Research ArticlePain

Selective neuronal silencing using synthetic botulinum molecules alleviates chronic pain in mice

See allHide authors and affiliations

Science Translational Medicine  18 Jul 2018:
Vol. 10, Issue 450, eaar7384
DOI: 10.1126/scitranslmed.aar7384

Relieving pain with botox

Chronic pain affects more than 25 million Americans and is associated with reduced life span, anxiety, and depression. Opioid administration is often effective in relieving pain but, unfortunately, opioids have serious side effects, including risk of addiction and overdose. In a new study, Maiarù et al. have leveraged the inhibitory effects of botulinum toxin on neuronal activity. They developed two botulinum-conjugated molecules (SP-BOT and Derm-BOT) that were able to silence subpopulations of pain-related spinal neurons in several mouse models of chronic pain. Intrathecal administration of one dose of SP-BOT or Derm-BOT produced long-term pain relief in the mouse models that was comparable to the effects of opioid treatment. The results suggest that botulinum-conjugated molecules could be an opioid-free alternative for treating chronic pain.


Chronic pain is a widespread debilitating condition affecting millions of people worldwide. Although several pharmacological treatments for relieving chronic pain have been developed, they require frequent chronic administration and are often associated with severe adverse events, including overdose and addiction. Persistent increased sensitization of neuronal subpopulations of the peripheral and central nervous system has been recognized as a central mechanism mediating chronic pain, suggesting that inhibition of specific neuronal subpopulations might produce antinociceptive effects. We leveraged the neurotoxic properties of the botulinum toxin to specifically silence key pain-processing neurons in the spinal cords of mice. We show that a single intrathecal injection of botulinum toxin conjugates produced long-lasting pain relief in mouse models of inflammatory and neuropathic pain without toxic side effects. Our results suggest that this strategy might be a safe and effective approach for relieving chronic pain while avoiding the adverse events associated with repeated chronic drug administration.


Noxious stimuli of sufficient intensity to induce tissue damage lead to increased excitability of peripheral and central neuronal circuits that heightens pain experience and serves to protect damaged tissue from further trauma (14). In some cases, ongoing disease or the failure of potentiated pain signaling networks to return to preinjury levels leads to persistent or chronic pain conditions (5). Persistent pain is highly prevalent and extremely difficult to treat (6, 7) with widely prescribed drugs such as opioids having significant unwanted side effects (79). Although research into developing new analgesic drug therapies has been intense, translating knowledge from preclinical observations in animal models to new therapies in the clinic has been challenging (6). Research into the control of chronic pain states has, however, identified pathways connecting the spinal cord and brain that are keys to the regulation of on-going pain states (1013). Pioneering studies in rats and companion dogs (11, 14) showed that persistent pain states can be ameliorated by using a saporin–substance P (SP) conjugate to ablate a small population of spinal SP receptor [neurokinin-1 receptor (NK1R)] expressing projection neurons that convey pain-related information to the brain. To circumvent of the problem of killing spinal neurons with saporin, we designed botulinum conjugates that were safe to construct, nontoxic, and acted relatively quickly after intrathecal injection to silence pain-processing neurons in the spinal cord (15, 16).

Botulinum neurotoxin serotype A (BoNT/A) is made up of a light-chain zinc endopeptidase and a heavy chain that is responsible for binding the toxin to neuronal receptors and promoting essential light-chain translocation across the endosomal membrane (17). Once internalized within the neuron, the light chain has the capacity to silence neurons for several months via the specific proteolytic cleavage of synaptosomal-associated protein 25 (SNAP25), a protein essential for synaptic release (15, 16, 18). This inhibition is slowly reversed as the endopeptidase loses activity (17). Cleaved SNAP25 (cSNAP25) is found in neurons but not in glial cells and is the unique substrate for botulinum protease cleavage (19, 20). We exploited a recently introduced “protein stapling” method (15, 18) using SNARE (soluble N-ethylmaleimide–sensitive factor attachment protein receptor complex) proteins to link the light-chain/translocation domain (LcTd) of botulinum neurotoxin type A (BOT) to neurotransmitter ligands SP and dermorphin that target pain-processing neurons in the dorsal horn. To silence NK1R-expressing neurons, we used an SP-botulinum (SP-BOT) construct previously developed (15), whereas mu opiate receptor (MOR)–expressing neurons were silenced using a dermorphin-botulinum (Derm-BOT) construct. We found that the new constructs were selectively internalized after binding to their target receptors, silenced neurons, and produced a long-term amelioration of pain states.


SP-BOT conjugate induces long-term reduction of inflammatory and neuropathic pain sensitivity in mice

To silence NK1R-expressing neurons, we used an SP-BOT construct previously developed (15). SP-BOT (fig. S1) was injected intrathecally over the lumbar spinal cord of adult C57BL6/J male mice. Hind paw mechanical withdrawal thresholds measured with von Frey filaments were used as an indicator of analgesia. The intrathecal injection of SP-BOT had no effect on baseline mechanical threshold in naïve mice tested more than 7 days (Fig. 1A) and produced no signs of motor impairment assessed by plantar spreading or rotarod performance (Fig. 1B). However, in two models of inflammatory pain induced by ankle or hind paw injection of complete Freund’s adjuvant (CFA), intrathecal injection of SP-BOT 2 days after CFA (when mice showed increased pain sensitivity) produced a substantial reduction in mechanical hypersensitivity that accompanied inflammation (Fig. 1, C and D). One single intrathecal injection was effective in reducing pain sensitivity for the duration of the experiment (21 days for the ankle model and 12 days for the hind paw model). As internal control, in the hind paw model, we showed that threshold of mechanical allodynia in the contralateral paw was unchanged (fig. S2A). Dose-response experiments in animals that received an ankle injection of CFA showed that maximal reduction of pain sensitivity was obtained with intrathecal injection of 100 ng of SP-BOT (fig. S3A). Intrathecal injections of the unconjugated BOT without a receptor-binding domain (LcTd) had no effect on mechanical hypersensitivity after injection of CFA in the ankle (fig. S4).

