Abstract
Chemotherapy-induced peripheral neuropathy (CIPN) affects ∼68% of patients undergoing chemotherapy, causing debilitating neuropathic pain and reducing quality of life. Cisplatin is a commonly used platinum-based chemotherapeutic drug known to cause CIPN, possibly by causing oxidative stress damage to primary sensory neurons. Metabotropic glutamate receptors (mGluRs) are widely hypothesized to be involved in pain processing and pain mitigation. Meclizine is an H1 histamine receptor antagonist known to have neuroprotective effects, including an anti-oxidative effect. Here, we used a mouse model of cisplatin-induced CIPN using male and female mice to test agonists of mGluR8 and Group II mGluR as well as meclizine as interventions to reduce cisplatin-induced pain. We performed behavioral pain tests, and we imaged Ca2+ activity of the large population of dorsal root ganglia (DRG) neurons in vivo. For the latter, we used a genetically-encoded Ca2+ indicator, Pirt-GCaMP3, which enabled us to monitor different drug interventions at the level of the intact DRG neuronal ensemble. We found that CIPN increased spontaneous Ca2+ activity in DRG neurons, increased number of Ca2+ transients, and increased hyper-responses to mechanical, thermal, and chemical stimuli. We found that mechanical and thermal pain caused by CIPN was significantly attenuated by the mGluR8 agonist, (S)−3,4-DCPG, the Group II mGluR agonist, LY379268, and the H1 histamine receptor antagonist, meclizine. DRG neuronal Ca2+ activity elevated by CIPN was attenuated by LY379268 and meclizine, but not by (S)−3,4-DCPG. Furthermore, meclizine and LY379268 attenuated cisplatin-induced weight loss. These results suggest that Group II mGluR agonist, mGluR8 agonist, and meclizine are promising candidates as new treatment options for CIPN, and studies of their mechanisms are warranted.
SIGNIFICANCE STATEMENT Chemotherapy-induced peripheral neuropathy (CIPN) is a painful condition that affects most chemotherapy patients and persists several months or longer after treatment ends. Research on CIPN mechanism is extensive but has produced only few clinically useful treatments. Using in vivo GCaMP Ca2+ imaging in live animals over 1800 neurons/dorsal root ganglia (DRG) at once, we have characterized the effects of the chemotherapeutic drug, cisplatin and three treatments that decrease CIPN pain. Cisplatin increases sensory neuronal Ca2+ activity and develops various sensitization. Metabotropic glutamate receptor (mGluR) agonist, LY379268 or the H1 histamine receptor antagonist, meclizine decreases cisplatin's effects on neuronal Ca2+ activity and reduces pain hypersensitivity. Our results and experiments provide insights into cellular effects of cisplatin and drugs preventing CIPN pain.
- chemotherapy-induced neuropathy
- chronic pain
- GCaMP calcium imaging
- in vivo imaging
- meclizine
- primary sensory neuron
Introduction
Chemotherapy-induced peripheral neuropathy (CIPN) is a side effect of chemotherapy, which affects ∼68% of patients, causing severe neuropathic pain that may persist for six months or longer in ∼30% of patients who have undergone chemotherapy (Quasthoff and Hartung, 2002; Seretny et al., 2014; Addington and Freimer, 2016; Cioroiu and Weimer, 2017; Flatters et al., 2017). Cisplatin is a platinum-based chemotherapeutic agent that inhibits tumor growth by crosslinking DNA nucleotides (Greystoke et al., 2017; Armstrong-Gordon et al., 2018), and which is widely used to treat cancer in various locations including lung, stomach, and head (Seretny et al., 2014; Calls et al., 2020). There have been extensive studies on strategies to reduce neuropathic pain and to understand its mechanisms at the level of individual cells, using in vitro dorsal root ganglia (DRG) explants and histologic analysis in animal models (Quasthoff and Hartung, 2002; Jaggi and Singh, 2012; Addington and Freimer, 2016; Cioroiu and Weimer, 2017). However, because of a lack of suitable tools and techniques, CIPN has not been studied in vivo at the level of an intact DRG neuronal ensemble, leaving an important gap in the understanding of CIPN mechanisms necessary to develop better therapeutics.
There are numerous nonmutually exclusive proposed mechanisms for CIPN caused by platinum-based chemotherapeutic drugs (Jaggi and Singh, 2012). These potential mechanisms including mitochondrial malfunction and cytoplasmic calcium unbalance (Melli et al., 2008), alteration of potassium channel subtype expression (Descoeur et al., 2011), upregulation of transient receptor potential vanilloid receptors (Anand et al., 2010; Ta et al., 2010), endoplasmic reticulum (ER) stress (Nawrocki et al., 2005; Yi et al., 2020), primary sensory neuron senescence (Calls et al., 2021), and oxidative stress via mitochondrial disruption generating reactive oxygen species (ROS; Joseph et al., 2008; Gorgun et al., 2017; Khasabova et al., 2019). ROS produced by mitochondria are normally broken down by the superoxide dismutase pathway (McCord and Fridovich, 1988). An SOD mimetic drug, calmangafodipir, reduced CIPN symptoms caused by oxaliplatin, a platinum-based drug (Glimelius et al., 2018). Of particular relevance to our work here, physiological changes associated with cisplatin-induced CIPN overlap extensively with known mechanisms of inflammatory pain (Hu et al., 2018; Quintão et al., 2019; Y. Zhu et al., 2019).
Concomitant administration of anti-inflammatory drugs with cisplatin has reduced CIPN-induced neuropathic pain both in animal models and in clinical trials (Glimelius et al., 2018; Khasabova et al., 2019). In vivo DRG imaging may be a particularly useful strategy for studying how concomitant drug administration modifies the effect of cisplatin on primary sensory neurons. All mGluRs except mGluR6 are found on central and/or peripheral neurons known to be involved in nociception, and are known to affect pain (Yang and Gereau, 2002, 2003; Marabese et al., 2007; Chiechio et al., 2010; Jaggi and Singh, 2012; Mazzitelli et al., 2018; Pereira and Goudet, 2018). In particular, inflammatory pain is attenuated by Group II mGluRs (mGluR2 and mGluR3) and Group III mGluRs (mGlurR4, mGluR7, and mGluR8; Zammataro et al., 2014; Mazzitelli et al., 2018; Pereira and Goudet, 2018), and is enhanced or modulated by Group I mGluRs (mGluR1 and mGluR5; Bhave et al., 2001; Karim et al., 2001; Zammataro et al., 2014; Radwani et al., 2017; Mazzitelli et al., 2018; Pereira and Goudet, 2018; Masuoka et al., 2020). Thus, metabotropic glutamate receptors (mGluRs), especially Groups II and III, are promising therapeutic targets for CIPN-induced neuropathic pain (Mazzitelli et al., 2018; Pereira and Goudet, 2018).
Meclizine is also an attractive candidate for reducing CIPN caused by platinum-based drugs. Meclizine is an H1 receptor antagonist and a pregnane X receptor agonist in mice shown to be neuroprotective in animal models of conditions as diverse as Huntington's (Gohil et al., 2011), Parkinson's (Hong et al., 2016), and hypoxia (Zhuo et al., 2016). Meclizine protects cultured DRG primary sensory neurons from cisplatin-induced damage by enhancing pentose phosphate pathway activity, enhancing NADPH production, and improving clearance of DNA damage (Gorgun et al., 2017). Furthermore, meclizine shifts retinal ganglion cells toward the glycolysis metabolism pathway and away from the mitochondrial respiratory pathway (J. Zhu et al., 2020), suggesting that meclizine could protect against cisplatin-induced oxidative stress and mitochondrial toxicity.
