Vincristine preferentially induces behavioural sensitization to mechanical as opposed to thermal stimulation
Activation of cannabinoid CB1 and CB2 receptor subtypes attenuates vincristine-induced mechanical hypersensitivity. Using the vincristine injection paradigm employed here, animals remained in relatively good health, as characterized by the absence of mortality observed with higher dosing paradigms (Authier et al., 1999, 2003a). Vincristine induced a failure of normal weight gain relative to saline-treated controls, similar to previous reports (Weng et al., 2003). A small percentage of animals (<5%) exhibited gastrointestinal bleeding, a common problem for chemotherapy patients (Sandler et al., 1969; Jackson et al., 1988; Tolstoi, 2002; Ozcay et al., 2003), during later stages of the experiment (that is, days 5–12). Weng et al. (2003) reported no similar symptoms and normal stool in the same vincristine-dosing paradigm. Differences may be attributed to the large number of subjects evaluated in our study coupled with the low frequency of symptom occurrence.
Changes in mechanical withdrawal thresholds observed here cannot be attributed to the development of sensitization to repeated testing. Mechanical allodynia developed in vincristine-treated animals, but not in their saline-treated counterparts who were tested at the same time. Mechanical hypersensitivity developed by day 3 post-vincristine, reaching its lowest level on day 7 and remained stable until day 12. Other studies similarly report that mechanical hypersensitivity is maximal by day 8 post-vincristine (Nozaki-Taguchi et al., 2001; Weng et al., 2003). Vincristine-induced mechanical allodynia resolved completely by day 31 in our study, although lack of recovery has been reported with other dosing paradigms (Nozaki-Taguchi et al., 2001).
Hypersensitivity to thermal stimulation (or thermal hyperalgesia) was notably absent in vincristine-treated rats that nonetheless exhibited robust mechanical allodynia. By contrast, paclitaxel induces thermal hyperalgesia or thermal hypoalgesia (depending upon the dosing schedule), which may be absent in vincristine and cisplatin models of chemotherapy-induced neuropathy (Authier et al., 2000, 2003a, 2003b; Nozaki-Taguchi et al., 2001; Weng et al., 2003; Lynch et al., 2004; Cata et al., 2006a). Thermal hyperalgesia has been observed in mice using a different vincristine dosing paradigm beginning at 4 weeks following initial vincristine treatment (Kamei et al., 2005). Nonetheless, vincristine may induce cold allodynia/hyperalgesia (Authier et al., 2003b; Lynch et al., 2004), consistent with clinical reports (Cata et al., 2006b).
An upregulation of neuropeptide Y (NPY) in medium and large diameter dorsal root ganglion cells has been postulated to underlie development of mechanical allodynia (in the absence of thermal hyperalgesia) following spinal nerve ligation (Ossipov et al., 2002). More work is necessary to determine whether similar neurochemical changes accompany the development of vincristine-evoked mechanical allodynia in our study.
Subtype specificity of cannabinoid anti-allodynic actions
WIN55,212-2 (2.5 mg kg−1 i.p.) restored mechanical withdrawal thresholds to >100% of previncristine levels. WIN55,212-2 (1.5 mg kg−1 i.p.) reversed both mechanical and thermal hypersensitivity in a paclitaxel-induced neuropathy model (Pascual et al., 2005) but did not reverse vincristine-induced mechanical hypersensitivity in our study. Doses of WIN55,212-2 that eliminated vincristine-induced mechanical allodynia in our study did not induce motor deficits in the bar test. Thus, WIN55,212-2-induced anti-allodynic effects are independent of any motor effects of cannabinoids. Similar or higher doses of WIN55,212-2 (2.5−5 mg kg−1 i.p.) also attenuate mechanical allodynia in models of traumatic nerve injury (Herzberg et al., 1997; Bridges et al., 2001; Fox et al., 2001; Ibrahim et al., 2003; Walczak et al., 2005; LaBuda and Little, 2005) and diabetic neuropathy (Ulugol et al., 2004). WIN55,212-2 also attenuates deep tissue hyperalgesia in a murine model of cancer pain through a CB1 mechanism (Kehl et al., 2003).
