Long-term Effects of Botulinum Toxin on Neuromuscular Function

Approval was obtained from the Subcommittee on Animal Care Research at the Massachusetts General Hospital, Boston, Massachusetts. Male Sprague-Dawley rats, weighing 200–260 g, were used. Animals were allowed to accommodate to the standard conditions of our animal facility with free access to rat chow and water for at least a week.

The animals were randomly allocated to three experimental groups, each group receiving differing doses of botulinum toxin (0.625 U, n = 12; 2.5 U, n = 12; and 10 U, n = 12). Botulinum toxin (Botox®; Allergan Inc., Irvine, CA) was reconstituted with 0.9% sterile saline. The stock solution of 100 U was further diluted with 0.9% sterile saline according to the manufacturer’s instructions. Controls (n = 12) received an equivalent volume of saline.

For the injection, the rats were anesthetized with pentobarbital (60 mg/kg intraperitoneal), the limb shaved and disinfected. One group of animals received injection of botulinum toxin only. The total volume (0.5 ml) of diluted botulinum toxin was aliquoted into two equal parts (0.25 ml) and then injected into the medial and lateral aspects of the middle of the tibialis cranialis muscle belly, where the neuromuscular junction is usually located. The contralateral (noninjected) side that received no injection served to study the distant effects of the toxin. A separate control group of animals received an equivalent volume of saline using the same injection technique. The contralateral (noninjected) side that received no injection served as naive control. After injection of toxin or saline, the animals were monitored until recovery from anesthesia, and then returned to their cages. Clinical studies indicate that the effects of botulinum toxin last for approximately 3 months. This study, therefore, looked at the neuromuscular effects beyond that period, at 128 days after injection.

Anesthetized with pentobarbital (60 mg/kg intraperitoneal) at 128 days, the animals were tracheostomized and ventilated using a Harvard ventilator. The right jugular vein was catheterized for drug and fluid administration. The right carotid artery was cannulated to measure arterial blood pressure and perform blood gas analyses. Heart rate and mean arterial pressure were continuously monitored to ensure stable hemodynamic conditions throughout the experiment. Arterial oxygen tension (Pao²), arterial carbon dioxide tension (Paco²), and acid–base status were intermittently measured and corrected if necessary. Body temperature was monitored with a rectal probe and maintained at 36°–38°C by a heat lamp. Surgical anesthesia was maintained with repetitive doses of pentobarbital. The administration of repeat doses of pentobarbital was based on cardiovascular signs of inadequate anesthesia.

Rats were excluded from the experiment if they were hemodynamically unstable (mean arterial blood pressure < 80 mmHg) at the beginning of the experiments or if their blood gas status throughout the experiment was not within the defined predefined ranges (Pao²> 100 mmHg; pH 7.36–7.44; Paco²= 36–44 mmHg; base excess −2 ± 2 mEq).

Neuromuscular transmission and function were monitored by evoked mechanomyography using a peripheral nerve stimulator (NS252; Fisher & Paykel Health Care, Irvine, CA) along with a Grass Force transducer and software (Grass Instruments, Quincy, MA). For this purpose, rats were placed in the dorsal recumbent position. The tendon of the insertion of tibialis cranialis muscle was surgically exposed on each side and individually attached to separate Grass FT03 force displacement transducer (Grass Instruments). Both sciatic nerves were exposed at its exit from the lumbosacral plexus at the thigh and tied with ligatures. Distal to the ligatures, stimulation electrodes were attached for nerve-mediated stimulation of the tibialis cranialis. The knees were stabilized rigidly with a clamp. A baseline tension of approximately 50 g was applied to the tendon of each tibialis cranialis muscle, which yielded optimal evoked tensions. The nerve-evoked tension of the respective tibialis muscle was recorded via  a Grass P122 amplifier and displayed using the Grass Polyview Software (Grass Instruments).