In this study, we used mechanical and thermal pain behavioral assays and confocal microscopic imaging with the genetically-encoded fluorescent calcium indicator, Pirt-GCaMP3, to investigate cisplatin-based CIPN-induced neuropathic pain in vivo at the level of an intact ensemble comprising the large DRG primary sensory neuronal population. We found that cisplatin induced mechanical and thermal hypersensitivity in pain behavioral assays, increased levels of spontaneous Ca2+ activity in DRG primary sensory neurons, and increased Ca2+ responses to mechanical, thermal, and chemical stimuli. mGluR8 agonist (S)−3,4-DCPG, Group II mGluRs (mGluR2 and mGluR3) agonist LY379268, and histamine H1 receptor antagonist, meclizine, each attenuated cisplatin-induced pain in behavioral assays. Furthermore, LY379268 and meclizine attenuated spontaneous calcium activity in DRG neurons in vivo. Finally, meclizine and LY379268 reduced cisplatin-induced weight loss.
Materials and Methods
Animals
All experiments were performed in compliance with policies and procedures of the Institutional Animal Care and Use Committee at University of Texas Health Science Center at San Antonio (UTHSA) and in accordance with National Institutes of Health and American Association for Accreditation of Laboratory Animal Care. Pirt-GCaMP3 mice (Kim et al., 2014, 2016) on a C57BL/6J background or C57BL/6J mice (The Jackson Laboratory) were used in all experiments. Pirt-GCaMP3 mice used in experiments were all heterozygous. Both males and females were used. All animals were at least eight weeks old, were kept in a 14/10 h light/dark cycle, and were provided ad libitum access to food and water.
Drugs and drug treatment
Cisplatin, LY379268, and (S)−3,4-DCPG were purchased from Abcam. Meclizine, HC 030031, AMG 333, NF 110, and AMG 9810 were purchased from Tocris. Cisplatin stocks were dissolved in 1-methyl-2-pyrrolidinone at 8 mg/ml (Sigma-Aldrich). LY379268, (S)−3,4-DCPG, and NF 110 stocks were dissolved in water. Meclizine stocks were dissolved in DMSO. HC 030031, AMG 333, and AMG 9810 were suspended at 50 µm in water employing repeated vortexing and freeze-thaw cycles as necessary until no pellets could be observed after centrifuging at 18,000 rcf. Aliquots of all drugs except cisplatin were stored at −80°C until use. Cisplatin stock was kept at −20°C until use. All drugs were added to 0.9% saline immediately before injection, intraperitoneally. Meclizine was added to water immediately before injection, intraperitoneally. On days that both behavioral testing and injection occurred, animals were weighed before behavioral testing. Mouse weights were used to calculate drug dosage. Injections were performed after behavioral testing. For drugs used in imaging experiments, aliquots from 50 µm stock were vortexed vigorously and diluted to 10 µm final concentration in PBS. Isoflurane was purchased from the Piramal Group (Mumbai, India). Pentobarbital was purchased from Diamondback Drugs. Cisplatin and saline vehicle injections were performed on days 1, 3, 5, and 7. In experiments using multiple rounds of cisplatin injection, cisplatin was also injected on days 12, 14, 16, and 18. When (S)−3,4-DCPG or LY379268 were concomitantly administered to animals receiving cisplatin, injections were given 30 min before cisplatin. When meclizine was concomitantly administered with cisplatin, injections were given 3 h before cisplatin. In imaging experiments where drugs were applied directly to the DRG, the order of drug administration was HC 030031, AMG 333, NF 110, and AMG 9810.
Behavioral tests
von Frey mechanical test
Mice were placed in a 4.5 × 5 × 10 cm transparent container on a metal mesh and allowed to habituate for at least 30 min before testing. Each mouse was tested at a specific force eight times to determine the lowest force required to elicit a paw withdrawal response more than 50% of the time.
Hot plate thermal test
Mice were placed in a 4.5 × 5 × 10 cm transparent container on a temperature controlled hot plate set to 45°C. Latency to acute nocifensive behavior was determined by onset of hindpaw lifts, licking, jumping, or flinching.
DRG exposure surgery
L5 DRG exposure surgery and imaging was performed as previously described (Kim et al., 2016). Mice were anesthetized with pentobarbital (40–50 mg/kg). After deep anesthesia was achieved, the animal's back was shaved, and the shaved skin was aseptically prepared by cleaning with alcohol and iodine pad, and ophthalmic ointment was applied to keep the eyes moist (Lacrilube, Allergan Pharmaceuticals). Mice were kept on a heating pad and monitored with a rectal thermometer to maintain body temperature at 37 ± 0.5°C.
Dorsal laminectomy was performed in the L4–L6 area. A 2-cm midline incision was made in the lower back around the lumbar enlargement area. Paravertebral muscles were dissected away to expose the lower lumbar enlargement area, and the bones were cleaned. Small rongeurs were used to remove the surface aspect of the L5 DRG transverse process bone near the vertebra to expose the DRG without disrupting the neurons, axons, and other cells in the DRG.
In vivo DRG GCaMP Ca2+ imaging
In vivo GCaMP Ca2+ imaging of the large DRG was performed over 1–6 h following exposure surgery on day 8 after the last of four drug injections on days 1, 3, 5, and 7. Mice were kept on a heating pad and monitored with a rectal thermometer to maintain body temperature at 37 ± 0.5°C. Mice were laid abdomen down on a custom-designed imaging stage. Movement from breathing, heart beats, etc., was minimized by holding the head in a custom designed holder with an anesthesia/gas mask and custom designed vertebral clamps. Continuous anesthesia was maintained with 1−2% isoflurane in pure oxygen.
For imaging and analysis, the stage was fixed under a single photon confocal microscope (Carl Zeiss AG). Raw image stacks (512 × 512–1024 × 1024 pixels in the x-y plane and 20- to 30-µm voxel depth) were converted into time lapse movies (∼6.5–7 s were required to produce a single frame) and analyzed using Zeiss Zen 3.1 Blue Edition software (Carl Zeiss AG).
Putative responding cells were identified by visual observation of raw image time lapse movies. Ca2+ transient intensities were calculated by ΔF/F0 = (Ft – F0)/F0, where Ft is the pixel intensity in a region of interest (Cioroiu and Weimer) at the time point of interest, and F0 is the baseline intensity determined by averaging the intensities of the first two to six frames of the ROI in the experiment. For calcium responses to stimuli that produced multiple transients, only the first peak was analyzed. Cells showing calcium transients before the stimulus were assumed to be spontaneously active and were not included in analysis of stimuli-induced transients. Each peak of spontaneous Ca2+ transients was analyzed individually. Suitable transients (not too much movement of images, clear baseline fluorescence, no nearby cells with Ca2+ transients) were randomly sampled for analysis. An effort was made to sample the same number of cells from each DRG for calculating average and SEM for area under the curve (AUC) and ΔF/F0 intensity; no more than 40% of cells came from a single DRG.
