Skip to main content
NCBI home page
Journal of Neurophysiology logoLink to Journal of Neurophysiology
. 2019 Mar 27;121(6):2061–2070. doi: 10.1152/jn.00464.2018

Upper and lower limb motor axons demonstrate differential excitability and accommodation to strong hyperpolarizing currents during induced hyperthermia

Oliver R Marmoy 1,2,✉, Paul L Furlong 2, Christopher E G Moore 1
PMCID: PMC6620697  PMID: 30917073

Abstract

Length-dependent peripheral neuropathy typically involves the insidious onset of sensory loss in the lower limbs before later progressing proximally. Recent evidence proposes hyperpolarization-activated cyclic nucleotide-gated (HCN) channels as dysfunctional in rodent models of peripheral neuropathy, and therefore differential expression of HCN channels in the lower limbs was hypothesized as a pathophysiological mechanism accounting for the pattern of symptomatology within this study. We studied six healthy participants, using motor axon excitability including strong and long [−70% and −100% hyperpolarizing threshold electrotonus (TEh)] hyperpolarizing currents to preferably study HCN channel function from the median and tibial nerves from high (40%) and low (20%) threshold. This was recorded at normothermia (~32°C) and then repeated during hyperthermia (~40°C) as an artificial hyperpolarizing axon stress. Significant differences between recovery cycle, superexcitability, accommodation to small depolarizing currents, and alterations in late stages of the inward-rectifying currents of strongest (−70% and −100% TEh) currents were observed in the lower limbs during hyperthermia. We demonstrate differences in late IH current flow, which implies higher expression of HCN channel isoforms. The findings also indicate their potential inference in the symptomatology of length-dependent peripheral neuropathies and may be a unique target for minimizing symptomatology and pathogenesis in acquired disease.

NEW & NOTEWORTHY This study demonstrates nerve excitability differences between the upper and lower limbs during hyperthermia, an experimentally induced axonal stress. The findings indicate that there is differential expression of slow hyperpolarization-activated cyclic nucleotide-gated (HCN) channel isoforms between the upper and lower limbs, which was demonstrated through strong, long hyperpolarizing currents during hyperthermia. Such mechanisms may underlie postural control but render the lower limbs susceptible to dysfunction in disease states.

Keywords: axon, HCN channels, hyperpolarization, hyperthermia, nerve excitability, peripheral neuropathy

INTRODUCTION

Upper and lower limb nerve axons differ in their functional properties and demands as afferent and efferent sensorimotor pathways. This difference may result from primary or secondary neurological development due to length-dependent factors but is poorly understood. It is known that in disease states, such as uremic, diabetic, or chemotherapy-induced neuropathy, neuropathic symptoms predominate in the lower limbs and are therefore termed “length-dependent peripheral neuropathy” (LDPN). The term describes this common form of peripheral neuropathy, where symptoms typically begin with sensory disturbance in the lower limbs as tingling, hypoesthesia, and neuropathic pain, which is one of the most common comorbidities worldwide and significantly contributes to chronic pain, foot ulceration, and below-knee amputation (Boulton 2005). In time, there is progression proximally and later upper limb involvement (Dyck et al. 1985; Said 2007). The length-dependent difference of upper and lower limbs may contribute to the pathogenesis responsible for the distal generation of symptoms and may be a result of direct hyperglycemic damage, poorer axoplasmic flow and vascular supply, increase in neurotoxicity, decreased metabolite availability, or axon regrowth inhibition (Arnold et al. 2013; Low et al. 1989). There is, however, no single mechanism to explain the predisposition of the lower limb to earlier dysfunction as seen in LDPN (Quasthoff 1996; Tavaoli et al. 2008). Recent evidence suggests that hyperpolarization-activated cyclic nucleotide-gated (HCN) channels in rodents are pathologically modified in acquired disease states and may therefore contribute to the symptomatology seen in LDPN (Tsantoulas et al. 2017; Tu et al. 2010). The HCN channels are responsible for generating a small depolarization of the axonal membrane, through the hyperpolarization-activated current (IH), in response to hyperpolarization. They are unique in being the only nerve ion channel to be activated by strong hyperpolarization and serve to bring a hyperpolarized membrane back to the resting membrane potential before further activation. Evidence to demonstrate a biophysical difference between upper and lower limb axons that predisposes them to earlier dysfunction and symptomatology in healthy individuals, however, is scant (Krishnan et al. 2004).

