|Home | About | Journals | Submit | Contact Us | Français|
Understanding information flow in sensory pathways requires cell-selective approaches to manipulate the activity of defined neurones. Primary afferent nociceptors, which detect painful stimuli, are enriched in specific voltage-gated sodium channel (VGSC) subtypes. Toxins derived from venomous animals can be used to dissect the contributions of particular ion currents to cell physiology. Here we have used a transgenic approach to target a membrane-tethered isoform of the conotoxin MrVIa (t-MrVIa) only to nociceptive neurones in mice. T-MrVIa transgenic mice show a 44 ± 7% reduction of tetrodotoxin-resistant (TTX-R) VGSC current densities. This inhibition is permanent, reversible and does not result in functional upregulation of TTX-sensitive (TTX-S) VGSCs, voltage-gated calcium channels (VGCCs) or transient receptor potential (TRP) channels present in nociceptive neurones. As a consequence of the reduction of TTX-R VGSC currents, t-MrVIa transgenic mice display decreased inflammatory mechanical hypersensitivity, cold pain insensitivity and reduced firing of cutaneous C-fibres sensitive to noxious cold temperatures. These data validate the use of genetically encoded t-toxins as a powerful tool to manipulate VGSCs in specific cell types within the mammalian nervous system. This novel genetic methodology can be used for circuit mapping and has the key advantage that it enables the dissection of the contribution of specific ionic currents to neuronal function and to behaviour.
Neuronal communication relies on action potentials (APs) generated by the activity of voltage-gated sodium channels (VGSCs) following membrane depolarisation. The alkaloid toxin tetrodotoxin (TTX) has been exploited for more than 40 years due to its unique ability to block VGSCs and therefore to assess the contribution of these channels to cell excitability and AP propagation. Nociceptive sensory neurones (nociceptors) detect noxious peripheral stimuli; this information is then transmitted to the superficial dorsal horn of the spinal cord, relayed to the brain, and perceived as pain (Lewin & Moshourab, 2004). Nociceptors express two unusual VGSCs, Nav1.8 and Nav1.9, which are resistant to TTX (Dib-Hajj et al. 1998; Akopian et al. 1999). Nav1.8 generates sodium currents with a high activation threshold (−40 mV) and slow inactivation (Sangameswaran et al. 1997; Akopian et al. 1999; Renganathan et al. 2002), whereas Nav1.9 produces a persistent current with a more hyperpolarised voltage dependence and ultraslow recovery from inactivation (Baker, 2005; Cummins et al. 2007). In addition nociceptors are enriched in the Nav1.7 TTX-S VGSC subtype (Nassar et al. 2004) which produces the threshold currents (Matsutomi et al. 2006). Small molecules that specifically block the function of these VGSC subtypes include chemical tools (Jarvis et al. 2007), small interfering RNAs (Dong et al. 2007), and venom-derived toxins (Terlau & Olivera, 2004). The μO-conotoxins MrVIa and MrVIb were the first group of peptide toxins reported to inhibit VGSC currents in mammalian dorsal root ganglia (DRG) neurones (Daly et al. 2004). MrVIa was found to inhibit TTX-R VGSC currents with an IC50 value of ~80 nm and a ~10 times higher IC50 value (~1 μm) for TTX-S sodium currents (Daly et al. 2004; see Supplemental material associated with the current paper, Table 1, available online only).
In this study we have used the tethered toxin approach (Ibañez-Tallon et al. 2004; Holford et al. 2009) combined with cell-specific transgenesis, to deliver a genetically encoded tethered form of the neurotoxin MrVIa to nociceptors in mice. We show that this approach can be successfully used to manipulate VGSC currents in a cell-autonomous manner. Furthermore, the nociceptor-specific inhibition of VGSC currents in these transgenic mice leads to specific changes in the firing of noxious cold-sensitive nociceptors and reduction in inflammation-induced pain behaviour.
Mice were housed in the animal facility of the Max-Delbrück Center with ad libitum access to food and water in an air-conditioned room at 22–23°C with a standard 12 h light/dark cycle. Mice were killed by placement in a CO2-filled chamber for a 2–4 min followed by cervical dislocation. All procedures conformed to the German guidelines of animal experimentation laid down by the government. Animal housing and care, as well as protocols for killing mice, are registered with and approved by the appropriate German federal authorities (Landesamt für Gesundheit und Soziales), which also governed proper implementation.
The mouse BAC clone (RP23-214H2) encompassing the Scn10a gene (Invitrogen) was modified to include the t-MrVIa expression cassette using the two-step recombination system (Gong et al. 2002). After electroporation and homologous recombination, resolved clones were screened by PCR and Southern blotting. The modified BAC for injection was purified by CsCl, linearised and 3 ng μl−1 of BAC DNA was injected into fertilised oocytes of hybrid BDF mice.
