Snakebite is a major public health issue in the rural tropics. Antivenom is the only specific treatment currently available. We review the history, mechanism of action and current developments in snake antivenoms. In the late nineteenth century, snake antivenoms were first developed by raising hyperimmune serum in animals, such as horses, against snake venoms. Hyperimmune serum was then purified to produce whole immunoglobulin G (IgG) antivenoms. IgG was then fractionated to produce F(ab) and F(ab′)2 antivenoms to reduce adverse reactions and increase efficacy. Current commercial antivenoms are polyclonal mixtures of antibodies or their fractions raised against all toxin antigens in a venom(s), irrespective of clinical importance. Over the last few decades there have been small incremental improvements in antivenoms, to make them safer and more effective. A number of recent developments in biotechnology and toxinology have contributed to this. Proteomics and transcriptomics have been applied to venom toxin composition (venomics), improving our understanding of medically important toxins. In addition, it has become possible to identify toxins that contain epitopes recognized by antivenom molecules (antivenomics). Integration of the toxinological profile of a venom and its composition to identify medically relevant toxins improved this. Furthermore, camelid, humanized and fully human monoclonal antibodies and their fractions, as well as enzyme inhibitors have been experimentally developed against venom toxins. Translation of such technology into commercial antivenoms requires overcoming the high costs, limited knowledge of venom and antivenom pharmacology, and lack of reliable animal models. Addressing such should be the focus of antivenom research.
There are 5.8 billion people living in areas where venomous snakebites are a relatively common occurrence, and it is estimated that 1.8 to 2.7 million envenomings and 90 000 to 137 880 deaths occur each year. Snakebite is therefore a major public health issue, particularly in South and South-east Asia, Sub-Saharan Africa and Latin America [1–3]. Those mostly affected from snakebite are young and middle-aged farmers from rural and poor farming communities in the above regions . The World Health Organization now recognizes snakebite as a neglected tropical disease of category A, considering the enormous suffering associated with the condition and the politically and socio-economically disadvantaged nature of communities that are affected [5,6].
Snake venom is a mixture of many toxins, non-enzymatic and enzymatic proteins, with a broad range of pharmacological properties. In humans these result in a number of different acute envenoming syndromes, including local tissue damage at the bite site due to the spread of venom locally, and effects due to systemic spread of venom such as venom-induced consumption coagulopathy, neuromuscular paralysis, acute kidney injury, cardiovascular collapse and myotoxicity [7–11]. Some of the acute effects may lead to permanent effects such as amputations, blindness and hypopituitarism, while the traumatic experience of the snakebite can lead to psychological effects . Many factors such as the venom composition, injected venom amount, characteristics of the bite, first-aid and co-morbidities of the snakebite patient determine the severity of envenoming syndromes.
Snake antivenom is currently the only available specific treatment for snakebite. It was initially developed in the late nineteenth century. Despite the lack of high quality evidence for clinical effectiveness in treating the major effects of snake envenoming [13,14], it is considered the most important intervention for snakebite globally [6,15]. However, there are major issues in the adequate supply of antivenom , so it is included in the World Health Organization's list of essential drugs. Snake antivenoms are traditionally developed as polyclonal antibodies from large mammals (horse, sheep and goats), which are raised against the venom of one snake species (monovalent) or a mixture of the venom of several snake species (polyvalent) .
This mini-review focuses on the current developments in snake antivenoms and the mechanism of action of antivenoms as therapeutic interventions for snakebite.
Early and current antivenoms
Behring and Kitasato developed the theory of humoral immunity with their work on tetanus immunity in 1890 . Following this, Phisalix and Bertrand identified an ‘anti-toxic’ substance in the blood of guinea pigs that had been vaccinated against heat-treated viper venoms. They concluded that the ‘anti-toxic’ substance found in the blood resulted from a reaction within the animal against the venom [19,20]. They further suggested the possibility of ‘anti-venomous serotherapy’ in humans in 1894. Possibly inspired by this previous work on the antisera against tetanus and diphtheria toxins, Albert Calmette studied ways of raising serum against monocellate cobra (Naja kaouthia) venom [20,21]. In 1895, Calmette successfully produced the first commercial antiserum for humans against Indian cobra (Naja naja) venom in horses. However, Calmette believed that a single mode of toxic effects of snake venom exists and believed that his antiserum was able to treat envenoming by any snake.
