We recently developed a novel test to classify sensory fibers in terms of their EPF responses to different frequencies of electrical stimuli using a Neurometer® . This Neurometer® device has been validated by many clinical studies as a useful tool in the clinical evaluation and management of neurologic disorders [11–13]. Wallace et al.  reported that the patient described the 2000 Hz stimuli as a "light tickle", the 250 Hz stimuli as a "prickly feeling" and the 5 Hz stimuli as a deep warmth or coolness. Therefore, the flexor responses (EPF test) or the withdrawal responses (EPW test) observed in mice seem to be the result of escape from the unpleasant (but not nociceptive) 2000 Hz stimuli, or the nociceptive 250 and 5 Hz stimuli. Recently, Koga et al.  reported that these three different frequencies of electrical stimuli applied with a Neurometer® selectively activate Aβ-, Aδ- or C-fibers in an electrophysiological study using rats. The fiber-specificity of the Neurometer® was supported by the present study, in which spinal transmission was pharmacologically characterized in terms of nociceptive behavior and phosphorylation of extracellular signal-regulated kinase (pERK) in the spinal dorsal horn. The 250 Hz (Aδ-fiber) stimuli induced pERK signals in medium/large-sized DRG neurons and in neurons of the lamina I layer of the dorsal horn. The intrathecal treatments with NMDA receptor antagonists largely inhibited both the pERK signals in the dorsal horn and nociceptive behavior. The 5 Hz (C-fiber) stimuli induced pERK signals in small-sized DRG neurons and in lamina I and II neurons. Both NK1 receptor antagonist and NMDA receptor antagonists blocked spinal pERK signals and nociceptive behavior. On the other hand, the 2000 Hz (Aβ-fiber) stimuli also induced pERK signals in less abundant medium/large-sized DRG neurons, but not in the spinal dorsal horn. The lack of pERK signals induced by 2000 Hz stimuli was consistent with the report by Ji et al. , who demonstrated that no pERK signals were observed in the dorsal horn following Aβ-fiber stimulation, which was verified electrophysiologically. This finding is contrast with our behavioral study, in which 2000-Hz stimuli caused significant paw withdrawal responses that were blocked by intrathecally administered CNQX, an AMPA/kainate antagonist. From the report that rapid pERK signals are caused by an increase in cellular Ca2+ concentration [2, 14], the lack of pERK signals in spinal neurons after Aβ-fiber stimulation might be explained by the fact that AMPA/kainate receptors have lower Ca2+ permeability than NMDA receptors . AMPA/kainate receptor antagonist-sensitive Aβ-fiber-mediated neurotransmission is also supported by previous reports [16, 17]. Although there is an inconsistency in the current intensities and in the duration of electrical stimulation between the behavioral EPW test and the pERK experiment, the fiber-specificity seems to be conformed, since the pharmacological characterization of C-(5 Hz) and Aδ-fiber stimuli (250 Hz)-induced pERK signals is consistent with the behavioral studies in the present study. Furthermore, as Koga et al.  reported that Aβ-specificity is retained in DRG electrophysiological studies as far as the current intensity of 2000 Hz below 1300 μA is used, the Aβ-fiber specificity by 2000 Hz, 1000 μA stimuli in the present study also seems to be retained, though this stimulation failed to induce ERK activation in naïve mice. Therefore, it is evident that analysis using pERK as a biochemical marker is useful for the characterization of spinal pain transmission.