Genomic DNA (5 μg) from each transgenic founder line was digested with NcoI and separated by gel electrophoresis. After gel denaturation (0.5 m NaOH, 1.5 m NaCl) and equilibration (10 × saline–sodium citrate buffer (SSC)), membrane blotting (Hybond N+ from Amersham) was performed in 10 × SSC. The membrane was quickly washed in 2 × SSC, dried for 2 h and crosslinked with UV light (Hoefer). DNA probes were radioactively labelled with α-32P-dCTP (NEN) using the Prime-It RmT Random Primer Labeling kit (Stratagene). Twenty-five nanograms of the probe was purified via Probe-Quant G50 micro-columns (Amersham) and denatured at 95°C for 5 min before hybridisation. Hybridisation was performed at 65°C overnight in 5 × Denhardt's solution, 6 × SSC, 0.5% SDS and 100 μg ml−1 salmon sperm DNA. The membrane was washed in 2 × SSC/0.1% SDS followed by stringent washes in 0.1 × SSC/0.1% SDS. Autoradiograms were exposed at −70°C for 3 days.
Mouse tissues were rapidly dissected, collected in Trizol and homogenised. Total RNA was extracted by phenol/chloroform and treated with DNaseI (Invitrogen). RT-PCRs were performed using one-step platinum RT-PCR (Invitrogen). Primer pairs used were: t-MrVIa-F: 5′-TGGGAGTACTGCATAGTGCCGAT-3′, t-MrVIa-R: 5′-CCCGTAGTTCCATCCTTCCTTCAA-3′; Nav1.8-F: 5′-TAACGTGTGGGTCTCTGTGC-3′, Nav1.8-R: 5′-AGGTATGGAGCCAGGTCCTC-3′; β-actin-F: 5′-TCGTGCGTGACATCAAAGAGAAGC-3′, β-actin-R: 5′-ATGGATGCCACAGGATTCCATACC-3′.
RNA probes were synthesised with T7 RNA polymerase using DIG labelling (Roche) and purified with the RNeasy kit (Qiagen). Dorsal root ganglia (DRG) from neonatal mice were collected in ice-cold PBS and frozen in OCT (Sakura). Cryosections were cut with 12 μm thickness. In situ hybridisation was performed with DIG-labelled probes as described (Muller et al. 2005).
Whole-cell patch-clamp recordings were performed on DRG cultures from P7–P15 mice as previously described (Hu & Lewin, 2006) and 100 ng ml−1 nerve growth factor was added to the culture medium. Recordings were done 4–24 h after plating, using borosilicate capillaries with a resistance of 5–8 MΩ. For VGSC recordings, cells were perfused with extracellular solution containing (in mm): 140 NaCl, 4 KCl, 1 MgCl2, 2 CaCl2, 4 glucose, 10 Hepes (pH 7.4) and pipettes were filled with solution containing (in mm): 122 KCl, 10 NaCl, 1 MgCl2, 1 EGTA, 10 Hepes (pH 7.3). PI-PLC (Prozyme) was applied with 10 μg ml−1 for 1 h at 37°C. TTX experiments were performed with 0.5 μm TTX. Membrane potentials were kept at −60 mV, prepulsed to −120 mV for 150 ms and depolarised from −50 to +40 mV for 50 ms in increments of 10 mV. For current clamp analyses, neurones were held at resting membrane potential and current injection was increased from 80 pA in 80 pA increments. For VGCC recordings, the bath solution was replaced with (mm): 140 TEA-Cl, 2 CaCl2, 10 Hepes, 10 glucose and 100 nm TTX, pH 7.4 with TEA-OH. The pipette solution contained (mm): 120 CsCl, 4 MgCl2, 10 Hepes, 10 EGTA, pH 7.3 with CsOH. Currents were activated by 100 ms depolarisation pulses from a holding potential of −80 to 0 mV with 10 s intervals. Recordings were performed using an EPC10 amplifier (HEKA) at a sampling rate of 50 kHz. The series resistance was compensated for 40–60%. Data acquisition and analysis were performed with Patchmaster and Fitmaster software (HEKA Electronics, Darmstadt, Germany).
Single afferent units of the saphenous nerve from adult mice were identified using a mechanical stimulus (Milenkovic et al. 2008). C-fibres were subsequently examined for responsiveness to a cold ramp in which the temperature decreased from 30 to 0°C. Experiments were carried out blind, i.e. the genotypes of the mice were not known to the experimenter. Cold stimuli were delivered with a computer-controlled Peltier device in contact with the skin receptive field.