Stimuli were applied carefully so as to not cause movement during imaging. Brush stimuli were applied by repeated gentle brushing from heel to toes approximately once per second using small (5 mm) and large (40 mm) brush bristles. Press (SMALGO Algometer, Bioseb Instruments) and von Frey filaments were applied directly to the hindpaw ipsilateral to the DRG being imaged for 15–20 s. Thermal stimuli were applied by immersing the hindpaw in water at a specific temperature (0°C or 45°C) for 15–20 s. Stimuli were applied after 35–40 s of baseline imaging. HC 030031, AMG 333, NF 110, and AMG 9810 were applied during imaging by topically pipetting onto DRG neurons and allowed to sit for at least 5 min. Between drug applications, neurons were rinsed three to four times in PBS, and the final wash was allowed to sit on the DRG for at least 5 min.
Statistical analysis
Statistics were performed on GraphPad Prism 9.0.1. Ca2+ transient intensities and numbers of activated cells showing spontaneous Ca2+ activity or Ca2+ activity following stimuli were analyzed by Student's t test or one-way or two-way ANOVA followed by post hoc Tukey's test or post hoc Dunnett's test, as appropriate. Specific statistical tests performed are indicated in the text of the results and figure and table legends.
Results
Cisplatin induces mechanical and thermal hyperalgesia, increases spontaneous Ca2+ activity in the DRG neurons, and increases the sensitization of DRG neurons to mechanical and thermal stimuli
To determine the effects of cisplatin on DRG Ca2+ signaling, animals were injected with cisplatin (3.5 mg/kg, i.p.) every other day for four total injections (Deng et al., 2012). Compared with saline-treated controls, cisplatin-injected animals developed mechanical hyperalgesia after the second injection of cisplatin (Fig. 1A, Movie 1; drug effect F(1,48) = 475.7, p < 0.001; time effect F(3,48) = 34.4, p < 0.001; interaction F(3,48) = 40.43, p < 0.001; n = 7 per group; Tukey's df = 48, day 4 q = 17.52, p < 0.001; day 8 q = 20.88, p < 0.001; day 11 q = 21.12, p < 0.001) and thermal hyperalgesia after the fourth cisplatin injection (Fig. 1B; drug effect F(1,48) = 35.32, p < 0.001; time effect F(3,48) = 15.69, p < 0.001; interaction F(3,48) = 5.378, p = 0.0028; n = 7 per group; Tukey's df = 48, day 8 q = 6.722, p = 0.0005; day 11 q = 6.974, p = 0.0003). Ca2+ imaging experiments were performed as previously described (Kim et al., 2016). Animals were anesthetized and DRGs were surgically exposed and imaged. In cisplatin-injected mice, we found significant increases in the numbers of cells with spontaneous Ca2+ oscillation (t(18) = 5.150, p = 0.000067, n = 11 saline and 9 cisplatin) and in total spontaneous Ca2+ activity (total of spontaneous Ca2+ oscillation and steady-state high Ca2+ levels; t(18) = 4.380, p = 0.000361, n = 11 saline and 9 cisplatin) compared with saline-injected mice, but no significant increase in the number of cells exhibiting steady-state high Ca2+ activity compared with saline-injected mice (t(18) = 1.800, p = 0.088631, n = 11 saline and 9 cisplatin; Fig. 1C, Movie 2), indicating that the increase was primarily because of spontaneous Ca2+ oscillation. The numbers of DRG neurons responding to press (100, 300, and 600 g), 10-g von Frey, and 45°C stimuli were increased compared with saline controls (Figs. 1D, 2D, 2G, Movie 1; 100 g t(15) = 3.457, p = 0.003523, n = 8 saline and 9 cisplatin; 300 g t(6) = 3.179, p = 0.01910, n = 4 each; 600 g t(6) = 2.945, p = 0.02,577, n = 4 each; 10 g t(18) = 2.386, p = 0.02,825, n = 11 saline and 9 cisplatin; 45°C t(15) = 4.625, p = 0.000330, n = 11 saline and 9 cisplatin).
Stimulus-induced Ca2+ transients were monitored in saline-treated and cisplatin-treated animals. The numbers of cells showing spontaneous Ca2+ activity and Ca2+ transients in response to stimuli were higher in DRGs of cisplatin-treated animals than in DRGs of saline-treated animals (F(1,47) = 49.21, p < 0.001 by two-way ANOVA). Average AUC of Ca2+ transients of cisplatin group was decreased by a large number of low-amplitude Ca2+ transients (many neurons are activated; Fig. 4C). Analysis of spontaneous and press-induced Ca2+ transients showed that spontaneous Ca2+ transients in the DRG of saline-treated controls took longer to return to baseline than in cisplatin-treated animals (Fig. 1Eb,Ec; drug effect F(1,684) = 17.38, p < 0.001; time effect F(2,684) = 147.6, p < 0.001; interaction F(2,684) = 1.207, p = 0.2997; n = 43 saline, 189 cisplatin; Tukey's df = 684, 0 s q = 4.256, p = 0.0323; 7 s q = 1.621, p = 0.8617; 14 s q = 6.263, p = 0.0002). At 100- and 300-g press, cisplatin-treated animals had greater AUC and were slower to return to baseline than saline-treated animals (Fig. 1Fb,Fc,Gb,Gc; 100 g, AUC t(233) = 3.533, p = 0.000499; drug effect F(1,1068) = 16.88, p < 0.001; time effect F(5,1068) = 13.33, p < 0.001; interaction F(5,1068) = 2.788, p = 0.0165; n = 88 saline, 137 cisplatin; Tukey's df = 1068, 7 s q = 2.276, p = 0.9057; 14 s q = 6.274, p = 0.0006; 21 s q = 7.668, p < 0.0001; 28 s q = 0.5069, p > 0.9999; 35 s q = 6.544, p = 0.0003; 300 g, AUC t(292) = 3.878, p = 0.000130; drug effect F(1,1554) = 15.04, p = 0.001; time effect F(5,1554) = 48.06, p < 0.001; interaction F(5,1554) = 1.779, p = 0.1140; n = 105 saline, 189 cisplatin; Tukey's df = 1554, 7 s q = 3.724, p = 0.2619; 14 s q = 5.281, p = 0.0105; 21 s q = 4.777, p = 0.0360; 28 s q = 6.345, p 0.0005). Ca2+ transients at 600-g press were not significantly different between saline-treated and cisplatin-treated animals (Fig. 1H; drug effect F(1,1073) = 1.277, p = 0.2587; time effect F(5,1073) = 29.62, p < 0.0001; interaction F(5,1073) = 1.066, p = 0.3774). Compared with Ca2+ transients in DRG neurons of saline-treated mice, cisplatin treatment increased AUC of Ca2+ transients in response to small brush stimulus and decreased AUC of Ca2+ transients in response to paw immersion in 45°C water (Fig. 2Bb,Bc,Fb,Fc; small brush AUC t(18) = 6.069, p = 0.00001; drug effect F(1,104) = 23.54, p < 0.0001; time effect F(5,104) = 498.9, p < 0.0001; interaction F(5,104) = 123.8, p < 0.0001; n = 10 saline, 10 cisplatin; Tukey's df = 104, 7 s q = 34.87, p < 0.0001; 14 s q = 5.798, p = 0.0045; 45°C drug effect F(1,623) = 5.675, p = 0.0175; time effect F(5,623) = 62.38, p < 0.0001; interaction F(5,623) = 2.559, p = 0.0264; n = 61 saline, 51 cisplatin; Tukey's df = 623, 7 s q = 0.511, p > 0.9999; 14 s q = 5.185, p = 0.0141). Compared with saline-treated controls, more Ca2+ transients were delayed in cisplatin-treated animals in response to large brush stimulus (Fig. 2Cc; drug effect F(1,136) = 0.5561, p = 0.4571; time effect F(5,136) = 37.74, p < 0.0001; interaction F(5,136) = 10.31, p < 0.0001; n = 11 saline, 16 cisplatin; Tukey's df = 136, 7 s q = 2.479, p = 0.8400; 14 s q = 2.726, p = 0.7398; 21 s q = 7.500, p < 0.0001; 28 s q = 5.980, p = 0.0024; 35 s q = 3.181, p = 0.5175). Amplitude of Ca2+ transients in response to paw immersion in 45°C water was significantly higher in saline-treated animals than in cisplatin-treated animals (Fig. 2Fc; AUC t(110) = 2.568, p = 0.01156; drug effect F(1,623) = 5.675, p = 0.0175; time effect F(5,623) = 62.38, p < 0.0001; interaction F(5,623) = 2.559, p = 0.0264; n = 61 saline, 51 cisplatin; Tukey's df = 623, 7 s q = 0.5110, p > 0.9999; 14 s q = 5.815, p = 0.0141; 21 s q = 3.223, p = 0.4928; 28 s q = 0.4958, p > 0.9999; 35 s q = 0.9730, p > 0.9999). Paw immersion in 0°C water, or paw stimulus with 0.07- and 10-g von Frey, did not show significant differences between Ca2+ transients of two groups (Fig. 2Eb,Ec,Hb,Hc,Ib,Ic; drug effect 0°C F(1,364) = 0.9285, p = 0.3359; 0.07 g F(1,73) = 0.3066, p = 0.5815; 10 g F(1,525) = 1.199, p = 0.2740).