Fig. 1 SP-BOT administered intrathecally reduced the mechanical hypersensitivity that developed in long-term inflammatory and neuropathic pain states.

(A) Mechanical threshold assessed using von Frey filaments in naïve mice before (B1) and 1 to 7 days (D1 to D7) after intrathecal injection of SP-BOT (100 ng/3 μl; n = 4 per group). (B) Time on rotarod apparatus after SP-BOT intrathecal injection (n = 4 per group). (C) von Frey filaments were used to measure mechanical hypersensitivity in mice injected with 5 μl of CFA in the ankle joint and injected 3 days later with intrathecal SP-BOT (100 ng/3 μl). Mice were tested at baseline and up to 21 days after CFA injection (n = 5 to 6 per group). (D) CFA (20 μl) was also injected into the plantar surface of the hind paw, and 4 days later, mice received intrathecal SP-BOT (100 ng/3 μl; n = 7 per group). (E) SP-BOT was injected intrathecally 5 days after SNI and alleviated the neuropathic mechanical sensitivity that had developed (n = 8 per group). (F) NK1R−/− mice and their WT littermates received intrathecal SP-BOT 5 days after SNI (n = 8 per group). Data show means ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001. Difference in sensitivity was assessed using repeated measures two-way followed by one-way analysis of variance (ANOVA). For complete statistical analyses, please refer to table S1, and for maximum possible effect (%MPE), please refer to table S2.

We next investigated the effect of SP-BOT on neuropathic pain by testing the mechanical sensitivity in the unilateral (left) spared nerve injury (SNI) model of neuropathic pain (pain that is derived from peripheral nerve damage). The lesion induced hypersensitivity in the lateral area of the paw on the left side, which is innervated by the spared sural nerve. SP-BOT was injected intrathecally when the mechanical hypersensitivity was fully developed and we observed a reduction in mechanical hypersensitivity that began around 3 days after SP-BOT injection and lasted for the duration of the experiment (22 days; Fig. 1E). Mechanical thresholds for the contralateral paw (right) were unchanged (fig. S2B). To confirm the essential role of NK1R in mediating the effects of SP-BOT–induced reduction of mechanical pain sensitivity, we used NK1R knockout (KO) (NK1R−/−) mice (21). Neuropathic mechanical hypersensitivity was similar in NK1R−/− and wild-type (WT) littermates after SNI. Intrathecal injection of SP-BOT was effective in alleviating mechanical hypersensitivity only in WT animals, whereas in NK1R−/− mice, mechanical allodynia was not affected by SP-BOT injection (Fig. 1F). The results indicate that the NK1R is essential for SP-BOT–mediated reduction of mechanical pain hypersensitivity.

SP-BOT is internalized only by NK1R-expressing neurons but does not cause cell death

The specificity of the targeted toxin was investigated by examining the distribution of cSNAP25 by immunohistochemistry in spinal cord tissue sections using an antibody specific for cSNAP25 (22). Tissue was taken from CFA-treated mice that had received intrathecal injections of SP-BOT, CFA-treated animals that received intrathecal saline injection, and naïve animals (n = 4 per group). Double-fluorescent immunohistochemistry for cSNAP25 and NK1R indicated that the SP-BOT construct was expressed in cell bodies and axonal and dendritic branches of NK1R-positive neurons (Fig. 2, A and B). Cell bodies were first seen 96 hours after intrathecal injection of the construct, and the numbers and distribution of labeled cell bodies within the superficial dorsal horn remained unchanged for the duration of the experiment and were unaffected by peripheral treatment (fig. S5). Analysis of the parabrachial nucleus of the hindbrain—the major site of termination of NK1R-positive spinal projection neurons (23, 24)—revealed cSNAP25-positive putative axons in all mice injected 14 days previously with intrathecal SP-BOT but not in saline-injected controls (Fig. 2C). Because NK1R is not found in axons (25), the results suggest that there had been axonal transport of cSNAP25 and/or botulinum protease after uptake of the SP-BOT conjugate by spinal NK1R-positive dendrites and cell bodies. Immunohistochemistry measuring c-Fos expression, a marker of cell activity (26), showed that in CFA-injected mice, the activity of neurons had been reduced in the parabrachial area of mice that had received an intrathecal injection of SP-BOT 3 days previously, suggesting that SP-BOT successfully silenced spinal NK1R+ cells (Fig. 2D). In naïve mice, there was no evidence of microglial or astrocytic activation after SP-BOT treatment (fig. S6). In addition, no changes in the extent of NK1R-positive immunofluorescence were found in the dorsal horn of mice that had been treated with SP-BOT, suggesting lack of construct-induced cytotoxicity or receptor down-regulation (Fig. 2E).

Fig. 2 SP-BOT was internalized by NK1R-positive neurons without toxicity.

