Subcutaneous ω-Conotoxins Alleviate Mechanical Pain in Rodent Models of Acute Peripheral Neuropathy

The peripheral effects of ω-conotoxins, selective blockers of N-type voltage-gated calcium channels (CaV2.2), have not been characterised across different clinically relevant pain models. This study examines the effects of locally administered ω-conotoxin MVIIA, GVIA, and CVIF on mechanical and thermal paw withdrawal threshold (PWT) in postsurgical pain (PSP), cisplatin-induced neuropathy (CisIPN), and oxaliplatin-induced neuropathy (OIPN) rodent models. Intraplantar injection of 300, 100 and 30 nM MVIIA significantly (p < 0.0001, p < 0.0001, and p < 0.05, respectively) alleviated mechanical allodynia of mice in PSP model compared to vehicle control group. Similarly, intraplantar injection of 300, 100, and 30 nM MVIIA (p < 0.0001, p < 0.01, and p < 0.05, respectively), and 300 nM and 100 nM GVIA (p < 0.0001 and p < 0.05, respectively) significantly increased mechanical thresholds of mice in OIPN model. The ED50 of GVIA and MVIIA in OIPN was found to be 1.8 pmol/paw and 0.8 pmol/paw, respectively. However, none of the ω-conotoxins were effective in a mouse model of CisIPN. The intraplantar administration of 300 nM GVIA, MVIIA, and CVIF did not cause any locomotor side effects. The intraplantar administration of MVIIA can alleviate incision-induced mechanical allodynia, and GVIA and MVIIA effectively reduce OIPN associated mechanical pain, without locomotor side effects, in rodent models. In contrast, CVIF was inactive in these pain models, suggesting it is unable to block a subset of N-type voltage-gated calcium channels associated with nociceptors in the skin.

The CaV2.2 channels play a crucial role in transmitting nociceptive pain signals from the peripheral nervous system (PNS) to the central nervous system (CNS) [10] and are considered as promising analgesic targets. The ability of certain ω-conotoxins to selectively block CaV2.2 channels led to their development as a new class of intrathecal analgesics for severe pain. GVIA, isolated from Conus geographus, was reported to be the first ω-conotoxin irreversibly blocking CaV channel [11]. In 2004, Ziconotide, a synthetic version of ω-conotoxin MVIIA found in Conus magus, gained approval from US Food and Drug Administration (FDA) to treat chronic refractory pain [12,13], albeit with dose limiting side effects including dizziness, nystagmus, sedation, blurred vision, and orthostatic hypotension [14][15][16]. Poor oral bioavailability and CNS access combined with a short plasma half-life [17] restrict the use of MVIIA to intrathecal administration with incumbent complications of headaches, pump-associated infections, and bleedingassociated diathesis [18,19]. Other highly potent ω-conotoxins, e.g., CVID and CVIF isolated form Conus catus, showed potential as an analgesic with a better safety profile in preclinical studies, with CVID showing uncertain therapeutic advantage in limited clinical studies [20][21][22][23].
Acute postsurgical pain (PSP) and chemotherapy-induced peripheral neuropathy (CIPN) are two major health issues worldwide, related to peripheral pain. Despite receiving acute pain medications, up to 71% of patients report moderate to extreme pain after surgery [24][25][26]. Currently available treatment options to manage PSP include opioid analgesics, nonsteroidal anti-inflammatory drugs, local anesthetics, paracetamol, and physiotherapy. These analgesics have abuse potential, side effects, and even lack of efficacy. On the other hand, CIPN is one of the major side effects of anticancer drugs, with a very high prevalence of 70-100% for platinum-based antineoplastic agents [27]. At present, there is no safe and fully effective treatment for CIPN, with partial relief from pain provided by anticonvulsants, e.g., carbamazepine, gabapentin, and antidepressants, e.g., amitriptyline and duloxetine [28]. Due to increasingly effective anticancer treatments, the number of cancer survivors that suffer from CIPN are increasing every year [29]. To improve the quality of their lives, it is crucial to develop effective therapeutic strategies to treat CIPN.
The central analgesic effects of intrathecal ω-conotoxins have been characterised in a range of rodent pain models [20,[30][31][32]. Although previous studies suggest that calcium channels contribute to peripheral neural mechanisms of hyperalgesia [33], the peripheral analgesic effects of ω-conotoxins remain to be fully characterised. In this study, we investigated the analgesic effects of subcutaneously administered ω-conotoxins GVIA, MVIIA, and CVIF in rodent models of postsurgical pain (PSP) and cisplatin-and oxaliplatin-induced peripheral neuropathy (CIPN).

