Paclitaxel and vincristine are chemotherapeutic drugs that are used widely for the treatment of solid tumors. Both evoke a dose-limiting peripheral neuropathy that often presents with a distal, symmetrical pain syndrome. Even with cessation of chemotherapy, the pain is variably reversible and may persist for months to years. The cause of this chemotherapy-evoked neuropathic pain is unknown. There are no confirmed prophylactic or therapeutic treatments (1).
Calcium is a key regulator of major cellular processes. Its cytosolic concentration is determined mainly by extracellular Ca2+ influx, release of Ca2+ from internal stores, and mitochondrial uptake. We have shown that ethosuximide, a relatively selective T-type Ca2+-channel blocker, and gabapentin, an antagonist of Ca2+ channels containing the α2δ subunit, significantly reduce paclitaxel- and vincristine-evoked neuropathic pain (2,3). Recently, we have obtained evidence that rats with paclitaxel-evoked neuropathic pain have an abnormality of mitochondria within myelinated and unmyelinated sensory axons (4). A mitochondrial abnormality can potentially lead to dysregulation of intracellular Ca2+. Similarly, there is evidence that the painful peripheral neuropathy produced by anti-human immunodeficiency virus (HIV) nucleoside analog therapy is associated with disrupted Ca2+ homeostasis secondary to mitochondrial dysfunction (5).
The experiments reported here explore the hypothesis that impaired Ca2+ regulation is associated with paclitaxel- and vincristine-evoked neuropathic pain in rats. In addition, we compared the chemotherapy-evoked pain syndrome with those produced by traumatic nerve injury and anti-HIV nucleoside analog therapy.
These experiments conformed to the ethics guidelines of the International Association for the Study of Pain (6), the National Institutes of Health, and the Canadian Institutes of Health Research. Experimental protocols were approved by the Facility Animal Care Committee of the Faculty of Medicine, McGill University.
Adult male Sprague-Dawley rats (Harlan Sprague-Dawley Inc., Indianapolis, IN; breeding colony, Frederick, MD), weighing 250–270 g at the start of the experiments, were housed with sawdust bedding under a 12-h dark/light cycle and had access to food and water ad libitum.
Paclitaxel (Taxol®; Bristol-Myers-Squibb, Montreal, Quebec, Canada) was injected intraperitoneally (IP; 2 mg/kg) every other day for 4 days (7); it was prepared from a stock solution of 6 mg/mL in Cremophor EL and ethanol (50:50) diluted to 2 mg/mL with normal saline. Vincristine (Novopharm Ltd, Toronto, Ontario, Canada) in saline vehicle was injected IP (50 μg/kg) for 10 consecutive days (3). 2′,3′-dideoxycytidine (ddC; saline vehicle; Sigma-Aldrich, Oakville, Ontario, Canada) was administered as a single IV bolus of 50 mg/kg via the tail vein (5). The chronic constriction injury (CCI) model was produced as described by Bennett and Xie (8). CCI rats received a contralateral sham operation (the nerve was similarly manipulated but not ligated).
Drugs that modify cellular Ca2+ levels were injected intrathecally while the rats were anesthetized with 2% isoflurane in oxygen. The rat was placed prone with its spinal column arched over a support. Lumbar puncture was performed at the level of L5-6 intervertebral space using a 30-gauge needle. We verified the accuracy of these injections in pilot experiments in which the injection of 2% lidocaine produced a brief bilateral paralysis of the hindlimbs. A transient lateral flick of the rat’s tail confirmed intrathecal needle placement. The test solution was then injected over 5 s via a 50-μL Hamilton syringe. For rats with neuropathic pain, the effects of all Ca2+-decreasing drugs were tested at the approximate time of the models’ peak symptom severity (22–28 days after paclitaxel injection, 13–19 days after vincristine injection, 8–12 days after CCI, and 8 days after ddC) in rats with confirmed pain hypersensitivity.