Initially, during supramaximal stimuli, baseline twitch tension was established. The baseline mechanomyographic responses were stabilized over a period of 10 min using the train-of-four (TOF) stimulation pattern, every 20 s. The evoked muscle tension developed during TOF stimulation was recorded at the end of the 10-min period. This was followed by tetanic stimulation at 50 Hz for 5 s to assess the maximal tetanic muscle tension and the muscle fade associated with this stimulus. After an interval of 30 min with ongoing TOF stimulation every 20 s, to allow muscle recovery from the effects of the preceding tetanus, the potency of atracurium, a nondepolarizing muscle relaxant, was tested on both sides by the cumulative dose–response method. Bolus doses of atracurium were given intravenously in increments of 0.1–0.4 mg/kg until the first twitch height (T1) of the TOF was below 5% of the baseline twitch tension (95% twitch depression) on both sides. Each incremental dose was only given when the previous dose had produced maximal effect, as indicated by three equal consecutive T1 twitches. After the last dose of atracurium, the twitch was allowed to recover to baseline values.

After complete recovery of T1 (> 95% recovery of twitch height), a continuous infusion of atracurium was started, and the infusion rate was adjusted to achieve a constant 50% neuromuscular block on the botulinum toxin–injected or saline-injected side. After 10 min of stable 50% twitch depression on the injected side at that stable infusion rate, steady state conditions were assumed to be present. At this point, the twitch height on the noninjected side was also noted during the steady state 50% twitch suppression on the injected side. One milliliter of heparinized blood was withdrawn for later determination of plasma concentrations of atracurium. The blood was immediately transferred to an Eppendorf tube containing 20 μl H²SO⁴, 1 m, and centrifuged (3,500 rpm, 10 min, 4°C). The plasma was collected, and 0.2-ml portions were aliquoted into Eppendorf tubes containing 0.8 ml H²SO⁴, 15 mm. The samples were immediately frozen at −80°C. After blood sampling, tibialis cranialis, gastrocnemius, and soleus muscles on both sides were harvested, weighed, snap frozen on dry ice, and stored at −80°C for later biochemical analysis of nAChR concentration.

The nAChR protein expression in the tibialis cranialis muscles was assayed using the ¹²⁵I-α-bungarotoxin binding assay, as previously described,24and expressed as femtomole nAChRs per milligram protein. The protein concentration of the muscle extract was assayed using the Bio-Rad DC protein assay kit (Bio-Rad Laboratories, Herkules, CA). The plasma concentration of atracurium was determined by high-performance liquid chromatography as described previously.25 

All values are expressed as mean ± SEM. Assuming that the twitch response or relation of the neuromuscular block to the dose of atracurium is governed by the Hill equation, linear regressions of the degree of block in logit scale and the respective cumulative dose of atracurium in log scale were calculated for every rat on each leg. These linear relations are typically characterized by slope (Hill coefficient) and intercept. Individual effective doses (EDs) were calculated by simple retransformation into linear scale. Values were compared by two-way repeated-measurement analyses of variance, using dose (between groups) and leg (intragroup and intergroup) as independent factors. Post hoc  testing was performed with paired or unpaired t  tests (P < 0.05).

All injected animals survived for 128 days. Originally, 12 rats were injected in each group. Hemodynamic and metabolic variables were monitored throughout the functional and pharmacologic studies. Nine rats in total were excluded from the experiment because of hemodynamically instability (mean arterial blood pressure < 80 mmHg) at the beginning of the experiments or because their blood gas status throughout the experiment was not within the defined predefined ranges (Pao²> 100 mmHg; pH 7.36–7.44; Paco²= 36–44 mmHg; base excess −2 ± 2 mEq) (n = 9). Therefore, the final statistical analyses included n = 9 in the saline group, n = 11 in the 0.652-U group, n = 10 in the 2.5-U group, and n = 9 in the 10-U group. In these animals, hemodynamic and metabolic values were not different between groups. During the cumulative dose–response, the mean Pco²was 37.8 ± 1.3, arterial blood pH ranged between 7.36 and 7.44, and base excess was −0.5 ± 0.6.

