Animals for RNA extraction: 4 day old C57BL/6J mice (stock number 000664, Jackson Laboratory, Bar Harbor, ME, USA) were euthanized, and their inner ears were quickly removed from the temporal bones. All experimental procedures involving animals were approved by the Johns Hopkins University Animal Care and Use Committee. Animals for oxaliplatin experiments: Wild-type mice were CBA/CaJ (Jackson Labs, Bar Harbor ME). Germline α9 knockout (KO) mice (Abazeed et al., 2013) originally on a 129Sv/Ev and CBA/CaJ background were crossed using an accelerated backcrossing program (Jackson Labs) until 99.5% identity with wild-type CBA/CaJ mice was achieved. Mice were then further backcrossed with wild-type CBA/CaJ mice an additional three generations. Experiments involving animals have been reported according to ARRIVE guidelines (Kilkenny et al., 2010). All efforts were made to minimize animal suffering and to reduce the number of animals used. All experimental procedures were in accordance with the National Institutes of Health guidelines for the care and use of laboratory animals and were performed under approved protocols at the University of Utah.

Mouse cochlear neuroepithelia were isolated in RNAlater stabilization solution (ThermoFisher Scientific, Waltham, MA, USA) and quickly frozen. Total RNA was extracted using Trizol reagent (Invitrogen, ThermoFisher Scientific). Reverse transcriptase reactions were performed using Superscript III (ThermoFisher Scientific) with random primers p(dN)₆ (Roche, Sigma-Aldrich, St Louis, MO, USA) and polymerase chain reaction (PCR) using the high fidelity TaKaRa LA Taq DNA Polymerase (TaKaRa Bio, Mountain View, CA, USA).

Primers 5′-cgACTAGTgttgggaaaggATGaaccggccccatccc-3′ and 5′-cgCTCGAGctaatctgctcttgctatgatcaagacgg-3′ were used to amplify Chrna9 coding sequence, and primers 5′-cgACTAGTccagcagggcctgttgctttacatctcc-3′ and 5′-cgCTCGAGttacagggcttgcaccagtaccaggaggc-3′, Chrna10 coding sequence. In both cases, these primers were designed to add SpeI and XhoI restriction sites 5′ and 3′ respectively of the coding sequences and to amplify part of the 5′UTR.

The 1,467 and 1,423 bp fragments amplified were isolated and purified after agarose gel separation and inserted into pGEMT vector (pGEMT Vector System, Promega). Sequences of the clones obtained were compared to the direct sequences of pooled independent PCR amplifications using cochlear neuroepithelia RT-PCR as template, to reference sequences (NM_001081104, NM_001081424), and to C57BL/6J genomic sequences using BLAT—UCSC Genome Browser (University of California—Santa Cruz, CA, USA) (GRCm38/mm10).

The sequence alignments presented in Figure 1 were obtained using Clustal Omega (Sievers et al., 2011) (MegAlign Pro, DNASTAR Lasergene, Madison, WI, USA). The following sequences were used as reference: Mus musculus (NP_001074573, NP_001074893), Rattus norvegicus (NP_075219, NP_072161), Homo sapiens (NP_060051, NP_065135) for α9 and α10 subunits, respectively.

Chrna9 and Chrna10 cDNAs were further subcloned into a modified pSGEM vector (originally developed by Dr. Michael Hollmann, Ruhr University, Bochum, Germany), using SpeI and XhoI restriction sites. This vector allows the expression of a transcript including the 5'UTR of the β-globin of Xenopus laevis, the coding sequence of the nAChR subunit preceded by its native 5'UTR, and followed by the 3'UTR of the β-globin of X. laevis. It was optimized to increase the expression (translational efficiency) of the nAChR subunit in oocytes (Filchakova and McIntosh, 2013).

Plasmid DNA were isolated from bacteria using Maxiprep Qiagen plasmid kits (Qiagen, Hilden, Germany). The sequences of the expression plasmids were confirmed by automated DNA sequencing in the Johns Hopkins University School of Medicine DNA sequencing Core facility.