DRG neurones were isolated from an adult mouse and cultured as previously described (Hu & Lewin, 2006). Two hundred micrograms per millilitre of biotinylated isolectin-B4 (IB4; Vector Labs) were used for IB4 live-stainings. DRGs were electroporated (Amaxa) with 5 μg of the pCS2+ eukaryotic expression plasmid containing t-MrVIA (CS2-t-MrVIA) and 0.5 μg of pCS2+ plasmid encoding the green fluorescent protein (CS2-EGFP) and plated on poly-l-lysine–laminin-coated coverslips. One day later, cells were fixed with 4% paraformaldehyde (PFA) and washed in PBS. Three per cent goat serum was used for blocking. The following antibodies and markers were used: mouse anti-FLAG (Sigma) at 1:1000, wheat germ agglutinin (Molecular Probes) at 1:200 (5 μg ml−1), goat anti-mouse-Alexa 633 (Molecular Probes) at 1:400. Pictures were taken on a Zeiss LSM confocal microscope.
HEK293-T cells were transfected with the CS2-t-MrVIa expression vector using lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions. Immunoprecipitation with FLAG agarose beads (Sigma) and Western blot assays were performed as described (Ibañez-Tallon et al. 2002).
Mice (n= 1 per genotype) were anaesthetized by intraperitoneal injection of 500 μl of 5 mg Ketavet/200 μg Rompun in saline. Reflex responses such as limb withdrawal to paw pressure were used to indicate the need for a further dose of anaesthetic. Mice were then perfused with 4% PFA and cryosections were cut with 14 μm thickness. Mouse sections were blocked with 3% goat serum in PBS. The following antibodies and marker were used for immunostaining: rabbit anti-subP (Zymed Laboratories) at 1:1000, isolectin-B4–Alexa 488 (Vector Labs) at 1:500 and horse anti-rabbit–Alexa 594 (Molecular Probes) at 1:1000.
Standard fura-2-based ratiometric calcium imaging was used to record responses of DRG neurones cultivated from either wild-type or Tg-t-MrVIa mice (2–5 weeks of age) to application of different TRP channel agonists: capsaicin (TRPV1, 100 nm, 30 s), menthol (TRPM8, 100 mm, 30 s) and cinnamaldehyde (TRPA1, 100 mm, 30 s). KCl (40 mm, 10 s) was applied at the end of experiments and only cells exhibiting a robust response were analysed. An inverted microscope (Zeiss Observer A1) equipped with a MetaFluor photonics imaging system, including Polychromator V, and a CoolSNAP ES camera (Visitron) was used for cell imaging. Mouse genotype was unknown to the experimenter.
Xenopus laevis female frogs (n= 5) were anaesthetized by immersion for 30 min in a 0.35% tricaine (Roth, Karlsruhe, Germany) solution until the animals were completely immobile and no further reflex limb withdrawal was detected. Frogs were rinsed in water before surgery and kept on ice during the surgery procedure. A relatively small piece of ovary was taken after performing a 1–2 cm-long abdominal incision through the overlying skin and a second incision through the muscles. The muscular incision was closed with DS16-5USP re-absorbable sterile surgical thread (Catgut GmbH). The skin incision was closed with DS16-5USP non-absorbable sterile surgical thread (Catgut GmbH) and removed at the next surgery. After surgery frogs were kept for one day of observation in the laboratory before being returned to the main aquarium. The preparation of Xenopus laevis oocytes, cRNA transcripts and two-electrode voltage clamp recordings were conducted as described previously (Ibañez-Tallon et al. 2002). VGSC subunit cRNAs alone or with tethered toxins (ratio channel:toxin was 1:4) were injected into oocytes and recorded 3–5 days later.
Mice (n= 2 per genotype) were anaesthetized by intra-peritoneal injection of 500 μl of 5 mg Ketavet/200 μg Rompun in saline. Reflex responses such as limb withdrawal to paw pressure were used to indicate the need for a further dose of anaesthetic. Saphenous nerves were isolated from mice which had been perfused with 4% paraformaldehyde and postfixed in 4% paraformaldehyde/2.5% glutaraldehyde in PBS. Following treatment with 1% OsO4 in 0.1 m phosphate buffer for 2 h for osmication, tissue samples were dehydrated in a graded ethanol series and propylene oxide and embedded in Poly/BedR 812 (Polysciences, Inc., Eppelheim, Germany). Semi-thin sections (1 μm) were stained with toluidine blue. Ultra-thin sections (70 nm) were contrasted with uranyl acetate and lead citrate. The numbers of non-myelinated and myelinated axons in the saphenous nerve of wild-type and transgenic mice were determined by the counting of 12 random fields under the electron microscope and normalising to total cross-sectional nerve area observed in semi-thin sections.