Nerve excitability studies are now an established technique for the study of membrane and nodal properties of an axon, using alterations in prestimulus conditioning currents or temporal characteristics of the stimulus to track changes in a determined target response (Bostock et al. 1998; Kiernan et al. 2000). Five measures of excitability are typically recorded, including the stimulus response (SR), strength duration (SD), current-threshold (I/V) curve, threshold electrotonus (TE), and recovery cycle (RC); these recording indexes are detailed in the review by Kiernan et al. (2000). These methods have previously employed ischemia as a nerve stressor to show that slow K+ conductance is better expressed in the upper limb (Krishnan et al. 2004). Within-nerve length dependence of the peroneal nerve corroborated these findings (Kuwabara et al. 2000, 2001a). Temperature stress of the axon at the point of stimulus showed differential changes in excitability between fibers of different activation thresholds and corroborated the theory of higher HCN channels in low-threshold fibers (Moore et al. 2016; Trevillion et al. 2010). The differential effect of marked hyperpolarization and hyperthermia on HCN channel function between the upper and lower limbs has, however, not yet been studied (Burke et al. 1999; Kiernan et al. 2001).

Experimental excitability in rodents suggests that HCN channel function is compromised in acquired neuropathy states (Tsantoulas et al. 2017; Tu et al. 2010). This is consistent with early studies showing greater inward rectification in diabetic patients (Horn et al. 1996; Shimatani et al. 2015). The contribution of HCN channels in human LDPN has not been studied, but first the particular distribution of HCN channels between the upper and lower limbs is not known nor how HCN channels may differ in response to physiological stress such as hyperthermia, which is examined in this study. This may lower the safety margin for activation and be a likely mechanism of HCN channel inability to counteract the effects of axon hyperpolarization induced through hyperthermia. It has been demonstrated that HCN2 isoforms significantly contribute to neuropathic pain, also seen in animal models of acquired neuropathies, which would indicate their potential pathogenesis (Emery et al. 2012; Tsantoulas et al. 2017; Young et al. 2014). Therefore, because of its interest in this study, hyperthermia was chosen to induce functional hyperpolarizing changes in axons of the upper and lower limbs to identify differences in HCN channel expression or explain how activation contributes to the earlier dysfunction in LDPN. Low-threshold fibers (20%) were particularly targeted because of their known increased IH and to improve participant tolerance during lower limb stimulation (Trevillion et al. 2010).

The aim of this study was to identify functional differences in nerve excitability between the upper and lower limbs, particularly with regard to HCN channels, in healthy individuals. We hypothesized that there may be different physiological mechanisms between the upper and lower limbs that may render the lower limbs more susceptible to dysfunction in disease states.

METHODS

All participants provided written informed consent with Ethics Committee approval from the study sponsor (Aston University), with local National Health Service research and innovation approval provided by the host organization (Portsmouth Hospitals) before registering on the National Institute for Health Research portfolio database. The study conformed to the standards set in the Declaration of Helsinki except for registration in a database.

Nine healthy participants were studied, with six qualifying for data analysis (age range 18–49 yr; 3 men, 3 women). Three participants were excluded because of poor stimulus tolerance (n = 2) and asymptomatic carpal tunnel syndrome (n = 1). Nerve excitability was stimulated and recorded at two sites. First stimulation and recording were performed at the wrist with the stimulating cathode at the distal wrist crease of the dominant hand and the anode placed proximally on the forearm, with the compound muscle action potential (CMAP) recorded from the abductor pollicis brevis of the thenar eminence. Second stimulation and recording were performed with the stimulating cathode at the ankle (1–2 cm posterior and inferior to the medial malleolus) over the course of the tibial nerve and the anode placed 8–10 cm proximally over the bony protrusion of the tibia, with the active recording electrode placed over the abductor hallucis (AH) muscle and the reference electrode placed 6 cm distally over the bony metatarsophalangeal joint (Fig. 1). Stimulation at the ankle required compression with a small plastic cylinder to ensure effective stimulation of the tibial nerve. The ground electrode was placed over the dorsum of the hand or foot respectively.

