in neurotoxicity and antivenom efficacy across the phylogeny. Note that we only included five of 76 species of New World coral snakes (~7%) and these were widely dispersed in the phylogeny, so accurate estimates of ancestral states are unlikely to be estimated here. Nevertheless, such reconstructions still provide a visualization of the minimum amount of evolutionary change in these traits across the clade. The continuous functional traits were reconstructed by maximum likelihood in the contMap function in phytools (Revell 2012).

We then fit pGLS models (Symonds and Blomberg 2014) in caper (Orme et al. 2015) to test for relationships. We restricted the venom-antivenom model to a single explanatory variable, and we arcsine square-root-transformed the response variable to improve model fit (since the functional activity values are bound between 0 and 1).

Results and Discussion

The venoms of M. euryxanthus, M. fulvius, M. pyrrhocryptus, M. spixii, and M. tener (10 μg/mL) caused a concentration-dependent blockade of nerve-mediated twitches in the chick biventer cervicis nerve-muscle preparation (Figs. 2-7, n = 3). All venoms (10 μg/mL) significantly inhibited contractile responses to exogenous acetylcholine (ACh, 1 mM) and carbachol (CCh, 20 μM), but not KCl (40 mM) (Figs. 2-7, n = 3, P < 0.05).

Fig. 2
figure 2

Comparative effects of Micruroides euryxanthus, Micrurus fulvius, Micrurus pyrrhocryptus, Micrurus tener, and Micrurus spixii venoms (10 μg/mL) on nerve-mediated twitches in the chick biventer cervicis nerve-muscle preparation (n = 3)

Full size image
Fig. 3
figure 3

Effects of Micruroides euryxanthus venom (5–10 μg/mL, n = 3) on a nerve-mediated twitches and b responses to exogenous ACh (1 mM), CCh (20 μM), and KCl (40 mM) of the isolated chick biventer cervicis nerve-muscle preparation in the presence and absence of Coralmyn® coral snake antivenom (10 units/mL added prior to 10-μg/mL venom concentration, n = 3). *P < 0.05, significantly different from contractile response of the initial, paired t test

Full size image
Fig. 4
figure 4

Effects of Micrurus fulvius venom (5–10 μg/mL, n = 3) on a nerve-mediated twitches and b responses to exogenous ACh (1 mM), CCh (20 μM), and KCl (40 mM) of the isolated chick biventer cervicis nerve-muscle preparation in the presence and absence of Coralmyn® coral snake antivenom (10 units/mL added prior to 10-μg/mL venom concentration, n = 3). *P < 0.05, significantly different from 10-μg/mL venom alone, two-way ANOVA. *P < 0.05, significantly different from contractile response of the initial, paired t test

Full size image
Fig. 5
figure 5

Effects of Micrurus pyrrhocryptus venom (5–10 μg/mL, n = 3) on a nerve-mediated twitches and b responses to exogenous ACh (1 mM), CCh (20 μM), and KCl (40 mM) of the isolated chick biventer cervicis nerve-muscle preparation in the presence and absence of Coralmyn® coral snake antivenom (10 units/mL added prior to 10-μg/mL venom concentration, n = 3). *P < 0.05, significantly different from contractile response of the initial, paired t test

Full size image
Fig. 6
figure 6

Effects of Micrurus spixii venom (5–10 μg/mL, n = 3) on a nerve-mediated twitches and b responses to exogenous ACh (1 mM), CCh (20 μM), and KCl (40 mM) of the isolated chick biventer cervicis nerve-muscle preparation in the presence and absence of Coralmyn® coral snake antivenom (10 units/mL added prior to 10-μg/mL venom concentration, n = 3). *P < 0.05, significantly different from contractile response of the initial, paired t test

Full size image
Fig. 7
figure 7

Effects of Micrurus tener venom (5–10 μg/mL, n = 3) on a nerve-mediated twitches and b responses to exogenous ACh (1 mM), CCh (20 μM), and KCl (40 mM) of the isolated chick biventer cervicis nerve-muscle preparation in the presence and absence of Coralmyn® coral snake antivenom (10 units/mL added prior to 10-μg/mL venom concentration, n = 3). *P < 0.05, significantly different from contractile response of the initial, paired t test