Mice were backcrossed to C57/BL6/J for at least five generations. Motor coordination: motor coordination was assessed by placing the mice on an accelerating rotarod apparatus (Ugo Basile, Italy). Animals were scored for time to fall. Thermal nociception: thermal latency was determined using the Plantar test (Ugo Basile) (Hargreaves et al. 1988). Fifteen minutes after placing the mice in the Plexiglas chamber a high intensity light beam was directed at the plantar surface of the hindpaw. Movement of the paw stopped the light beam and the latency was indicated on a digital screen. Measurements were made five times on each paw and mean values were used for statistical analysis. Mechanical nociception: mechanical sensitivity was determined using a dynamic plantar aesthesiometer (Ugo Basile, Milan, Italy). All animals were acclimatised to the behavioural testing apparatus for 15 min. An increasing mechanical force (1 g s−1) was applied to the plantar surface of the mouse hindpaw. Each paw was measured five times and mean values of the applied forces were determined and used for statistical analysis. Cold nociception: mice were placed onto a Peltier-cooled plate maintained at 0 ± 1°C. The number of lickings and jumpings were determined during 90 s. Preferential heat test: mice had free access to two plates kept at 10°C and 30°C. The time spent on each plate was measured. Inflammatory pain: mice (n= 10 per genotype) were anaesthetized by isoflurane inhalation. Thirty microlitres of 2% carrageenan in saline (Fluka) was injected into the plantar surface of the mouse right hindpaw to induce inflammatory hyperalgesia. On the following 3 days mice were tested for responses to heat and mechanical stimuli as described for the acute pain studies. Mice were killed by placing in a box filled with 100% CO2 at the termination of the tests.
Results are presented as means ±s.e.m. Unpaired t tests or two-way ANOVAs were used when indicated. P < 0.05 was considered statistically significant.
To generate a membrane-tethered isoform of MrVIa (t-MrVIa) we inserted a nucleotide sequence coding for the mature toxin between the sequences coding for a secretion signal and a glycophosphatidylinositol (GPI) anchor signal (Ibañez-Tallon et al. 2004) (Fig. 1A). Tethered toxin expression in nociceptors was achieved by (BAC) transgenic methodology (Gong et al. 2002) using genomic DNA that contains the Scn10a nociceptor-specific gene encoding for the Nav1.8 VGSC α-subunit. Figure 1B shows the diagram of the BAC containing the Scn10a gene and the homology boxes for recombination with the modified shuttle vector containing the t-toxin cassette. Four transgenic founder lines were identified by Southern blotting (Fig. 1C). Comparison of the endogenous Scn10a allele to the transgene band indicated that transgenic lines 2 and 27 had the highest BAC transgene copy number. We encountered breeding problems with transgenic line 27 in that very few transgenic mice were born, which may have been a non-specific effect due to the size or the site of transgene insertion. We thus used line 2 (Tg(Scn10a-t-MrVIa)2Iit referred to hereafter as Tg-t-MrVIa mice) for all further experiments.
Tg-t-MrVIa mice express t-MrVIa and Scn10a in DRGs, but not in other tissues along the nociceptive pathway whereas wild-type littermates showed comparable levels of Scn10a transcripts in DRGs, but lacked t-MrVIa expression (Fig. 2A). As expected from the developmental pattern of Scn10a (Benn et al. 2001), in situ hybridisation indicated that t-MrVIa transcripts were present during embryonic development (Fig. 2B). In adult mice, t-MrVIa transcripts were detected in small-diameter nociceptors, but not in large-diameter sensory neurones (Fig. 2C).