The following compounds, all delivered in a volume of 20 μL, were used to manipulate neuronal Ca2+ levels: (a) TMB-8 (8-(dimethylamino) octyl 3,4,5-trimethoxybenzoate (46 nmol; Sigma-Aldrich, Oakville, Ontario, Canada). Intrathecal TMB-8 (up to 46 nmol) reduced anti-HIV nucleoside analog-evoked neuropathic pain in rats (5). (b) Quin-2 (2-([2-bis-(carboxymethyl)amino-5methylphenoxy] methyl)-6-methoxy-8-bis(carboxymethyl) aminoquinoline potassium (1.8 nmol; EMD Biosciences [Calbiochem], San Diego, CA). This intrathecal dose of Quin-2 has been used to investigate the role of Ca2+channels in formalin-evoked pain (9). (c) EGTA (ethylene glycol-bis(β-aminoethyl ether)-N,N,N′,N′-tetraacetic acid tetrasodium salt (0.1 μmol; Sigma-Aldrich). Intracerebroventricular EGTA (up to 1 μmol) has been used in experiments on opioid analgesia in mice (10). In pilot experiments, we investigated the intrathecal dose of EGTA at 0.01, 0.1, and 1 μmol to determine a dose that did not elicit ataxia in rats. The largest tolerated dose was 0.1 μmol. (d) EGTA-am (ethylene glycol-bis(β-aminoethyl)-N,N,N′,N′-tetraacetoxymethyl ester (0.1 μmol; EMD Biosciences [Calbiochem]). EGTA-am is the acetoxymethyl ester of EGTA. EGTA-am and TMB-8 act intracellularly, whereas Quin-2 and EGTA act extracellularly. Their effects were examined 30 min after the intrathecal injection.
TMB-8, Quin-2, and EGTA were freshly prepared in normal saline; EGTA-am was freshly prepared in 20% Pluronic F-127 in dimethyl sulfoxide (Molecular Probes, Eugene, OR). For each drug, a group of control rats received an intrathecal injection of the same volume of the relevant vehicle.
The effects of intradermal injections of TMB-8 (116 nmol/10 μL) or an equal volume of saline were examined in ddC- and paclitaxel-treated rats. These injections were made while the rats were under isoflurane anesthesia into the dorsal skin of one hindpaw using a 30-gauge needle, and the rats were tested 30 min later.
The rats were assessed for mechano-allodynia and mechano-hyperalgesia. In CCI rats, we also examined heat hyperalgesia. Paclitaxel-treated rats have little or no heat hyperalgesia (7). Rats were habituated to the testing environment for 15 min on three separate occasions before data collection. Naïve baselines were noted before the establishment of the pain models.
Mechanical hypersensitivity was evaluated using von Frey hairs (VFH; Stoelting, Wood Dale, IL). Mechano-allodynia was assessed as an increased incidence of response to stimulation with a 4-g VFH; naïve rats almost never withdraw from this stimulus. Subsequently, mechano-hyperalgesia was assessed with a 15-g VFH, which evokes a relatively infrequent incidence of withdrawals in naïve rats. The rats were confined under inverted plastic cages and placed on an elevated wire mesh platform, which allowed access to the plantar surface of the hindpaws. Each VFH was applied to the mid-plantar hindpaw (the region in the center of the ring formed by the tori at the base of each of the digits) and held for 5 s. This was repeated five times for each paw, and the sum of withdrawal responses from both paws to each VFH was expressed as a percentage response (number of hindpaw withdrawals/10 × 100). For the CCI rats, the response frequencies to each VFH for the ipsilateral and contralateral paws were calculated separately.
Pain hypersensitivity to mechanical stimuli was also assessed with the Randall-Selitto paw-pressure test using an Analgesy-meter® (Ugo Basile, Comerio-Varese, Italy). Testing began approximately 5 min after VFH testing. Rats were wrapped loosely in a fabric towel with the hindpaw placed between the Analgesy-meter platform and stylus. A linearly increasing mechanical force was then applied until the paw was withdrawn, and the corresponding force was recorded as the withdrawal threshold. Three readings for each paw were taken at 5-min intervals and averaged. With the pain models used here, there is a decrease in the paw-pressure threshold of approximately 25%; we classify this decrease as mechano-hyperalgesia. However, we note that the paw-pressure stimulus may activate both cutaneous and deep-tissue sensory afferents, whereas the 15-g VFH stimulus delivers most of its energy to the cutaneous innervation.
