Here we show, that ATX-II enhances persistent and resurgent currents in large diameter DRGs (Figure 2), but fails to do so in small DRGs (Figure 3). This is most likely due to a lack of sufficient β4-subunit expression in these small neurons (Figure 3d), which is thought to be essential for the generation of resurgent currents. Addition of the β4-peptide to the pipette solution enabled us to record ATX-II-inducible resurgent currents in small DRGs and HEK cells stably expressing Nav1.7. Large DRGs are linked to A-fibers, and supporting our in vitro findings, injection of small amounts of ATX-II into the skin of human volunteers evoked responses in accordance with A-fiber activation.
Large DRGs exhibit endogenous resurgent currents that can be enhanced by ATX-II. This is not the case for small DRGs, even though small DRGs are known to express ATX-II sensitive sodium channels (such as Nav1.1, Nav1.2 and Nav1.6 [16, 18, 20, 33]). Mice lacking functional Nav1.6 (med mice) display reduced resurgent currents , and there is evidence, that the interaction of the β4-subunit with sodium channels is a major precondition for resurgent currents to occur (e.g. ). Our RT-qPCR data of the FACS sorted DRGs show that large DRGs have higher levels of β4 mRNA than small neurons (Figure 3d and [21, 34]). Considering the primer efficiencies, DRGs of both sizes have ~5 times more β4 than Nav1.6 (5 and 4.5 times for small and large DRGs, respectively), which would be ample for a simple 1:1 interaction. In small DRGs Nav1.7 is expressed ~7 times stronger than Nav1.6, making it more likely for the β4-subunit to interact with Nav1.7 than with Nav1.6 in small DRGs. It may be possible that Nav1.7 has a lower affinity for β4, and therefore needs higher concentrations for a successful interaction, which could be provided by addition of β4-peptide to the recording solution (Figures 4 and 5).
Compared to Nav1.7, Nav1.6 shows resurgent currents in a larger population of cells under control condition with the peptide in the pipette (53.1% vs. 26.7%, respectively). Nevertheless it might be hard to compare channels expressed in a cell line and draw conclusions to the extent of its modification in its native environment as for example in DRGs. If the peptide was present, seven out of 17 cells transfected with Nav1.6 show resurgent currents, whereas only three out of 15 cells expressing Nav1.7 showed these currents by themselves. Although the cellular background may play a role, these data suggest that Nav1.6 is more prone to interact with the β4-peptide and to produce resurgent currents. It may be possible that the amount of endogenous β4-subunit in small DRGs is simply not sufficient for a robust interaction with any endogenous TTXs sodium channel expressed in these cells and therefore no resurgent currents can be recorded.
An increase in ATX-II concentration up to 25 nM failed to induce resurgent currents in small DRGs ( Additional file 3: Figure S3), but only enhanced persistent currents. This finding is comparable to that of the application of the wasp venom β-Pompilidotoxin to CA3 neurons , which induced tail and persistent, but no resurgent currents unless the β4-peptide was present in the recording pipette. Our data suggest that small DRGs share this characteristic with CA3 neurons.
An enhanced persistent current may very well increase excitability by its own, and indeed application of 1 μM ATX-II to IB4 negative small DRG neurons resulted in a broader AP . Interestingly, ATX-II sensitivity seems to be distinct between the different subsets of small DRG neurons, as IB4 positive small DRGs did not change their AP properties in response to ATX-II application. The toxin concentrations that we used in our study induced a shift of TTXs current activation and steady-state fast inactivation ( Additional file 2: Figure S2), but failed to have effects on persistent or resurgent currents (Figure 3). Therefore, it is likely that the toxin concentrations used in our experiments remained below threshold for an effect on persistent or resurgent currents in small DRGs, and that in small DRGs persistent and resurgent currents are less sensitive to small amounts of ATX-II than activation and steady-state fast inactivation.
