Over the past three decades, there has been a continuing fascination with botulinum neurotoxin (BoNT) and the opportunities for use of this potentially deadly neurotoxin as a treatment in certain clinical settings. BoNT interferes with the liberation of acetylcholine from presynaptic nerve endings, and clinically the therapy results in the prevention of muscle contractions and/or gland secretion. Since BoNT is used primarily as a focal injectable treatment, resulting in fewer side effects than systemically administered medications, it has evolved to become the preferred treatment of choice for management of many forms of dystonia, limb spasticity, cosmetic glabellar lines, hyperhidrosis, and sialorrhea [1]. The use of OnabotulinumtoxinA to address migraines was approved by the FDA in 2010 [2], and its use to treat bladder hyperactivity was approved in 2011 [3,4,5]. BoNT also has a broad spectrum of off-label indications, including focal hand dystonia, lower limb spasticity [6,7], management of chronic anal fissures [4], and tension headaches [8]. The commercially available neurotoxins include three main brands of BoNT/A: OnabotulinumtoxinA (BOTOX^®), AbobotulinumtoxinA (DYSPORT^®), IncobotulinumtoxinA (XEOMIN^®), and also a brand of BoNT type B, RimabotulinumtoxinB (MYOBLOC^® or NEUROBLOC^®) [9]. Each formulation has specific clinical recommendations which have been formulated based on the available science, and on differing biochemical profiles. Much of the literature has focused on the A and B neurotoxin types, with little mention of the other seven serologic types. Our goal in this review paper was to update the latest biomolecular findings in BoNT science. This review will cover the clinical scope, including mechanisms of action, and also will summarize the neurotoxin properties of the seven different toxin serotypes (A-G). Finally, we will discuss the potential opportunities for utilizing the various serotypes in clinical settings.

Botulinum neurotoxins (BoNT) are protein complexes produced by anaerobic gram positive Bacilligrouped under the name of Clostridium botulinum. The first description of botulism was penned in 1822 by Justin Kerner, who described a group of people suffering muscular weakness following the ingestion of contaminated sausages [10].

C. botulinum strains/species produce spores that are ubiquitous in the environment, surviving for extended periods of time under adverse environmental conditions. Under appropriate anaerobic conditions, spores may germinate, and the resultant vegetative cells may produce neurotoxin. Spore germination and neurotoxin production has been described in foods and in association with improper canning (particularly home canning); inappropriate fermentation of meat/sausage, bean curd, and fish; and preservation of a variety of products under oil or conditions that promote an anaerobic environment [11]. Much of the current food regulatory structure in the United States is based on assuring that conditions under which food is prepared and stored do not facilitate germination of C. botulinum spores and neurotoxin production.

The ingestion of food contaminated with pre-formed BoNT, or, more rarely, infection by C. botulinum of an exposed wound or growth in an immunologically immature intestinal tract (i.e., infant botulism) [12], can produce the clinical syndrome of botulism. Human botulism has been associated with BoNT types A, B, E and F, with BoNT/F being the least frequently reported [13]. Types C and D have been associated with botulism in animals. In accordance with the observed geographic distribution of botulinum spores in the environment, type A cases in the United States have come predominantly from areas west of the Mississippi, while type B cases predominate in the eastern United States. Type E cases are generally associated with fish or seafood, and, in the United States, tend to be associated with the consumption of fermented fish from Alaska [14].

Botulism is characterized by progressive flaccid paralysis of motor and autonomic nerves, and it usually occurs in a proximal to distal pattern, starting from muscles innervated by the cranial nerves, and then moving sequentially to affect the upper limbs, followed by respiratory muscles, and ultimately the lower limbs. Severe cases of botulism can include respiratory muscle paralysis, leading to ventilatory failure and death, if immediate supportive care is not provided [15]. Once clinical symptoms appear, the neurotoxin is considered to have irreversibly bound to receptors and entered the cell, blocking acetylcholine release. The administration of antitoxin can neutralize unbound neurotoxin, but cannot reverse existing symptoms [13].