In 1897, Charles Martin demonstrated the inability of Calmette's cobra antisera to neutralize Australian red-bellied black snake (Pseudechis porphyriacus) and common tiger snake (Notechis scutatus) venoms, which suggested Calmette's initial claim of a universal antivenom was not true. Furthermore, Martin thought that the interaction between the venom and the antiserum is direct (later understood as antibody-antigen interaction), while Calmette believed that administration of the antiserum indirectly provides protection, by inducing a reaction in cells of the body [19,22,23]. The next year, Vital Brazil found that Calmette's cobra antisera was also ineffective in neutralizing the venoms of south American pit-vipers, Crotalus durissus terrificus and Bothrops jararaca. This demonstrated the specificity of antiserum and the need for antivenoms to be raised against the snakes in question polyvalent . He also observed two patterns of toxic effects in these two snake species. Envenoming by C. d. terrificus caused neurotoxicity and less severe local effects, while B. jararaca envenoming caused local effects without neurotoxicity. At the Butantan Institute, Vital Brazil produced anti-bothroptic and anti-crotalic monovalent antisera, as well as antisera against a mixture of both venoms (first polyvalent antivenom), in mules and horses [19,24,25]. He subsequently introduced a successful program for distributing antivenoms across Brazil by setting up an antivenom-for-snakes exchange program, in which the farmers can exchange snakes for antivenom [24,26]. Subsequently, the United States, Australia, Costa Rica and South Africa, and later several other countries, started antivenom production. The complete history of the development of snake antivenoms is described in detail by Pucca et. al. .
Almost 125 years after the development of the first commercial antivenom, there are still no other specific therapies available for treating snakebite. Early antivenoms were unpurified animal serum raised against snake venoms (serum antivenimeux). Since then, the purification procedures for commercial antivenoms have improved with the expansion of human knowledge on biotechnology and immunology. This included purification of hyperimmune serum to produce whole immunoglobulin G (IgG) antivenoms and the fractionation of IgG to produce F(ab) and F(ab′)2 antivenoms (Figure 1). These developments were mainly aimed at reducing early adverse reactions and serum sickness to antivenoms, because of their foreign animal origin. In addition to the more common horse-derived antivenoms, avian antivenoms from chicken eggs (IgY) have been investigated due to the more ethical and cost-effective production procedures, and potentially increased safety [27,28]. Camelid IgG and non-IgG components have also been considered as potential alternative antivenoms, because they are considered less reactogenic and more thermostable (Figure 1) [29–33].
Schematic representation of various antibodies and fragments used in antibody research: mammalian IgG and avian IgY antibodies consist of two heavy chains (shades of blue) and two light chains (shades of green) and the camelid heavy antibodies consist of two heavy chains (shades of blue).
Recent developments towards future antivenoms
Over the last few decades, our understanding of venom composition has improved significantly . With the development of improved pharmacological techniques, the functional aspects of different toxin groups relevant to envenoming syndromes in humans have been better understood [37–41]. In addition, proteomic and transcriptomic approaches have been used to investigate venom composition and variation (venomics), which has revolutionized our understanding of snake venom [42–47].
Conventionally, the selection of immunogen venoms for development of antivenoms has been based on the toxic-profile of the particular snake envenoming and the taxonomic distribution of the target snakes . Careful selection of the venoms in the immunization mixture is essential for antivenom production in order to maximize the coverage of clinically important toxin groups and snakes in a particular region [34,35,48–50]. Proteomic techniques have been used to identify the various venom proteins that contain the epitopes recognized by the antivenom molecules (antivenomics). This has made it possible to understand the immunorecognition profile of antivenoms [51,52]. This has lead to the integration of the toxicological profile of a venom and its composition to identify and rank its medically relevant toxins (toxicovenomics) [53–55].
There are some major limitations with current animal-derived antivenoms, such as poor effectiveness against venom-induced local effects, containing a large proportion of clinically irrelevant antibodies, frequent allergic reactions, complex manufacturing processes that depend on two biological systems (snake and the immunized animal), and high batch-to-batch variation [56,57]. Developments in immunotherapy used for modern vaccines and targeted cancer therapy could potentially revolutionise antivenom therapy. Unfortunately, this is unlikely to change in the near future, because of the lack of funding to support the initial expense in incorporating these new recombinant technologies into antivenom production . However, there have been a number of promising developments in antivenom research and it will be important to test these in clinical trials.