The relative proportions of small, medium, and large diameter neurons that exhibited spontaneous Ca2+ oscillation and response to stimuli were indistinguishable between cisplatin-treated and saline-treated mice. Nonetheless, following cell activation, spontaneous Ca2+ transients and Ca2+ transients following most stimuli predominantly occurred in small diameter neurons (<20 µm), followed by medium diameter neurons (20–25 µm), and finally, relatively few large diameter neurons (>25 µm). However, more large diameter neurons were activated by stronger mechanical stimuli (10-g von Frey, 300- and 600-g press) than by other stimuli, albeit still a lower proportion than small and medium diameter neurons (Tables 1, 2).
Cisplatin-induced hyperalgesia is attenuated by concomitant administration of mGluR8 agonist, (S)−3,4-DCPG
Since mGluR8 agonists are known to attenuate some kinds of inflammatory pain (Marabese et al., 2007; Palazzo et al., 2011), we tested mGluR8 agonist, (S)−3,4-DCPG, in the cisplatin-induced pain model by injecting (S)−3,4-DCPG (30 mg/kg, i.p.) 30 min before each cisplatin injection. (S)−3,4-DCPG attenuated cisplatin-induced mechanical and thermal hyperalgesia (Fig. 3A,B; mechanical drug effect F(2,221) = 463.7, p < 0.0001; time effect F(12,221) = 6.148, p < 0.0001; interaction F(24,221) = 3.563, p < 0.001; n = 7 saline, 7 cisplatin, 6 cisplatin plus (S)−3,4-DCPG; Tukey's df = 221 cisplatin vs cisplatin plus (S)−3,4-DCPG df = 221, day 0 q = 1.120, p = 0.7085; day 4 q = 1.120, p < 0.0001; cisplatin vs saline day 0 q = 1.311, p = 0.6238, day 4 q = 10.63, p < 0.0001; thermal drug effect F(2,234) = 40.91, p < 0.0001; time effect F(12,234) = 8.010, p < 0.001; interaction F(24,234) = 2.424, p = 0.0004; n = 6 saline, 6 cisplatin, 9 cisplatin + (S)−3,4-DCPG; Tukey's cisplatin vs cisplatin plus (S)−3,4-DCPG df = 234, day 0 q = 1.007, p = 0.7566; day 8 q = 4.964, p = 0.0016). However, neither the number of activated neurons showing spontaneous Ca2+ activity nor the number of activated neurons responding to 45°C, 0.07-g von Frey, 10-g von Frey, or 100-g press stimuli were different between cisplatin-treated and cisplatin plus (S)−3,4-DCPG-treated animals (Fig. 3C; 45°C t(16) = 1.448, p = 0.1696, n = 8 each; 0.07 g t(8) = 1.177, p = 0.2731, n = 5 each; 10 g t(15) = 1.305, p = 0.2114, n = 9 cisplatin, 8 cisplatin plus (S)−3,4-DCPG; 100 g t(11) = 0.4550, p = 0.6580, n = 9 cisplatin, 4 cisplatin + (S)−3,4-DCPG). The number of activated neurons following 0°C cold stimulus was increased in cisplatin plus (S)−3,4-DCPG animals compared with cisplatin-only controls (Fig. 3C; t(5) = 4.845, p = 0.04903, n = 4 cisplatin, 3 cisplatin plus (S)−3,4-DCPG). (S)−3,4-DCPG slowed spontaneous decay time of Ca2+ transients (return to baseline; Fig. 3Db,Dc; drug effect F(1,655) = 34.58, p < 0.001; time effect F(2,684) = 454.8, p < 0.001; interaction F(2,684) = 1.987, p = 0.1379; n = 189 cisplatin, 33 (S)−3,4-DCPG + cisplatin; Tukey's df = 655, 0 s q = 3.889, p = 0.0671; 7 s q = 7.092, p < 0.001; 14 s q = 3.425, p = 0.1502). (S)−3,4-DCPG also increased the neuronal number of delayed Ca2+ transients in response to 0°C cold (Fig. 3Eb,Ec; drug effect F(1,366) = 6.630, p = 0.0104; time effect F(5,366) = 20.26, p < 0.001; interaction F(5,366) = 6.794, p < 0.001; n = 40 cisplatin, 25 cisplatin plus (S)−3,4-DCPG; Tukey's df = 366, 7 s q = 3.887, p = 0.2072; 14 s q = 0.2487, p > 0.9999; 21 s q = 6.720, p = 0.0002; 28 s q = 4.831, p = 0.0338) but decreased AUC of Ca2+ transients in response to 10-g von Frey stimulus (Fig. 3Hb,Hc; t(73) = 3.469, p = 0.000879, n = 55 cisplatin, 20 cisplatin plus (S)−3,4-DCPG).