Local Effects of ω-Conotoxins on Surgery-Induced Mechanical and Thermal Allodynia
Surgery of the right hind paw elicited mechanical allodynia in mice, with PWTs to mechanical stimulation of mice in the surgery group being significantly lower compared to naïve mice at 24 h after surgery (naïve, 3.4 ± 0.1 g; post-surgery, 1.5 ± 0.1 g; p < 0.0001; Figure 2A). All animals maintained body weight and a healthy appearance. The animals were monitored during the whole time and no sign of adverse effects of the surgery was observed.

Effect of ω-Conotoxins in Oxaliplatin-Induced Mechanical Allodynia
The intraplantar administration of a single dose (40 µg/paw) of oxaliplatin into the right hind paw elicited strong mechanical allodynia 24 h after injection ( Figure 3A) (control, 3.3 ± 0.1 g; oxaliplatin, 1.4 ± 0.1 g; p < 0.0001; Figure 3A). Oxaliplatin injected mice maintained healthy appearance and body weight with no sign of licking, shaking, or lameness of the injected paw.

Effect of ω-Conotoxins in Cisplatin-Induced Mechanical Allodynia
Intraplantar administration of 40 µg/paw cisplatin into the right hind paw elicited strong mechanical allodynia in mice 24 h after the injection (Figure 4). The mechanical paw withdrawal thresholds of mice injected with cisplatin were significantly lower compared to naïve mice when tested 24 h after injection (naïve, 3.3 ± 0.1 g; 40 µg/paw cisplatin, 1.9 ± 0.1 g; p < 0.0001; Figure 4). Cisplatin injected mice did not lose weight and maintained a healthy appearance. No sign of adverse effects including, licking, shaking, or lameness of the injected paws was observed.