An alternate approach to polyclonal antibodies, is developing monoclonal antibodies against the key toxins in snake venoms. Studies of high-affinity rodent monoclonal antibodies against snake toxins were first undertaken three decades ago . Unfortunately, these animal-derived monoclonal antibodies have many of the same drawbacks as traditional antivenoms, being of animal origin.
Improved biotechniques over the last few decades has led to the development of chimeric, humanized and human antibodies. This has in turn influenced the development of antibodies against snake toxins . Morine et al.  developed human monoclonal antibodies (HuMAb) based on hybridomas against metalloproteinase HR1a from Protobothrops flavoviridi, using transgenic mice. Using phage display technology, Laustsen et al.  developed fully human monoclonal IgGs against dendrotoxins from Dendroaspis polylepis (the black mamba), which provided full protection in mice when the antibodies were co-administered with toxins. A problem with monoclonal antibodies is that being directed at single targets (monoclonal), they are not able to neutralize the large number of different toxins in snake venoms, an important reason for the efficacy of polyclonal antibodies .
In addition to IgG, F(ab′)2 and Fab, recombinant antibody fragments, such as camelid nanobodies (VHHs) [61–63] and single-chain variable fragments (scFvs) [64–66] have demonstrated efficacy in neutralizing snake venoms (Figure 1). The advantage of antibody fragments, such as scFvs is that they are less-immunogenic and have higher tissue penetration. These fragments may be useful in treating paralysis and local effects in envenoming, due to their high tissue penetration. Furthermore, to reduce the immunogenicity, human scFvs, instead of murine scFvs, have been successfully tested against venom toxins [67–69]. However, they have a shorter serum half-life in vivo , meaning they are likely to require repeat doses when used in humans.
A different approach has been to investigate the inhibition of venom enzymes, such as phospholipases A2 and protease inhibitors. Varespladib (LY315920), a phospholipase A2 (PLA2) inhibitor, increased the survival of mice against Micrurus fulvius and Vipera berus venom compared with control mice not treated with varespladib . Other venom enzyme inhibitors have been tested against snake venom toxins. Laustsen et. al. provide a comprehensive list of various non-antibody proteins tested against various snake toxins . Since these small molecule inhibitors target only one group of toxins, they are unlikely to be a stand-alone therapy, similar to monoclonal antibodies, because of the range of toxic effects of snake venoms. Incorporating them into ‘oligoclonal antibody mixes’, together with other mixtures of selected human monoclonal antibodies may be a novel approach to producing future antivenoms that overcome some of the drawbacks of current antivenoms .
With many types of monoclonal antibodies and their fragments being tested successfully against snake toxins, this is a promising way forward for future development of antivenoms. Since the production costs of recombinant antibodies continue to decrease with the introduction of newer cost effective systems, antivenoms based on recombinant antibodies could be practically affordable in the future . One major problem in improving antivenoms is that snake envenoming is mainly a health problem in poorer rural areas of the world, where large pharmaceutical companies are unlikely to invest in the initial high cost of development.
Pharmacokinetics of antivenoms
The major issue with the treatment of snake envenoming is that there is a time lag between the snakebite and the administration of antivenom. A number of factors contribute to this lag time, the main delay being the time for the patient to get from where the bite occurs to a health care facility that stocks antivenom. There is then an additional delay once the patient arrives in hospital, between the time of arrival and the time when the decision is made to administer antivenom. The former delay is difficult to reduce and is due to the remoteness of many bites and poor transport being available to patients. In this instance, the focus needs to be on improved first aid to delay venom absorption and distribution . In contrast, the latter delay can be reduced by improved early diagnosis with bedside testing for systemic envenomation and stream-lined treatment protocols . One factor in the patient's favor is that venom takes time to reach the systemic circulation from the bite site, and thence to the target site of the toxin (e.g. neuromuscular junction).
In most cases, snakebites result in subcutaneous/sub-dermal venom injection (short fangs — elapids) or intramuscular venom injection (longer fangs — vipers) [74,75]. However, more recent evidence suggests that for some snakes/toxins, there is intravenous injection [76,77]. Because snake venoms contain toxins with a range of molecular sizes, the rates of absorption of these toxins into the circulation will vary significantly . Currently, there is little known about the variable absorption processes for snake venoms.