AM1241 (2.5 mg kg−1 i.p.) induced a CB2-mediated suppression of vincristine-induced mechanical allodynia without inducing antinociception. Metabolism of AM1241 may limit the duration of CB2-mediated anti-allodynia observed here. Nonetheless, CB2 agonists may represent preferred therapeutic agents relative to CB1 agonists due to their limited profile of CNS side-effects (Hanus et al., 1999; Malan et al., 2001). AM1241 is an effective anti-hyperalgesic agent in animal models of traumatic nerve injury (Ibrahim et al., 2003) and inflammation (Quartilho et al., 2003; Hohmann et al., 2004; Nackley et al., 2003, 2004). Our studies suggest that CB2 is also a novel target for the treatment of chemotherapy-induced neuropathy.
Activation of either CB1 or CB2 receptors suppressed the maintenance of vincristine-evoked mechanical allodynia. The anti-allodynic effects of WIN55,212-2 were partially blocked by each antagonist alone at 30 min post-injection whereas complete blockade was observed at 60 min post-drug. Moreover, i.t. administration of both antagonists concurrently completely blocked the anti-allodynic effects of spinally administered WIN55,212-2. Our data also raise the possibility that targeting multiple cannabinoid receptor subtypes simultaneously may act synergistically to suppress chemotherapy-induced neuropathy.
Effects of cannabinoids and morphine on vincristine-induced neuropathy
Opiates are commonly administered to cancer patients experiencing chemotherapy-induced neuropathy (Lynch et al., 2004; Cata et al., 2006b). In our study, a leftward shift in the dose–response curve for mechanical withdrawal thresholds was observed for WIN55,212-2 relative to morphine. WIN55,212-2, at a dose of 2.5 mg kg−1, exhibited effects of approximately the same magnitude as morphine at a dose of 8 mg kg−1. Additional doses are required to enable calculations of the ED50 for each drug and verify differences in agonist potency. Our low dose of morphine (2.5 mg kg−1 i.p.) suppressed neuropathic nociception induced by spinal nerve ligation (LaBuda and Little, 2005; Joshi et al., 2006) and induced antinociception (Ibrahim et al., 2006), but failed to suppress vincristine-induced allodynia in our study. The high dose of morphine (8 mg kg−1 i.p.) normalized paw withdrawal thresholds in our study but only partially (50%) reversed paclitaxel-evoked mechanical hypersensitivity (Flatters and Bennett, 2004). Cannabinoids show enhanced antihyperalgesic efficacy relative to opiates in other neuropathic pain models (Mao et al., 1995, 2000). Lower efficacy of morphine in reducing abnormal sensations related to myelinated as opposed to unmyelinated fibre activation (Taddese et al., 1995) is consistent with the differential neuroanatomical distribution of μ-opioid and cannabinoid receptors at spinal and primary afferent levels (Hohmann and Herkenham, 1998a; Hohmann et al., 1999; Bridges et al., 2001). Thus, cannabinoids may be more potent and efficacious than opiates in suppressing diverse forms of neuropathic and deafferentation-induced pain.