(A) Images of NK1R and cSNAP25 immunoreactivity in the superficial dorsal horn of mice 14 days after intrathecal injection of SP-BOT. Green, cSNAP25; red, NK1R. Scale bars, 100 μm. (B) Images of selective targeting of NK1R-expressing neurons in the superficial dorsal horn 96 hours (top) or 14 days (bottom) after intrathecal injection of SP-BOT. Green, cSNAP25; red, NK1R. Scale bars, 20 μm (top) and 10 μm (bottom). (C) Schematic illustration and images of the lateral parabrachial (LPb) area of mice 25 days after intrathecal injection of SP-BOT or saline. Green, cSNAP25 in spinoparabrachial axons. Scale bar, 80 μm. DRG, dorsal root ganglia. (D) Bar graph illustrating the number of c-Fos–immunostained nuclei in the PB from both saline and SP-BOT–injected mice. Mice received intrathecal SP-BOT, and 3 days later, they were injected with CFA into the plantar surface of the hind paw. Tissue was taken 6 hours later. Values reported are the mean number of c-Fos+ nuclei (±SEM) normalized to the mean of c-Fos+ nuclei in naïve control mice (n = 4 per group). (E) Quantification of NK1R fluorescence intensity in the contralateral superficial dorsal horn of mice 18 days after intraplantar CFA injection and 14 days after intrathecal injection of SP-BOT or saline. All data were normalized to laminae I/II saline-treated mice (n = 4 per group). *P < 0.05. The comparison of three groups was determined using one-way ANOVA.

Derm-BOT conjugate alleviates long-term pain states

Opioids such as morphine are effective in relieving chronic pain. Their analgesic properties are mostly mediated by the MOR (27). In the dorsal horn, MOR is expressed by interneurons and some primary afferents and by some projection neurons (2830). To test the possibility that inhibiting MOR-expressing neurons could promote analgesic effects, we conjugated the botulinum toxin to the MOR agonist dermorphin (Derm-BOT) (31, 32) and compared the analgesic efficacy of Derm-BOT with morphine.

Derm-BOT has been injected intrathecally at the optimal dose of 100 ng/3 μl in naïve mice, and in mice previously injected with CFA in the ankle joint or in the hind paw after increased mechanical, hypersensitivity was established. Derm-BOT injection did not affect mechanical pain sensitivity in naïve control mice (Fig. 3A); in contrast, we observed a reduction in the mechanical hypersensitivity that lasted until the end of the experiments (up to 18 days) in both models of inflammatory pain (Fig. 3, B and C). Furthermore, when Derm-BOT was injected after SP-BOT, the reduction in pain sensitivity induced by SP-BOT was not further augmented (Fig. 3D). We then investigated the effect of Derm-BOT on the hypersensitivity that develops after SNI surgery and found that a single intrathecal injection of the construct alleviated the mechanical hypersensitivity for the duration of the experiment (23 days; Fig. 3E).

Fig. 3 Derm-BOT reduced the mechanical hypersensitivity in inflammatory and neuropathic pain models in mice.

(A) Mechanical threshold assessed using von Frey filaments in naïve mice before (B1) and after (D1 to D7) intrathecal injection of Derm-BOT (100 ng/3 μl; n = 4 per group). (B) Mechanical threshold was measured in mice before and after CFA injection (5 μl) in the ankle joint. Four days later, mice were injected intrathecal with Derm-BOT (100 ng/3 μl). Mice were tested at baseline and up to 14 days after CFA injection (n = 5 per group). (C) CFA (20 μl) was injected into the plantar surface of the hind paw, and 4 days later, mice received intrathecal Derm-BOT (100 ng/3 μl; n = 8 per group). (D) Mechanical threshold measured using von Frey filaments in mice injected with CFA (5 μl) in the ankle joint and injected 3 days later with intrathecal SP-BOT (100 ng/3 μl). Two weeks later, mice injected with SP-BOT were reinjected with intrathecal Derm-BOT (n = 4 per group). (E) Derm-BOT was injected intrathecally in mice 5 days after SNI surgery (n = 9 per group). Data show means ± SEM. *P < 0.05, **P < 0.01, ***P ≤ 0.001. Difference in sensitivity was assessed using repeated-measures two-way, followed by one-way ANOVA.

Derm-BOT conjugate was internalized by MOR-positive neurons and did not induce toxicity

Immunohistochemical analysis of spinal cord sections from Derm-BOT–injected mice showed that all cSNAP25-positive cell bodies and many neuronal processes throughout the dorsal horn were stained with MOR antibody (Fig. 4, A and B) but there was no labeling of axons in the dorsal roots. Cell bodies were first seen 96 hours after intrathecal injection of the construct, and the numbers and distribution of cSNAP25-labeled cell bodies remained unchanged for the duration of the experiment (fig. S5). These results indicated that cSNAP25-positive cell bodies and fibers were likely to be MOR-positive local neurons (Fig. 4, A and B) and that MOR-positive primary afferents did not internalize the construct. We also failed to find evidence for glial activation in naïve mice treated with Derm-BOT (fig. S6). As with SP-BOT, no indication of toxicity was found after Derm-BOT injection (fig. S7).

Fig. 4 Derm-BOT was internalized by MOR-expressing neurons.

(A) Images of cSNAP25 and MOR immunoreactivity in the superficial dorsal horn of mice 14 days after injection of intrathecal Derm-BOT. Green, cSNAP25; red, MOR. Scale bar, 100 μm. (B) Images of selective targeting of cSNAP25 to MOR-expressing neurons in the superficial dorsal horn 96 hours (top and bottom) or 14 days after intrathecal injection of Derm-BOT. Green, cSNAP25; red, MOR. Scale bars, 20 μm (top) and 10 μm (middle and bottom).

Derm-BOT conjugates replicate the analgesic actions of morphine

Finally, we compared the effects on mechanical pain sensitivity of Derm-BOT with morphine (5 nmol) (33) in the SNI mouse model. Intrathecal Derm-BOT reduced mechanical sensitivity in SNI mice to the same extent as intrathecal morphine, and no additive effects were seen when morphine was injected intrathecally into mice pretreated with Derm-BOT (Fig. 5A). This implied that pretreatment with Derm-BOT silenced many of the MOR-expressing neurons in the lumbar dorsal horn. Derm-BOT also generated analgesia in NK1R−/− mice (Fig. 5B), whereas SP-BOT was ineffective, confirming the specificity of the botulinum constructs.