Full size image

Based on the t 90 values obtained, the observed rank order of neurotoxicity for the New World coral snakes is M. tener > M. spixii ≥ M. pyrrhocryptus > M. euryxanthus > M. fulvius (Table 2; Fig. 2). This contrasts with the rank order of lethality based on the murine LD50 values previously reported by Sánchez et al. (2008); M. spixii > M. fulvius > M. pyrrhocryptus > M. tener. However, it is important to note the use of t 90 values as a measure of toxicity differs from that of LD50 values in several important ways. Murine LD50 values determine the concentration required to kill 50% of a population of mice (i.e., quantity) over an extended period of time (24 h), whereas t 90 values focus on the time required for the venom to exert its flaccid paralysis effects (i.e., time). As coral snakes mostly consume reptiles (diapsid) and use their venom to rapidly subjugate their prey (Roze 1996), murine (synapsid) LD50s may not be particularly informative from an evolutionary point of view; however, they are likely a better measure of potential lethality to humans (also synapsid) in untreated envenomings (Bénard-Valle et al. 2014). Differences between the two measures of toxicity are likely to be due to the fact that the crude venom possesses other toxic components apart from neurotoxins; such as cardiotoxins, myotoxins, phospholipases, and pro- and anticoagulant factors, all of which act either synergistically or independently on their respective targets to contribute to the lethality of the venom over such a prolonged period of study. This is not the case in t 90 studies as the in vitro neuromuscular preparations’ only factor in neurotoxins and myotoxins, the two types of toxins likely to inhibit twitch height. Despite this, the knowledge obtained from t 90 values is extremely valuable given that the primary symptom of envenoming by many elapids, including coral snakes, is neuromuscular paralysis.

Table 2 Snake venom neurotoxicity in the chick biventer cervicis ranked by t 90 (data from other studies shown for comparison)
Full size table

To the best of our knowledge, this study represents the first analysis of the neuromuscular activity of the venom from M. euryxanthus. Here, the relatively low t 90 value as well as the complete abolition of the response to exogenous ACh and CCh (Fig. 3) suggest a postsynaptic effect.

Prior incubation of Coralmyn® coral snake antivenom (10 units/mL) did not significantly neutralize the neurotoxic effects of M. euryxanthus (P = 0.5662), M. pyrrhocryptus (P = 0.5747), M. spixii (P = 0.9520), and M. tener (P = 0.8218) venoms (Figs. 3, 4, 5, 6, and 7, n = 3). However, Coralmyn® coral snake antivenom significantly neutralized the effects of M. fulvius venom (P = 0.0005) (Fig. 4a, n = 3, P < 0.05) and prevented the inhibition of contractile responses to exogenous ACh (P = 0.0001) and CCh (P = 0.0001) (Fig. 4b, n = 3, P < 0.05). While it is important to note that pre-incubation of antivenom in an in vitro setting does not necessarily correlate to clinical success, failure to neutralize the neurotoxic effects of these venoms in such ideal conditions may potentially indicate clinical failure.

Against the venoms of the two South American coral snakes (M. spixii and M. pyrrhocryptus) and the Arizona coral snake (M. euryxanthus), Coralmyn® was ineffective in neutralizing their neurotoxicity. Although the response to the venom of M. pyrrhocryptus was marginally delayed in the presence of antivenom, the venom was not sufficiently neutralized for it to be deemed effective. However, even adjusting for relative neurotoxicity, the antivenom was clearly most effective against M. fulvius than all other species (Fig. 8).