Heat hypersensitivity was assessed using the method of Hargreaves et al. (11). Rats were confined in a plastic enclosure on the glass surface of a thermal stimulator (UCSD, San Diego, CA). A focused beam of radiant heat under the glass floor was aimed at the fat part of the heel. The latency to paw withdrawal was taken as an index of the heat-pain threshold. The heat intensity was standardized such that mean naïve baseline latencies were between 8 and12 s and remained constant throughout the experiment. A cutoff latency of 20 s was imposed to avoid possible tissue damage. Testing was repeated 3 times for each paw at 10-min intervals, and the scores were averaged.
All data are expressed as mean ± sem. The effects of intrathecal or intradermal drug versus vehicle were analyzed with unpaired t-tests. Differences were considered significant at P < 0.05.
Treatment with paclitaxel, vincristine, or ddC produced the expected bilateral neuropathic pain, with statistically significant increases in pain sensitivity as measured with the 4-g and 15-g VFH tests and the paw-pressure test. The CCI procedure produced the expected statistically significant increases in pain sensitivity to VFH, paw pressure, and heat on the nerve-injured side; no significant changes were found on the sham-operated side with any of the four tests.
In paclitaxel- and vincristine-treated rats tested with VFH (Fig. 1), intrathecal TMB-8, Quin-2, EGTA, and EGTA-am all produced statistically significant reductions in mechano-allodynia and mechano-hyperalgesia, whereas intrathecal injections of their respective vehicles had no effect. The degree of pain reduction was similar for each of the four drugs. In addition, all four drugs produced statistically significant reduction of mechano-hyperalgesia, as assessed in the paw-pressure test (Fig. 2). None of the four drugs had any effect on responses of naïve animals tested with 4-g or 15-g VFH and the paw-pressure test using the same doses (n = 6 per group; data not shown).
In CCI rats, neither TMB-8 nor Quin-2 had any effect on the mechano-allodynia and mechano-hyperalgesia assessed with VFH, mechano-hyperalgesia assessed with the paw-pressure test, or on heat hyperalgesia (Fig. 3). Neither drug had any effect in any of the four tests on the response sensitivity of the contralateral hindpaw (Fig. 3).
As shown in Figure 4, both intradermal and intrathecal injections of TMB-8 reduced mechano-hyperalgesia (paw-pressure test) in ddC-treated rats, as previously reported (5). Although intrathecal TMB-8 was effective in paclitaxel-treated rats, the intradermal TMB-8 dose was without effect.
An increase in cytosolic Ca2+ concentration ([Ca2+]c) mediates a wide range of neuronal functions including membrane excitability, neurotransmitter release, synaptic plasticity, gene expression, and excitotoxicity (12). It is probable that [Ca2+]c also plays a critical role in glial function. Altered [Ca2+]c has been demonstrated after formalin-evoked pain (9), posttraumatic painful neuropathy (13), and anti-HIV nucleoside analog-evoked painful peripheral neuropathy (5).
Our data show that the neuropathic pain (mechano-allodynia and mechano-hyperalgesia) produced by paclitaxel and vincristine is significantly ameliorated by drugs that decrease the extracellular and intracellular availability of Ca2+. We confirmed this effect with 4 different drugs (TMB-8, Quin-2, EGTA, and EGTA-am). Three of the drugs (TMB-8, Quin-2, and EGTA) are chemically distinct, and 3 have distinct mechanisms or sites of action. Quin-2 is a highly selective Ca2+ chelator (14), as is EGTA. Both are membrane impermeable and will preferentially chelate extracellular Ca2+, reducing Ca2+ influx. We used intracellular Ca2+-decreasing drugs to further explore the regulation of Ca2+ downstream to extracellular Ca2+ influx. TMB-8 is membrane permeable and works intracellularly to prevent Ca2+-induced Ca2+ release from the endoplasmic reticulum via binding to ryanodine receptors (15). EGTA-am is a nonpolar ester of EGTA, making it membrane permeable. It is inactive until de-esterification by intracellular esterases and, thus, preferentially chelates, free intracellular Ca2+. None of these drugs had any effect on the responsiveness of naïve rats, showing that their effects were not caused by a general analgesic action or by an impairment of normal Ca2+ signaling.