Fig. 5 Derm-BOT precludes the effect of morphine and retains efficacy in NK1R−/− mice.

(A) Mechanical threshold using von Frey filaments in mice injected intrathecally with Derm-BOT 5 days after SNI surgery. Twenty-nine days later, mice were injected with intrathecal morphine (5 nM; n = 9 per group). (B) Mechanical threshold measured using von Frey filaments in NK1R−/− mice before and after SNI surgery. Five days after surgery, mice were injected with intrathecal SP-BOT and were injected with intrathecal Derm-BOT 2 weeks later (n = 8 per group). Data show means ± SEM. */#P < 0.05, **P < 0.01, ***/###P ≤ 0.001. Difference in sensitivity was assessed using repeated-measures two-way followed by one-way ANOVA.


There is an urgent need for new pain-relieving therapies (34). Here, we used animal models of inflammatory and neuropathic pain to show that a single injection of compounds derived from botulinum toxin can silence pain-processing neurons in the spinal cord and decrease pain hypersensitivity. In two sets of experiments, we targeted NK1R-expressing neurons that relay pain-related information from the spinal cord to the brain and the MOR-expressing spinal cells that modulate activity of NK1R-expressing output neurons (10, 30, 35). We describe a long-term effect on mechanical pain sensitivity on both inflammatory and neuropathic pain states after a single injection of the constructs and demonstrate in vivo receptor specificity. We found no additive effects of SP-BOT and Derm-BOT, suggesting that, although the constructs silence different neurons, they are likely to be part of the same neural pain network with MOR-expressing excitatory neurons modulating NK1R-positive projection neurons. Hence, these new botulinum constructs would appear to be equally useful in reducing pain hypersensitivity.

Among the seven types (A to G) of botulinum toxin that target neurons, because of its long-lasting activity and high efficiency, BoNT/A has been approved by the U.S. Food and Drug Administration for treating a variety of disorders (3640). In neuronal cultures, the proteolytic activity of BoNT/A persists beyond 80 days, whereas other subtypes of BoNT have shorter half-lives (37, 41). Peripheral injections of botulinum neurotoxins have been shown to reduce both neuropathic pain and the frequency of migraine attacks in human patients (4244). This antinociceptive action has been exploited by a number of groups (44, 45); more recently, using a synthetic procedure, it was possible to separate the pain relieving from the paralyzing actions by synthesizing peptide components of BoNT/A and “restapling” them into a unique configuration (15, 16). Systemic administration of these reassembled molecules was shown to inhibit neuronal activity without causing toxicity (18). To generate the botulinum-based molecules, we used a synthesis procedure that allowed nonchemical linking of recombinantly produced proteins using core components of the SNARE complex to achieve irreversible linkage of two separate peptide fragments into a functional unit (16). This approach was particularly important because the production of functional botulinum-based molecules has significant safety issues due to protein toxicity. The assembly of the functional toxin from innocuous parts is therefore an important advance because safety issues have severely restricted the development of botulinum-derived molecules for medical use.

We generated new molecules by substituting the nonspecific neuronal binding targeting domain of BoNT/A with ligands that recognize the key G protein–coupled neurotransmitter receptors NK1R and MOR. Binding to these receptors was followed by internalization of the construct and, because of the inclusion of the translocation domain into our constructs, release of the protease domain of the toxin into the cytoplasm and inhibition of synaptic release. The synthesis of SP-BOT has been previously described (15), but synthesizing the Derm-BOT construct required further synthetic steps. Dermorphin is a potent selective MOR agonist (31, 32, 46) and has been successfully used previously in saporin conjugates to selectively ablate MOR-expressing neurons (13). Conjugation of dermorphin to the LcTd portion of botulinum was complicated because dermorphin binds to the receptor through its N terminus, the portion of the molecule generally used for the botulinum conjugation procedure (16). To circumvent this problem, we introduced a synthetic inversion procedure (see Methods) that allowed conjugation of dermorphin to the LcTd portion of botulinum toxin while retaining the free N terminus of dermorphin for binding to the MOR, followed by internalization and SNAP25 cleavage.

It is likely that the separation of the botulinum translocation domain from the neuropeptide ligands using the “stapling” mechanism allowed sufficient freedom for the translocation domain to perform the pH-dependent structural transition necessary to facilitate transfer of the botulinum protease from the luminal space of vesicle into the neuronal cytosol. However, it has been reported (47) that attachment of SP directly to botulinum protease allowed entry into neurons and SNAP25 cleavage. Omission of the obligatory translocation domain from the construct suggests that the activity would have been suboptimal and may account for the short in vivo efficacy (47).

SNAP25, the unique substrate for botulinum peptidase activity, is found throughout dendrites, where a role in spine morphogenesis has been proposed (48), and in cell bodies and axons. NK1Rs are located on dendrites and cell bodies, whereas MOR is also found on axons and axon terminals and binding and internalization would be expected at most receptor binding sites (25, 49, 50). Given the lack of axonal NK1R expression, the presence of cSNAP25 in spinal to brainstem axons after spinal treatment with SP-BOT was most likely the result of cleavage of SNAP25 in NK1R-positive dendrites and cell bodies in the dorsal horn, followed by axonal transport of cSNAP25 and/or the protease to the synaptic terminals within the brainstem.