Surprisingly, mRNA for Nav1.7 is expressed in large and small DRGs at comparable levels (Figure 3d, ). Patients with the inherited pain syndromes erythromelalgia and PEPD, both due to mutations in Nav1.7, describe their pain mainly as burning - a typical C-fiber associated characteristic. As Nav1.7 seems to be evenly distributed between large and small DRGs, it is likely that associated proteins, such as the β4-subunit, or cell-background specific channel modifications, may play a major role in the generation of Nav1.7 related pain. The discrepancy may also be due to species differences: We investigated DRGs from mice, but erythromelalgia and PEPD manifest themselves in humans. Knowledge about the developmental regulation of β4 expression is still lacking. PEPD mutations of Nav1.7 were shown to induce resurgent currents by slowing fast inactivation [7, 8]. We show in this study, however, that even in the presence of ATX-II no resurgent current could be evoked in small DRGs or Nav1.7 expressed in HEK293 cells unless the β4-peptide is present. Patients with PEPD are severely affected during their first year of life but symptoms become somewhat milder later in life. It remains speculative if this might be due to a down regulation of β4 with age, for which data are currently not available.
ATX-II alone was unable to induce resurgent currents in heterologously transfected Nav1.7 or Nav1.6, which was reported before . Up to now, only one toxin, the β-scorpion peptide Cn2, was able to induce resurgent currents in a heterologous expression system without the addition of an intracellular peptide in Nav1.6 . These currents were enhanced by ATX-II. This finding suggests that sodium channels need some kind of modification, like treatment with Cn2 or potentially a disease causing mutation, such as PEPD, in order to be capable to produce resurgent currents.
In the 70s and early 80s of the last century, it was observed that ATX-II broadens the AP of crustea peripheral axons, and not much later it became evident that this was due to a slowing of inactivation properties of voltage-gated sodium channels [10, 15, 39, 40]. When applied to large DRGs in high concentrations (10 μM), it significantly slowed τh of Nav current inactivation . Here, we used concentrations that are an order of magnitude lower and although we used recombinant ATX-II (purity >98%), which may be more specific in its effects than purified ATX-II, the effect on the peak current decay time constants was no longer detectable ( Additional file 2: Figure S2 b,c), but nevertheless resurgent currents were altered significantly. Also, we observed shifts in activation and steady-state fast inactivation. This suggests that even though the effects on inactivation kinetics were small, the concentration of ATX was high enough to enhance resurgent currents, which in large DRGs seem very sensitive to even small channel alterations. Higher concentrations of ATX-II evoked resurgent currents in large DRGs, which were too large to be properly recorded under our conditions. This suggests that 5 nM ATX-II is a concentration at which effects may be evoked in large, but not in small DRGs, thereby allowing a separation of the two groups.
Molecular studies discriminated between sodium channel subtypes in heterologous expression systems for their sensitivity to ATX-II, showing that mainly Nav1.1, Nav1.2 and Nav1.6 are modified at very low concentrations [16, 33, 38]. These three channels can be found in large and small DRGs [18, 20] and in large, but not small DRGs their modification by ATX-II enhances resurgent currents.
ATX-II slows inactivation kinetics for both channels investigated in this study: Nav1.6, which is supposed to be the main carrier of the AP in large DRGs and Nav1.7, which is suggested to play this role in small DRGs . The slower fast inactivation occurs the higher the probability is for a potential open channel blocker to interact with the pore. Consequently, in the presence of the β4-peptide, we were able to significantly increase resurgent currents in both channel subtypes (Figure 5). These experiments suggest that two requirements are necessary for ATX-II to increase resurgent currents: I) an ATX-II sensitive sodium channel (such as Nav1.7 or Nav1.6, but also other subtypes, [16, 33, 38]) and II) the availability of an intracellular open channel blocker. Both seem to be present in large DRGs, whereas our experiments suggest that small DRGs are missing at least one of the components. As the TTXs sodium channels that were shown to be expressed in small DRGs are almost all sensitive to ATX-II (Nav1.1, 2, 6 and 7, [16, 18, 20], Figures 5 and 6), even at the low concentrations that we used here, it is quite likely that small DRGs are lacking sufficient open channel blocker. Accordingly, when we added the β4-peptide to the intracellular recording solution, resurgent currents were present in small DRGs (Figure 4).