Our results are consistent with a previous report (5) that intrathecal TMB-8 is also effective against the neuropathic pain produced by anti-HIV nucleoside analog therapy. We confirm that there is a peripheral site of action for TMB-8 in the ddC syndrome. However, using the same dose of TMB-8, we found no evidence for a peripheral site of action for paclitaxel-evoked pain.
Although the intrathecal dosages of TMB-8 and Quin-2 used were effective in paclitaxel-, vincristine-, and ddC-evoked pain, they were ineffective against the mechano-allodynia, mechano-hyperalgesia, and heat hyperalgesia seen with posttraumatic painful peripheral neuropathy. We suspect that the absence of an effect in the CCI rats may reflect a true difference in underlying pathogenesis. Activation of the N-methyl-d-aspartate receptor (NMDAR) leads to an influx of extracellular Ca2+, which increases [Ca2+]c and initiates neuronal events that contribute to central sensitization (9). Increased intracellular Ca2+ from NMDAR activation has been demonstrated in the ipsilateral dorsal horn of spinal cords of CCI rats (16). Thus, a larger dose of the Ca2+-decreasing drugs may be required to show an effect in CCI rats.
Paclitaxel and vincristine are thought to exert their antitumor activity largely by binding to β-tubulin and disrupting mitotic spindle formation in actively dividing cells. Paclitaxel stabilizes β-tubulin polymerization, whereas vincristine inhibits spindle assembly (17). Axonal microtubules are composed largely of β-tubulin, and it has generally been accepted that the neurotoxicity caused by paclitaxel and vincristine is caused by disruption of microtubule structure that impairs axoplasmic transport and leads to a dying-back neuropathy. Although this hypothesis may be germane to the axonal degeneration caused by large doses of paclitaxel and vincristine, there are reasons to question that it can account for the neuropathic pain produced by small doses of paclitaxel and vincristine. Importantly, electron microscopy studies of sensory nerves from rats with paclitaxel- and vincristine-evoked neuropathic pain have found no evidence of dying-back and only subtle changes in the organization of axonal microtubules (4,7,18).
Vincristine has been reported to affect Ca2+ movement through the mitochondrial membrane, reducing both the amount and rate of Ca2+ uptake and decreasing Ca2+ efflux (19). Paclitaxel has been shown to abolish the normal oscillations in [Ca2+]c and to evoke a spike increase in [Ca2+]c that is independent of extracellular Ca2+ levels and intracellular inositol triphosphate-sensitive Ca2+ stores (20). The same study also demonstrated that paclitaxel causes a rapid decline in mitochondrial membrane potential and, notably, a loss of mitochondrial Ca2+ via the mitochondrial permeability transition pore. It is associated with β-tubulin (21) and is a likely binding site for paclitaxel and vincristine on mitochondria. In support of this alternative hypothesis of impaired mitochondrial function in paclitaxel- and vincristine-evoked neuropathic pain, we have reported that sensory primary afferent axons from rats with paclitaxel-evoked neuropathic pain have a greatly increased incidence of mitochondria that are swollen, vacuolated, and have severely disrupted cristae, which are seen to have collapsed, fragmented, and puddled at the periphery of the organelle (4). Mitochondrial function is critically dependent on molecular exchange between its outer and inner membranes; thus, disrupted cristae indicate impaired function. These structural abnormalities suggest that there would be pathological changes in mitochondrial function.