At 128 days after injection of toxin, there were no differences in the tibialis muscle tensions between the saline-injected and contralateral side, and also the side contralateral to botulinum toxin injection. Relative to these muscles, there was a significant decrease in evoked muscle tension on the botulinum toxin–injected side in a dose-dependent manner. Animals injected with 0.625, 2.5, or 10 U showed a significant (P < 0.05) decrease in evoked single twitch tension (T1 of TOF) ranging from 77.4% to 46.4%, respectively, compared with the contralateral (noninjected) side. The muscle tension was also significantly (P < 0.05) reduced to 60.7% (2.5 U) and 46.4% (10 U) when compared with the salineinjected controls (fig. 1A). Although the animals injected with 0.625 U had lower twitch tensions compared with the contralateral side, the tension in limb injected with 0.625 U was not different compared with the saline-injected control. The specific muscle tensions (twitch tension per mg wet tibialis muscle mass) were not different between the botulinum toxin–injected side, the contralateral (noninjected) side, and the saline-injected controls (fig. 1B). Response to tetanic stimulation and the calculated specific tetanic tensions followed a pattern, similar to that of single twitch tension in that the absolute tetanic tensions were decreased on the toxin side but not the specific tetanic tensions (figs. 2A and B). Fade after TOF or tetanus was not different between groups.

The ED¹⁰of atracurium was significantly reduced in the leg receiving 10 U botulinum toxin (table 1). No differences, however, were observed in the ED⁵⁰values between the experimental and the contralateral (noninjected) side or the saline-injected and contralateral muscles. The plasma concentration of atracurium to achieve steady state 50% neuromuscular paralysis on the toxin injected side was significantly lower in the highest botulinum toxin dose group (10 U) compared with the saline-injected group or the medium-dose (2.5-U) and low-dose (0.625-U) botulinum toxin groups (table 1). The plasma atracurium concentrations to achieve 50% neuromuscular paralysis on the injected side in the medium- and small-dose toxin groups were similar to that of controls. (The infusion rate was titrated to achieve a 50% neuromuscular block on the injected side.) The twitch suppression obtained on the contralateral side was similar to the injected side at all times, except in the 10-U group. The measured block achieved on the contralateral (noninjected) side was significantly smaller in the highest toxin dose (10-U) group (table 1). In contrast to ED¹⁰or ED⁵⁰, the ED⁹⁰in the 10-U group was significantly increased on both the botulinum toxin and the contralateral noninjected side relative to the saline-injected and the contralateral leg. The trend for increased ED⁹⁰on the 2.5 U–injected side did not reach statistical significance (P > 0.05). Consistently, the slopes of the dose–response curve were significantly reduced in the groups receiving botulinum toxin doses of 2.5 and 10 U compared with the contralateral (noninjected) leg or the saline-injected leg (table 1).

There was a significant (P < 0.05) dose-dependent decline in the tibialis cranialis muscle mass to 80.0% (0.625 U), 66.6% (2.5 U), and 50.0% (10 U) on the injected side compared with the contralateral (noninjected) leg. The tibialis cranialis muscle on botulinum toxin–injected side also weighed 88.8% (0.625 U), 66.6% (2.5 U), and 55.5% (10 U) of the saline-injected controls (table 2). Muscle mass of the adjacent gastrocnemius muscle was significantly (P < 0.05) reduced in the 10-U group when compared with the contralateral (noninjected) leg and the 0.625 U– and 2.5 U–injected legs. There were no differences when compared with the saline-injected leg. Muscle mass of the adjacent soleus muscle was also significantly (P < 0.05) decreased in the high-dose group compared with the contralateral side, the saline-injected leg, and the 0.625 U– and 2.5 U–injected legs (table 2).

The concentrations of membrane nAChRs in the tibialis anterior muscle were significantly (P < 0.05) increased compared with that of contralateral (noninjected) leg and to that of the saline-injected controls (fig. 3).