Recently we showed that the anticancer agent oxaliplatin selectively enhances persistent and resurgent currents in large diameter DRGs at low temperatures . Oxaliplatin is highly effective but unfortunately has dose limiting side-effects: acute and chronic neuropathies, which are both characterized by cold evoked very unpleasant dysaesthesias. These sensations are similar to those experienced by the subjects of the present study following intradermal ATX-II injection. Besides an increased cold perception 20 and 30 min after ATX-II injection a pricking sensation outlasting the cold stimulus was observed (see Figure 6f), indicating ongoing activity in the affected nerve fibers.
ATX-II did not display a marked temperature dependence in its enhancement of resurgent or persistent currents, which is in contrast to oxaliplatin (Figure 2 and ). Low temperatures induce a general increase in membrane resistance, which would then allow a small depolarizing current, like resurgent currents, to be potentially sufficient to initiate a following AP. This may explain why the subjects report cold enhanced sensations after ATX-II injection (Figure 7f); even though resurgent currents were not significantly different at 22°C compared to 30°C (Figure 2).
Oxaliplatin failed to affect C-fibers or small diameter DRGs, suggesting that this drug, too, relies on the presence of the β4-subunit for its effects, or displays some degree of subtype specificity. The presence or absence of the endogenous blocking particle might thus render certain types of neurons susceptible for activation by toxins or drugs which act via voltage-gated sodium channels. Thus, focusing only on specific sodium channel subtypes for the treatment e.g. of pain might not be reaching far enough, but the channels functions in their cellular context need to be considered.
Pain sensations in vivo may be transmitted via two main types of fibers: the myelinated fast conducting A-fibers, which are connected to large DRGs, and the slower and unmyelinated C-fibers which are linked to small diameter DRGs . We have shown that low concentrations of ATX-II activate large DRGs suggesting that A-fibers convey the ATX-II evoked sensations of the volunteers. Accordingly, mechanical nerve fiber block abolished the ATX-II evoked sensations. Most likely Aδ-fibers mediate the ATX-II induced pain symptoms but may also participate in the itch-like sensations. Recently it was shown that the effects of intradermal insertion of spicules from the pods of a cowhage plant induce intense itching. Aδ-fibers seem to be involved in transmitting this itch . In some subjects, these sensations were markedly reduced by mechanical nerve block. Intradermal injection of ATX-II, on the other hand, evoked itch-like sensations, that were not described as common itch, but unpleasant sensation, for which there is no adequate word. “Tingling” was used by some subjects, which maybe an overlapping sensation induced by ATX-II and cowhage spicules. Apart from Aδ-fibers it is likely that Aß-fibers are activated by ATX-II since sensations of mechanical character were described, such as "like punctured by many thin needles". Possibly, the uncommon simultaneous activation of Aδ- and Aβ-fibers may generate the unique ATX-II mediated perception.
In our in vivo experiments the A-fiber block may induce paraesthesias itself, both during and upon recovering from the block. However no paraesthesias were described prior to and after ATX-II injection. Nonetheless, the reported sensations after ATX-II injections and relief of block were comparable to those described when no prior block was applied. The fact that cold evoked pain, which strongly increased after removal of the block, was closely restricted to the injection site (Figure 7f) argues against a general block-induced fiber irritation. Nevertheless, an irritation of the fibers cannot be ruled out, but we estimate it to be rather small in comparison to the effects evoked by ATX-II alone.
Intradermal injection of ATX-II did not evoke a large axon reflex flare (Figure 7e), arguing against a major activation of mechano-insensitive C-fibers . Nevertheless, mechano-sensitive C-fibers may mediate itch without the induction of a major axon reflex flare [43, 44]. Subjects tested with cowhage spicules, which activate only mechano-sensitive C-fibers vigorously and Aδ-fibers, report quite different sensations than those injected with ATX-II (burning pain, severe itch, which was also clearly labeled as such). As mechano-sensitive C-fibers are also resistant to differential nerve block and should therefore have transmitted the ATX-II-mediated sensations, we assume that mechano-sensitive C-fibers are not a major target of ATX-II in humans.