Mitochondria play a key role in intracellular Ca2+ homeostasis. The influx of extracellular Ca2+ via activated voltage- and ligand-gated Ca2+ channels on the cell membrane and the release of Ca2+ from endoplasmic reticulum can generate large, localized increases in [Ca2+]c. Mitochondria, with their large buffering capacity and close proximity to these structures, can rapidly sequester this free Ca2+ (12,22). Impaired mitochondrial Ca2+ uptake (or increased leakage of mitochondrial Ca2+) would thus alter the spatio-temporal changes of [Ca2+]c, increase propagation of [Ca2+]c signals (23), and modulate Ca2+-dependent processes such as increased exocytosis of neurotransmitters (22). All these changes may result in heightened neuronal excitability and impaired glial function. Reducing the availability of extracellular Ca2+, blocking Ca2+ release from intracellular stores, or chelating cytoplasmic-free Ca2+ would thus be expected to reverse some of the adverse consequences of impaired mitochondrial Ca2+ regulation.
The increase in neuronal [Ca2+]c after NMDAR activation is usually buffered by mitochondrial uptake (24,25). When mitochondrial Ca2+ uptake is impaired, an abnormal accumulation of cytoplasmic Ca2+ would occur in the vicinity of the NMDAR. This engages regulatory proteins such as Ca2+-dependent phosphatase, which inactivate the NMDAR and shorten the duration of Ca2+ influx (26). Experiments using cortical brain-slice preparation have shown a reduction of Ca2+ influx and intracellular Ca2+ load caused by NMDAR activation in the presence of impaired mitochondrial Ca2+ uptake (25). This mechanism would reduce the contribution of NMDAR to central sensitization in chemotherapy-evoked pain. This could (further) account for the effectiveness of the Ca2+-decreasing drugs in reducing paclitaxel-, vincristine-, and ddC-evoked pain, but not trauma-evoked pain, in CCI rats. Additionally, it would also explain the lack of effect of MK801 in paclitaxel-treated animals (2) when NMDAR antagonists are invariably effective in models of posttraumatic painful peripheral neuropathy.
It is possible that the Ca2+-decreasing drugs used in this experiment exerted their effects on altered Ca2+-mediated neuronal and glial functions other than mitochondrial buffering. Of course, mitochondria also perform other cellular functions, including energy production and the regulation of free radicals. Our results do not exclude these additional possible mechanisms for the generation of chemotherapy-evoked neuropathic pain. Moreover, our results do not resolve the locus of the observed effect. Intrathecally administered drugs might act in the spinal cord dorsal horn, the dorsal nerve root, or the dorsal root ganglion. Further evaluation of the role of mitochondria and Ca2+ dysregulation in chemotherapy-evoked pain may have a significant impact to developing successful clinical strategies for the management of this painful complication of cancer treatment.
In summary, our findings suggest that chemotherapy-evoked neuropathic pain is, at least in part, mediated by altered cellular Ca2+ homeostasis. Given the critical role mitochondria play in modulating [Ca2+]c signals and the effects of paclitaxel and vincristine on mitochondria, this dysregulation of Ca2+ could possibly be secondary to impaired mitochondria function.
We thank Ms. Lina Naso for assistance and Wenhua Xiao, MD, for reviewing the manuscript.
1. Quasthoff S, Hartung HP. Chemotherapy-induced peripheral neuropathy. J Neurol 2002;249:9–17.
2. Flatters SJ, Bennett GJ. Ethosuximide reverses paclitaxel- and vincristine-induced painful peripheral neuropathy. Pain 2004;109:150–161
3. Xiao WH Siau C Bennett GJ. Effect of gabapentin in chemotherapy-evoked neuropathic pain in the rat. Abstr Soc Neurosci 2004;982–5.
4. Flatters SJ, Bennett GJ.. Studies of peripheral sensory nerves in paclitaxel-induced peripheral neuropathy: evidence for mitochondrial dysfunction. Pain.
5. Joseph EK, Chen X, Khasar SG, Levine JD. Novel mechanism of enhanced nociception in a model of AIDS therapy-induced painful peripheral neuropathy in the rat. Pain 2004;107:147–58.