NK1R-expressing spinal projection neurons have been shown to be essential for the maintenance of pain states (11). Information related to injury reaches the brain largely through NK1R-positive projection neurons of the superficial dorsal horn that terminate massively in the parabrachial area of the brainstem and, to a lesser extent, within the thalamus (35, 51). The parabrachial area is crucial for supplying information to forebrain areas that generate the affective-motivational component of pain (52, 53), whereas thalamic afferents terminate within cortical areas concerned with both pain discrimination and affect (51). Forebrain activation can, in turn, regulate dorsal horn sensitivity by activating descending controls from the brainstem to the spinal cord (3, 12, 54). Thus, a shift in the balance between pain inhibiting and facilitating controls from the brainstem, informed by NK1R-positive dorsal horn projection neurons, plays a role in setting spinal nociceptive thresholds required by on-going behavioral priorities and may ultimately contribute to pathological pain states (54). It follows that the inflammatory and neuropathic mechanical allodynia are disrupted by intrathecal ablation or silencing of these NK1R-expressing projection neurons with SP-saporin (11) or SP-BOT constructs, respectively. Recent work has shown that chemotoxic ablation of the NK1R-positive pain pathway in companion dogs can relieve bone cancer pain (14), demonstrating the applicability of the approach to higher mammals in different pain subtypes.

The disadvantage of the SP-saporin procedure is that it takes several weeks to become effective and kills neurons (11, 14). Our intention was to design a reversible and nontoxic molecule that would achieve the same analgesia rapidly and without cell death. The approach described here using SP-BOT silences NK1R-expressing neurons without cell death and is effective in days rather than weeks; in addition, SP-BOT is relatively easy to synthesize. As expected, the analgesic effect of SP-BOT constructs was completely lost in NK1R−/− mice.

MOR is expressed by dorsal horn interneurons and found in some small-diameter primary afferent sensory fibers (49, 55, 56). However, previous research has implied that the opioid tolerance and opioid-induced hyperalgesia that follow repeated injections of morphine are mediated by primary afferent MORs (55). It was also shown that intrathecal morphine produced strong mechanical and thermal antinociception in naïve mice but that was lost in mice in which MOR had been deleted only from primary afferents (55), suggesting that spinal neurons expressing MORs did not play a role in setting baseline mechanical thresholds or the generation of analgesic tolerance after repeated injections of morphine. However, intrathecal Derm-BOT in naïve mice reported here had no effect on baseline mechanical pain sensitivity but only on mechanical thresholds in injury-induced pain states. This suggests that the target for Derm-BOT–mediated analgesia was not primary afferents expressing MOR but MOR-positive dorsal horn neurons. A similar result was reported in rats after the partial ablation of MOR-expressing neurons with dermorphin-saporin (Derm-SAP) conjugates (57), raising the possibility that presynaptic opiate receptors may not internalize after opiate agonist administration (58).

Currently, new approaches to the control of chronic pain have adopted both central intrathecal and peripheral systemic approaches. Intrathecal opioids and other drugs are often given in clinical practice to relieve chronic pain when other treatment routes are exhausted or to circumvent the inherent risks of long-term systemic opioid treatment. However, intrathecal administration requires a surgically embedded pump to administer a prolonged infusion of the drug to the spinal cord (59, 60). Intrathecal treatments primarily target and inhibit central sensitization, the driving force behind chronic pain states. Unfortunately, long-term intrathecal opioid administration can result in respiratory depression, intrathecal granuloma, opioid tolerance, and other serious side effects (61). Moreover, although systemic opioids remain the gold standard for pain control, there are major concerns around the problems of drug overdose and addiction in part due to the relaxation of prescribing of opioids for nonterminal chronic pain (9). Conjugates of the silencing domain of botulinum toxin with SP or dermorphin provide substantial analgesia without evident toxic effects and over long periods of time after a single intrathecal injection. Complete analgesia is not entirely desirable. As clinical studies with antinerve growth factor, antibodies have demonstrated that encouraging the use of an already damaged limb may have resulted in further damage leading to hip or knee replacement (62, 63). The successful use of SP-saporin in rodents and dogs also opens up the possibility that silencing of this pathway with SP-BOT might be sufficient to control chronic pain states in human patients without permanent damage to the spinal cord (11, 14). In addition, the side effects of chronic opioid use including analgesic tolerance, paradoxical opioid-induced hyperalgesia, and addiction (64) might be avoided by a single intrathecal injection of the Derm-BOT construct.

Translating knowledge from preclinical observations in animal models of pain states to new therapies in the clinic has been difficult and has met with limited success. Differences between animal behavioral tests and human chronic pain features, particularly the assessment of both sensory and affective features of the pain state, and measurements of long-term efficacy and species variability may have been confounding factors (6). Nevertheless, the successful translation of the SP-saporin treatment from rats to companion dogs with bone cancer pain suggests (11, 14) that there is potential for the introduction of botulinum-based silencing approaches for the control of pain without cytotoxicity or recourse to repeated treatment of analgesics that can produce adverse behavioral effects.


Study design

This study was designed to evaluate the effect of SP-BOT and Derm-BOT on pain sensitivity. In behavioral studies, mice were randomly assigned to experimental groups. The experimenter was always blind to treatment and genotype. We could not predict a priori the effect size for the botulinum constructs, and we were guided by Mead’s resource equation. Therefore, we aimed to use at least 6 mice in each group and no more than 11. Occasionally, mice were excluded from the study if they were found to have bodily damage from fighting with cage mates (5 of 206 total mice were discarded). We did perform statistical analysis at the end of each round of experiments to satisfy the 3Rs (replacement, reduction, and refinement), which dictates that “The number of animals used should be the minimum number that is consistent with the aim of the experiment” ( Raw data for all experiments is presented in table S3.