It should be noted that the CNQX-insensitive but NMDA or NK1 receptor antagonists-sensitive spinal pain transmission was observed in the present study, since many electrophysiological studies show predominant contribution of AMPA/kainate receptor in the generation of EPSC in spinal neurons [17, 18]. In the study by Miller and Woolf , CNQX reduced the fast component (Aβ) of EPSC to ~30%, while did the slow component (Aδ +C) only to ~50%. As NMDA receptor antagonist inhibited the slow (Aδ +C), but not fast (Aβ) component of EPSC, it is suggested that there exists the CNQX-insensitive NMDA-receptor-mediated slow component (Aδ +C) in the pain transmission. This view is further supported by the finding that the addition of NMDA elicited an inward current at holding potentials of -60 mV in the presence of Mg2+ . Although the machinery underlying CNQX-insensitive NMDA current remains to be determined, there is a possibility for the predominant NMDA receptor activation. Takasu et al.  reported that CNQX-insensitive NMDA receptor-mediated Ca2+ influx in the presence of ephrinB2-EphB2 activation, which activates NMDA receptor through a Src-mediated phosphorylation. In nerve-injured mice, the EPW thresholds in response to 2000 and 250 Hz stimuli were significantly decreased (hypersensitized), while the threshold in response to a 5 Hz stimulus was increased (hyposensitized). These results are also consistent with our previous study using the algogenic-induced paw flexion (APF) test, showing that prostaglandin I2 (PGI2) agonist-induced A-fiber responses were hypersensitization and that substance P-induced C-fiber responses were diminished in nerve-injured mice [21–23]. It should be noted that the 2000 Hz stimuli-induced hypersensitization in injured mice was completely abolished by intrathecal administration of NMDA receptor antagonists, but not by CNQX. This fact suggests that non-nociceptive information transmitted through Aβ-fibers via AMPA/kainate receptor-mediated spinal neurotransmission is converted into NMDA receptor-sensitive nociceptive information in nerve-injured mice. By contrast, Aδ-fiber-mediated nociceptive responses were also hypersensitized, but the spinal antagonism remained unchanged.
Consistent with the C-fiber hypoalgesia and the Aδ-fiber hyperalgesia observed in the EPW test, the numbers of pERK-positive neurons throughout C-fibers and Aδ-fibers were decreased and increased, respectively, in nerve-injured mice. It is interesting that pERK-signals in smaller DRG neurons (<18 μm) were selectively lost following nerve injury, but the underlying machineries remain to be determined. By contrast, the number of Aδ-fiber stimuli-induced pERK-signals in spinal neurons was increased in injured mice, although no significant change was observed in DRG neurons. The selective activation of spinal neurons in terms of pERK-signals may be explained by the up-regulation of the voltage-dependent calcium channel α2δ-1 (Caα2δ-1) subunit [24, 25], which enhances spinal neurotransmission, leading to an activation of post-synaptic neurons.
The important finding in this study is that Aβ-fiber-induced ERK activation, which is not observed in the naïve state, was detected in the neuropathic pain state, significantly, in the superficial laminae of the dorsal horn, which normally receive nociceptive neural projections. The results obtained using transcutaneous Aβ-fiber stimulation by the Neurometer® are highly consistent with previous studies using low-threshold electrical stimulation to the sciatic nerves of nerve-injured rats [26, 27]. In the present study, we succeeded in the pharmacological characterization of Aβ-fiber-induced ERK activation caused by nerve injury. The Aβ-fiber-induced pERK activation in spinal neurons of nerve-injured mice was blocked by NMDA receptor antagonists (AP-5 and MK-801), but not by an AMPA/kainate receptor antagonist (CNQX), consistent with the present findings that Aβ-fiber hypersensitization was blocked by NMDA antagonists in the EPW test. These results strongly suggest that the Aβ-fiber stimulation may activate spinal neurons that were originally innervated by nociceptive C-fibers or Aδ-fibers. As nerve injury causes hyposensitivity to C-fiber stimulation, the interaction between Aβ-fibers and Aδ-fibers seems to be more important for the neural plasticity observed in the neuropathic pain state. There is an alternative possibility that Aβ-fibers become hypersensitive through an alteration of gene expression, but such a mechanism can not explain the plasticity, since the present study showed no significant ERK activation in laminae III–V, which are expected to be innervated by Aβ-fibers in naïve animals .
Regarding the mechanisms underlying the neural plasticity in neuropathic pain, there are reports that ephaptic discharges, that is, abnormal neural sensitization in pre-synaptic neurons or neural sprouting in the spinal dorsal horn, occurs [29–31]. From the recent finding that demyelination of A-fibers and allodynia occur through mechanisms associated with the lysophosphatidic acid receptor in the dorsal roots of nerve-injured mice , we have proposed that demyelination-induced loss of insulation may cause abnormal cross-talk among A-fibers and sprouting, resulting in a functional switch of innocuous stimulus to painful perception [8, 23].