Fig. 1.

Fig. 1.

Open in a new tab

Stimulating and recording scenarios for recording of upper and lower limb nerve excitability. A: recording for the upper limb median nerve, with the active and recording electrodes over the belly and distal point of abductor pollicis brevis, respectively. The median nerve was stimulated through a cathode at the wrist over the course of the median nerve, with the anode placed at the mid-forearm proximally and laterally. The temperature probe is seen next to the cathode, housed in a cardboard shell. B: recording for the lower limb tibial nerve, with the active and recording electrodes over the belly and distal point of abductor hallucis, respectively. The tibial nerve was stimulated through a cathode at the ankle inferior to the medial malleolus over the course of the tibial nerve, with the anode placed over the bony protrusion of the tibia. A cylindrical plastic tube was secured over the cathode to compress the electrode and minimize the volume between stimulating electrode and nerve for stimulation. The temperature probe is seen slightly proximally to the cathode housed in a cardboard shell. C: the warming procedure for the upper limb. The arm was submerged into a thermostatically controlled water bath after being encapsulated within a polythene plastic sleeve. The arm was suspended with a material sling inside the water bath that allowed water turbulence and ensured that the limb was at the midpoint in the water, not next to the heating element, which would have caused focal heating. D: the same heating procedure for the lower limb, with the addition of the foam edging to ensure patient comfort during this time in this position.

Excitability indexes of SR, SD, TE, I/V curve, and refractory period on high (40%)- or low (20%)-threshold fibers were generated (Bostock et al. 1998). SD was sampled at stimulus widths of 0.2, 0.4, 0.6, 0.8, and 1 ms and TE recorded in conditions of +20%, +40%, −20%, −40%, −70%, and −100%. Stimulation and recording were automatically controlled via a PC running the QTRAC nerve excitability program and a subroutine named TRONDNF, which records the five excitability indexes listed above (Bostock et al. 1998). Stimuli were delivered via a Digitimer DS5 (Digitimer, Welwyn Garden City, UK), with a Grass LP511 AC Amplifier (Grass Technologies, Natus Neurology) used to amplify the CMAP with band-pass filters (2 Hz to 10 kHz) before it was digitized with a sampling rate of 20 kHz. The HumBug (Digitimer) was used to remove 50-Hz noise before the signal was digitized through the analog-to-digital converter (National Instruments PCI-6221; National Instruments, Austin, TX). Skin was prepared with abrasive gel (NuPrep; Weaver) and an alcohol wipe to minimize skin impedance. Disposable sticky electrocardiographic electrodes were attached to deliver the stimulus current to the desired nerve (Ambu BlueSensor ECG electrodes; Ambu), with small polarizable sticky recording electrodes used for recording the CMAP (Ambu Neuroline Surface electrodes; Ambu).

The procedure followed a protocol in which “high” (40%) and “low” (20%) target thresholds were measured first during normothermia from the median and tibial nerves, respectively. The benefit of hyperthermia as an artificial stress was demonstrated by Howells et al. (2013), where it was found that slow K+ channels “dampen” the excitability of axons and, more importantly, IH becomes more hyperpolarized in hyperthermia. A minimum of 15 min between recording periods was given to minimize effects of activity-dependent hyperpolarization (Kuwabara et al. 2002), after which the arm was placed within a polythene sheath and submerged in a temperature-controlled water bath at 41–42°C for a minimum of 20 min until the limb temperature remained stable for >5 min to ensure that the nerve was adequately heated. During pilot studies a rise in systemic temperature was seen after limb warming, so to minimize confounding temperature variables normothermic measurements were all made before warming of any limb. The water bath was installed with a custom-made material sling to improve comfort and avoid blockage of water turbulence. Temperature was recorded with a thermometer calibrated against a thermistor thermometer accurate to 0.1°C, with a cardboard housing placed over the probe to ensure that this only recorded skin temperature and not external temperature (Drager HNICU-36; DeRoyal Industries). Nerve excitability was repeated during hyperthermia with low-threshold fibers. The experimental protocol was then repeated for the tibial nerve at the ankle. Throughout testing a minimum of 5 min between recordings was implemented to avoid any activity-dependent axon hyperpolarization (Kuwabara et al. 2001b).