Subjects in all experiments were adult mice (8 to 12 weeks old). WT mice were C57BL6/J from Envigo. NK1R−/− and WT littermates were obtained from a colony of mice derived from a 129/Sv × C57BL/6 genetic background (21). NK1R−/− mice were backcrossed with a WT C57BL6/J mouse for several generations. Experiments were always carried out using littermates from heterozygous breeding pairs. All mice were kept in their home cage in a temperature-controlled (20° ± 1°C) environment, with a light-dark cycle of 12 hours (lights on at 7:30 a.m.). Food and water were provided ad libitum. All efforts were made to minimize animal suffering and to reduce the number of animals used (UK Animal Act, 1986).


For genotyping, DNA was extracted from ear tissue, and the following primers were used for polymerase chain reaction (PCR): NK1R primer, 5′-CTGTGGACTCTAATCTCTTCC-3′ (forward) and 5′-ACAGCTGTCATGGAGTAGATAC-3′ (reverse); neomycin-resistant gene (NeoF) primer, 5′-GCAGCGATCGCCTTCTATC-3′. Samples from WT mice showed a single PCR product of 350 base pairs (bp); samples from NKR1−/− mice showed a single PCR product of 260 bp; and samples from heterozygous mice would present both bands (21).

Design and purification of botulinum constructs

Each BoNT/A consists of three domains: the binding domain, the translocation domain, and the catalytic light-chain domain, a zinc metallopeptidase. We used a protein stapling technique to produce LcTd conjugated to SP or dermorphin, a naturally occurring mu-opioid agonist that carries an unnatural d-amino acid, making it resistant to internal proteolysis. The synthesis that has been described previously for SP with in vitro controls for specificity is detailed in (15). Briefly, to synthesize the constructs, first, fusion protein consisting of the LcTd of the botulinum type A1 strain was fused to SNAP25 (LcTd-S25) and was prepared as previously described (16, 65). The chemically synthesized syntaxin-SP peptide had the sequence Ac-EIIKLENSIRELHDMFMDMAMLVESQGEMIDRIEYNVEHAVDYVE-Ahx-Ahx-RPKPQQFFGLM-NH2, where Ahx stands for aminohexanoic acid. Because of the need for the N terminus of dermorphin to be accessible for binding to the MOR, the syntaxin-dermorphin peptide was synthesized in two parts, syntaxin-maleimide and dermorphin-cysteine, which were then bio-orthogonally conjugated through two reactive C termini. The dermorphin and syntaxin sequences were YaFGYPS and EIIRLENSIRELHDMFMDMAMLVESQGEMIDRIEYNVEHAVDYVEK, respectively.

Second, the protein “staple” was prepared recombinantly from the rat vesicle-associated membrane protein 2 (VAMP2) sequence (amino acids 3 to 84) inserted into the XhoI site of pGEX-KG. Oriented attachment of peptides to protein was achieved by the SNARE assembly reaction. LcTd-S25, VAMP2 (3 to 84), and either syntaxin-dermorphin or syntaxin-SP were mixed at a molar ratio of 1:1:1 in 100 mM NaCl (sodium chloride), 20 mM Hepes, and 0.4% n-octylglucoside at pH 7.4 (buffer A). Reactions were left at 20°C for 1 hour to allow formation of the SNARE ternary complex. SDS-resistant and irreversibly assembled protein conjugates were visualized using Novex NuPAGE 12% bis-tris SDS–PAGE (polyacrylamide gel electrophoresis) gels (Invitrogen) run at 4°C in a NuPAGE MES SDS running buffer (Invitrogen). All recombinant proteins were expressed in the BL21-Gold (DE3)pLyss strain of Escherichia coli (Agilent) in pGEX-KG vectors as glutathione S-transferase (GST) C-terminal fusion proteins cleavable by thrombin. GST fusion constructs were purified by glutathione affinity chromatography and cleaved by thrombin. Synthetic peptides were made by Peptide Synthetics Ltd.

Cortical cultures

To confirm construct efficacy, rat cortical neurons were dissected from 8 to 12 embryonic day 17 rat pups and washed in Hanks’ balanced salt solution (HBSS) before being treated with trypsin for 15 min at 37°C, followed by addition of deoxyribonuclease (DNase; Sigma-Aldrich). Cells were resuspended in 1 ml of triturating solution [1% AlbuMAX (Gibco), trypsin inhibitor (0.5 mg/ml; Sigma-Aldrich), and DNase in HBSS (1 μg/ml)]. Cells were triturated using three progressively smaller glass pipettes before being diluted to 5 ml by the addition of cortical medium. Fifty thousand cells in 150 μl of medium were plated on 96-well plates coated with poly-d-lysine. Cells were maintained in a neurobasal medium (Gibco) supplemented with 1% B27 (Gibco), 1% penicillin/streptomycin, and 1% GlutaMAX (Gibco). Half the medium was changed every 3 to 4 days, and cultures were tested between 1 and 3 weeks after plating.