Emile Piere Van Ermergen isolated the crystalline form of the neurotoxin type A in 1895, and the clinical product of BoNT/A was not produced until 1946 by Edward Schantz [10]. The neurotoxin was first used in 1981, by Alan Scott (an ophthalmologist) for a patient in his practice with strabismus [10]. In subsequent years, these uses have multiplied, and now include dystonia, spasticity, sweating, cosmetics, and many other indications.

The gold standard measurement of BoNT potency has been defined as the median lethal dose of the neurotoxin needed to cause death of 50% of a population of Swiss Webster Mice (LD50) [43,44,45]. The LD50 values for mice range from 0.5 to 5 ng/kg depending on the serotype [17]. Alternatively, potency may be reported as LD50’s per mg of toxin. However, minor differences in testing procedures, such as the strain of mice used, or the choice of dilution buffer, can result in different LD50 values for an essentially identical sample. Because of the differences, the units used in labeling the neurotoxin potency are specific for each marketed product, and are considered not to be interchangeable.

Another way to test neurotoxin potency is the mouse abdominal ptosis assay [45,46]. This assay measures the dose of neurotoxin injected subcutaneously into the top of a mouse hind leg. The procedure is designed to produce weakness in the animal. The animals are scored using a five point scale, according to the abdominal bulge present at 24 and 48 h. The potency of neurotoxin can then be estimated using a parallel line method.

Physiologically, the neuromuscular transmission begins with an action potential arriving at the presynaptic endplate, which then transiently depolarizes the axon terminal, and briefly opens the calcium channel. The increase of intracellular calcium leads to exocytosis of Ach containing vesicles into the neuromuscular junction. Two molecules of Ach are required to bind Nm nAChR to open its central ion channel and to permit sodium to diffuse into the muscle cell, thereby depolarizing the endplate. The endplate potential spreads to nearby muscle fiber membranes and depolarizes them to threshold, and this initiates a muscle cell action potential. The propagation of this action potential along the terminal axon of the motoneuron induces the release of acetylcholine from the cytosol into synaptic clefts by the Soluble N-ethylmaleimide-sensitive factor activating protein receptor complex (SNARE). Once in the synaptic cleft, the acetylcholine molecules bind to post synaptic membrane and muscle fiber contraction is the resultant effect [47].

Biomolecular research has been providing important clues to assist in the understanding of the mechanism of BoNT action. In general terms, the carboxyl terminal of the heavy chain (Hcc) binds to specific receptors from the presynaptic terminal. BoNT is endocytosed and then the amino domain of the heavy chain (Hcn) translocates the Light chain (Lc) into the cytoplasm. Once in the cytoplasm, the light chain (Lc) is directed to cleave a determinate membrane or vesicle associated proteins, thus preventing acetylcholine release.

The process can be subdivided into four stages: (1) binding; (2) internalization; (3) membrane translocation; and (4) proteolysis of specific SNARE proteins.

There are currently only two serotypes of Neurotoxin available—BoNT/A (OnabotulinumtoxinA, AbobotulinumtoxinA, and IncobotulinumtoxinA) and a BoNT/B, patented by RimabotulinumtoxinB (Table 4). Each brand has been approved by the US FDA to treat specific pathologies (Table 5). Every commercial brand of BoNT should be titrated by the clinician as an individual drug, and should be used for specific symptoms. The doses are not interchangeable between available serotypes, and between each formulation. Each formulation of BoNT/A, for example, is not interchangeable with another dose of BoNT/A if from a different formulation [90].

After BoNT injection, clinical effects can be detected at approximately one week following injection, and the peak benefit is commonly observed approximately two weeks following injection. The timing of injection sessions varies by diagnosis, but is typically every three to four months, and booster injections before the next visit are not recommended because of the fear of neutralizing antibodies. The BoNTs are typically diluted in saline without preservatives. The amount of dilution, dosages and injection patterns can vary according the experience of the specialist, and the experience of the center. RimabotulinumtoxinB does not require dilution [93]. The reported side effects of the neurotoxins mimic botulism and may include blurred vision, eyelid drooping, dry mouth, dysphagia, muscle weakness, and rarely respiratory impairment [94,95,96]. Symptoms are usually focal and transitory, lasting no more than twelve weeks [93]. Other local side effects can include injection site hematomas, infections and pain. It should be noted that side effects directly correlate to the region injected. Autonomic reactions have been reported in more than one third of patients with cervical dystonia treated with BoNT/A [97], and autonomic resactions have been reported to be more frequent after BoNT/B injection compared with BoNT/A [98,99].