Cisplatin-induced hyperalgesia and the numbers of spontaneous Ca2+-activated cells are attenuated by Group II mGluR agonist, LY379268
Group II mGluRs are strong candidate analgesic targets for inflammatory pain (Mazzitelli et al., 2018). We tested effects of the mGlu2/3 receptor agonist, LY379268, on cisplatin-induced pain in the cisplatin-induced animal model. We found that injection of LY379268 (10 mg/kg, i.p.) 30 min before each cisplatin injection reduced cisplatin-induced mechanical and thermal hyperalgesia by day 19 (Fig. 4A,B; mechanical drug effect F(2,109) = 60.27, p < 0.0001; time effect F(3,109) = 38.62, p < 0.0001; interaction F(6,109) = 8.617, p < 0.0001; n = 12 saline, 12 cisplatin, 7 cisplatin plus LY379268; Dunnett's df = 109, cisplatin vs cisplatin plus LY379268, day 0 q = 0.5559, p = 0.8075; day 8 q = 1.149, p = 0.4264; day 12 q = 1.380, p = 0.3016; day 19 q = 3.404, p = 0.0019; cisplatin vs saline, df = 109 d 0 q = 0.08549, p = 0.9949; day 8 q = 8.339, p < 0.0001; day 12 q = 6.727, p < 0.0001; day 19 q = 6.847, p < 0.0001; thermal drug effect F(2,67) = 17.68, p < 0.0001; time effect F(3,67) = 11.57, p < 0.0001; interaction F(6,67) = 1.889, p = 0.0955; n = 7 per group; Dunnett's cisplatin vs cisplatin plus LY379268, df = 67, day 0 q = 0.6970, p = 0.7072; day 8 q = 0.5504, p = 0.8049; day 12 q = 2.213, p = 0.0550; day 19 q = 2.551, p = 0.0244; cisplatin vs saline, df = 67, day 0 q = 0.4930, p = 0.8378; day 8 q = 3.180, p = 0.0043; day 12 q = 3.507, p = 0.0016; day 19 q = 4.819, p < 0.0001). The numbers of activated neurons following 45°C hot stimulus was decreased in cisplatin plus LY379268-treated animals compared with cisplatin-treated controls (Fig. 4D; t(9) = 2.279, p = 0.04868, n = 8 cisplatin, 3 cisplatin plus LY379268). LY379268 increased AUC and slowed decay time of spontaneous Ca2+ transients (Fig. 4Ec; AUC t(264) = 3.641, p = 0.000327, n = 189 cisplatin, 77 cisplatin plus LY379268 drug effect F(1,783) = 40.79, p < 0.0001; time effect F(2,783) = 609.2, p < 0.0001; interaction F(2,783) = 0.2899, p = 0.7484; Tukey's df = 783 0 s q = 3.767, p = 0.0005; 7 s q = 4.234, p < 0.0001; 14 s q = 2.684, p = 0.0221) Spontaneous amplitude was not statistically distinguishable from saline (Dunnett's saline vs cisplatin plus LY379268 df = 340 q = 0.8542 p > 0.999). LY379268 slowed decay time of Ca2+ transients following 45°C (Fig. 4Fc; drug effect F(1,543) = 20.27, p < 0.0001; time effect F(5,543) = 14.28, p < 0.0001; interaction F(5,543) = 19.43, p < 0.0001; n = 51 cisplatin, 66 cisplatin plus LY379268; Tukey's df = 543 7 s q = 5.766, p = 0.0030; 14 s q = 4.781, p = 0.0369; 21 s q = 2.2800, p = 0.9042; 28 s q = 8.299, p < 0.0001; 35 s q = 6.328, p = 0.0006; 42 s q = 6.198, p = 0.0009).
Cisplatin-induced mechanical hyperalgesia, spontaneous DRG neuronal Ca2+ activity, and Ca2+ response to 45°C hot stimulus are attenuated by histamine receptor 1 antagonist, meclizine
Since meclizine has a protective role against DNA damage by cisplatin (Gorgun et al., 2017), we tested meclizine in cisplatin-treated animals for its effects on behavior and Ca2+ activity. Meclizine was injected (16 mg/kg, i.p.) 3 h before cisplatin injection (3.5 mg/kg, i.p.) four times over 8 d. Meclizine reduced cisplatin-induced mechanical hypersensitivity (Fig. 5A; drug effect F(2,129) = 45.55, p < 0.0001; time effect F(3,109) = 27.65, p < 0.0001; interaction F(6,109) = 5.729, p < 0.0001; n = 12 saline, 15 cisplatin, 16 cisplatin plus meclizine; Dunnett's cisplatin vs cisplatin plus meclizine, df = 129 d 0 q = 0.004792, p > 0.9999; day 8 q = 3.078, p = 0.0050; day 12 q = 3.052, p = 0.0054; day 19 q = 2.958, p = 0.0073; cisplatin vs saline, df = 129 d 0 q = 0.07429, p = 0.9960; day 8 q = 7.246, p < 0.0001; day 12 q = 5.845, p < 0.0001; day 19 q = 5.949, p < 0.0001) and reduced cisplatin-induced thermal hypersensitivity (Fig. 5B; drug effect F(2,63) = 19.04, p < 0.0001; time effect F(3,63) = 7.403, p = 0.0003; interaction F(6,63) = 2.958, p = 0.0131; n = 7 saline, 7 cisplatin, 6 cisplatin plus meclizine; Dunnett's cisplatin vs cisplatin plus meclizine, df = 63 d 0 q = 0.004792, p > 0.9999; day 8 q = 3.078, p = 0.0050; day 12 q = 3.052, p = 0.0054; day 19 q = 2.958, p = 0.0073; cisplatin vs saline, df = 129 d 0 q = 0.07429, p = 0.9960; day 8 q = 7.246, p < 0.0001; day 12 q = 5.845, p < 0.0001; day 19 q = 5.949, p < 0.0001). Meclizine also attenuated spontaneous Ca2+ activity in cisplatin-treated animals, reducing the numbers of neurons exhibiting spontaneous Ca2+ activity (Ca2+ oscillation plus steady-state high Ca2+) down to numbers comparable to numbers seen in saline-treated animals (Fig. 5C,D; oscillation t(15) = 2.872, p = 0.01163, n = 9 cisplatin, 8 cisplatin plus meclizine; steady high t(14) = 2.134, p = 0.05098, n = 9 cisplatin, 7 cisplatin plus meclizine; total spontaneous activity t(14) = 4.112, p = 0.01056, n = 9 cisplatin, 7 cisplatin plus meclizine), and reduced the numbers of Ca2+-activated neurons in response to 100-g press (mechanical stimulus) and 45°C water (thermal stimulus; Fig. 5C,Da, Movie 3; 100 g t(14) = 2.197, p = 0.04532, n = 9 cisplatin, 7 cisplatin plus meclizine; 45°C t(15) = 3.500, p = 0.00322, n = 9 cisplatin, 8 cisplatin plus meclizine). In addition, meclizine treatment significantly increased the AUC of Ca2+ transients and decay time (return to baseline) in response to 0°C water (cold stimulus; Fig. 5Eb,Ec; 0°C AUC t(87) = 2.187, p = 0.03146, drug effect F(1,505) = 9.078, p = 0.0027; time effect F(5,505) = 35.90, p < 0.0001; interaction F(5,505) = 7.338, p < 0.0001; n = 40 cisplatin, 49 cisplatin plus meclizine; Tukey's df = 505 7 s q = 4.679, p = 0.0463; 14 s q = 1.973, p = 0.9642; 21 s q = 6.654, p = 0.0002; 28 s q = 4.284, p = 0.1032; 35 s q = 1.589, p = 0.9935; 42 s q = 0.5821, p > 0.9999). In contrast, meclizine reduced AUC of Ca2+ transients in response to 45°C water (Fig. 5Fb,Fc; AUC t(59) = 2.519, p = 0.01451, drug effect F(1,501) = 0.3214, p = 0.5710; time effect F(5,501) = 45.65, p < 0.0001; interaction F(5,501) = 0.1126, p = 0.1126; n = 40 cisplatin, 21 cisplatin plus meclizine; Tukey's df = 501 7 s q = 5.132, p = 0.0163; 14 s q = 1.513, p = 0.9958; 21 s q = 0.7778, p > 0.9999; 28 s q = 0.07723, p > 0.9999; 35 s q = 2.394, p = 0.8709; 42 s q = 0.0000, p > 0.9999), 10-g von Frey (Fig. 5Hb,Hc; 10 g AUC t(80) = 3.580, p = 0.000587, drug effect F(1,424) = 12.13, p = 0.0005; time effect F(5,424) = 8.542, p < 0.0001; interaction F(5,424) = 2.385, p = 0.0376; n = 55 cisplatin, 27 cisplatin plus meclizine; Tukey's df = 424 7 s q = 5.415, p = 0.0081; 14 s q = 1.739, p = 0.9864, 21 s q = 4.117, p = 0.1406, 28 s q = 2.583, p = 0.8026, 35 s q = 1.523, p = 0.9955, 42 s q = 1.019, p = 0.9999), and 100-g press (Fig. 5Ib,Ic; AUC t(257) = 4.063, p = 0.000065, drug effect F(1,1089) = 33.17, p < 0.0001; time effect F(5,1089) = 9.116, p < 0.0001; interaction F(5,1089) = 4.919, p = 0.0001; n = 137 cisplatin, 122 cisplatin plus meclizine; Tukey's df = 1089 7 s q = 8.499, p < 0.0001; 14 s q = 5.082, p = 0.0176, 21 s q = 4.6700, p = 0.0460, 28 s q = 0.5574, p > 0.9999, 35 s q = 5.8300, p = 0.0024, 42 s q = 0.1629, p > 0.9999). Meclizine caused no detectable change on Ca2+ transients in response to 0.07-g von Frey (p > 0.5; Fig. 5Gb,Gc; drug effect F(1,85) = 0.08211, p = 0.7752; time effect F(5,85) = 6.552, p < 0.0001; interaction F(5,85) = 0.9157, p = 0.4749; n = 8 cisplatin, 10 cisplatin plus meclizine).