Western analysis of botulinum activity

Derm-BOT and SP-BOT [400 nM in buffer A (100 mM NaCl and 20 mM HEPES)] were added to the plated cortical cells at a 1:20 dilution to achieve the final concentration of 20 nM. Cells were incubated at 37°C, 5% CO2 for 65 hours before culture media was removed, and 20 μl of a loading buffer [56 mM sodium dodecyl sulfate, 0.05 M tris-HCl (pH 6.8), 1.6 mM UltraPure EDTA (Gibco), 6.25% glycerol, 0.0001% bromophenol blue, 10 mM MgCl2, benzonase (26 U/ml; Novagen)] was added to each well. Plates were shaken at 900 rpm for 10 min at 20°C, and samples were transferred to a fresh 0.5-ml tube. Samples were boiled for 3 min at 95°C and then run on Novex NuPAGE 12% bis-tris SDS-PAGE gels (Invitrogen). After separation, proteins were transferred onto immobilin-P polyvinylidene difluoride membranes and then incubated for 30 min in blotting solution [5% milk and 0.1% Tween 20 in phosphate-buffered saline (PBS)]. Mouse monoclonal SMI81 antibody (anti-SNAP25) was added at 1:2000 dilution to the blotting solution at 4°C for overnight incubation. Membranes were washed three times in 0.1% Tween 20 in PBS for 5 min and then incubated for 30 min in the blotting solution containing secondary peroxidase-conjugated donkey anti-rabbit antibody (Amersham) at a 1:2400 dilution. Membranes were washed three times for 5 min in 0.1% Tween 20 in PBS. Immunoreactive protein bands were visualized using SuperSignal West Dura Extended Duration solution (Thermo Fisher Scientific) with exposure to Fuji Medical X-ray Films (Fuji).

Behavioral testing

von Frey filament test. The experimenter was always blind to genotype and treatment group for all behavioral tests. Animals were placed in Plexiglas chambers, located on an elevated wire grid, and allowed to habituate for at least 1 hour. After this time, the plantar surface of the paw was stimulated with a series of calibrated von Frey’s monofilaments. The threshold was determined by using the up-down method (66). The data are expressed as log of the mean of the 50% pain threshold ± SEM. In some cases, the data were plotted as force (gram; figs. S8 and S9).

Rotarod test. Motor performance was evaluated by an accelerating rotarod apparatus with a 3-cm-diameter rod starting at an initial rotation of 4 rpm and slowly accelerating to 40 rpm over 100 s. Mice were expected to walk at the speed of rod rotation to keep from falling. The time spent on the rod during each of two trials per day was measured and expressed in seconds. Animals were tested only once at baseline to minimize the number of tests on the rotarod. Testing was completed when the mouse fell off the rod (that is, from a height of 12 cm).

Pain models

Mouse inflammatory models.
CFA-induced ankle joint inflammation

Inflammation was induced by injection of 5 μl of CFA (Sigma-Aldrich) into the left ankle joint under isoflurane anesthesia induced in a chamber delivering 2% isoflurane combined with 100% O2 and maintained during injection via a face mask. The needle entered the ankle joint from the anterior and lateral posterior position, with the ankle held in plantar flexion to open the joint (67).

CFA-induced hind paw inflammation

CFA (20 μl) was injected subcutaneously into the plantar surface of the left hind paw using a microsyringe with a 27-gauge needle. Mice were maintained under isoflurane anesthesia during the injection.

Mouse neuropathic model: SNI

The SNI was performed as previously described (68). Briefly, under isoflurane anesthesia, the skin on the lateral surface of the thigh was incised, and a section made directly though the biceps femoris muscle exposing the sciatic nerve and its three terminal branches: the sural, the common peroneal, and the tibial nerves. The common peroneal and the tibial nerves were tight-ligated with 5-0 silk and sectioned distal to the ligation. Great care was taken to avoid any contact with the spared sural nerve. Complete hemostasis was confirmed, and the wound was sutured.

Intrathecal injections

Intrathecal injections were performed under anesthesia (69). The mice were held firmly but gently by the pelvic girdle using thumb and forefinger of the nondominant hand. The skin above the iliac crest was pulled tautly to create a horizontal plane where the needle was inserted. Using the other hand, the experimenter traced the spinal column of the mouse, rounding or curving the column slightly to open the spaces between vertebrae. A 30-gauge needle connected to a 10-μl Hamilton syringe was used to enter between the vertebrae. After injection, the syringe was rotated and removed, and posture and locomotion were checked. All intrathecally delivered drugs were injected in a 3-μl volume.


Mice were anesthetized with pentobarbital and perfused transcardially with physiological saline containing heparin (5000 IU/ml), followed by 4% paraformaldehyde (PFA) in a 0.1 M phosphate buffer (PB; 25 ml per adult mouse). Lumbar spinal cords were dissected out, fixed in 4% PFA for an additional 2 hours, and transferred into a 30% sucrose solution in a PB containing 0.01% azide at 4°C for a minimum of 24 hours. Spinal cord sections were cut on a freezing microtome set at 40 μm. For fluorescent immunohistochemistry, sections were left to incubate with primary antibodies overnight at room temperature (anti-cSNAP25 antibody recognizing the cleaved end of SNAP25 1:50,000 ref, TRIDEANQ; anti-NK1, guinea pig, 1:10,000, Neuromics; anti-MOR, rabbit, 1:10,000, Neuromics). For NK1R and MOR immunohistochemistry, direct secondary antibody was used at a concentration of 1:500 (Alexa Fluor). For cSNAP25 staining, appropriate biotinylated secondary antibody was used at the concentration of 1:400 and left for 90 min. Sections were then incubated with avidin-biotin complex (1:250 Vectastain A plus 1:250 Vectastain B; ABC Elite, Vector Laboratory) for 30 min, followed by a signal amplification step with biotinylated tyramide solution (1:75 for 7 min; PerkinElmer). Finally, sections were incubated with fluorescein isothiocyanate–avidin for 2 hours (1:600). An antibody against Iba1 (ionized calcium binding adaptor molecule 1; goat, 1:500, overnight, Abcam) was used to identify microglia and an anti-GFAP (glial fibrillary acidic protein) antibody to stain for astrocytes (rabbit, 1:4000, overnight, Dako) by immunohistochemistry. The direct secondary antibody was used at a concentration of 1:500 (Alexa Fluor). All fluorescent sections were transferred to glass slides and cover slips applied with Gel Mount Aqueous Mounting Medium (Sigma-Aldrich) to prevent fading and stored in dark boxes at 4°C. In colabeling studies, controls included omission of the second primary antibody.