X. laevis oocytes were injected with 25 ng of each cRNA encoding α9 and α10 subunits as previously described (Cartier et al., 1996) and incubated for 48 h at 17°C in ND96 prior to use. Two-electrode voltage clamp electrophysiology was conducted to assess for functional expression of α9α10 nAChRs as previously described (Azam et al., 2008). The oocyte membranes were clamped at a holding potential of −70 mV and stimulated with 1 s pulses of 1 mM ACh once every 60 s. After a steady-state response baseline was observed, the oocytes were then stimulated with ascending concentrations of ACh, choline, or nicotine. Each oocyte was stimulated first with ACh followed by choline then nicotine. The maximal response value for activation by ACh was determined using the Hill equation: Y = Bottom + (Top-Bottom)/(1+10^∧((LogEC₅₀−X)*Hill slope)). The responses to all concentrations of ACh, choline, and nicotine were then normalized to this value and calculated as a % response.

Xenopus laevis oocytes were injected with 15–30 ng of cRNA (equal amounts for each nAChR subunit) and incubated for 48 h at 17°C in ND96 prior to use. The mouse α7 subunit was co-injected with an equal amount of ric3 to increase expression. The mouse α9 and α10 subunits were made as a part of this study;all other mouse clones were kindly provided by Dr. Jerry Stitzel (University of Colorado, Boulder, CO, USA). Two-electrode voltage clamp electrophysiology was used with oocyte membranes clamped at a holding potential of –70 mV. Oocytes were stimulated with 1 s pulses of ACh every 60 s at the following concentrations: for adult and fetal muscle 10 μM was used; for mouse α7, 200 μM; and for all other subunit combinations 100 μM ACh was used. After a steady baseline of ACh pulses was achieved using ND96, the solution was switched to ND96 containing various toxin concentrations and ACh pulses were observed for toxin response. All dose response curves, including IC₅₀ and hillslope data, were calculated using non-linear regression (curve fit) sigmoidal dose response (variable slope), (GraphPad Prism, San Diego, CA, USA).

Oxaliplatin (MedChem Express, Monmouth Junction, NJ) was dissolved at 0.6 μg/μl in 0.9% NaCl. RgIA4 was dissolved at 0.01 μg/μl in 0.9% NaCl. Mice were injected i.p. daily with oxaliplatin (3.5 mg/kg) or 0.9% saline (vehicle). Additionally, mice were injected s.c. with RgIA4 (40 μg/kg) or vehicle. Treatment weeks included injection on Thurs, Fri, and Mon–Wed. On Thurs, 24 h after last injection, mice were assessed for cold allodynia. Injections stopped on day 21, with assessment occurring on day 22 and subsequent testing occurring once a week for three additional weeks. Testing was conducted using a cold plate test chamber (IITC, Inc Life Science, Woodland, CA). Animals were allowed to acclimate in the chamber at room temperature (23°C) for 5 min. Temperature was then lowered at a rate of 10°C per minute. The testing was stopped when the animal lifted both forepaws and shaking or licking occurred. Alternating lifting of forepaws was not scored. Throughout the study period, experimenters were blinded as to the identity of the injected compounds.

A competition binding assay on intact HEK293 cells was performed using the Tag-lite™ technology (Cisbio bioassays) (Zwier et al., 2010). HEK293 cells were transfected with plasmids encoding the rat GABA_B1 subunit fused to a Snap tag at the N-terminus and the rat GABA_B2 subunit and seeded in 96-well black-walled plates at a density of 100,000 cells per well. Twenty-four hours after transfection, the cells were labeled with 300 nM SNAP-Lumi4Tb (Cisbio bioassays, Codolet, France) for 1 h at 37°C in Tag-lite buffer. After extensive washing with Tag-lite buffer, the cells were incubated with the indicated α-conotoxins. Non-fluorescent ligands (GABA or CGP54626) or buffer together with 10 nM fluorescent CGP54626-Red (Cisbio bioassays, Codolet, France) were incubated for 3 h at 4°C prior to signal detection. The fluorescence was collected at 620 and 665 nm using a Pherastar plate reader (BMG Labtech, Ortenberg, Germany), 50 ms after laser excitation at 337 nm. The FRET signal was then calculated as the ratio (signal at 665 nm)/(signal at 620 nm) × 10⁴ and normalized to specific binding. Non-specific binding was determined in the presence of a high concentration of unlabeled CGP54626 (1 μM).