Discussion
This study has demonstrated that the intraplantar administration of ω-conotoxins effectively alleviate acute peripheral neuropathy in different mouse models of acute and neuropathic pain. We found that the administration of up to 0.6 pmol/paw MVIIA (i.pl.) elicited an increase of mechanical PWTs in mice, which had undergone surgery on the plantar foot through the skin and fascia. Small but significant heat allodynia was elicited at 24 h after the incision which was consistent with previous findings [34]. However, none of the ω-conotoxins changed thermal PWTs compared to vehicle control after surgery. This lack of activity was observed in previous studies, where intrathecal administrations of MVIIA and CVID were found to be ineffective to reduce thermal allodynia in rodent model of neuropathic pain [21]. These findings were not surprising, considering the differences in the physiological properties of the nociceptors contributing to heat or mechanical hyperalgesia after incision, [35]. Heat hyperalgesia depends mainly on the activation of spinal N-methyl-D-aspartate (NMDA) receptors, transfer of protein kinase C, and production of cGMP and nitric oxide. In contrast, mechanical hyperalgesia is elicited by co-activation of metabotropic glutamate receptors and spinal α-amino-3hydroxy-5-methy-4-isoxazole propionate (AMPA), activation of phospholipase A2 (PLA2), and production of cyclooxygenase products [36][37][38][39][40].
In our study, the intraplantar application of GVIA and MVIIA significantly alleviated mechanical pain in OIPN model but was ineffective in CisIPN model suggesting differential mechanisms leading to development of OIPN and CisIPN. Although both oxaliplatin and cisplatin are platinum based antineoplastic drugs, and both act by DNA alkylation [41][42][43], they differ markedly in their ability to elicit pain syndromes [44]. Oxaliplatin causes both mechanical and cold allodynia, while cisplatin only evokes mechanical allodynia [45][46][47]. The most distinctive feature of oxaliplatin is its fast, nonenzymatic transformation to oxalate and dichloro 1,2-diaminocyclohexyl-platinum complex, which do not contribute to the cold but mechanical allodynia [48]. Moreover, intraplantar administration of oxaliplatin and cisplatin causes the over-expression of different sets of genes that could be involved in development of neuropathic pain following oxaliplatin and cisplatin administration in mouse dorsal root ganglia [44]. However, the exact mechanisms of CisIPN and OIPN are still unknown and differential characteristics of cisplatin-, and oxaliplatin-induced pain indicates different mechanisms might be involved [45].
The time course of action of subcutaneously administered ω-conotoxins has previously been shown to differ from intrathecally administered ω-conotoxins in spinal nerve ligation (SNL) model of neuropathic pain [21]. In our study, the peak response after intraplantar administration of GVIA and MVIIA (300 nM) was achieved within 10 min and it returned to pre-injection levels in less than 2 h. However, in the SNL model, the peak response was achieved 2 h after the intrathecal administration of GVIA and MVIIA, and residual effect was present at 24 to 48 h, especially at higher dose (1 µg/kg) [21]. This difference is probably due to the different pharmacokinetics followed by the ω-conotoxins after intraplantar and intrathecal administration. However, the contralateral administration of GVIA, MVIIA, and CVIF (300 nM; i.pl.) did not reverse mechanical allodynia in our OIPN model, suggesting that subcutaneously administered ω-conotoxins lack systemic analgesic effects within 10 min in the OIPN model at the doses administered. Supporting a lack of systemic effects, intraplantar administration of 300 nM GVIA, MVIIA or CVIF did not elicit ataxia or locomotion-related side effects e.g., paw slips, and distance travelled determined at 1 h after administration in mice. Previously, subcutaneous administration of ω-conotoxin CVID, CVIE and CVIF near the scruff of the neck did not cause side effects, although it partially reversed the incapacitance of mouse inflammatory pain [27,45,46,[49][50][51].
Although we have not investigated the site of action in this study, considering the high selectivity to CaV2.2 channels, it is likely that the peripheral analgesic activity of ωconotoxins in PSP and CIPN is due to the inhibition of CaV2.2 channels in peripheral nociceptors. Earlier studies have shown the involvement of peripheral CaV channels in pain signaling. Previously, Sann et al. demonstrated that, in Xenopus laevis embryos, CaV2.2 channels are expressed in sensory axons innervating the skin [50] and recent studies have shown that in mice, CaV channels are present along peripheral nerve axons, and Ca2+ currents in peripheral nerves are mediated mainly by CaV2.2 and CaV1.1 channels [51]. Moreover, it has been reported that increased local concentrations of Ca 2+ are involved in nerve injury-induced hyperalgesia [52,53]. This hypothesis is consistent with the study of Cousins et al. where nerve injury-induced hyperalgesia in mice was attenuated by subcutaneous application of MVIIA, while the selective CaV1.1 blocker nifedipine or the selective CaV2.1 blocker MVIIC were ineffective [33].
In the current study, oxaliplatin-induced mechanical pain was effectively reduced by both GVIA and MVIIA (i.pl.), but surgery-induced mechanical pain was alleviated only by MVIIA (i.pl.). Interestingly, CVIF (i.pl.) was not found effective in any of the peripheral pain models. This contrasting action of -conotoxins is particularly intriguing as all the ωconotoxins blocked CaV2.2 responses from SH-SY5Y cells with similar potency in FLIPR assay (IC50 for GVIA, MVIIA, and CVIF inhibition was 6.4 ± 2.2 nM, 9.1 ± 3.1 nM, and 8.6 ± 2.9 nM, respectively). In a previous study, different concentration-dependent inhibition of CaV2.2 response from different ω-conotoxins was reported by Adams et al., where CVID but not MVIIA, inhibited a distinctive N-type calcium channel to stop nerve-evoked transmitter release from preganglionic cholinergic nerves innervating the rat submandibular ganglia [54].
Although the exact reason behind the difference in CaV2.2 channel inhibition by ωconotoxins is unclear, there are several possibilities. Previously, Mould et al. reported that the presence of auxiliary CaV subunit, α2δ, reduce the affinity of the N-type channels for the ω-conotoxins, varying from 150-680-fold, where the effect was more prominent for CVID compared with MVIIA [55]. Several studies have shown that the α2δ subunit is critically involved in neuropathic pain as the expression of α2δ1 subunit increase significantly in dorsal root ganglia (DRG) in neuropathic pain models, [56][57][58][59]. In addition, the α2δ subunit is highly glycosylated [60], and the glycosylated α2 domain may cover the binding site of ω-conotoxin near the channel by creating electrostatic shielding [61]. Even though the effect of glycosylation in ω-conotoxin binding has not been thoroughly studied, differential glycosylation might contribute to the differences observed. Another possible reason might be linked to the presence of a different splice variant of CaV2.2 in peripheral nerve axons near skin, given single residues in domain III extracellular loop of CaV2.2 can affect ω-conotoxin access [62]. GVIA and MVIIA block resting, open, and inactivated channels [63], whereas CVIF has a higher affinity for channels in the inactivated state [23]. Thus, state dependence of channel block might influence the effectiveness of ω-conotoxins in different pain models.
The finding that local (subcutaneous) application of ω-conotoxins have analgesic potential in mouse models of PSP and OIPN without locomotor side effects, indicates, this class of analgesic has clinical potential for the treatment of postsurgical pain and OIPN. Peptide based drugs are increasingly gaining acceptance as clinical leads due to their high specificity and potency, although often lower bioavailability, metabolic instability and membrane permeability can limit their use [64]. In our study we found that onset of analgesic action of GVIA and MVIIA on OIPN was rapid (< 10 min) although the duration of action at the dose investigated was also short (60 min). However, to prolong the analgesic activity, controlled drug delivery systems like microspheres, hydrogels, liposomes, or solid-lipid nanoparticles can be implemented [65]. In addition, further studies are required to understand the mechanism behind differential effects of GVIA, MVIIA and CVIF in the skin in different models of pain.