An important feature of membrane t-toxins is their cell-autonomous action. To assess, whether this property was maintained in vivo, we prepared DRG primary cultures, which contain a mixture of sensory neurone types e.g. nociceptors and mechanoreceptors. Nociceptors were discriminated from mechanoreceptors by measuring AP configuration using the whole-cell patch clamp technique (Fig. 2L; AP half-peak width >3.1 ms for nociceptors and <0.8 ms for mechanoreceptors; see online Supplemental Table 2) and cell soma size (<25 μm soma diameter for nociceptors and >25 μm for mechanoreceptors; Supplemental Table 2) as previously described (Stucky & Lewin, 1999; Lawson, 2002). Representative traces of VGSC currents, evoked by step depolarisation in voltage clamp mode, in nociceptors and mechanoreceptors of both Tg-t-MrVIa mice and wild-type littermates are shown in Fig. 2D and G. Current–voltage relations indicated a significant reduction of VGSC currents in nociceptors of Tg-t-MrVIa mice compared to wild-type littermates (Fig. 2E). In contrast, VGSC currents in mechanoreceptors of Tg-t-MrVIa mice were indistinguishable from those measured in mechanoreceptors of wild-type (Wt) littermates (Fig. 2G and H). Normalisation of peak VGSC currents to cell capacitance indicated a 41.1 ± 9% reduction of peak current densities in nociceptors of Tg-t-MrVIa mice compared to controls (Wt: 413.7 ± 37.2 pA pF−1, Tg-t-MrVIa: 243.5 ± 37.2 pA pF−1, n= 32, P < 0.05). However, no difference was observed in mechanoreceptors (Wt: 306.9 ± 57.4 pA pF−1, Tg-t-MrVIa: 332.8 ± 26.8 pA pF−1, n= 9) (Fig. 2I; Supplemental Table 2). To exclude the possibility that the t-MrVIa toxin could be affecting other ion channels present in nociceptors besides VGSCs, we tested the functionality of VGCC (Fig. 2J and K) and TRP channels (Fig. 1 in online Supplemental material). Patch-clamp recordings of VGCCs performed by step depolarisations indicated no differences between nociceptors of Tg-t-MrVIa mice compared to wild-type littermates (Fig. 2J). Current–voltage relationships did not differ between nociceptors of Tg-t-MrVIa and wild-type mice with currents measured with increasing voltage steps from −50 to +40 mV, and maximal peak currents obtained at a voltage depolarisation from −50 to −10 mV (Wt: −1.11 ± 0.9 nA, Tg-t-MrVIa: −1.13 ± 0.8 nA, n= 10 each) (Fig. 2K). The functionality and incidence of the TRP channels: TRPV1, TRPM8 and TRPA1 was measured by fura-2-based imaging of increases in [Ca2+]i in DRG neurones through the sequential application of the specific activators capsaicin, menthol and cinnamaldehyde (Supplemental Fig. 1A and C). No significant differences in the percentage of DRG neurones responding to each agonist were observed between Tg-t-MrVIa mice compared to wild-type littermates (capsaicin: Wt 35.8 ± 8.1%, Tg-t-MrVIa 35.5 ± 5.1; menthol: Wt 15.3 ± 5.6%, Tg-t-MrVIa 13.8 ± 4.0; cinnamaldehyde: Wt 10.1 ± 6.6%, Tg-t-MrVIa 11.6 ± 6.2, n= 4 for both genotypes) (Supplemental Fig. 1B and D). These results show that these ion channels are not non-specifically affected by genetic expression of the t-toxin in these neurones. Other parameters, such as voltage-gated potassium outward currents (Fig. 2F) and the resting membrane potential (Wt −51.9. ± 1.95 mV, Tg-t-MrVIa −52.7 ± 1.1 mV, n= 32; Supplemental Table 2), were also unaltered in nociceptors of Tg-t-MrVIa mice compared to controls. These results indicate that t-MrVIa is cell autonomous and specific for VGSCs in vivo.
Nociceptors, electroporated with t-MrVIa and EGFP, showed FLAG immunolabelling at the cell membrane, which was abolished after treatment with phosphatidylinositol-specific phospholipase C (PI-PLC) that cleaves GPI-anchored proteins from the cell surface (Fig. 3A). Immunoprecipitation assays indicated that t-MrVIa is enriched in membrane fractions and cannot be detected in the culture medium (Fig. 3B). VGSC currents were measured before and after treatment with PI-PLC. Total peak current densities in nociceptors of Tg-t-MrVIa mice were restored to 87.5 ± 8.6% of control currents after PI-PLC (Fig. 3C). Taken together, these data demonstrate that t-MrVIa acts at the cell surface to block VGSCs.
A number of studies have reported that native MrVIa and its close relative MrVIb preferentially block TTX-R VGSCs, but at higher concentrations a subset of TTX-S channels are also affected by these toxins (Daly et al. 2004; online Supplemental Table 1). In Supplemental Table 1 we have presented an overview of the reported VGSC blocking capabilities of the native soluble forms of μO-conotoxins MrVIa and MrVIb in different experimental systems. These studies show that while the native conotoxins MrVIa/b are not highly selective for only one type of VGSC, they can discriminate between TTX-R channels present in nociceptors over TTX-S. Although the relative extent of the inhibition of VGSC current depends on the assay, and there is no direct comparison of all α-subunits within one experimental set of data, it would seem that native MrVIa/b preferentially blocks Nav1.8 followed by Nav1.7, Nav1.4 and Nav1.2. To determine whether the modified MrVIa used in this study maintains the characteristics of the native soluble MrVIa form, we performed oocyte electrophysiological assays with t-MrVIa and different VGSC α-subunits. These experiments show significant block of Nav1.2, but not of Nav1.1, Nav1.3 and Nav1.6 VGSCs by t-MrVIa (Supplemental Fig. 2) in agreement with the reported affinities of the soluble toxin (Supplemental Table 1). Thus our experimental evidence suggests that the tether does not change the specificity of the MrVIa toxin.