Our study documents that a single injection of the toxin can have long-lasting effects, even at 128 days after the initial injury. Despite lack of obvious anatomical denervation, our study confirms that the toxin can produce dose-dependent muscle weakness and a denervation-like state, evidenced as up-regulation of nAChRs. This observation is consistent with previous observations of the short-term neuromuscular effects of botulinum toxin.22,23The absence of complete anatomical denervation was electrophysiologically confirmed by the presence of nerve-evoked muscle contraction in the tibialis anterior muscle. Contrary to expectations, a normal sensitivity to the neuromuscular effect of the nondepolarizing muscle relaxant, atracurium was observed in the 0.625-U and 2.5-U toxin groups, despite up-regulation of nAChRs. Another novel and unexpected finding was that the highest dose of toxin (10 U) had a variable response depending on the ED value examined; ED¹⁰of atracurium was smaller, ED⁵⁰was normal, and ED⁹⁰was increased on the botulinum toxin–injected side relative to the saline-injected controls. An increased sensitivity to atracurium, evidenced as a decrease in plasma concentration requirement for a 50% paralysis, was also observed in the highest toxin (10-U) group, regardless of the profound up-regulation of nAChRs. The expression of nAChRs on the contralateral, noninjected side was similar to controls, suggesting that injection of the toxin does not have distant effects. Muscle mass of soleus and gastrocnemius on the ipsilateral side to botulinum toxin was decreased after injection of 10 U, suggesting spread to immediate local areas.

It is well known that botulinum toxin has almost exclusive effects at the nerve terminal, preventing the release of acetylcholine. Therefore, after botulinum toxin poisoning, despite normal nerve conduction up to the nerve terminal, release of acetylcholine does not occur.12,13The toxin, taken up by the acetylcholine vesicles, cleaves the SNARE proteins (SNAP-25, synaptobrevin, and syntaxin), preventing the assembly of the fusion complex and thus blocking the release of acetylcholine.26,27Recovery from the effects of the neurotoxin is partly due to axonal sprouting around toxin-blocked receptors and the formation of new motor endplates. Rate, frequency, and success of nerve sprouting seem to play an important role in determining recovery.28So far, it has not been determined how the toxin is metabolized or otherwise eliminated from the nerve ending. The effects of botulinum toxin are, therefore, similar to the effects of sodium channel blocker, tetrodotoxin, in that both prevent the release of acetylcholine at the nerve terminal leading to paresis.29,30An unexpected finding was that the botulinum toxin group did not show any fade during TOF or tetanus. Tetrodotoxin-induced decreased acetylcholine release leads to increased quantal content and prolongation of endplate current increasing synaptic strength in that muscle, most likely due to up-regulated nAChRs.29,30It is tempting to speculate that a similar mechanism was operative during botulinum toxin–mediated paresis; an increase in synaptic strength, due to up-regulated nAChRs, may have prevented the potential for fade during botulinum toxin–induced paresis.

Although the toxin has no direct effect on the muscle, loss of muscle mass or muscle wasting was observed even 128 days after toxin exposure. This observation is consistent with the results of Billiante et al .,21who observed long-lasting electromyographic and histologically atrophic changes up to 144 days after injection of the toxin. In our study, despite the presence of nerve-mediated contraction, muscle tension did not return to normal levels at 128 days. The muscle weakness or decreased tension-generating capacity was most likely due to loss of muscle mass because the specific single twitch and specific tetanic tensions (muscle tension per gram muscle weight) were not different between the experimental group and controls. At the highest dose of toxin, the effects could be seen even in the immediate vicinity beyond the tibialis, where the muscle mass was decreased in the soleus and gastrocnemius on the same side, but not at distant sides. The ED⁹⁰was, however, increased on the contralateral, noninjected side after the highest dose of toxin. This finding is difficult to explain, particularly because no changes in nAChRs or muscle mass were seen on the contralateral side. Nevertheless, anatomical denervation of one leg is known to have distant effects such as the diaphragma.31 

Classic theory of receptor pharmacology suggests that in the presence of increased receptor number (up-regulated nAChRs), the antagonist concentration or dose requirement to achieve a given effect (e.g ., paralysis) by a competitive antagonist (e.g ., atracurium) is increased while there is hypersensitivity to agonists such as succinylcholine.32–33Clinical reports do confirm that patients with botulinum toxin infection demonstrate a hyperkalemic response to succinylcholine.7,8In the current study, however, a trimodal response to the neuromuscular effects of the nondepolarizing muscle relaxant, atracurium, was observed. A leftward shift in the ED¹⁰, a normal ED⁵⁰, and a rightward shift in the ED⁹⁰were observed for atracurium. The unaltered ED⁵⁰and a higher ED for a more profound block (e.g ., ED⁹⁵) has been observed previously during partial denervation,31another condition with up-regulation and altered isoform expression.