Chemicals
Cisplatin and oxaliplatin were obtained from Sigma Aldrich (Castle Hill, New South Wales, Australia). The ω-conotoxins GVIA, MVIIA, and CVIF were obtained from Alomone Labs (Jerusalem, Israel). All reagents were from Sigma Aldrich unless otherwise stated.

Animals
All behavioral experiments (CisIPN, OIPN, and PSP models) were performed using adult male C57BL/6J mice (age 8-10 weeks). The mice were housed in groups of 3 or 4 per cage under 12-h light-dark cycles with access to water and standard rodent chow ad libitum. Ethical approvals for in vivo experiments in animals were obtained from the University of Queensland animal ethics committee. The animal ethics approval numbers for postsurgical pain model and cisplatin-and oxaliplatin-induced neuropathy model were PHARM/170/16 and PHARM/PACE/368/16. The experiments were conducted in accordance with the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes, eighth edition (2013) and the International Association for the Study of Pain Guidelines for the Use of Animals in Research. Animals were kept in the experimental room for at least 30 min before behavioral testing was performed. All measurements were performed at room temperature (ambient temperature of 21-23 °C) by a blinded observer kept unaware of the treatment each animal received. Sample sizes of each experiment are detailed in the figure legends of the corresponding figure. Statistical justification for n = 6 per group is as follows: 80% power to detect a 30% difference relative to baseline readings with a 25% interindividual variability between animals with a statistical significance criterion of p < 0.05.