We next used TTX to discriminate between TTX-R and TTX-S currents present in nociceptors. TTX-R currents in nociceptors of t-MrVIa transgenic mice show a 44 ± 7% reduction compared to wild-type mice (Wt: 284.3 ± 34.1 pA pF−1, n= 30; Tg-t-MrVIa: 158.6 ± 20.4 pA pF−1, n= 31, P < 0.5). This block was reversed by PI-PLC treatment thus demonstrating that the tethered toxin was responsible for this inhibition (PI-PLC: 243.6 ± 31.5 pA pF−1, n= 12) (Fig. 3C and D). We observed a non-statistically significant inhibition (31 ± 13%; Wt: 129.6 ± 22.1 pA pF−1, Tg-t-MrVIa: 89.3 ± 16.6 pA pF−1, P= 0.23) of TTX-S currents in nociceptors with a very large sample size (n= 32 nociceptors recorded) (Fig. 3C). In addition the recovery of TTX-S currents after PI-PLC enzymatic cleavage of the tether was not statistically significant as it was for TTX-R currents. Thus the changes in TTX-S currents observed in transgenic mice might be indicative of partial (but not statistically significant) antagonistic activity of the t-MrVIa on these channels (Fig. 3C). Furthermore, we used a pulse protocol that avoids any substantial contamination by Nav1.9, which undergoes ultraslow inactivation such that clamping the membrane potential for several minutes to at least −100 mV is required to release it from inactivation (Cummins et al. 1999; Amaya et al. 2006). Given that we used a holding potential of −60 mV while measuring TTX-R currents, similar to the experiments performed with the soluble toxin (Daly et al. 2004), our results demonstrate a block of Nav1.8 TTX-R channels by t-MrVIa in vivo.
We next tested the effect of t-MrVIa on isolectin-B4 positive (IB4+ve) nociceptors, which have a higher density of TTX-R currents (Dib-Hajj et al. 1998; Stucky & Lewin, 1999; Snape et al. 2010). Voltage clamp analyses in DRG neurones live-labelled with fluorescently coupled IB4 indicated that the t-MrVIa-mediated block was 48 ± 8% in the IB4+ve neurone population (Wt: 422.7 ± 57.0 pA pF−1, n= 10, Tg-t-MrVIa: 221.8 ± 33.2 pA pF−1, n= 12) and 48 ± 9% in the IB4–ve population (Wt: 215.1 ± 33.6 pA pF−1, n= 20; Tg-t-MrVIa: 111.6 ± 19.2 pA pF−1, n= 19) (Fig. 3E; Supplemental Table 2). TTX-R peak current densities were indeed higher in IB4+ve neurones (Fig. 3E) consistent with the reported higher content of Nav1.8 in this cell population (Stucky & Lewin, 1999; Snape et al. 2010). We hypothesized that the t-toxin might have a stronger effect on IB4+ve nociceptors since the Scn10a gene promoter was used to genetically drive expression of the t-toxin. Therefore, we measured AP amplitudes in IB4+ve and IB4–ve nociceptors (Fig. 3F). Current clamp assays revealed that IB4+ve but not IB4–ve nociceptors from Tg-t-MrVIa mice had significantly reduced AP amplitudes (Δ= 14.9 mV) (Fig. 3F; Supplemental Table 2). No significant changes in the number of myelinated or unmyelinated fibres (A-fibres: Wt 711.8 ± 3.1, Tg-t-MrVIa 714.8 ± 50.7; C-fibres: Wt 2920.6 ± 197.0, Tg-t-MrVIa 2660.7 ± 173.1; n= 3 mice; online Supplemental Fig. 3), and in the distribution of IB4+ve and IB4–ve nociceptor populations (n= 2 mice, online Supplemental Fig. 4) were observed. The central terminations in the outer laminae of the dorsal horn were also unchanged (Fig. 3G). In summary, t-MrVIa selectively reduces TTX-R VGSC currents in nociceptors in the absence of detectable morphological changes.
