Objectives. We aim to determine the effects of Botulinum toxin type A (BTX-A) on neuropathic pain behavior and the expression of P2X3 receptor in dorsal root ganglion (DRG) in rats with neuropathic pain induced by L5 ventral root transection (L5 VRT).
Methods. Neuropathic pain was induced by L5 VRT in male Sprague-Dawley rats. Either saline or BTX-A was administered to the plantar surface. Behavioral tests were conducted preoperatively and at predefined postoperative days. The expression of P2X3 receptors in DRG neurons was detected by immunoreactivity at postoperative days 3, 7, 14, and 21.
Results. The number of positive P2X3 neurons in the ipsilateral L5 DRG increased significantly after L5 VRT (P < 0.001). This increase persisted for at least 3 weeks after the operation. No significant changes in P2X3 expression were detected in the contralateral L5, or in the L4 DRGs bilaterally. Subcutaneous administration of BTX-A, performed on the left hindpaw at days 4, 8, or 16 post VRT surgery, significantly reduced mechanical allodynia bilaterally and inhibited P2X3 over-expression induced by L5 VRT.
Conclusions. L5 VRT led to over-expression of P2X3 receptors in the L5 DRG and bilateral mechanical allodynia in rats. Subcutaneous injection of BTX-A significantly reversed the neuropathic pain behavior and the over-expression of P2X3 receptor in nociceptive neurons. These data not only show over-expression of purinergic receptors in the VRT model of neuropathic pain but also reveal a novel mechanism of botulinum toxin action on nociceptive neurons.
Neuropathic pain affects millions of Americans and often diminishes the quality of life of the patients . The management of persistent neuropathic pain is difficult and current treatment options too often fail. Botulinum toxin type A (BTX-A) has been used to treat a number of pain conditions including neuropathic pain . BTX-A is a neurotoxic protein produced by the bacterium Clostridium botulinum. This two-chain polypeptide has a protease unit that cleaves synaptosomal associated protein of 25 kDa, one of the Soluble NSF Attachment REceptor (SNARE) proteins essential for neurotransmitter release [4,5]. It is commonly used to treat neuromuscular hyperactivity diseases such as dystonia and spasticity [6–9]. It has also been used to treat several different types of headaches with positive outcomes, including tension-type headaches, cervicogenic headaches, and migraine [10–15]. More recent clinical studies have suggested that BTX-A might be useful in treating neuropathic pain such as trigeminal neuralgia and chronic neuropathic pain, diabetic neuropathy, and complex regional pain syndrome [2,16–19]. Our recent study further demonstrates that BTX-A is effective in the treatment of postherpetic neuralgia .
Progress has been made to understand the mechanisms of BTX-A action on the nervous system. Animal experiments demonstrate that BTX-A can inhibit not only the exocytosis of acetylcholine but also other neurotransmitters such as glutamate, substance P, and calcitonin gene-related peptide [21–23]. Inhibition of the release of these neurotransmitter is believed to be related to the relief of neuropathic pain symptoms as they have been indicated in pain transmission [4,24–29]. Human studies further demonstrate that BTX-A reduces pain and neurogenic inflammation induced by capsaicin [30,31]. However, it is not clear whether mechanisms beyond neurotransmission play a role in the analgesic effects of BTX-A. We are particularly interested in the potential interaction between BTX-A and the transmembrane channels/receptors that are tightly linked to pain. Specifically, we wanted to determine if BTX-A affects the cellular expression of the purinergic receptors that have been indicated in pain transmission. Adenosine triphosphate (ATP) elicits pain in humans upon infusion into the skin through activation of the ATP-gated ion channels of the P2X family that have been identified and characterized in sensory neurons [32–35]. Seven P2X subunits have been identified so far, each with two transmembrane segments separated by a large ectodomain. ATP is believed to bind the ectodomain and cause significant calcium permeability through these channels . Only P2X3 is expressed exclusively in small diameter nociceptive sensory neurons among the seven subunits. Pain behaviors following injury are indeed attenuated in the P2X3 knockout mice . The role of P2X receptors in pain facilitation is further supported by the facts that both sensory neurons and interneurons in the spinal cord dorsal horn could be depolarized by ATP in culture [38,39]. Peripheral injections of P2X receptor agonists enhance nociception in animal models of inflammatory pain and mechanical allodynia [40,41]. Furthermore, the number of P2X3 receptor immunoreactivity in dorsal root ganglion (DRG) neurons increased after a chronic constriction injury (CCI) of the sciatic nerve . Clearly, the P2X3 receptor contributes to transmission of nociceptive pain signals and plays a role in both acute and chronic pain [43–52].