Patients who do not benefit from BoNT are considered primary non-responders. This occurs in about 30% of patients with cervical dystonia (CD) as a result of inadequate dosing, inaccurate muscle selection or inaccessible muscles. In contrast, secondary unresponsiveness to BoNT occurs in 16% of CD patients, of which 35% have antibodies against BoNT by the mouse neutralization assay [100]. Later, studies with AbobotulinumtoxinA and OnabotulinumtoxinA in CD patients reported ranges of neutralizing antibodies from 4% to 17% depending of the formulation used [97,101,102,103]. Factors which increase the risk of immunogenicity include large molecules of non-human origin and aggregated forms of the protein. The large size of the complex and the neurotoxin subunits increases the chance of antibody development against the BoNT preparation [104]. However, only a fraction of patients present with antibodies which can block neurotoxin effects (Neutralizing antibodies, Figure 6). Recent studies have reported the generation of neutralizing antibodies which act against epitopes from neurotoxin Hc and also from Lc [105].

The development of antibodies also depends on antigenic extrinsic factors; this includes the BoNT formulation, its physiochemical properties, the protein load, and the doses used per treatment cycle. The total cumulative doses, and the repeated treatment with booster doses (besides the regular three-month injections), could potentially increase the risk of antibody development [108]. Finally, host genetic factors may also be contributory [106].

When the OnabotulinumtoxinA formulation contained 25 ng of protein complex per 100 U of neurotoxin was administered to patients, 4%–17% were reported to develop a neutralizing antibody. The protein content has subsequently been decreased by the manufacturers to 5 ng per 100 U of neurotoxin (in 1998), with a reduction in the rate of neutralizing antibody production to 1.2% [109]. The incidence of neutralizing antibodies using AbobotulinumtoxinA has been reported as 0%–3%. The rate of neutralizing antibody with RimabotulinumtoxinB was reported as 10% at one year, and 18% at 18 months [110]. While reported in the literature, these results have not been replicated.

The IncobotulinumtoxinA, which is free of proteins complexes, has not stimulated antibody development in rabbits when compared to OnabotulinumtoxinA and AbobotulinumtoxinA [111]. Additionally, there is a study in patients with upper limb spasticity that revealed no neutralizing antibodies following injections of IncobotulinumtoxinA or a placebo [112]. Approximately 1% of 1080 subjects from an IncobotulinumtoxinA development program presented with neutralizing antibodies; however, each of the patients had been previously exposed to BoNT which contained complexing proteins [22].

More comparative studies and longer term follow-up will be necessary to clarify the advantages and disadvantages of BoNTs lacking protein complexes compared with the conventional BoNTs preparations. Each neurotoxin can potentially generate an antibody response and produce cross-reacting antibodies, based on similar antigenic structures. Cross-reacting antibodies can recognize BoNT/A, and BoNT/B, but may not affect neurotoxin binding [106]. However, a recent study identified two clonally related human monoclonal antibodies designated as 1B18 and 4E17 with potential neutralizing effects. Antibody 1B18 was reported to bind BoNT/A, and BoNT/B, and antibody 4E17 to bind BoNT/A, B, E and F. Both antibodies bound to a conserved epitope at the tip of the BoNT translocation domain. The immunoglobulin G constructed from variants of 1B18 was found to neutralize BoNT/B in vivo, and 4E17 were found to neutralize in vivo BoNT/E, suggesting the functional importance of this epitope in the intoxication pathway [107].

An antibody neutralization experiment was performed in mice and the equivalent amount of BoNT/A1 or BoNT/A5 antibody was directed against mice injected with BoNT/A1. The mice injected with BoNT/A5 antibodies died faster than those injected BoNT/A1 antibodies [35], suggesting different antibody neutralization profiles for BoNT/A1 and BoNT/A5. A major improvement in the understanding in BoNT immunogenicity could contribute both to an optimization of BoNT treatment, and to development of better human vaccines for botulism.

