The NaV1.3, NaV1.8, and NaV1.9 sodium channels are differentially regulated after nerve injury , and the TTX-resistant NaV1.8 and NaV1.9 are selectively distributed in peripheral sensory neurons ; these properties have identified them as potential molecular targets for neuropathic pain. Antisense oligodeoxynucleotide and siRNA studies suggest that sensory neuron NaV1.8 expression and spinal cord NaV1.3 expression are important for behavioral phenotypes and aberrant neurophysiology in rats with peripheral nerve injuries [4–7, 25]. Rat loss-of-function studies have not yet been performed for NaV1.9, but studies of NaV1.9-null transgenic mice suggest involvement in inflammatory rather than neuropathic pain [26, 27].
The requirement of NaV1.8 and NaV1.3 for neuropathic pain remains controversial, however, as neuropathic pain behavior develops normally in NaV1.8- and NaV1.3-null mice [20, 21, 28], as well as in mice treated with diptheria toxin to selectively destroy all postmitotic sensory neurons expressing NaV1.8 . It is possible that the discrepancy between these findings and those of rat loss-of-function studies could arise from compensation for NaV1.8 or NaV1.3 deletion, as well as species differences, although it has also been shown that intrathecal antisense knockdown of NaV1.3 did not reverse pain symptoms in a rat spared nerve injury (SNI) model .
We hypothesized that SNE injury would produce sensory neuron hyperexcitability which would be related to altered local expression of NaV1.8. We found that SNE injury in rats induced a unique electrophysiological phenotype characterized by increased sciatic nerve CAP amplitude in both C- and A-fiber populations (Fig. 1). The C-fiber increase likely contributed to the thermal hypersensitivity we have previously observed in this model, while Aδ-fiber excitability would be expected to contribute to the observed mechanical hypersensitivity . However, the overall magnitude of the A-fiber CAP increase and lack of change in conduction velocity suggest increased excitability of Aβ-fibers instead of or in addition to Aδ-fibers. Aβ-fibers are not normally involved in nociception  with the exception of a small subpopulation , but phenotypic changes in Aβ-fibers have been reported after nerve injuries including SNE [34, 35] that result in nociceptive afterdischarge in spinal wide dynamic range neurons . It is possible that such a phenotypic switch might have been involved in the observed A-fiber CAP enhancement and concomitant mechanical or thermal hypersensitivity.
It is notable that increased A- and C-fiber CAP amplitudes persisted even after the stimulus-output plot reached a plateau (Fig. 1). Because increased CAP amplitude is typically interpreted as a recruitment of more axons, a stimulus-output plateau suggests that all available axons have been recruited. A further CAP increase could potentially arise from a change in the total number of axons; however, after SNE the numbers of large- and small-diameter afferents do not change proximal to the injury site, where the recorded tissues were sampled . It is therefore more likely that the observed CAP increase represents peripherally increased Na+ conductance in individual axons. It is further likely that such an increase in conductance would be at least partially mediated by TTX-resistant channels, as CAP TTX resistance increased concomitantly with CAP amplitude for both C- and A-fibers after SNE injury (Fig. 2). An alternative explanation for these observations is that increased expression of TTX-sensitive Na+ channels could increase both the CAP amplitudes and the safety factor for axonal action potential conduction during TTX application. Another possibility we considered was the upregulation of a third TTX-R channel, Nav1.5, which is normally expressed in sensory neurons during development but declines to very low levels at birth and adulthood . Nav1.5 exhibits faster kinetics than Nav1.8 and is significantly more sensitive to TTX (IC50 ~2 μM) than Nav1.8 or Nav1.9 ; thus upregulation of Nav1.5 could potentially account for the observed increases in TTX-resistance, including the blockade of A and C-fiber CAP conduction by 1 μM TTX after SNE. However, we did not observe increased NaV1.5 mRNA expression in the DRG or the sciatic nerve after SNE.
Much larger increases in CAP TTX-resistance, and only in C-fibers (resistant to block by 100 μM TTX), were reported for the SNL model , underscoring major differences between the SNE and SNL models. SNL is thought to cause extensive Wallerian degeneration of injured axons from L5 and L6 DRG, leaving only uninjured L4 axons in the sciatic nerve . In contrast, CCI and SNE cause selective degeneration of Aβ-fibers distal to the constriction site with some remyelination occurring 2–4 weeks after injury [37, 38]; this partial, distal degeneration leaves proximal injured and uninjured axons structurally intact.