Fluorescence-Imaging Assays
The increase of intracellular calcium induced by the CaV2.2 channel activation was measured by using calcium-sensitive Calcium 4 no-wash kit (Molecular Devices, Sunnyvale, CA, USA). Fluorescence signals were detected using a fourth generation Fluorescence Imaging Plate Reader, FLIPR TETRA (Molecular Devices, Sunnyvale, CA, USA). Cells were seeded on 384-well black-walled clear flat-bottom imaging plates (Corning, Lowell, MA, USA) at a density of 50,000 cells/well, 24 h before the assay. The Calcium 4 dye was diluted in physiological salt solution (PSS; 140 mM NaCl, 11.5 mM glucose, 5.9 mM KCl, 1.4 mM MgCl2, 1.2 mM NaH2PO4, 5 mM NaHCO3, 1.8 mM CaCl2, 10 mM HEPES, pH 7.4) with 0.1% bovine serum albumin (BSA), with 10 µM nifedipine added to block the native L-type CaV response in SH-SY5Y cells. The media was removed, and the cells were loaded with 20 µL dye per well and incubated for 30 min at 37 °C in a 5% humidified CO2 incubator. Excitation and emission wavelength of FLIPR TETRA were set at 470-495 and 515-575 nM, respectively, and camera gain and intensity adjusted for each plate to yield a minimum of 1500-2000 arbitrary fluorescence units (AFU) baseline fluorescence. A two-addition protocol was used where 5 baseline fluorescence readings were taken prior to first addition of 10 µL ω-conotoxins (1-500 nM) followed by fluorescent readings for 10 min. In the second addition, 10 µL stimulation buffer containing 90 mM KCl and 5 mM CaCl2 was added, followed by a further 10 min of fluorescent readings. Although the addition of 90 mM KCl and 5 mM CaCl2 leads to changes in osmolarity, it does not significantly affect Ca 2+ transients, as illustrated by the lack of effect of addition of 90 mM NaCl and 5 mM CaCl2, which affects osmolarity to the same extent (Supplementary Figure S1). To normalize the data, the response from positive control (90 mM KCl and 5 mM CaCl2) was defined as 100% and the response from negative control (PSS buffer) was defined as 0%.

Cisplatin-and Oxaliplatin-Induced Neuropathy
Previously established models of cisplatin-and oxaliplatin-induced neuropathy [66] were used to assess the effect of GIVA, MVIIA, and CVID on mechanical allodynia. Briefly, a single dose of 40 µL cisplatin or oxaliplatin was administered by shallow intraplantar injection (i.pl.) to the right hind paw under light isoflurane anaesthesia to induce peripheral neuropathy in mice, as described previously [66].

Postsurgical Pain
Surgery of mice was performed following a previously described method [34]. Mice were anaesthetized by inhalation with 3% isoflurane for induction, followed by 1-1.5% isoflurane for maintained anaesthesia via a nose cone. Seventy percent ethanol was used to disinfect the plantar (glabrous) part of the right hind paw before a longitudinal incision of 7 mm was made on the plantar surface of the hind paw through the skin and fascia using a number 11 surgical scalpel. The incision started from about 3 mm proximal point of the heel and extended toward the toes. Keeping the origin and insertion intact, the plantaris muscle was incised longitudinally and elevated with sterile forceps to mimic muscle retraction. Gentle pressure with sterile gauze was applied to the wound to stop bleeding. The incision was closed with two sterile sutures after hemostasis, using the simple interrupted suture technique, and Betadine (5% povidone-iodine solution) was spread over the closed wound. After stopping inhalation anaesthesia, the mice were observed for abnormal behaviours, including signs of spontaneous pain and paw guarding for 10 min. Finally, mice were moved back to their cages for recovery prior to behavioural assessments.