Cisplatin-induced weight loss is attenuated by meclizine and LY379268 but unaffected by (S)−3,4-DCPG
Cachexia, a condition associated with cancer and characterized by weight loss, muscle loss, and adipose loss, increases cancer mortality (Fearon et al., 2013), and is exacerbated by nearly every form of chemotherapy (Langer et al., 2002; Kazemi-Bajestani et al., 2016). Here, we monitored cisplatin-associated weight loss over the time course of experiments. Cisplatin treatment was associated with rapid, persistent weight loss (Fig. 6A–C). Concomitant (S)−3,4-DCPG treatment had no attenuating effect on weight loss (Fig. 6A). Cisplatin-associated body weight loss among (S)−3,4-DCPG-treated animals returned to baseline levels 32 d after the last cisplatin injection (Fig. 6A; Dunnett's df = 425 cisplatin vs cisplatin plus (S)−3,4-DCPG day 58 q = 1.020, p = 0.4924). Animals that received concomitant LY379268 experienced less weight loss over the course of injections than cisplatin-treated controls (Fig. 6B; drug effect F(2,191) = 124.9, p < 0.0001; time effect F(10,191) = 25.30, p < 0.0001; interaction F(20,191) = 3.375, p < 0.0001; n = 7 each, Dunnett's df = 191 cisplatin vs cisplatin plus LY379268 day 18 q = 2.294, p = 0.421). Animals that received concomitant meclizine administration experienced moderate weight loss (drug effect F(2,154) = 50.01, p < 0.0001; time effect F(10,154) = 3.595, p = 0.0003; interaction F(20,154) = 1.753, p = 0.0306; n = 7 saline, 7 cisplatin, 6 cisplatin plus meclizine) and returned to baseline within 10 d after the fourth injection. We also observed a late effect, as body weight of meclizine-treated animals began decreasing 50 d after the last injection (Fig. 6C; Dunnett's df = 191 cisplatin vs cisplatin plus meclizine day 18 q = 1.385, p = 0.2922).
TRP channel inhibitors administered directly onto DRGs during imaging experiments reduces the number of cells producing spontaneous Ca2+ transients and NF 110 and AMG 9810 sensitize saline-treated animals to 45°C stimulus
In order to study somatosensory ion channels or receptors involved in cisplatin-induced drug hypersensitivity, we applied four TRP channel inhibitors to DRG to study their effects on spontaneous activity and four stimuli: large brush, 10-g von Frey, 100-g press, and hindpaw immersion in 45°C water. The TRP channel inhibitors used and their targets were HC 030031 (TRPA1; McNamara et al., 2007), AMG 333 (TRPM8; Horne et al., 2018), NF 110 (P2X3 and P2X2/3 heteromultimers; Hausmann et al., 2006), and AMG 9810 (TRPV1; Doherty et al., 2005).
The number of cells with Ca2+ activity in response to stimuli was largely unaffected by inhibitor application. Treatment with inhibitors reduced spontaneous Ca2+ transients (drug effect F(4,82) = 4.843, p = 0.0015) but this effect was indistinguishable between treatment groups (drug x treatment group interaction F(16,82) = 0.4225, p = 0.9725). The treatment group of the animals was more likely to have an effect on the number of cells with Ca2+ activity (for a complete list of F and p values, see Table 3), but the treatment group was not associated with a significant effect of any TRP channel inhibitor.
Application of NF 110 to DRG had a striking effect on Ca2+ transients in response to 45°C (Fig. 7Ia,Ib,Ic; saline controls AUC t(102) = 3.242, p = 0.001607; inhibitor effect F(1,387) = 16.17, p < 0.001; time effect F(5,387) = 4.493, p = 0.0005; interaction F(5,387) = 2.204, p = 0.0532; n = 44 baseline, 60 NF 110; Tukey's df = 387, 7 s q = 0.635, p > 0.9999; 14 s q = 7.050, p < 0.0001; Cisplatin AUC t(211) = 4.407, p = 0.000017; inhibitor effect F(1,874) = 12.62, p = 0.0004; time effect F(5,874) = 9.165, p < 0.0001; interaction F(5,874) = 3.070, p = 0.0094; n = 126 baseline, 79 NF 110; Tukey's df = 874, 7 s q = 1.079, p = 0.9998; 14 s q = 6.297, p = 0.0006; 21 s q = 5.469, p = 0.0066). No other TRP channel inhibitor had a significant effect on Ca2+ transient fluorescence intensity in response to 45°C (data not shown). NF 110 had no detectable effect on Ca2+ transient fluorescence intensity in response to mechanical stimuli (data not shown). In addition, in saline-treated animals, NF 110 and AMG 9810 increased the number of cells producing Ca2+ transients in response to 45°C, but there were no effects on any other treatment group (Fig. 7H; saline F(4,19) = 4.506, p = 0.0099; n = baseline 9, HC 030031 4, AMG 333 4, NF 110 3, AMG 9810 3; Dunnett's df = 18, HC 030031 q = 0.2484, p = 0.9979; AMG 333 q = 0.9985, p = 0.7607; NF 110 q = 3.060, p = 0.0238; AMG 9810 q = 3.329, p = 0.0132).