6. Zimmermann M. Ethical guidelines for investigations of experimental pain in conscious animals. Pain 1983;16:109–10.
7. Polomano RC, Mannes AJ, Clark US, Bennett GJ. A painful peripheral neuropathy in the rat produced by the chemotherapeutic drug, paclitaxel. Pain 2001;94:293–304.
8. Bennett GJ, Xie YK. A peripheral mononeuropathy in rat that produces disorders of pain sensation like those seen in man. Pain 1988;33:87–101.
9. Coderre TJ, Melzack R. The role of NMDA receptor-operated calcium channels in persistent nociception after formalin-induced tissue injury. J Neurosci 1992;12:3671–5.
10. Ben-Sreti MM, Gonzalez JP, Sewell RD. Effects of elevated calcium and calcium antagonists on 6,7-benzomorphan-induced analgesia. Eur J Pharmacol 1983;90:385–91.
11. Hargreaves K, Dubner R, Brown F, et al A new and sensitive method for measuring thermal nociception in cutaneous hyperalgesia. Pain 1988;32:77–88.
12. Berridge MJ, Lipp P, Bootman MD. The versatility and universality of calcium signalling. Nat Rev Mol Cell Biol 2000;1:11–21.
13. White DM, Cousins MJ. Effect of subcutaneous administration of calcium channel blockers on nerve injury-induced hyperalgesia. Brain Res 1998;801:50–8.
14. Tsien RY, Pozzan T, Rink TJ. Calcium homeostasis in intact lymphocytes: cytoplasmic free calcium monitored with a new, intracellularly trapped fluorescent indicator. J Cell Biol 1982;94:325–34.
15. Chiou CY, Malagodi MH. Studies on the mechanism of action of a new Ca-2+ antagonist, 8-(N,N-diethylamino)octyl 3,4,5-trimethoxybenzoate hydrochloride in smooth and skeletal muscles. Br J Pharmacol 1975;53:279–85.
16. Kawamata M, Omote K. Involvement of increased excitatory amino acids and intracellular Ca2+
concentration in the spinal dorsal horn in an animal model of neuropathic pain. Pain 1996;68:85–96.
17. Jordan MA, Wilson L. Microtubules as a target for anticancer drugs. Nat Rev Cancer 2004;4:253–65.
18. Dougherty PM, Cata JP, Cordella JV, et al. Taxol-induced sensory disturbance is characterized by preferential impairment of myelinated fiber function in cancer patients. Pain 2004;109:132–42.
19. Tari C, Fournier N, Briand C, et al. Action of vinca alkaloides on calcium movements through mitochondrial membrane. Pharmacol Res Commun 1986;18:519–28.
20. Kidd JF, Pilkington MF, Schell MJ, et al. Paclitaxel affects cytosolic calcium signals by opening the mitochondrial permeability transition pore. J Biol Chem 2002;277:6504–10.
21. Carre M, Andre N, Carles G, et al. Tubulin is an inherent component of mitochondrial membranes that interacts with the voltage-dependent anion channel. J Biol Chem 2002;277:33664–9.
22. Montero M, Alonso MT, Carnicero E, et al. Chromaffin-cell stimulation triggers fast millimolar mitochondrial Ca2+
transients that modulate secretion. Nat Cell Biol 2000;2:57–61.
23. Boitier E, Rea R, Duchen MR. Mitochondria exert a negative feedback on the propagation of intracellular Ca2+ waves in rat cortical astrocytes. J Cell Biol 1999;145:795–808.
24. Budd SL, Nicholls DG. Mitochondria, calcium regulation, and acute glutamate excitotoxicity in cultured cerebellar granule cells. J Neurochem 1996;67:2282–91.
25. Kannurpatti SS, Joshi PG, Joshi NB. Calcium sequestering ability of mitochondria modulates influx of calcium through glutamate receptor channel. Neurochem Res 2000;25:1527–36.
26. Lieberman DN, Mody I. Regulation of NMDA channel function by endogenous Ca2+
-dependent phosphatase. Nature 1994;369:235–9.