A battery of behavioural tests was performed on Tg-t-MrVIa mice. Motor activity and baseline sensitivities to noxious thermal and mechanical stimuli, as well as thermal hyperalgesia after inflammation, were normal (Supplemental Fig. 5; Fig. 4A). In contrast, mechanical hyperalgesia after inflammation, shown as the relative threshold for paw withdrawal before and after inflammation, was significantly reduced in Tg mice with respect to wild-type mice (n= 10 mice) (Fig. 4B), as was the nocifensive response to noxious cold (number of licks and jumps during 90 s on a 0°C plate: Wt 3.3 ± 0.9, Tg-t-MrVIa 1.2 ± 0.4) (n= 10 mice) (Fig. 4C; Supplemental Movies 1 and 2). No differences in behavioural thermoreception were observed (Fig. 4D). Thus, our data demonstrate that Tg-t-MrVIa mice have strong and specific deficits in nociception. We next asked whether the behavioural deficits we observed can be explained by changes in the stimulus–response properties of identified cutaneous nociceptors. We thus used the skin–nerve preparation to measure the firing frequencies of single cold-sensitive C-fibres. Two types of cold-activated fibres were recorded when subjected to a cooling gradient. A low-threshold population that started firing at innocuous temperatures (~25°C, Fig. 4E) and a high-threshold population that started to respond only below 10°C (Fig. 4F). While the firing frequency in the low-threshold population did not differ between wild-type and Tg-t-MrVIa mice (Fig. 4E), the high-threshold population had significantly decreased firing rates during the cold ramp (Fig. 4F). Maximal differences were observed around 0°C in agreement with our observations on noxious cold-evoked behavioural responses. The firing rates of C-fibres to noxious cold were significantly different in Tg-t-MrVIa mice, and were 4.5 ± 1.2 spikes s−1 (n= 8) in wild-type and 1.9 ± 0.5 spikes s−1 (n= 4) in Tg-t-MrVIa mice when the cold ramp reached 0°C (Fig. 4F). Thus our data indicate that TTX-R VGSC currents blocked by t-MrVIa in nociceptors are required both for nociception in the cold and for detection of noxious cold per se.
This study establishes the utility of genetically encoded, cell membrane-anchored peptide toxins to dissect the contribution of VGSC conductances in a defined neuronal type in vivo, and thereby establish a causal link between cellular physiology and behaviour. We show that VGSC currents in nociceptors are cell-autonomously and selectively inactivated by t-MrVIa toxin and that this block can be reversed by enzymatic cleavage of the tether (Fig. 3). Within the targeted neuronal population, the t-toxin shows a high degree of specificity for TTX-R VGSCs and there was no evidence for compensatory changes in the activity of other ion channels present in the same neurones such as TTX-S VGSCs, VGCCs and TRP channels.
Detailed studies of Nav1.8 and Nav1.7 ion channel mutations have demonstrated the distinct roles of these channels in vivo, and compensatory mechanisms that are activated in abnormal physiological states (Akopian et al. 1999; Nassar et al. 2004; Rush et al. 2006). For example, analyses of Nav1.8 knockout mice have shown that functional inactivation of Nav1.8 is accompanied by a significant upregulation (about twofold) of TTX-S currents and a compensatory shift of the steady-state inactivation due to over-expression of Scn9a transcripts encoding the TTX-S VGSC Nav1.7 (Akopian et al. 1999; Matsutomi et al. 2006). Here we asked whether a similar compensatory mechanism operates in Tg-t-MrVIa mice. As described above, partial inhibition of TTX-R currents in Tg-t-MrVIa was observed with no evidence of compensation by TTX-S currents. Interestingly the selective, but partial, block of TTX-R currents in the Tg-t-MrVIa mice described here resulted in pain-related phenotypes such as reduced mechanical hyperalgesia and insensitivity to noxious cold, which have been previously observed in Nav1.8 null mutant mice and in in vivo models with selective inactivation of Nav1.8 (Akopian et al. 1999; Jarvis et al. 2007; Zimmermann et al. 2007). However, while cutaneous fibres of Nav1.8 null mutants show a significant reduction of menthol-sensitized cold responses (Zimmermann et al., their Fig. 3G) and reduced firing rates when stimulated by constant mechanical pressure at low temperatures (Zimmermann et al., their Fig. 3E and F) they do not show differences in cold-induced discharge activity (Zimmermann et al., their Fig. S6A). In our study we show that cutaneous fibres of Tg-t-MrVIa mice have significantly reduced firing rates after stimulation with cold ramps (from 25°C to 0°C) without the concurrent application of menthol, indicating that TTX-R VGSC currents sustain the ability of neurones to fire APs at very low temperatures. Thus the presence of TTX-R VGSC current is not only responsible for the detection of mechanical stimuli in the cold, but also for the detection of noxious cold per se. The apparent discrepancy between our study and that of Zimmermann and colleagues might be attributed to the differences in the protocols used to measure cold-induced responses (a cooling ramp from 30°C to 10°C was used by Zimmermann et al., while we used a ramp from 25°C to 0°C). Nonetheless when the cooling gradient reached 10°C we already observed reduced firing discharges in Tg-t-MrVIa mice (Fig. 4E). Since studies on Nav1.9−/− null mutants have not indicated any contribution of these channels to cold sensitivity (Amaya et al. 2006), the t-MrVIa-mediated block of TTX-R currents resulting in cold insensitivity is presumably accounted for by blockade of Nav1.8 exclusively. Studies on corneal C-fibres have shown that many cold-sensitive afferents sustain discharge to thermal stimulation even when TTX is applied to the preparation (Brock et al. 1998). It is thus quite possible that the residual TTX-insensitive current generated in the cutaneous endings of cold-sensitive fibres in the skin of transgenic mice is sufficient to preserve afferent responses to mild cooling. Environmental changes to innocuous cold temperatures are detected by the menthol TRPM8 receptor (Bautista et al. 2007). However, when temperatures become painful Nav1.8 seems to be not only essential to maintain the excitability of nociceptors (Zimmermann et al. 2007), but based on our results the activity of Nav1.8 is also required for nociceptors to detect noxious cold temperatures.