The specific function of acetylcholine at a particular cholinergic synapse is largely determined by its Ach receptor subtype [47], classified as nicotinic and muscarinic. These receptors project extensively throughout the central nervous system, innervating a broad range of structures within the brain; however, their mutual interactions remain unknown. The Nm subtype of nicotinic receptor dominates at the neuromuscular junction, and the Nn subtype is known to be present at autonomic ganglia, and in the central nervous system. At the same time the muscarinic receptors (M1 to M5) are distributed in a variety of locations; the muscarinic receptor M1 is found in pyramidal cells from brain cortex and hippocampus, and this receptor is mainly excitatory [113]; M2 is expressed in the nucleus basalis, occipital cortex and in lower levels in hippocampus, caudate and putamen [114], and also seems to participate in glutamatergic and gabaergic transmission; M3 has a similar distribution to M1, but with lower expression; M4 receptors are highly expressed in caudate and putamen and are associated with dopaminergic receptors [115], and the M5 receptor is expressed in low levels in the hippocampus, substantia nigra and ventral tegmental area [116]. Ach neurotransmission in the CNS has been implicated in the pathophysiology of several psychiatric and neurological diseases including schizophrenia, bipolar disorder, substance abuse, Alzheimer’s disease and Parkinson’s disease [117]. Therefore, an improved understanding of the Ach network and its modulator role in the nervous system could lead to the engineering of new neurotoxins to address pathologies beyond the current indications.

Biomolecular research is continually expanding the study of neurotoxins’ serotypes, subtypes and variants. It was discovered recently that there was a specific strain of C. botulinum which produced a DC mosaic botulinum neurotoxin formed by 2/3 of BoNT/D and 1/3 of BoNT/C [118]. This new neurotoxin was called BoNT/DC, and resulted in high lethality in mice (1.1 × 10⁻⁹ LD50/mg protein) compared with other types of neurotoxins [70]. This lethality could be related to particular ganglioside binding properties of BoNT/D and BoNT/C. The ganglioside binding pattern is also influenced by the conservative motif SXWY. Most of the botulinum neurotoxins which do not have the SXWY motif, bind to gangliosides using only the ganglioside binding site (GBS). BoNT/D and C do not have a SXWY motif; However, BoNT/D can recognize gangliosides using either GBS or a protein binding site (PBS), and BoNT C can recognize gangliosides through GBS and also has an auxiliary contribution of a ganglioside binding loop (GBL). The ganglioside recognition seems to lead to the binding activity of the neurotoxin to the specific cell [71].

The progress in biomolecular technology has made designs of recombinant neurotoxins possible. For example, replacing the amino acid histidine in the position 1241 of BoNT/F with a corresponding lysine residue of BoNT/E increases the affinity of the modified BoNT/F for GD1a. This change also allows a further ability of this recombinant neurotoxin to bind ganglioside GM1a, something BoNT/E could not do [71]. Another significant target for molecular engineering is the recognition of SNARE protein substrate by neurotoxins. There can be a potential manipulation of the neurotoxins main pockets and Lc residues in order to increase the affinity to the specific SNARE protein target. The enhanced understanding of the molecular structure and the conformation of the different neurotoxin serotypes could help researchers to design new hybrid toxins. These capabilities will further improve, as the science continues to expand.