Discussion
CIPN is a painful, debilitating, and complex phenomenon that occurs commonly in patients treated with chemotherapeutic drugs. Various chemotherapeutic drugs act through different molecular and cellular mechanisms, suggesting that causes of CIPN are similarly diverse (Quasthoff and Hartung, 2002; Seretny et al., 2014; Addington and Freimer, 2016; Cioroiu and Weimer, 2017; Flatters et al., 2017). This suggests that the treatment of CIPN should be guided by the chemotherapeutic drug used. The DRG is a site of peripheral sensory neuropathy. Therefore, we sought to study the effects of specific chemotherapeutic agents and specific therapies on peripheral neurons and to understand how specific therapies affect responses to different mechanical, thermal, and chemical stimuli. To accomplish this, we used in vivo GCaMP Ca2+ imaging of the large population of DRG neurons (>1800/DRG), which enabled us to analyze neuronal firing and activity at a population level as an ensemble. Such in vivo imaging allows for the detection and study of physiologically important phenomena that would be a challenge to detect using more conventional methods such as electrophysiological recordings (Kim et al., 2016).
Prior research has shown that cisplatin treatment induces spontaneous activity and hyperexcitability in dissociated primary sensory neurons in vitro (Laumet et al., 2020). Here, using in vivo ensemble analysis of the large array of DRG neurons in an animal model of CIPN, we found that in response to several mechanical and thermal stimuli Ca2+ activity and Ca2+ transients were increased by cisplatin treatment. The increased numbers of DRG neurons showing Ca2+ activity in response to mechanical and thermal stimuli are consistent with mechanical and thermal hypersensitivity observed in behavioral tests. As expected, strong mechanical stimuli caused larger diameter (>25 µm) neurons to activate more often, while thermal, cold, and weak mechanical stimuli (0.07-g von Frey) activated few if any large diameter neurons, consistent with other reports (Kim et al., 2016; Chisholm et al., 2018; Ishida et al., 2021). Previous work has shown activation of increased numbers of DRG neurons in response to stimuli in a variety of pain models (Kim et al., 2016; Chisholm et al., 2018; Kucharczyk et al., 2020; Ishida et al., 2021). Taken together, these results show that the numbers of DRG neurons activated in response to stimuli measured by Pirt-GCaMP3 fluorescence is a valid indicator of peripheral pain hypersensitivity. Our results show that concomitant administration of (S)−3,4-DCPG, LY379268, or meclizine attenuated CIPN mechanical and thermal pain hypersensitivity. Our results also show that two of three drugs (meclizine, LY379268) which attenuated CIPN pain hypersensitivity also attenuated the number of neurons exhibiting spontaneous Ca2+ activation. Furthermore, concomitant meclizine or LY379268 administration reduced cisplatin-induced weight loss.
Cisplatin increased the fluorescence intensity of Ca2+ transients for several mechanical stimuli (100- or 300-g press or small brush). Interpretation of AUC and plots of fluorescence intensity must be done carefully. Increases in average AUC and amplitude of Ca2+ transients should indicate a stronger response to stimuli (e.g., neurons firing more action potentials, upregulation of voltage-gated and other Ca2+ channels, lowered thresholds for Ca2+ channel activation, increased efflux of Ca2+ from ER or mitochondria, etc.). However, these same factors that cause increased numbers of neurons to produce Ca2+ transients with higher amplitudes can increase baseline cytosolic Ca2+ concentrations and, therefore, increased baseline fluorescence. It is important to note that plots of fluorescence depend on the change in fluorescent intensity relative to baseline. Cisplatin-treated animals generally exhibit higher overall baseline fluorescence intensity (increased F0). This reduces the ΔF/F0 ratio, reducing the calculated AUC. Despite the apparently elevated baseline intensities, the calculated AUC in cisplatin-treated animals was nearly always at least as high as in saline vehicle-treated control animals and often higher. An alternative, nonmutually exclusive possibility, is that the increased number of cells producing Ca2+ transients in response to cisplatin treatment produce low amplitude transients, which would reduce the average amplitude (Fig. 4C).
The effect of (S)−3,4-DCPG on cisplatin-induced CIPN pain does not appear to be mediated by DRG neurons. mGluR8 is present in most DRG (Carlton and Hargett, 2007; Govea et al., 2012) and trigeminal ganglia (Boye Larsen et al., 2014) neurons. However, the role of mGluR8 in the DRG in CIPN has not been determined. Peripheral administration of mGluR8 agonists (such as (S)−3,4-DCPG) can attenuate thermal and mechanical hyperalgesia caused by carrageenan, formalin, and capsaicin (Marabese et al., 2007; Govea et al., 2012; Pereira and Goudet, 2018). (S)−3,4-DCPG can modulate pain centrally, as well. Infusion of (S)−3,4-DCPG into the periaqueductal gray area was sufficient to reduce nociceptive responses and thermal and mechanical hypersensitivity caused by formalin and carrageenan injection. In addition, infusion of an antagonist of Group III mGluRs to the periaqueductal gray area reduced the analgesic effect of peripheral (S)−3,4-DCPG, showing that mGluR8 can act fully or partially through a central anti-nociceptive pathway (Marabese et al., 2007). Administration of mGluR8 agonist (S)−3,4-DCPG 30 min before cisplatin produced a striking reduction in cisplatin-induced mechanical hypersensitivity and reduced thermal hypersensitivity; yet, (S)−3,4-DCPG-treated animals were not distinguishable in terms of numbers of cells with spontaneous Ca2+ activity or numbers of activated cells responding to any stimulus analyzed in this study except 0°C, where (S)−3,4-DCPG increased the number of responding neurons. While there were some subtle differences in Ca2+ transients, (S)−3,4-DCPG failed to produce the striking effects on DRG neuronal Ca2+ activity seen with LY379268 and meclizine. Since (S)−3,4-DCPG administration was systemic, it was not possible to distinguish central and peripheral effects. The lack of striking decreases in numbers of spontaneously Ca2+-activated neurons or activation from stimuli combined with known effects of (S)−3,4-DCPG in the periaqueductal gray area suggests that attenuation of cisplatin-induced mechanical and thermal hypersensitivity by (S)−3,4-DCPG is central rather than peripheral or that (S)−3,4-DCPG works peripherally through a mechanism that does not affect DRG neuronal Ca2+ activity. mGluR8 may be a target for CIPN pain although it does not seem to act peripherally.
Group II mGluR agonists attenuate hyperalgesia and allodynia from a broad range of causes (Carlton et al., 2009; Davidson et al., 2016; Johnson et al., 2017; Mazzitelli et al., 2018; Sheahan et al., 2018), and antagonists of Group II receptors may aggravate hyperalgesia and allodynia, block analgesia, and prolong recovery time from these conditions (Yang and Gereau, 2002, 2003; Zhuo et al., 2016), and may increase the number of action potentials in primary sensory neurons in response to capsaicin or heat (Carlton et al., 2011). Group II mGluRs (mGluR2/3) are broadly expressed in primary sensory neurons (Carlton and Hargett, 2007; Boye Larsen et al., 2014; Sheahan et al., 2018). Our data show that in this cisplatin-induced CIPN model, mGluR2/3 agonist LY379268 reduces mechanical and thermal hyperalgesia and decreases the number of DRG neurons exhibiting spontaneous Ca2+ activity as well as the number of DRG neurons producing Ca2+ transients in response to 45°C. Mechanical and thermal sensitivity assays were based on systemic administration, so it is not possible to determine the extent to which the effect was peripheral or central. However, reduced numbers of DRG Ca2+-activated neurons indicates that analgesia via primary sensory neuron activity is one mechanism. Other studies have shown peripheral mGluR2 or mGluR2/3 agonists can block inflammatory hyperalgesia (Yang and Gereau, 2002, 2003; Yamamoto et al., 2007). Side effects remain a major concern for Group II mGluR agonists, as convulsions have been observed in preclinical animal models (Dunayevich et al., 2008). Limiting a mGluR2/3 drug to the periphery may reduce side effects (Chishty et al., 2001; Gao et al., 2017; Fan et al., 2018). Our results combined with earlier research suggest that a mGluR2/3 agonist with poor blood brain barrier penetration could treat CIPN pain without disrupting mGluR2/3 signaling in the central nervous system.