Bioluminescence Resonance Energy Transfer (BRET) assay was performed on HEK293 cells by recording heterotrimeric G protein dissociation following receptor activation as previously described (Gales et al., 2006). Briefly, HEK293 cells were transfected with plasmids encoding the rat receptor subunits (GABA_B1a and GABA_B2) and the G protein heterotrimer fused to BRET chromophores [Gα_i1-Rluc or Gα_oA−Rluc, Gβ2 and Gγ2-Venus (Comps-Agrar et al., 2011)]. Cells were seeded in a 96-well white plates at a density of 100,000 cells per well. Twenty-four hours after transfection, the cells were washed with phosphate buffered saline prior to Coelenterazine h substrate (5 μM) and GABA (1 mM) or α-conotoxin (1 μM) addition. The BRET signal was recorded over time using the Mithras LB 940 plate reader (Berthold Biotechnologies, Bad Wildbad, Germany) that allows the sequential integration of light signals detected with two filter settings (Rluc filter, 485 ± 20 nm and Venus/YPF filter, 530 ± 25 nm). The BRET signal is determined as the ratio between Venus and Rluc emissions.

In the developing mouse cochlea, an ACh response and cholinergic efferent activity with pharmacology consistent with activation of α9 and α10 nAChR subunits, can be detected in cochlear inner hair cells at P4 (Roux et al., 2016). We cloned the cDNAs of α9 and α10 subunits expressed in P4 C57BL/6J mouse cochlear neuroepithelium (see Materials and Methods). At the protein level, α9 isolated from mouse showed 97% identity with rat, and 91% identity with human homologs. α10 showed 98% identity with rat and 92% identity with human homologs. Sequences of the extracellular domains of mouse α9 and α10 are shown in Figure 1. Residues previously determined to be critical for α-conotoxin RgIA binding (Azam et al., 2015) are indicated by arrowheads in Figure 1B. Note the complete conservation of these residues between rat and mouse sequences.

Each cDNA insert was subcloned into the pSGEM Xenopus oocyte expression vector and used for cRNA preparation. The cRNAs were then injected into oocytes to assess whether mouse α9 and α10 subunits could assemble into functional receptors. The oocytes were subjected to two-electrode voltage clamp electrophysiology and assessed for functional responses to nicotinic agonists (Figure 2). Under these conditions, the oocytes responded robustly to 1 mM ACh as evidenced by current amplitudes that were often in excess of 10 μA. Choline also evoked currents in these oocytes and behaved as a partial agonist relative to ACh. Nic on the other hand failed to evoke currents at concentration from 100 nM to 10 mM. This pharmacological profile is similar to that reported for heterologously expressed human and rat α9-containing nAChRs (Elgoyhen et al., 1994, 2001; Verbitsky et al., 2000; Sgard et al., 2002).

α-conotoxins that block the α9α10 nAChR were previously shown to vary widely in their potency for rat vs. human nAChRs, with these peptides being orders of magnitude less potent on the human α9α10 nAChR. We therefore assessed the potency of RgIA4 on mouse α9α10 nAChRs. RgIA4 potently blocked the ACh response of the α9α10 nAChR subtype with an IC₅₀ of 1.2 nM (Figure 3). α-conotoxin [V11L;V16D]ArIB is a potent antagonist of homomeric α7 nAChRs (Whiteaker et al., 2007). Other antagonists of α7 nAChRs, including methyllycaconitine (MLA) and α-bungartoxin also potently block α9α10 nAChRs. We therefore assessed the activity of α-conotoxin [V11L;V16D]ArIB on mouse α9α10 nAChRs. In contrast to RgIA4, [V11L;V16D]ArIB (10 μM) blocked only 12 ± 8.9% of 100 μM ACh-induced current. Thus, RgIA4 is > 10,000-fold more potent than α-conotoxin [V11L;V16D]ArIB on mouse α9α10 nAChRs.

The activity of RgIA4 was further assessed on additional mouse nAChR subtypes including the fetal and adult muscle subtypes α1β1δγ, α1β1δϵ, and neuronal subtypes α2β2, α3β2, α3β4, α2β4, α4β2, α4β4, and α7 as described in Materials and Methods. For each of the α/β subunit heteromers, 10 μM peptide blocked less than 50% of current (Table 1). The IC₅₀ for α7 was 4.5 μM (95% confidence interval 3.1–6.4 μM). Block of the α9α10 nAChR by a low concentration (33 nM) RgIA4 abolished ACh-induced current. Recovery from block following peptide washout was very slow (~2% at 15 min, Figure 3). In contrast, recovery from block of α7 was rapid. After two min of washout of 5 μM RgIA4 (the approximate IC₅₀ concentration), the ACh current recovered to 101.6 ± 6.9% of pre-compound application baseline (n = 3, data not shown).