We also found that sciatic nerve NaV1.8 immunoreactivity was increased after SNE injury, while immunoreactivity within cell bodies of DRG sensory neurons ipsilateral to SNE was reduced. Similar findings were previously reported for the CCI model , underscoring simmilarities with SNE. This peripheral translocation of Nav1.8 together with increased TTX-resistance suggests involvement of peripherally accumulated NaV1.8 in sciatic nerve hyperexcitability after SNE injury.
Several mechanisms may underly the peripheral translocation of NaV1.8 protein observed in multiple injury models. In the present study, we detected ipsilaterally decreased DRG NaV1.8 immunoreactivity (Fig. 3), but NaV1.8 mRNA levels in the DRG were not affected by SNE injury (Figs. 4 &5). Similar post-CCI immunohistochemical redistribution of NaV1.8 to sciatic nerve axons without increased DRG mRNA expression previously led Novakovic et al. to suggest peripheral trafficking of existing NaV1.8 protein with increased local insertion . Decreased NaV1.8 mRNA levels and attenuated NaV1.8 currents that could be compatible with such a mechanism were also observed in injured small-diameter DRG somata cultured from CCI-injured rats ; NaV1.8 protein levels in injured neurons within the DRG were also shown to decrease in rats after SNL, CCI, SNI, and axotomy injuries . However, while the majority of de novo protein synthesis is believed to occur in DRG somata, an increasing number of studies suggest that axons contain mRNAs such as β-actin, cofilin and RhoA that are locally translated in response to guidance cues during nervous system development [40–42]. Recently, intra-axonal CREB mRNA translation was shown to be essential for neuronal survival in response to nerve growth factor application . We thus hypothesized that axonal NaV1.8 mRNA accumulation might contribute to the increased immunoreactivity in the nerve proximal to SNE injury.
To address this hypothesis, we first determined the presence of NaV1.8 mRNA in the sciatic nerve using RT-PCR (Fig. 4 &5; also see . The 3'RACE sequencing identified a polyA tail in NaV1.8 PCR amplicons, suggesting that our PCR data did not reflect genomic DNA contamination. In addition, we determined the cellular origin of NaV1.8 transcripts because sciatic nerve contains non-neuronal cells such as Schwann cells and fibroblasts in addition to neuronal axons. Cells cultured from both naïve DRG and sciatic nerve showed expression of s100B, confirming the presence of Schwann and/or satellite 'glial' cells. However, NaV1.8 expression was not detected in the nerve cultures (Fig. 6), suggesting that NaV1.8 transcripts detected in intact nerve tissue were likely derived from axons.
Significant up-regulation of NaV1.8 mRNA was demonstrated in the ipsilateral sciatic nerve when compared to the contralateral or uninjured nerve (Fig. 4). To explore the mechanism of the peripheral NaV1.8 mRNA increase, we examined the polyadenylation state. There were no differences in maximum NaV1.8 polyA tail length or smearing pattern between injured nerve and uninjured DRG samples (Fig. 7). Therefore, intra-axonal polyadenylation resulting in reduced turnover and increased steady-state mRNA levels  was ruled out as a mechanism of increased NaV1.8 mRNA accumulation in SNE axons.
Strikingly, simultaneous analysis of multiple sodium channels further demonstrated that after SNE injury, steady-state NaV1.8 mRNA was significantly increased in the ipsilateral nerve by nearly 11-fold when compared to the contralateral nerve. In contrast, we were unable to observe any changes in ipsilateral mRNA levels of NaV1.3, NaV1.6, NaV1.7, or NaV1.9 (Fig. 5). It is unlikely that the selective NaV1.8 mRNA increase was the indirect result of ultrastructural changes, as both the fiber number and composition remain unaltered at the sampling location proximal to the injury site . These findings thus suggest a novel mechanism of selective NaV1.8 mRNA accumulation in injured axons.
A recent study of rat saphenous nerve neuromas revealed increased local translation of structural and metabolic proteins, although no changes were observed in the steady-state protein levels of NaV1.8, NaV1.3, or the α 2δ calcium channel . It has also been shown that axonally transported importin β and vanilloid receptor mRNAs can be locally translated in sciatic nerve axons after nerve injury and dorsal horn synaptic terminals during inflammation, repsectively [47, 48]. Various mRNAs coding cytoskeletal and metabolic proteins have been found to be axonally transported and locally translated in a cultured axonal preparation . Thus, we propose here that enhanced and selective peripheral transport is the most likely mechanism for axonal NaV1.8 mRNA up-regulation, and that local mRNA accumulation could contribute to the observed increase in sciatic nerve NaV1.8 immunoreactivity. Our findings also support the involvement of peripheral NaV1.8 protein in sciatic nerve hyperexcitability after neuropathic injury.