Snake antivenoms are almost always administered intravenously and the antibodies reach the systemic circulation almost immediately. Other routes of administration, such as intramuscular and subcutaneous, significantly delay the absorption of antivenom into the circulation and should not be used [79,80]. However, the distribution of antivenom to toxin target sites is much slower and varies, because of the size of the antibodies . Different types of antivenoms [such as whole IgG, Fab and F(ab′)2] have different molecular masses, and hence have different pharmacokinetics . The larger IgG molecule has poor distribution to peripheral sites, but a longer elimination half-life. In contrast, the smaller Fab fragments distribute better, but have short elimination half-lives. To date, most information on antivenom pharmacokinetics comes from animal studies,  and there is limited information on the pharmacokinetics of antivenoms in humans [81,83,84]. From the available studies, antivenom concentrations have a biphasic decline after intravenous administration of whole IgG and F(ab′)2 antivenoms, with an initial rapid distribution phase and a slower elimination phase. Recently, a study based on population pharmacokinetic approach that incorporates variability of individual patients, demonstrated similar pharmacokinetic of a F(ab′)2 antivenom .
Pharmacodynamics of antivenoms
Toxins exert their actions via molecular interactions between their toxic site and the target molecule or enzyme pathway. The principle action of antivenoms is based on antibodies binding to antigenic sites on the toxin molecules, that interfere with toxin action . Toxin neutralization is only achieved if the antibodies or antibody fragments interfere directly with the part of the toxin that interacts with the target/substrate . However, neutralization of the toxin activity is not the only mechanism of stopping the effects of the toxin. Preventing the toxin reaching the target molecule is just as important, which can occur by a number of different mechanisms — steric hindrance, trapping the toxin in the central compartment and enhancing elimination of the toxin. The different mechanisms of antivenom action are as follows: [81,86]
Disrupting toxin–target interaction: a, Antibodies and their fragments bind to epitopes at the ‘pharmacological site’ of the toxin or, to epitopes located in close proximity to the ‘pharmacological site’ of the toxin, disrupt the toxin–target interaction. b, Antibodies and their fragments may bind to an epitope more distant from the ‘pharmacological site’ of a toxin, leading to conformational changes in the target-bound toxin, decreasing its affinity for the target. Commercially used antivenoms and high-affinity monoclonal antibodies have been shown to accelerate the dissociation of the Naja nigricollis toxin α from nicotinic acetylcholine receptor toxin α complex [87–89]. Here, the monoclonal antibody binding site of the toxin was shown to be distant from the toxic site . Such antibodies could not only neutralize and reverse the toxic effect, but also could possibly lead to redistribution of toxins from their target sites. It is unlikely that the antibodies or their fragments could act in this way on intracellularly acting toxins, such as pre-synaptic neurotoxins.
Trapping: The simple act of antibodies binding to toxins and preventing them reaching their peripheral sites is likely to be an important mechanism of action. Once whole IgG or F(ab′)2 antivenoms are administered, the antibody molecules will form immunocomplexes with toxins in the circulation, trapping them in the central compartment. Toxins are then unable to reach their target site and potentially toxins already distributed to extra-vascular sites will re-distribute back to the central compartment .
Steric hindrance: Polyclonal antibodies, by binding to a number of antigenic sites, may surround a large toxin and prevent it binding to its target molecule [91,92]. This is likely to be the mechanism of antivenom neutralizing the effect of procoagulant toxins that act in the central compartment. There is evidence to support this for the prothrombin activator in Australian brown snake (Pseudonaja textilis) venom .
Elimination: Antibodies and F(ab′)2 form multivalent immunocomplexes with toxins, facilitating such complexes being removed by phagocytic cells, eliminating the toxin from the circulation or from other tissues. Fab, scFVs and VHHs are unable to form immunocomplexes, and appear to be mainly eliminated renally.
Antivenom research: human versus experimental animal
Animal models have been essential throughout the development and testing of antivenoms. To date, the major test of antivenom efficacy against overall toxic effects of venoms has been rodent lethality prevention studies (ED50) . The toxicological profiling of the venoms and toxins using lethality models is based on the general argument that the ‘toxins possessing the highest rodent lethality are the most medically important toxins in human envenoming’ [52–54]. In addition, many other in vitro tests for individual toxic effects such as neurotoxicity and VICC, are based on the animal models, predominately of rodent origin . Recent studies of the differential susceptibility of humans and rodents to toxic effects of snake venoms, such as post-synaptic neurotoxicity  and VICC,  challenge the relevance of rodent models for some envenoming syndromes in humans. Lack of reliable animal models could potentially affect more promising steps in antivenom development such as toxico-venomic approaches, which currently rely on rodent lethality tests [53,54]. In this regard, careful selection of in vitro tests for specific toxic effects of venoms and toxins will be essential in establishing the efficacy of newer antivenoms, avoiding the problems with rodent lethality testing .