Mechanisms and site of action
In our study, WIN55,212-2 suppressed vincristine-induced mechanical allodynia when administered i.t. but not when administered locally into the paw. In fact, local injections of either vehicle or WIN55,212-2 (30 μg i.pl.) in our study enhanced mechanical allodynia in the injected paw relative to preinjection levels. Changes in weight bearing due to sensitization at the site of i.pl. injection may contribute to the increases in paw withdrawal thresholds observed in all groups (including vehicle) in the non-injected paw. The same local dose employed here (30 μg i.pl.) suppressed mechanical allodynia in models of diabetic neuropathy (Ulugol et al., 2004) and traumatic nerve injury (Fox et al., 2001) but failed to attenuate paclitaxel neuropathy (Pascual et al., 2005) or suppress vincristine-induced neuropathy in our study. Local injection of WIN55,212-2 (30 μg i.pl.) also elevated paw withdrawal thresholds in the non-injected paw above baseline (previncristine) levels, but failed to reverse the hypersensitivity observed at the site of the i.pl. injection. Leakage of the cannabinoid into the systemic circulation may contribute to changes in paw withdrawal thresholds observed in the non-injected paw. A higher local WIN55,212-2 dose (150 μg i.pl.) that induces clear systemic effects (Fox et al., 2001) eliminated the hypersensitivity observed at the site of the i.pl. injection. However, this dose nonetheless failed to suppress vincristine-evoked mechanical allodynia relative to preinjection levels and did not normalize paw withdrawal thresholds to previncristine levels.
Our data provide direct evidence that spinal sites of action are implicated in both CB1 and CB2 receptor-mediated suppressions of chemotherapy-induced neuropathy. Interestingly, CB2 receptor mRNA and protein are upregulated in spinal cord of rats subjected to traumatic nerve injury (Zhang et al., 2003; Walczak et al., 2005; Wotherspoon et al., 2005). Direct spinal administration of a CB2 agonist also suppresses mechanically evoked responses in wide dynamic range neurons in neuropathic but not in sham-operated rats (Sagar et al., 2005), suggesting a functional role for spinal CB2 receptors in neuropathic pain states.
Vincristine induces central sensitization in spinal wide dynamic range neurons, including abnormal spontaneous activity, wind-up and afterdischarge responses to suprathreshold mechanical stimulation (Weng et al., 2003). These aberrant neurophysiological responses may mediate the observed chemotherapy-induced neuropathy. Cannabinoids suppress C-fibre-mediated responses and wind-up of spinal wide dynamic range neurons through either CB1 (Strangman and Walker, 1999; Drew et al., 2000) or CB2 (Nackley et al., 2004)-specific mechanisms. Further studies are required to determine the neurophysiological basis for cannabinoid-mediated suppression of chemotherapy-induced neuropathy (see Hohmann, 2005).
Enhanced primary afferent glutamate release (presynaptic facilitation) may also contribute to the abnormal behavioural phenotype and central sensitization induced by chemotherapeutic treatment. Consistent with this hypothesis, decreased protein levels for the glutamate-aspartate transporter (GLAST), glial glutamate transporter-1 (GLT-1) and excitatory amino-acid carrier-1 (EAAC1) are observed following paclitaxel treatment (Cata et al., 2006a). It is worth noting, however, that glutamate and NMDA receptor antagonists reverse hyperalgesia in a nerve-injury model (Mao et al., 1995), but not in chemotherapy-induced neuropathy models (Aley and Levine, 2002; Flatters and Bennett, 2004). Thus, distinct mechanisms may be implicated in the development of neuropathic nociception induced by traumatic nerve injury and chemotherapeutic treatment, respectively.
Abnormal primary afferent input, presynaptic and/or descending (Porreca et al., 2001; Vera-Portocarrero et al., 2006) facilitation and chemotherapy-induced dysregulation of calcium homoeostasis (Siau and Bennett, 2006) may enhance neuronal excitability, thereby increasing intracellular Ca2+ (Kawamata and Omote, 1996). Ethosuximide, a T-type calcium antagonist and other drugs which reduce intra- and extracellular Ca2+, also reduce vincristine-induced mechanical hypersensitivity (Flatters and Bennett, 2004; Siau and Bennett, 2006). Additional studies are required to determine if cannabinoid suppression of chemotherapy-induced neuropathy is related to cannabinoid suppression of Ca2+ conductance (Mackie and Hille, 1992; Mackie et al., 1995) and central sensitization.