Mechanical Paw Withdrawal Threshold Measurements
Mechanical paw withdrawal thresholds (PWT) were assessed using electronic von Frey apparatus (MouseMet Electronic von Frey, Topcat Metrology Ltd., Little Downham, United Kingdom) as described previously [62,[67][68][69]. Briefly, animals were habituated in individual mouse runs for at least 30 min before the measurements. Using a soft-tipped von Frey filament, the pressure against the foot pad of the right hind paw was slowly increased through rotation of the device handle with the force increasing at a rate of 1 g/s, and the force (g) causing paw withdrawal (lift, shake, lick) was displayed by the apparatus. For the PSP model, the von Frey probe tip was placed adjacent to the wound, against the glabrous plantar surface. One biological replicate was determined by averaging three repeated measurements at 5 min intervals for each mouse. Three repeats were consistent and there was no sign of sensitization or desensitization. Mechanical allodynia was assessed 24 h after injection of cisplatin and oxaliplatin, and 24 h after surgery. To normalize the data for the dose response curves of ω-conotoxins in OIPN, the response from vehicle control was defined as 0% and the response from healthy control was defined as 100%. No constraints were added to define the bottom of the curves, but the top of the curve was set to "must be less than 100" considering ω-conotoxins typically do not show analgesic effects in addition to their antinociceptive actions.

Thermal Paw Withdrawal Threshold Measurements
Thermal PWT was assessed using the MouseMet Thermal apparatus (Topcat Metrology Ltd., Little Downham, United Kingdom) as described previously [70]. Briefly, at least 30 min prior testing, mice were placed in individual mouse runs to acclimatize to the room conditions. The plantar side of the incised right hind paw was touched by the tip of the probe which was preheated to 37 °C. The heat was automatically increased at a rate of 2.5 °C/s, and the temperature limit was set to 55 °C. One biological replicate was determined by averaging three repeated measurements at 5 min intervals for each mouse, and heat thresholds were measured 24 h after surgery. Three repeats were consistent and there was no sign of sensitization or desensitization.

Locomotor Performance Assessment
Motor coordination and locomotor effects of intraplantar administration of ωconotoxins were measured using a Parallel Rod Floor apparatus (Stoelting Co, Wood Dale, IL, USA) by assessing foot slips and distance covered as described previously [68]. The number of foot slips and the distance travelled (m) were assessed for 2 min at 1 h after the intraplantar injection of ω-conotoxins. The ataxia index was calculated by dividing the number of foot slips by the distance travelled (m), and the ANY-Maze software (Stoelting Co) was used to records the foot slips and the distance travelled.

Treatments
To avoid spontaneous hydrolysis caused by Cl − in physiological solutions, cisplatin and oxaliplatin were prepared on the day of the experiment in 5% glucose/H2O solution. On the day of the experiment, ω-conotoxin GVIA, MVIIA, and CVIF were prepared in phosphate-buffered saline with 0.1% bovine serum albumin (BSA) to keep them free from adsorption to plastics. Forty microliters of 1 µg/µL cisplatin and oxaliplatin were administered by intraplantar injection into the right hind paw through a sterile 30 G needle, as described previously [68]. To investigate local effects, the ω-conotoxins were administered ipsilateral in a volume of 20 µL/paw using a sterile 30 G needle, 24 h after intraplantar injection of cisplatin and oxaliplatin, and 24 h after surgery. To investigate systemic effects, GVIA, MVIIA, and CVIF were administered by intraplantar injection, contralateral to oxaliplatin administration. Mechanical and thermal allodynia were measured 10 min following the intraplantar administrations of ω-conotoxins. Locomotor side effects are at least partially controlled by CNS [70], and previous studies have already proved that intrathecally administered ω-conotoxins can show analgesic effects for more than 1 h [35], therefore, the locomotor side effects were measured 1 h after the administration of ω-conotoxins.

Data and Statistical Analysis
Data analysis was performed using GraphPad Prism v8.0 (GraphPad Software Inc., San Diego, CA, USA). Sigmoidal curves for the calculation of half-maximal inhibitory concentration (IC50) values were fitted to individual data points for concentrationresponses by a four-parameter logistic Hill equation. The response over baseline from calcium influx fluorescence in FLIPR TETRA was calculated using Screen Works 3.2.0.14. Statistical significance was determined by one-way ANOVA with Dunnett's post-test, or two-way ANOVA with Sidak's post-test as appropriate, and p < 0.05 was considered significant, unless stated otherwise. Data are presented as the mean ± SEM.