We chose to use the L5 ventral root transection (L5 VRT) model of neuropathic pain in this study because the selective transection of motor fibers without sensory neuron injury produced bilateral behavioral signs of neuropathic pain for a long duration (56 days) and to a similar extent as L5 spinal nerve transection in rats [53–55]. Preservation of the integrity of the sensory neurons is particularly advantageous for studies of the DRG neurons . We thus conducted experiments designed to determine  whether L5 VRT increases the expression of P2X3 receptors in the DRG neurons in rats; whether BTX-A can inhibit the expression of P2X3 receptors; and  whether BTX-A can reduce pain behaviors induced by L5 VRT.
Materials and Methods
Seventy-eight male Sprague-Dawley rats weighting 180–250 g were used. The rats were housed in separated cages and the room was kept at 24 ± 1°C temperature and 50–60% humidity, under a 12:12 light–dark cycle and with free access to food and water ad libitum. All experimental procedures were approved by the Local Animal Care Committee and were carried out in accordance with the guideline of the National Institutes of Health on animal care and the ethical guidelines for investigation of experimental pain in conscious animal .
The L5 VRT Model
Rats were anesthetized with sodium pentobarbital (50 mg/kg body weight, i.p.). Additional doses of the anesthetics were given as needed. All manipulations were done on the left side of spinal column. Special care was paid to prevent infection and to minimize the influence of inflammation. The hair of the rat's lower back was shaved, and the skin was sterilized with 0.5% chlorhexidine and covered with clean gauze. Sterile operating instruments were used.
The procedures of L5 VRT were performed as described by Li et al. . Briefly, after a midline skin incision in the lumbar region, the L5 vertebra was freed of its muscular attachment. A left-sided L5 hemilaminectomy was performed, and the transverse process of the L5 vertebra was removed to expose the L5 ventral root. The dura mater and arachnoid membrane were incised, and the L5 ventral root was identified where it lay at the most lateral side of spinal canal and just beneath the L5 dorsal root. The ventral root was gently pulled up with fine forceps and carefully transected 2–3 mm proximal to the DRG, and a small segment (approximately 2 mm) of the root was removed. Then, the wound was washed with saline and closed in layers with 3–0 silk threads. Great care was taken to avoid any damage to the nearby L5 dorsal root and its DRG. In the sham group, the identical operation was performed for exposure of L5 ventral root, but the root was not transected. At the end of each study, animals in L5 VRT groups were deeply anesthetized with intra-peritoneal 20% urethane and were dissected to verify that the lesions were done at the correct level. The transection of the roots at the correct level was confirmed by tracing the cut end of the root to the L5 DRG after an extensive dissection.
Aliquots of 100 IU/vial of BTX-A (Lanzhou Institute of Biological Products, Lanzhou, China) were reconstituted in adequate volume of 0.9% saline to obtain respective doses. BTX-A in a dose of 7 U/kg in a volume of 20 µl was injected with a 27½ gauge syringe into the plantar surface of the ipsilateral hindpaw (left) in conscious rats. BTX-A or saline control was injected at postoperative days 4, 8, or 16 after L5 VRT.
Mechanical sensitivity was measured bilaterally both preoperatively and postoperatively. The rats were accommodated to the testing environment by exposing the rats to the testing chambers for a period of 15 minutes on three separate days just prior to the preoperative testing. Mechanical sensitivity was assessed using eight calibrated von Frey hairs with bending forces of 0.5, 0.7, 0.9, 2.4, 4.6, 5.5, 7.0, or 9.0 g (Life Science Instruments, Woodland Hills, CA, USA) and four additional nylon filament of the same diameter (0.53 mm) but with different lengths and bending forces of 12, 16, 22, or 33 g. The up–down method was used as described previously . The von Frey hairs with bending forces of 0.5–9 g were purchased from Life Science Instruments (Woodland Hills, CA, USA), 9–33 g made from. Briefly, three rats were placed under separate transparent Plexiglas chambers positioned on a wire mesh floor. Five minutes were allowed for habituation. The tests were initiated with the use of 2.04 g hair and the stimuli were presented consecutively in an ascending or descending direction. Each stimulus consisted of a 2–3 s application of the von Frey hair to the middle of the plantar surface perpendicular to the foot, with 5-minute interval between stimuli. Five trials were performed on bilateral hind paws. Quick withdrawal or licking of the paw in response to the stimulus was considered a positive response. Response threshold was defined as the lowest force of von Frey Hair that produced a brisk withdrawal response to one of five stimuli. In cases where no consecutive response was observed, the strongest force was assigned as the threshold. Three experimenters performed the behavioral tests and only one of them knew the design of the study but did not determine the thresholds.