Quantification of fluorescence

For quantification of NK1R and MOR fluorescence, a region of interest (ROI) was located over laminae I/II. The ROI was 3087 μm2 for NK1R and 1617 μm2 for MOR immunostained tissue. Fluorescence was measured for six sections per animal using the same ROI. Readings were taken from the side of the spinal cord contralateral to the inflamed paw or nerve lesion. Contrast enhancement and fluorescence threshold were kept constant. Readings from saline and botulinum construct intrathecal-injected mice were compared.

c-Fos immunohistochemistry

c-Fos immunohistochemistry was used to assess the silencing of the lamina I NK1R-positive neurons. Preemptive intrathecal treatment with SP-BOT in naïve mice was followed 3 days later by an injection of CFA into the left paw under isoflurane anesthetic. Six hours later, animals were perfused and processed for c-Fos expression in the lateral parabrachial area. For DAB (3,3′-diaminobenzidine), sections were blocked in a PB with 3% serum, 3% triton, and 2% H2O2 for 1 hour and then incubated over weekend with the primary antibody (anti–c-Fos, rabbit, 1:10000, Millipore Merck KGaA). The sections were then incubated in an appropriate secondary antibody at 1:500 for 2 hours, followed by incubation with avidin-biotin complex (1:1000 Vectastain A plus 1:1000 Vectastain B; ABC Elite, Vector Laboratory) for 1 hour. The substrate was then developed using a peroxidase substrate DAB kit (Vector #SK4100) at optimized times, and the sections were washed and mounted. The following day, the sections were dehydrated in increasing ethanol concentrations (70%, 70%, 95%, 95%, 100%, 100%, histoclear ×2) and coverslipped with DPX.

Five sections through the LPb from each mouse were analyzed for population density of c-Fos neurons. c-Fos–immunoreactive neurons were counted in the lateral parabrachial area bilaterally. Counts from the sections were averaged, and the mean was used for further statistical analysis. To quantify cSNAP-positive neuronal cell bodies, four spinal cord sections from each mouse were counted. Means were taken for each treatment for further analysis. Counts were from laminae I to III of the dorsal horn.

Statistical analysis

All statistical tests were performed using the IBM SPSS Statistics programme (version 20), and P < 0.05 was considered statistically significant. For the behavioral experiments, statistical analysis was performed on the data normalized by log transformation (von Frey data), as suggested by Mills et al. (70). Difference in sensitivity was assessed using repeated measures two-way or one-way ANOVA, as appropriate and as indicated. In all cases, a significant effect of the main factor(s), or interactions between them, was taken as the criterion for progressing to post hoc analysis. Bonferroni correction was the preferred post hoc approach when we had three groups or more; in this case, if the general ANOVA was significant but no Bonferroni significance was observed, then we also reported the results of the least significant difference post hoc analysis. When we had two groups, we report the result of the one-way ANOVA. In all cases, “time” was treated as a within-subjects factor, and “genotype” and “treatment” were treated as between-subject factors. The statistical significance in Fig. 2D was determined using one-way ANOVA, followed by Fisher’s least significant difference test.

The MPE was calculated as:Embedded Imagewhere log(0.6 g) is our maximum von Frey’s force applied. Please note that, as in our previous paper (67), we logged the data of the behavioral tests to ensure a normal distribution because the von Frey’s hairs are distributed on an exponential scale. Mills et al. recently demonstrated that log transformation makes more “mathematical and biological sense” (70).


Fig. S1. Synthesis of botulinum peptide conjugates using a stapling bridge.

Fig. S2. SP-BOT has no effect on mechanical threshold in the contralateral paw.

Fig. S3. Effect of different doses of intrathecal SP-BOT or Derm-BOT on CFA-induced hypersensitivity.

Fig. S4. Intrathecal injection of unconjugated BOT LcTd (Neg-BITOX) without a receptor binding domain had no effect on inflammatory hyperalgesia.

Fig. S5. cSNAP25-positive neurons after SP-BOT or Derm-BOT intrathecal injection.

Fig. S6. SP-BOT or Derm-BOT intrathecal injection does not induce glial activation in the dorsal horn.

Fig. S7. Quantification of MOR fluorescence intensity.

Fig. S8. Effect of SP-BOT injection on withdrawal threshold plotted as force.

Fig. S9. Effect of Derm-BOT injection on withdrawal threshold plotted as force.

Table S1. Statistical analysis for Figs. 1, 2, 3, and 5 and figs. S2 and S7.

Table S2. Maximum possible effect.

Table S3. Raw data (Excel file).


Acknowledgments: We thank S. M. Géranton for the helpful discussion during the preparation of the manuscript. We would like to thank D. Wheeler and J. Mullen for comments on the manuscript. We also thank K. de Vos and R. Bresnahan from the University of Sheffield for supplying the rat cortical neurons and S. Beggs for the instruction in image analysis. Funding: This work was supported by the Medical Research Council grant MR/K022539/1. Author contributions: M.M., B.D., and S.P.H. designed experiments. J.A., C.L., and B.D. designed and synthesized botulinum constructs. S.P.H., M.M., C.L., and B.D. wrote the manuscript. M.M. and M.C. conducted behavioral experiments. M.M., I.E.-A., and A.S.M. conducted immunohistochemical experiments. M.M. and S.P.H. analyzed data. Competing interests: The authors declare that they have no competing interests. Data and materials availability: All data associated with this study are present in the paper or the Supplementary Materials.
View Abstract

Navigate This Article