Centrally administered BoNTs can block the release of other neurotransmitters including glutamate, glycine, noradrenalin, dopamine, serotonin and neuropeptides [17]. BoNT/A and BoNT/E have been suggested as possible candidates for epilepsy treatment because of their excitatory/inhibitory signal modulation effect, with predominant sensibility from the glutamatergic terminals as compared with gabaergic terminals [48]. BoNT/E has been tested to control epileptic discharges in a mouse model of mesial temporal lobe epilepsy. In this model, it was observed that BoNT administration during chronic seizure phases reduced the seizure frequency (ANOVA and Holm-Sidak test p < 0.05); however, its benefit only lasted for 21 days [119]. Other studies have suggested a neuroprotective role for BoNT/E, before or shortly after status epilepticus [120,121,122]. Neurotoxins are also interesting candidates for pain management, e.g., BoNT/A has been proposed as an analgesic because of a proposed reduction in the inflammatory phase of pain. BoNT/A may inhibit the release of certain nociceptive mediators such as substance P [123], calcitonin gene-related peptide (CGRP) [124], and glutamate [125]. The proposed mechanism of neurotoxin action on nociception has been described as blocking neurotransmitter release at the peripheral nerves, preventing pain perception and also by preventing peripheral sensitization [50]. BoNT may also prevent an indirect central sensitization mechanism. BoNT has subsequently been considered as a candidate for chronic inflammatory pain treatment, and BoNT/A is suggested as the most suitable serotype based on its stable intracellular proteolytic activity [126]. BoNT has been approved by the FDA for chronic migraine treatment [2], and has been used off-label in tension headaches by some practitioners. BoNT is also being studied in trigeminal neuralgia. A recent publication suggested that BoNT/A preferentially targeted C fibers, blocked neurotransmitter release, and subsequently reduced pain, neurogenic inflammation, and the cutaneous heat pain threshold. These findings were documented in a capsaicin induced trigeminal neuralgia human model [127]. Some small studies treating trigeminal neuralgia patients with BoNT also revealed clinical benefits [128]. A randomized, double-blind placebo controlled study with BoNT/A was performed in a mouse trigeminal neuralgia pain model [129]. The results suggested that BoNT/A was effective in preventing inflammatory pain up to eight days after the first treatment; however, the effect was not reproduced during the second dose. Randomized and controlled studies in humans with trigeminal neuralgia are needed to assess the neurotoxins’ potency and duration in pain treatment. Furthermore, a recent study found that when selectively targeting syntaxin, the neurotoxin can block synaptic release in neurons, and norepinephrine release in neuroendocrine cells [130].

The development of secondary resistance has resulted in increasing attention being given to the immunogenicity of the associated proteins, the neurotoxin size and configuration, and the influence of serotype. Designing new neurotoxin molecules has the potential to expand the possibilities for treatment of resistant patients, expansion of clinical indications, and limitation of side effects. An initial target may be the presynaptic specific gangliosides and proteins, and the exposed vesicular membrane. A second target may be the specific affinity for different BoNTs serotypes to the SNAREt and SNAREv complex.

Modifying the sequence of carboxyl terminals on the neurotoxin Hc to match the predominant gangliosides or SNAREv proteins, and future studies of the intervention in the Lc translocation and proteolytic stage could contribute to the potency of neurotoxin effect in refractory patients. Additionally, identifying SNAREt isoforms and their localizations in CNS and PNS could aid in increasing the specificity of neurotoxin uses and could provide better control of adverse effects. However, this has proven to be a complex field, with evidence that there are multiple cofactors important in the presynaptic binding, as well as in the SNARE cleavage. Another important aspect which needs further study is the mechanism of neurotoxin penetration in different tissues, for example, hyperactive neurons of epileptic foci could over-expose receptors associated to the high rate of exocytosis and endocytosis.

An ideal botulinum neurotoxin preparation should be stable in the environment, have low immunogenicity, have a specific function in determined cells, and have an adequate potency with controllable effects, long-lasting benefits, and few side effects. The ability to design neurotoxin recombinant mixtures may also enhance Hc terminal binding to the presynaptic protein array, aiding in the control of specific vesicular and trans-membrane SNARE proteins. Recombinant neurotoxin molecules are currently being studied in basic sciences laboratories, and these may play a role in botulinum neurotoxin resistant cases associated with antibody production. The potential clinical uses for botulinum neurotoxins are still largely undefined, but there is expanding interest in the potential role of neurotoxin in multiple disorders. There will likely be many possibilities for BoNT use beyond movement disorders, pain and other current indications. The understanding of the pathophysiologic mechanisms by which neurotoxins affect specific neuronal populations, and the differential effects seen with different serotypes, will aid us in tailoring treatment for each diagnosis and each patient’s individual needs.