The Immunoreactivity of P2X3 Receptor after L5 VRT
The animals with sham or L5 VRT operations were deeply anesthetized with urethane at postoperative day 3, 7, 14, and 21 and perfused through the ascending aorta with 200 mL of normal saline, followed by 200 mL of 4% paraformaldehyde in 0.1 M phosphate buffer, pH 7.2–7.4 at 4°C. After perfusion, the L4 and L5 DRGs were dissected, and post-fixed in the same fixative for 3 hours, and transferred into 30% sucrose in 0.1 M PB overnight. P2X3 antibody was bought from Chemicon International Company of America. The expression of P2X3 receptor in L4 and L5 DRG was observed by immunohistochemistry according to the method described [58,59]. Briefly, DRG sections (16 µm) were cut in a cryostat (LEICA CM1900; Meyer Instruments, Houston, TX, USA). One section was selected randomly from six nonadjacent sections and four sections were counted for each DRG and expressed as the percentage of total neuronal profiles (Figure 2). Six rats were included for each group for quantification of immunohistochemistry results. The gray scale of P2X3 receptor expression was determined with an image analysis system (HMIAS-2000; Wuhan Glitter Electronics Co., China). The sections were examined with an Olympus IX71 (Olympus Optical, Tokyo, Japan) fluorescence microscope and images were captured with a charge coupled device (CCD) spot camera.
The Expression of P2X3 Receptor after BTX-A Injection in L5 VRT Rats
The P2X3 receptor expression was determined in ipsilateral L5 DRG in L5 VRT animals at day 8 after subcutaneous administration of BTX-A or saline.
Quantitative data are expressed as means ± standard error. Analysis of variances (ANOVA) was used to detect differences in positive P2X3 staining cells (%) or in withdrawal thresholds (g) using treatment and time as independent variables, followed by preplanned comparisons between individual treatment groups or between different time points using Tukey post hoc tests. Student t-test was used to compare the positive P2X3 staining cells (%) between the BTX-A injection group and the saline control group at day 8 post-injection. Statistical tests were performed with SPSS 10.0 (SPSS Inc., Chicago, IL, USA). A P value of <0.05 was considered significant in all the tests.
VRT Induced Mechanical Hyperalgesia and Upregulation of P2X3 Expression in L5 DRG Neurons
All animals developed robust and prolonged bilateral mechanical hyperalgesia within 24 hours after the L5 VRT. An upregulation of P2X3 receptor expression was observed in the ipsilateral L5 DRG neurons in the VRT but not the control animals. The P2X3 immunoreactivity was quantified by the percentage of positively labeled P2X3 neurons as shown in Figures 1 and 2. As shown in Figure 1, the P2X3 immunoreactivity in the ipsilateral L5 DRG was clearly increased in day 7, 14, and 21 after L5 VRT surgery (Figure 1C–E), compared with the sham surgery control (Figure 1A). The quantitative data (mean + standard deviation) are summarized in Figure 1 (F = 85.519, P < 0.001, N = 6 in each group). Compared with the sham group, the P2X3 immunoreactivity in the ipsilateral L5 DRG increased significantly at day 7 after surgery and maintained at a considerably higher level for at least 21 days (Figure 2C. P < 0.05, Tukey post hoc tests for comparisons between the control taken on day 21 and the VRT groups at various time points). There was no significant change in the P2X3 immunoreactivity in the L4 DRGs (Figure 2A and B) or in the contralateral L5 DRG (Figure 2D) (F = 0.053, P > 0.05).