Fig. 8
figure 8

Ancestral state reconstructions over branches for a neurotoxicity t 90, whereby lower values indicate higher potency and b relative antivenom neutralization based upon t 90 relative shift, whereby higher values indicate greater relative antivenom neutralization. Bars indicate 95% confidence intervals for the estimate at each node; note that due to the high dynamicity of venom evolution, these quickly become broad as you move down the tree. Phylogeny follows Lee et al. (2016)

Full size image

The neutralization of M. pyrrhocryptus venom by a South American antivenom raised against Micrurus frontalis and M. corallinus was tested using also a chick BC preparation and found to be very effective (Camargo et al. 2011). These results, together with the ones from the present study, suggest that PLA2-predominant venoms, such as M. nigrocinctus, are not effective immunogens when attempting to neutralize 3FTx-predominant venoms. This scenario has been tested and was shown to be equally ineffective by Lomonte et al. (2016a, b) using an in vivo approach.

The results of the antivenom study with M. tener venom were unexpected given that Coralmyn® coral snake antivenom was raised against the venom of M. nigrocinctus nigrocinctus, a species that is both phylogenetically and geographically nearby to M. tener (Campbell and Lamar 2004). However, note that reptile venom often evolves at very high rates (as reflected in the very broad confidence intervals in Fig. 8), and so it may not be surprising to find that close phylogenetic relatives may still have venoms that differ in important properties for antivenom efficacy.

A likely explanation for the observed disparity in neutralization comes from an analysis of the composition of these two venoms and the immunogen M. nigrocinctus. While M. nigrocinctus venom has a high proportion of both PLA2 and 3FTx molecules, a study using an equine monospecific antivenom raised against this same venom showed that recognition of antibodies is mainly directed towards PLA2 (Fernández et al. 2011, Lomonte et al. 2016b). Similarly, Coralmyn® has been shown to have low antibody titers against 3FTx predominant venoms such as Micrurus surinamensis (Table 3). Given that the lethality of M. fulvius venom is determined by two main neurotoxic PLA2s and has no lethal 3FTxs (Vergara et al. 2014), neutralization of these molecules is enough to neutralize the lethality of the whole venom. A similarly interesting case has been recently described for the venom of Micrurus clarki which is devoid of 3FTx that are lethal to mice and has been proven to be very well neutralized by the antivenom from Clodomiro Picado Institute in Costa Rica which is also raised against M. nigrocinctus (Lomonte et al. 2016b). Conversely, M. tener has at least two highly lethal 3FTx molecules that have been shown to have a relevant role on the toxicity of the whole venom (Bénard-Valle et al. 2014). A similar case of poor neutralization of coral snake venom due to a highly lethal 3FTx has been described for the venom of Micrurus laticollaris (Carbajal-Saucedo et al. 2013).

Table 3 Venom lethality to mice and Coralmyn antibody titers for various coral snake species
Full size table

A study by Sánchez et al. (2008) indicated effective neutralization of both M. fulvius and M. tener venoms with Coralmyn® coral snake antivenom. It was noted that Coralmyn® was more effective in neutralizing M. fulvius venom rather than M. tener venom on an LD50 basis. The venom of M. fulvius has been previously described to be well neutralized by Coralmyn and an antivenom against the venom of the tiger snake Notechis scutatus (Wisniewski et al. 2003). This was contrary to the results of the current study in which only the venom of M. fulvius was neutralized, potentially suggesting that an insufficient dosage of antivenom was utilized for the neutralization of M. tener in the present study. The small proportion of 3FTx recognizing antibodies in Coralmyn® could be enough to neutralize the lethal 3FTx of M. tener but a high dosage of antivenom would be needed, probably explaining these differences.

As mentioned before, M. euryxanthus appears to be a 3FTx predominant venom with a mainly postsynaptic activity. This observation is reinforced by the very low antibody titers observed on Coralmyn® against this venom (Table 3).

This study has shown that the neurotoxic potency of representative species of both North and South American coral snakes varies greatly and that Coralmyn® coral snake antivenom fails to effectively neutralize their neurotoxic effects with the exception of M. fulvius venom. These findings suggest that venom variation between coral snake species may be of even greater clinical importance than geographical location and the 3FTx/PLA2 venom dichotomy would predict.