Concomitant meclizine treatment reduces mechanical and thermal hyperalgesia and decreases the number of DRG neurons exhibiting spontaneous Ca2+ activity (both Ca2+ oscillation and total cells with steady-state high Ca2+ activity) and reduces the number of DRG neurons producing Ca2+ transients in response to 45°C. Meclizine also reduced the intensity of Ca2+ transients arising from two different mechanical stimuli: 10-g von Frey and 100-g press, suggesting that meclizine may act on mechanical pain by reducing the activity of neurons responding to stimuli. Because meclizine treatment was systemic, it is impossible to distinguish a central from peripheral effect. However, there are good reasons to believe that meclizine acts via peripheral mitochondrial repair and/or by reducing oxidative stress. Cisplatin binds mitochondrial DNA in DRG neurons, inhibiting mitochondrial transcription and replication, and causing mitochondrial DNA degradation (Podratz et al., 2011). Cisplatin directly inhibits mitochondrial respiration system in vivo and reduces transcription and increases degradation of mitochondrial RNA (Garrido et al., 2008). Meclizine improves clearance of damaged mitochondrial DNA (Gorgun et al., 2017) and shifts mitochondrial respiration toward glycolytic metabolism (Hong et al., 2016; Zhuo et al., 2016). Meclizine also maintains an elevated ratio of reduced:oxidized glutathione and NADPH (Zhuo et al., 2016; Gorgun et al., 2017). Thus, meclizine employs multiple mechanisms to protect neurons from oxidative damage. ROS and oxidative stress are known to activate and/or potentiate a variety of transient receptor potential channels (Mori et al., 2016; Carrasco et al., 2018), which may partially depolarize neurons, thus rendering them more excitable and resulting in increased spontaneous Ca2+ activity and decreased neuron firing threshold in response to stimuli. ROS also affect potassium and sodium channels (Annunziato et al., 2002; Sahoo et al., 2014) and can damage a wide range of other molecules in the cell (Juan et al., 2021). By improving the redox state, enhancing DNA repair, and shifting primary sensory neuron metabolism away from mitochondrial respiration and toward glycolysis, meclizine may attenuate hyperexcitability through multiple known mechanisms, and thereby reduce hypersensitivity to painful stimuli. Meclizine is also an approved drug for use in humans, making it an excellent candidate for possible use in treating CIPN.
One of the more common side effects of platinum-based chemotherapeutic drugs is weight loss (Langer et al., 2002; Fearon et al., 2013; Kazemi-Bajestani et al., 2016). Numerous mechanisms by which cisplatin and other platinum-based chemotherapeutic drugs cause weight loss have been proposed (Yakabi et al., 2010; Hiura et al., 2012; Garcia et al., 2013; Matsumura et al., 2013; Yoshimura et al., 2013; Woo et al., 2016; Guo et al., 2018; Wong et al., 2020). Meclizine reduced both cisplatin-induced hypersensitivity to mechanical pain and cisplatin-induced weight loss. There were no detectable effects of (S)−3,4-DCPG on body weight of cisplatin-treated animals. Of relevance, meclizine is commonly used for treatment of nausea and vomiting during pregnancy (Heitmann et al., 2016) and motion sickness (Paule et al., 2004), and cisplatin induces nausea in humans (Dilruba and Kalayda, 2016). Meclizine is known to be a strong antagonist of H1 histamine receptor, a weak antagonist of muscarinic acetylcholine receptor (Kubo et al., 1987), and, in mice but not in humans, a weak antagonist of constitutive androstane receptor (Huang et al., 2004). Histamine suppresses food intake in the central nervous system (Clineschmidt and Lotti, 1973; Provensi et al., 2016), and histaminergic neurons play a large role in food intake, metabolism, weight gain or loss, and foraging-related locomotor activities (Provensi et al., 2016). Mice lacking the muscarinic three acetylcholine receptor eat less and weigh less than wild-type mice (Yamada et al., 2001; Gautam et al., 2006). Stimulation of hypothalamic muscarinic acetylcholine receptors stimulates food intake (Jeong et al., 2017). These observations suggest that meclizine may attenuate weight loss through antagonism of anorexigenic H1 histamine receptors or muscarinic acetylcholine receptors.
We investigated specific receptor roles in our imaging system by applying specific TRP channel inhibitors to DRG cell bodies. Unsurprisingly, the inhibitors decreased the number of spontaneously activating cells. The P2X3 and P2X2/3 heteromultimer inhibitor, NF 110, appeared to sensitize DRGs to 45°C stimulus, increasing the number of cells responding in saline injected control animals and dramatically increasing the fluorescence intensity of Ca2+ transients in both cisplatin-injected and saline-injected animals. No TRP channel inhibitors decreased cisplatin-induced sensitivity to stimuli. TRPV1 inhibitor, AMG 9810, also increased the number of Ca2+-activated cells responding to 45°C stimulus. These 45°C stimulus results seem paradoxical, but the inhibitor screen was for indirect effects on stimuli. Inhibitors were applied to the cell bodies rather than into the hindpaw where the receptors that respond directly to the stimuli are found. Importantly, these effects were detectable in saline-treated control animals, but not in cisplatin-treated or cisplatin-plus-TRP channel inhibitor-treated animals. These results suggest that P2X3-injected and TRPV1-induced sensitization is either blocked by, downstream of, or redundant with the effects of cisplatin.
We have used a cisplatin-based CIPN model to study DRG neurons in vivo at the level of an intact neuronal ensemble. We found that two (LY379268 and meclizine) of three drug treatments tested attenuated CIPN-induced pain hypersensitivity, reduced spontaneous DRG neuron Ca2+ activity, and reduced the number of cells responding to 45°C stimulus. The third drug, (S)−3,4-DCPG, did not reduce spontaneous Ca2+ activity, but (S)−3,4-DCPG is known to reduce pain hypersensitivity centrally in some inflammatory pain models. Finally, our results demonstrate that in vivo Ca2+ imaging of DRG neurons can be used to study mechanisms of CIPN and peripherally-acting candidates for therapeutic intervention.
Footnotes
This work was supported by National Institutes of Health Grants R01DE026677, R35DE030045, and R01DE031477 (to Y.S.K.), UTHSCSA startup fund (Y.S.K.), and a Rising STAR Award from University of Texas system (Y.S.K.).
The authors declare no competing financial interests.
- Correspondence should be addressed to Yu Shin Kim at kimy1{at}uthscsa.edu