BTX-A Significantly Reversed Mechanical Allodynia in Rats with VRT
Left L5 VRT produced substantial and consistent reduction in the paw withdrawal thresholds upon mechanical stimulation of both the hind paws compared with preoperative baseline (Figure 3A–C). The reduction of the withdrawal thresholds was bilateral. There were no differences noted between the ipsilateral and contralateral sides (F value = 0.751, P > 0.05). This bilateral mechanical hyperalgesia was long lasting and sustained over the whole period of experimental observation of up to 50 days (Figure 3C). Subcutaneous injection of BTX-A significantly reversed, in part, the reduction of withdrawal thresholds bilaterally (F = 5.64, P < 0.01). In contrast, subcutaneous injection of saline did not affect the withdrawal thresholds, compared with the pre-injection baseline values (F = 1.045, P > 0.05). The anti-hyperalgesia effects of BTX-A were evident regardless whether the injection was conducted at postoperative day 4, 8, or 16 after the VRT surgery (Figure 3A–C respectively). The effects of one injection lasted for at least 20 days, rising within 5 days and peaking around day 14 post-injection.
BTX-A Inhibited Over-Expression of P2X3 Receptors Induced by VRT
The upregulation of P2X3 expression in the ipsilateral L5 DRG neurons, induced by L5 VRT, was inhibited by subcutaneous injection of BTX-A. Eight days after BTX-A injection, the P2X3 immunoreactivity was significantly lower than the control group injected with saline (Figure 4A and B). The positive P2X3 staining neurons were 54.32 ± 4.24% in the BTX-A group, compared with 72.65 ± 6.31% in the control group (P < 0.001, N = 6, Student t test, t = 10.51), as shown in Figure 4C.
This study clearly demonstrated that selective transection of the L5 ventral root significantly increased the expression of P2X3 receptors in the ipsilateral DRG neurons. Subcutaneous injection of BTX-A substantially reduced the pain behaviors and the upregulation of P2X3 receptor expression induced by L5 VRT. These observations provide critical insight into the cellular mechanisms of the analgesic effects of botulinum toxins. It also shed light on the mechanisms of neuropathic pain induced by the ventral root transection model.
An interesting observation is that the expression of P2X3 receptor/channel was upregulated in the VRT model of neuropathic pain (Figures 1 and 2). Previous studies have shown that the expression of P2X3 was enhanced after CCI of the sciatic nerve [42,43]. However, the interpretation of this observation is complicated by the fact that the sciatic nerve constriction injury directly affects the integrity of the axons of the nociceptive afferent neurons. In this study, we took advantage of the VRT model of neuropathic pain, in which the injury largely spares the sensory afferent neurons in the DRG. Thus, the upregulation of the P2X3 receptor/channel is attributable to changes in the spinal cord and/or the affected limb as a result of the ventral root transection and the subsequent Wallerian degeneration of the motoneurons. Wallerian degeneration usually occurs in the axon stump distal to the site of injury and begins within 24–36 hours of the lesion. This time course is consistent with the observation that hyperalgesia in the VRT model develops in 24 hours after transection (Figure 3). The distal axon stumps tend to remain electrically excitable prior to degeneration, constituting one potential source of signaling to the nociceptive neurons in the same anatomical distribution. After injury, the axonal skeleton disintegrates and the axonal membrane breaks apart, leading to denervation of the motor units. The axonal degeneration is followed by degradation of the myelin sheath and infiltration by macrophages, which serve to clear the debris from the degeneration. Thus, the upregulation of the P2X3 receptor/channel expression might be attributable to changes in the spinal cord and/or the affected limb as a result of the ventral root transection and the subsequent Wallerian degeneration of the motoneurons.
Previous studies of neuropathic pain induced by L5 VRT have revealed a potential link between Wallerian degeneration of motor nerve and inflammation . Whether the upregulation of the P2X3 receptors is mediated by neuroinflammation requires further investigation. Such studies may lead to the identification of molecular cues that are critically important to sensitize or activate nociceptive neurons and result in hyperalgesia and upregulation of P2X3 expression. The motor function and anatomical changes of the VRT model of neuropathic pain have been reported . Consistent with previous observations, rats with L5 VRT appeared healthy and well groomed. Atrophy of the affected muscles on the ipsilateral limb was observed 1 week after L5 VRT. Slight deficit in locomotion was noticeable, primarily in paw placement on the ground with reduced spread of the two lateral toes. However, no spontaneous paw elevation or preferred weight-bearing side was seen. Therefore, we did not find evidence that the BTX effects are attributable to unmasking motor deficits.