This study did not quantitate the altered isoform expression. Even if one quantitated isoform expression by immunohistochemistry, the contribution of new receptor isoforms to neurotransmission would be difficult to assess. That a denervation state was indeed produced by the toxin was confirmed by the up-regulation of nAChRs. The possible reasons for the trimodal (ED¹⁰, ED⁵⁰and ED⁹⁰) responses to atracurium are as follows: After botulinum toxin–induced denervation/immobilization, there is de novo  expression of fetal forms of receptors. These new isoforms are expressed not only at the extrajunctional areas, but also perijunctionally and more importantly even junctionally, contributing to neurotransmission. Therefore, the mature receptors decrease in number and fetal isoforms increase in number at the junctional area. The increased sensitivity to atracurium (ED¹⁰) is probably related to inhibition by atracurium of the mature receptors that are decreased in number, resembling a myasthenic state. The decreased plasma concentration requirement for atracurium for 50% paralysis may reflect a continuum of the increased sensitivity of mature receptors. The unaltered ED⁵⁰is related to the interaction of atracurium with both mature and fetal isoform of receptors at midpoint. The discrepancy between ED⁵⁰and plasma concentration of atracurium to produce 50% paralysis may be due to hysteresis between dose, blood level, effect site concentration, and the lag-time for interaction of atracurium with the receptor during the cumulative dose–response studies. The fetal form of receptors consisting of this γ subunit and the α⁷subunit containing AChRs have decreased affinity for (resistance to) nondepolarizing muscle relaxants.31–35Therefore, blocking these remaining fetal nAChRs probably required higher doses. This could explain the higher ED⁹⁰and the flatter slope. The resistance to atracurium (increased ED⁹⁰) may also be related to inactivity-induced increased synaptic strength, which is related to increased quantal content, prolonged endplate current, and increased nAChRs number.29,30 

Healthy mammalian subjects typically present a slope of 3.5–5.6, irrespective of the neuromuscular blocking agent36and species. The slopes of approximately 4 in all control and contralateral legs and in the low dose (0.625 U)–injected leg are therefore consistent with previous observations in normals.36After injection of 2.5 or 10 U of toxin, the Hill coefficient ranged between 2.6 and 2.9 (table 1). These values indicate a flatter dose–response curve; i.e ., in these experimental groups, the affinity between nAChRs and atracurium was decreased. If the increased number of receptors consisted of the same isoform, one would expect a parallel rightward shift. Therefore, the flatter line also suggests an altered isoform expression, most likely due to fetal isoforms.

The clinical implications of this study are as follows: Typically, infection and immobilization as seen in the intensive care unit leads to up-regulation of nAChRs with resistance to nondepolarizing muscle relaxants31–35and increased sensitivity to succinylcholine with hyperkalemia.34,37,38Botulinum toxin–mediated up-regulation of nAChRs can lead to hyperkalemia with succinylcholine as reported in humans.7Based on our study, it is reasonable to state that dose requirements for nondepolarizing muscle relaxants would be variable after botulinum toxin infection, depending on the neuromuscular block that is required. In some instances, increased sensitivity may be seen in the affected muscles. A recent case report39confirms the increased sensitivity to a nondepolarizer in a muscle (orbicularis oculi) after cosmetic botulinum toxin injections compared with adductor pollicis muscle. Conversely, resistance to nondepolarizers may also be seen as indicated by the increased ED⁹⁰in our study. Furthermore, recovery from neuromuscular dysfunction after botulinum toxin infection may take several months and would be dependent on the severity of the infection. Human botulism immune globulin, which neutralizes the toxin, is now available.11Whether treatment with this can attenuate neuromuscular changes needs further study.