Nociceptor-specific VGSCs are important therapeutic targets and major efforts have been made to screen for specific molecules to block their function. For instance, conotoxins are already used as analgesic drugs. The MrVIa-related conotoxin MVIIa (commercialised as Prialt or ziconotide) is a specific antagonist of N-type VGCCs, and is used to treat patients with extreme neuropathic pain that is insensitive to opioids (Thompson et al. 2006). In contrast to local application of such analgesic peptides, delivered intrathecally via infusion pumps to avoid severe side effects due to diffusion of the toxin into the CNS, genetic delivery of tethered toxins to defined cell populations could become a viable alternate to prevent toxic effects and repetitive treatment. Future developments in the safety of gene transfer in humans offers considerable promise for the application of this novel tethered toxin technology to treat such extreme pain in humans. Extension of this approach to other peptide toxins together with the use of cell-specific genetic methods, may allow for specific intervention in defined neural circuits in vivo and therefore potential treatment of diseases caused by channelopathies.
Given the importance of VGSC function for AP propagation, the extensive use of TTX for functional inactivation of CNS neurones in vivo, and our demonstration that t-MrVIa can cell-autonomously inhibit VGSC function without resulting in misregulation of other ion channels or alterations in membrane potential, the extension of this methodology to other cell types and ion channels could provide a valuable new tool for the dissection of neural circuits. For instance, in a complementary study we have used other peptide toxins to block calcium influx through the VGCCs Cav2.1 and Cav2.2 (Auer et al. 2010). Thus, depending on the specificity and mode of action of the natural toxin of choice, this approach can be used for neuronal silencing by either inhibiting VGCCs and blocking neurotransmitter release (Auer et al. 2010), or, as in this study, inhibiting VGSCs and reducing cell excitability. Novel peptide toxin blockers specific for VGSC types present in the CNS (Nav1.6 and Nav1.2) have yet to be discovered, but once such tools become available our strategy could be then applied to effectively inactivate AP generation in central neurones. Thus, our study extends the range of available methodologies that can be used for genetic studies of neural circuitry (Luo et al. 2008). Furthermore, our technique represents a validated means to address the contribution of specific ion channels and receptors to neuronal function and behaviour.
We thank A. L. Goldin for VGSC oocyte expression plasmids, C. Birchmeier and F. Spagnoli for discussion and B. Erdmann and C. Hanack for technical support. This work was supported by grants from the DFG to I.I.-T. and G.R.L. within the collaborative research centre (SFB 665) and from the Alexander von Humboldt Foundation to E.St.J.S.
A.S.S. performed the majority of the experiments and generated the transgenic mice. J.H. performed electroporation experiments, skin–nerve electrophysiology and assisted with patch clamp experiments. E.St.J.S. performed calcium imaging experiments and edited the manuscript. S.F. performed in situ experiments and helped with figure preparation. J.S.-T. performed DRG electrophysiology; B.K. helped with molecular cloning and behavioural tests; S.A. helped with molecular cloning, conducted the comparative analyses of t-toxin studies and helped with table and figure preparation; I.I.-T. and G.R.L. planned and supervised experiments. I.I.-T. conceived, designed and supervised the project. I.I.-T. wrote the manuscript with help from G.R.L. and A.S.S. All authors approved the final version of the manuscript.
J. Hu: Centrum für Integrative Neurowissenschaften (CIN), University of Tübingen, Paul-Ehrlich-Str. 15-17, 72076 Tübingen, Germany.
Supplemental Movies 1
Supplemental Movies 2
Supplementary Table 1
Supplementary Table 2
Supplementary Figure 1
Supplementary Figure 2
Supplementary Figure 3
Supplementary Figure 4
Supplementary Figure 5
As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer-reviewed and may be re-organized for online delivery, but are not copy-edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors