Botulinum toxins (BTXs) are neurotoxic substances produced by Clostridium botulinum, a gram-positive anaerobic bacterium. In botulism poisoning, flaccid paralysis occurs by inhibiting the release of neurotransmitters from the peripheral cholinergic nerve terminals of the skeletal and autonomic nervous system. Paralysis begins at the ocular muscles and then spreads to the muscles of the face, before reaching the respiratory muscles and causing respiratory failure.
BTX has traditionally been found in seven serotypes: A, B, C1, D, E, F, and G [1
]. They have similar molecular weights and common subunit structures, but differ in their reaction mechanisms, durations of effect, and side effects.
In recent years, using molecular genetic analysis, many genes have been discovered that encode new BTXs. Thus, subtypes, such as BTX/A1, BTX/A2, BTX/B1, and BTX/B2, and chimeric types, such as BTX/DC, BTX/CD and BTX/FA, have been found.
Clinical use of BTX began in 1973, when Scott demonstrated that injecting the toxin into orbicularis oculi muscles was effective in treating strabismus. Over the next several decades, its application expanded to a variety of diseases.
BTX has a molecular weight of 150 kDa, consists of an inactive single-chain polypeptide, and folds into a 3-domain structure. The light chain (50 kDa) is a zinc-dependent protease that constitutes an active toxin and cleaves the soluble N
-ethyl-maleimide-sensitive factor attachment receptor (SNARE) complex [2
]. The heavy chain (100 kDa) consists of an N-terminal translocation domain and a C-terminal receptor binding domain, and it acts in neuron-specific binding. The light and heavy chains are linked by disulfide bonds, which partially obscure the active sites of the toxin. When the single-chain disulfide bond is reduced, the light chain metalloprotease can be released to act as a toxin.
The main functional effect of BTX occurs in the neuromuscular junction, where it inhibits the release of acetylcholine from the presynaptic nerve ending, resulting in muscular and autonomic paralysis [4
]. The toxin-mediated muscle relaxation process proceeds in three phases: Binding, internalization, and the inhibition of neurotransmitter release. Specific binding to neurons is mediated by heavy chains [5
] and internalization is mediated by receptor-mediated endocytosis [6
]. Once BTX is internalized, the light chains within the vesicles are translocated across the vesicle membrane and released into the neuronal cytoplasm. SNARE is involved in the exocytosis of acetylcholine vesicles, located at the nerve endings, by attaching acetylcholine vesicles to the cell membrane, thus, allowing acetylcholine exocytosis to occur [8
]. BTX causes degradation of the SNARE protein, resulting in paralysis. Based on these mechanisms, BTX is clinically used to treat muscle spasticity associated with central nervous system (CNS) disorders, such as stroke, brain injury, spinal cord injury (SCI), cerebral palsy, and multiple sclerosis (MS).
However, two other functional effects of BTXs exist: The effects on the afferent limb of the motor nervous system and the analgesic effect on the sensory nerve system. Several preclinical studies have shown that BTX inhibits neuromodulator and transmitter secretion, which is important for neurotransmission in the sensory pathway, and, thus, BTX may reduce neuropathic pain.
Preclinical and clinical studies have reported the effects of BTX on peripheral neuropathic pain and, generally, demonstrated a high level of evidence for some diseases [9
]. However, few studies have reported its therapeutic effects on central neuropathic pain, and its effects have not been proven. The aim of this article is to review the mechanism of central neuropathic pain and to investigate the effect of BTX on central neuropathic pain.
A PubMed and EMBASE search (1980~March 2018) was performed as follows: ‘Botulinum toxins’, ‘neuropathic pain’, ‘neuropathy’, ‘pain’, ‘allodynia’, ‘hyperalgesia’ and ‘spinal cord injury’, ‘post-stroke pain’, ‘multiple sclerosis’, and ‘complex regional pain syndrome’. The results included animal studies, randomized controlled trials (RCTs), observational studies, case reports, and reviews. Editorials, guidelines, and trial protocols were excluded. Two reviewers individually assessed the abstracts to determine the eligibility of the studies. Articles not available in English and studies conducted in children (≤18 years of age) were also excluded.
2. Mechanism of Central Neuropathic Pain
The International Association for the Study of Pain (IASP) defines neuropathic pain as pain caused by a lesion or disease of the somatosensory nervous system [10
]. Neuropathic pain is a clinical description that requires a demonstrable lesion or a disease that satisfies the established neurological diagnostic criteria. It has two typical symptoms, allodynia and hyperalgesia. Allodynia describes a pain due to a stimulus that does not normally provoke pain and hyperalgesia refers to increased pain from a stimulus that normally provokes pain [10
Several molecular mechanisms are involved in the development of allodynia and hyperalgesia. After nerve injury, changes in the expression of sodium and calcium channels cause spontaneous activity in nerve endings, resulting in spontaneous pain. This is an important factor that causes sensitization. In addition, various cytokines, including glutamate, substance P, and proinflammatory cytokines, are involved in sensitization. Inflamed or ischemic tissues become acidified and this cellular environment causes pain by stimulating the release of neuropeptides from the primary afferent nerve tissue [11
]. When neuropeptides, such as calcitonin gene-related protein (CGRP) and substance P, are secreted into the endoneurium, they cause local blood flow and blood vessel leakage, leading to edema and pain.
Because central neuropathic pain is defined by IASP as a pain caused by a lesion or disease of the central somatosensory nervous system [10
], central neuropathic pain is a heterogenous group of neuropathic pain conditions. Major diagnostic conditions include: (1) Central pain associated with SCI; (2) central post-stroke pain; and (3) central pain associated with MS.
The mechanisms of neuropathic pain, following SCI, in various animal models have been published (Figure 1
). Cellular and molecular responses to SCI occur at various levels, from the distal terminals of primary afferent neurons to the cortical area along the nervous system, leading to neuropathic pain.
Several changes in the primary afferent neuron have been suggested. A definite factor that mediates SCI-related pain is an increased excitability of dorsal root ganglion (DRG) neurons. The spontaneous activity of nociceptive DRG neurons has been shown to increase after SCI [12
], and some authors have demonstrated that the expression of the capsaicin-sensitive cation channel transient receptor potential vanilloid type 1 (TRPV1) in DRG neurons increases after SCI, which is enhanced by capsaicin-evoked ion currents and calcium responses in DRG neurons [13
Neuronal hyperexcitability in the spinal dorsal horn is also associated with SCI-related neuropathic pain [14
]. This phenomenon might occur through chronic glial cell activation [16
], dendritic spine remodeling [18
], dysregulation of glutamate homeostasis [16
], glutamate receptor activation [20
], loss of GABAergic inhibitory interneurons [21
], interruption of descending inhibitory modulation by serotonin [23
], or upregulation of voltage-gated calcium channel alpha-2-delta-1 subunit proteins [25
In patients with central pain following SCI, neurons in the somatosensory thalamus fire in bursts of action potentials more frequently than do similar neurons in patients without pain [26
]. In rats with contusive SCI, thalamic ventralis postero-lateralis neurons exhibited a dysrhythmia in that a significantly higher proportion fired spontaneously when compared with neurons in uninjured rats [27
]. Based on these results, abnormal thalamic processes following SCI may mediate neuropathic pain. Additionally, in rats with SCI, neurons in the primary somatosensory cortex had significantly higher spontaneous firing rates, greater evoked responses to noxious mechanical stimulation, and a greater tendency to fire bursts of action potentials [28
]. Another study also revealed that phosphorylation of AMPA-type glutamate receptors in the primary somatosensory cortex play an important role in the development of hypersensitivity after SCI [29
The pathophysiology of central post-stroke pain, another important disease constituting central neuropathic pain, remains uncertain. Recent advances in brain imaging technology have increased the understanding of the role of specific anatomic locations. Several reports have suggested that central post-stroke pain commonly occurs in lesions affecting the thalamus, parietal cortex, dorsal putamen, posterior internal capsule, dorsal basal ganglia, brainstem, and lateral medulla [30
]. Particularly, spinothalamic tracts terminated with the ventral posterolateral thalamus and lesions of the ventral posteromedial thalamus and medial lemniscal thalamocortical pathway were found to be the major factors causing central post-stroke pain [34
]. These results have also been demonstrated in animal model experiments. Thermal hypersensitivity was observed in a rat model of the ventral posterior thalamic infarction [36
], with thermal and mechanical hyperalgesia developing in rats with a thalamic hemorrhagic lesion [37
In the middle cerebral artery occlusion rat model, increased N
-aspartate (NMDA) and AMPA receptor mediated excitatory transmission of the dorsal horn, decreases in GABA and glycine receptor mediated inhibitory transmission, and an increase in descending facilitation was proposed to be involved in the development of central post-stroke pain [38
]. In the ventral posterior thalamic lesion rat model, P2X7 expression in the medial thalamus was directly involved in nociceptive transmission, and short-term P2X7 inhibition led to a reduction of glutamatergic facilitation and neuronal hyperexcitability [39
3. Mechanism of BTX for Central Neuropathic Pain
The initial analgesic effect of BTX is caused by a decrease in muscle spasms. However, many preclinical and clinical studies suggest that a different mechanism underlies the analgesic effect of BTX. The hypothesis is that BTX inhibits the secretion of neuropeptides and suppresses inflammation and pain.
Several preclinical studies have shown that BTX-A inhibits the release of neurotransmitters that regulate pain and inflammation. McMahon et al. showed that BTX preferentially attenuates the slow phase of KCl-evoked glutamate release, which may be associated with synaptic vesicle mobilization according to a study that utilized a guinea pig formalin-induced pain model [40
]. Welch et al. reported that BTX inhibits potassium-evoked substance P secretion from cultured embryonic rat DRG neurons [41
], and Durham et al. demonstrated that BTX-A can directly decrease the release of CGRP from cultured rat trigeminal ganglion neurons [42
Xiao et al. demonstrated that BTX significantly reduces TRPV1 expression [43
]. A neuropathic pain model was induced by transection of the lumbar 5 ventral root in male rats. BTX-A or normal saline was administered to the plantar surface by subcutaneous injection. TRPV1 expression increased significantly in the lumbar 4–5 DRG after the transection of the lumbar 5 ventral root, and this increase persisted for at least 21 days. Subcutaneous injection of BTX-A significantly, and dose-dependently, reduced the expression of TRPV1 in the DRG neuron and significantly reduced hyperalgesia. A similar effect occurred on the expression of P2X3, one of the purinergic receptors associated with nociceptors, in a study that evaluated the effect of BTX on P2
expression, with the same method. Subcutaneous administration of BTX-A significantly, and bilaterally, reduced mechanical allodynia and inhibited the P2
overexpression induced by the transection of the lumbar 5 ventral root [44
One possible interpretation of these findings is that BTX reduces peripheral sensitization and afferent input to the spinal cord by inhibiting the release of neurotransmitters from peripheral nerve endings, thereby, indirectly a decreasing central sensitization. However, it has been hypothesized that the central effect may be direct by retrograde axonal transport of BTX along the branches of nociceptive neurons.
Immunohistochemical experiments have revealed that cleaved SNAP-25, a product of BTX-A action, is found not only in the peripheral region but also in the facial nucleus in the brain stem [45
], superior colliculus [46
], and motor region of the spinal cord [47
]. Antonucci et al. found cleaved SNAP-25 in the ipsilateral facial nucleus after a BTX-A injection into rat whisker muscles [45
]. Matak et al. found that cleaved SNAP-25 fragments were present in the ventral horn and dorsal horn of the spinal cord after low-dose toxin injections into the gastrocnemius muscle and sciatic nerve [47
]. These authors also identified that cleaved SNAP-25 was not detected in the spinal cord when they injected BTX-A into a sciatic nerve pretreated with colchicine, an axonal transport blocker. Therefore, they suggested that BTX-A showed a central effect by microtubule-dependent retrograde axonal transport [49
]. Wang et al., however, cited the possibility that disassociated cleaved SNAP-25 might have migrated from the terminal to the cell body, suggesting that the discovery of cleaved SNAP-25 in the CNS does not necessarily reflect the activity of BTX-A in the CNS [50
Retrograde axonal transport is well known as a transport pathway for various substances, such as tetanus toxin, and this fact suggests that BTX may also use the same pathway. Several studies have reported that the heavy chain, or the entire toxin, undergoes retrograde transport after a peripheral injection of BTX-A. Restani et al. directly monitored the endocytosis and axonal transport of BTX-A, and they showed that BTX-A was internalized by spinal cord motor neurons and underwent fast axonal retrograde transport [51
]. Wang et al. reported that fluorescently labeled BTX heavy chains were detected in spinal cord motor neurons after injection into the mouse hindlimb, which demonstrated retrograde transport of BTX [50
Several studies of the behavioral effects of BTX-A have demonstrated the central antinociceptive effect of BTX-A. Bilateral pain associated with experimental diabetes [52
], carrageenan-induced hyperalgesia, and paclitaxel-induced peripheral neuropathy [53
] or acidic saline-induced mirror pain, [54
] can be bilaterally reduced by the unilateral injection of BTX-A in rats. Bach-Rojecky et al. reported that mechanical and thermal hypersensitivity of the ipsilateral side, as well as the contralateral side, were decreased after subcutaneous unilateral BTX injection into the plantar surface of the hindlimb [52
]. In addition, Favre-Guilmard et al. reported a significant anti-hyperalgesic effect in the uninjected contralateral hindpaw after subcutaneous administration of BTX-A to the plantar surface in carrageenan-induced hyperalgesia and paclitaxel-induced peripheral neuropathy models. These results suggest that the antinociceptive effect of BTX-A cannot be explained by the peripheral action and it is possible that BTX-A has a central action through the retrograde axonal transport [53
]. This process is also expected to be a major mechanism in the BTX-A action on central neuropathic pain.
The mechanism of central neuropathic pain has been examined according to various hypotheses, including neuronal hyperexcitability and dysfunction of the spinothalamic tract. To date, various preclinical and clinical studies have been published on whether BTX may be effective for central neuropathic pain. BTX inhibits the secretion of substance P and CGRP from DRG, inhibits the expression of TRPV1 and P2X3, and induces a central effect through retrograde axonal transport. In addition, several studies have demonstrated its effect on central neuropathic pain associated with SCI, stroke, MS, and CRPS. Effects of BTX on neuropathic pain after SCI, post-stroke shoulder pain, and CRPS has been shown; therefore, it can be considered as one of the treatment options. In the future, well-designed studies will be necessary to assess the effects of BTX on central neuropathic pain, and, furthermore, effective injection sites, injection techniques, and adequate doses should be considered.