Understanding the therapeutic limitations of antivenom
Since antivenoms were first introduced, their effectiveness in treating snakebite was considered unequivocal and hence rarely challenged. While the pre-clinical research on antivenom development has progressed towards future antivenoms of human monoclonal antibody origin, high-quality clinical evidence on the existing antivenoms remains limited. This has compromised our understanding of the limitations of antivenom. Recent systematic reviews demonstrated a lack of randomized-controlled trial based evidence for antivenom in VICC  and in neuromuscular dysfunction . Since antivenom has become the standard practice worldwide, randomized-controlled trials of existing antivenoms are ethically difficult or often impossible to undertake. Well-designed observational clinical studies are required to better understand the limitations of antivenoms in treating different envenoming syndromes [14,96].
Some envenoming syndromes are irreversible, such as pre-synaptic neurotoxicity, VICC, cardiovascular collapse and local necrosis, which will limit the effectiveness of antivenom as a treatment. This is purely based on the nature of the irreversible pathophysiology of these venom effects, rather than it being an issue with antivenom itself. For example, the prothrombin activator complexes in Australasian elapid venom  and prothrombin activators in the carpet viper (Echis sp.) venoms  both result in VICC. However, antivenom appears to be effective for VICC in Echis envenoming, even days after the bite,  but not effective for Australasian snake envenoming VICC. The onset of VICC in Australasian elapid VICC is rapid, and there is potentially only a very limited time window for effectiveness of antivenom . This is likely because of the nature of the procoagulant toxins  in the Australasian venoms which are similar to the human clotting factors. This emphasizes the potential issues in generalizing the results of antivenom effectiveness for one snake to another.
Since the development of the first antivenoms over a century ago, they remain the only specific treatment for snake envenoming. The production procedures of current antivenoms have improved, but there has been no major change in the immune technology used to produce antivenoms, despite huge advances in antibodies used in vaccines and cancer therapy. There are some promising developments in humanized and human monoclonal antibodies and their fractions against snake toxins. In the light of the gradual reduction in the costs of such technologies, affordability of such future antivenoms should not be a significant barrier to their development. Oligoclonal antivenoms containing selected monoclonal antibodies and fragments with specific venom inhibitors may be a potential way forward in future antivenom therapy. However, translation of such technology to a snakebite patient being treated in rural Africa, Latin America or Asia requires significant investment, due to the considerable technological and economic gaps. An initial large investment in developing modern approaches to antibody therapy would have a sustained impact on antivenom improvements.
Lack of reliable animal models to represent envenoming in humans and the major gaps in the understanding of the clinical effectiveness and limitations of current antivenoms are likely to remain as barriers towards development of future antivenoms. This requires a re-examination of the pathophysiology of snake envenoming to identify for which toxic effects antivenoms (or other targeted therapies) might be effective. These should then form the focus for new antivenom therapies. In some cases, such as rapidly developing irreversible envenoming effects (e.g. early cardiovascular collapse ), antivenom may never be effective and public health or first aid measures need to be the focus.
Snakebite is a neglected tropical disease that causes major morbidity and mortality in the rural tropics. Antivenom is the only specific treatment for snake envenoming, but there has been no major change in the immune technology used to produce antivenoms for over 100 years.
Current antivenoms are a mixture of polyclonal antibodies raised against the toxins in an individual snake or a group of snake venoms, purified for human use. These continue to result in a high rate of adverse reactions and are not universally effective. Antivenoms could be improved with new immunotechnology used in vaccines and cancer treatments, to be more directed and less reactive.
Future research must aim to develop mixtures of targeted antibodies to cover the full range of snake venom toxicities, which are less reactive and potentially humanized, to decrease adverse reactions. There are some promising developments in humanized/human monoclonal antibodies and their fractions against snake toxins. With the gradual reduction in costs of such technologies, affordability of future improved antivenoms should not be a significant barrier to their development.
The authors declare that there are no competing interests associated with the manuscript.