Signaling from motoneuron injury to nociceptor neurons may also occur within the spinal cord. For example, L5 VRT induced a dramatic upregulation of expression of brain-derived neurotrophic factor (BDNF) in the spinal cord and in the DRG. The changes in the spinal cord, including an increase in the number of BDNF- and TrkB-immunoreactive nerve fibers in the dorsal column, may thus lead to the upregulation of P2X3 receptor expression in the nociceptive neurons. Our experiments not only confirmed the previous observation of long-lasting hyperalgesia after VRT  but also provided data that are important to the understanding of mechanisms of the this model of neuropathic pain.
As activation of the P2X3 receptors in the primary sensory neurons is involved in neuropathic pain, blocking or reducing P2X3 receptor may inhibit hyperalgesia or allodynia mediated by sensory neurons in neuropathic pain states. For example, sodium ferulate can decrease P2X3 receptor expression of DRG neurons and increase the thresholds of thermal or mechanical stimulation in CCI rats . The analgesic effect of BTX-A in this study may also be mediated by its inhibitory effects on P2X3 expression in the DRG neurons. However, the contralateral effects cannot be explained on this basis because we did not observe inhibition of P2X3 expression on the contralateral DRG neurons. While the potential link between P2X3 expression and the ventral root lesion-induced hyperalgesia cannot be completely dismissed, the temporal and spatial differences between P2X3 expression and pain behavior suggest that mechanisms other than P2X3 expression must be considered to understand the motoneuron injury-induced neuropathic pain and the analgesic effects of BTX-A.
Subcutaneous injection of botulinum toxin significantly reversed hyperalgesia induced by VRT. The withdrawal thresholds increased substantially within 5 days of BTX-A injection, peaked at day 14, and lasted for 20 days after a single injection (Figure 3). The same effects were also observed in the contralateral side. These effects were consistent regardless whether the injection was performed at 4, 8, and 16 days after the ventral root transection. This finding suggests that BTX-A can be effective in relieving neuropathic pain in conditions with various durations. Clearly, the use of the VRT model of neuropathic pain has made it possible to study this important variable because this model produces long-lasting and persistent hyperalgesia for up to 56 days , a much longer duration than that of the CCI model.
The fact that unilateral subcutaneous injection of BTX-A produced bilateral effects is consistent with the observation that BTX-A effectively reduced “mirror pain” induced by unilateral repeated intramuscular acidic injections . Also, unilateral BTX-A injection into the hindpaw pad in rats with diabetic neuropathy reduced mechanical hyperalgesia on both sides, starting in day 5 post-injection . Similar bilateral effect of unilateral BTX injection was observed in the paclitaxel-induced neuropathy in rats . It would be interesting to test if injection in the contralateral, healthy paw would produce the same results. Clinically, injection into the healthy side is presumably better tolerated, compared with the affected, hyperalgesic side.
Mechanistically, BTX-A likely reaches the spinal cord through the central terminals of the afferent neurons and inhibits neuropathic pain of both sides through a central mechanism that may involve supraspinal descending facilitatory pathways. This hypothesis is supported by the observation that colchicine pretreatment completely prevented the analgesic effects of botulinum toxins, suggesting that retrograde-axonal transport into the central nervous system is responsible for its central effects [61–63]. Although the action of BTX may be complex, it is clear that BTX can act centrally to produce analgesic effects that are independent of muscle relaxation.
This study demonstrates that there was an over-expression of P2X3 receptors as well as hyperalgesia after selective ventral root transection. Injection of BTX-A significantly reduced the pain behavior bilaterally and inhibited the over-expression of P2X3 receptors in the DRG neurons unilaterally. This observation is significant because it is the first demonstration that the botulinum toxin interacts with the nociceptive neurons through mechanisms other than neurotransmitter release. Instead, it inhibits the expression of the transmembrane purinergic receptors/channels, a key player in the development and maintenance of acute and chronic pain states. Although the molecular signaling pathways for this action remain to be determined, this observation provides new insight to the mechanisms of the analgesic effects of botulinum toxins. A better understanding of the molecular mechanisms is critically important in the development of new therapeutic agents and strategies for more effective management of chronic pain conditions, including various persistent neuropathic pain disorders.
The authors are grateful to Dr. Xianguo Liu and his team for providing technical support and helpful comments on this work.