Chemotherapy induced neuropathy is a common and dose limiting side effect of cancer treatment. Pain from relatively mild cold temperatures (cold allodynia) is particularly problematic following oxaliplatin treatment. There are no FDA approved medications to prevent this complication. RgIA4 was previously shown to prevent cold allodynia in oxaliplatin treated rats and mice. Effects lasted up to 72 h post-treatment (Romero et al., 2017). An unanswered question is whether the RgIA4 prevention of cold allodynia represented temporary delay in development of the oxaliplatin-induced side effect or was indicative of more long lasting effects. Here we evaluated the effects of 3 weeks of treatment with RgIA4 followed by a further 3 weeks of observation. Four groups of mice were treated with either saline + saline, oxaliplatin + saline, saline + RgIA4 or oxaliplatin + RgIA4. This treatment regimen continued for 3 weeks after which all treatments were stopped. Mice were assessed at weekly intervals during and after treatment for the presence of cold allodynia. Experimenters were blinded to treatment conditions. Oxaliplatin produced cold allodynia, an effect that significantly differed from control beginning by week three and continuing for 2 weeks post-oxaliplatin treatment. Oxaliplatin/RgIA4 treated animal differed significantly from oxaliplatin/saline treated animals as measured at weeks 2–5. RgIA4 treated animals did not differ from saline treated controls at any tested time point (Figure 4).

RgIA4 was chosen for the present study because it potently binds α9α10 nAChRs but not multiple other tested targets including opioid receptors or GABA_B receptors (GABA_BR) (Romero et al., 2017). In addition to RgIA4, other structurally related nAChR targeted α-conotoxins are analgesic in several models of neuropathic pain. Among these α-conotoxins, the α9α10 antagonist Vc1.1 has been repeatedly shown to be analgesic. Several studies have shown that Vc1.1 blocks N-type calcium channels in dorsal root ganglion, an effect blocked by GABA_B antagonists. In addition, GABA_B antagonists prevented analgesic effects of Vc1.1, strongly implicating GABA_B receptors as contributing to the therapeutic effects (Callaghan et al., 2008; Callaghan and Adams, 2010; Klimis et al., 2011; Huynh et al., 2015; Castro et al., 2016). However, other studies have failed to confirm GABA_B agonist activity (McIntosh et al., 2009; Napier et al., 2012; Wright et al., 2015), prompting debate as to whether the analgesic actions of Vc1.1 and related peptides occur via block of α9α10 nAChRs or stimulation of GABA_BR (Adams et al., 2012; Mohammadi and Christie, 2015). Separately, AuIB was shown to be analgesic, to stimulate GABA_BRs but not to block α9α10 nAChRs (Klimis et al., 2011). To further examine the structure, function relationship of conotoxins acting on GABA_BRs we first conducted a fluorescent ligand competition binding assay using GABA_BRs expressed in HEK293 cells. Lumi4-Tb-labeled GABA_BRs were incubated with the fluorescent ligand CGP54626-Red, together with test compounds as described in Material and Methods. Test compounds were assessed for their ability to modulate the homogeneous time resolved fluorescence resonance energy transfer (HTR-FRET) signal between Lumi4-TB and the red fluorescent acceptor. Vc1.1, ImI (which shares the first 8 amino acids of Vc1.1) (Johnson et al., 1995), AuIB, and the closely related AuIA (Luo et al., 1998) were examined (see Table 2 for peptide sequences). No displacement CGP54626-Red by any of the conotoxins was observed (Figure 5A), indicating that the toxins do not bind to the orthosteric GABA_B binding site. We next assessed for agonist activity of conotoxins at G-protein-coupled GABA_BRs using a bioluminescence resonance energy transfer (BRET) assay. Agonist binding to GABA_BR activates heterotrimeric G proteins. HEK cells expressing GABA_BR coupled to BRET chomophore labeled G-proteins (Gαi₁/αo_A-Rluc, Gβ2, and Gγ2-Venus) were utilized to monitor interaction between Gα_i or Gα_o and Gγ2. The association between Gαi/o-Rluc and Gγ2-Venus at basal state produces a high FRET signal. GABA_B binding to the GABA_BR induces G-protein activation and movement of the Gγ2 subunit away from the Gαi/o subunit leading to a decrease in BRET. GABA robustly decreased BRET, consistent with agonist activity whereas the conotoxins had no effect (Figure 5B).

Conotoxins, including some of those referenced in this paper have been patented by the University of Utah with JM listed as an inventor. JM has received funding from Kineta Inc. The other authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.