All patients were tested with traditional nerve conduction studies and a screening questionnaire to exclude a history of neurological or diabetic disease. Data collected were tested for normality with the Lilliefors test, using paired Student’s t-tests on all recorded variables between study groups of limb and temperature, following a Benjamini-Hochberg procedure with a false discovery rate of 10% to correct for multiple-comparison measures. Thus P value was dependent on its variable rank size and the number of comparisons made.

RESULTS

Nerve excitability was recorded from high (40%)- and low (20%)-threshold fibers of the median and tibial nerves during normothermia and from low-threshold fibers of the median and tibial nerves during hyperthermia. The mean resting and hyperthermic temperatures and temperature differences during each scenario are shown in Table 1.

Table 1.

Mean temperature recordings and change during normothermia and induced hyperthermia with thermostatically controlled water bath

Nerve Tested Normothermic Temperature of 40% + 20% Threshold Fibers, °C Hyperthermic Temperature of 20% Threshold Fibers, °C Mean Increase, °C
Median (wrist) 31.7 ± 0.5 39.3 ± 0.2 7.6 ± 0.6 (P = 0.0032)
Tibial (ankle) 29.1 ± 0.5 39.3 ± 0.2 10.5 ± 0.6 (P = 0.0032)
Open in a new tab

Values are means ± SE for n = 6 participants. All measurements were made with a thermistor-calibrated thermometer recording skin temperature within 3 cm of the stimulating cathode. Insulative shielding was used to minimize the confounding influence of external heat on recorded temperatures.

CMAP latency recorded from abductor pollicis brevis during median nerve stimulation significantly decreased by 0.94 ± 0.1 ms during hyperthermia (P = 0.0194). A nonsignificant small reduction of 4.4% was seen in the peak CMAP amplitude and an increase of 15.9% in the stimulus required to elicit 50% of the peak response in hyperthermia. CMAP latency recorded from AH during tibial nerve stimulation reduced significantly during hyperthermia by 1.63 ± 0.2 ms (P = 0.0004). A significant reduction of 15.9% in the peak CMAP was seen during hyperthermia (P = 0.004) alongside a nonsignificant decrease in the stimulus required to elicit 50% of the peak response. Therefore these changes were not thought to reflect any significant conduction block during hyperthermia, as all peak amplitudes remained within 20% of their normothermic measures.

Changes in Strength-Duration

SD properties were not significantly different between temperature variables in either median or tibial nerves (Fig. 2). However, axon rheobase exhibited a larger mean change of 1.1 mA in the lower limb compared with a mean change of 0.3 mA in the upper limb, seen as a slope change of SD indicating less threshold change to longer stimuli in the lower limb during hyperthermia.

Fig. 2.

Fig. 2.

Open in a new tab

Strength duration properties in normal temperature and during hyperthermia in the same limb. Data are plotted as group means (△) ± SE (error bars). A: strength duration properties of the median nerve at normal temperature (black) and during hyperthermia (gray). B: strength duration properties of the tibial nerve at normal temperature (black) and during hyperthermia (gray).

Changes in Recovery Cycle

The RC in the median nerve demonstrated significant reduction in the relative refractory period (RRP) of 0.484 ms (P = 0.019), supernormal period (SNP) (P = 0.0419; including at 5 and 7 ms), and refractoriness at 2.5 ms (P = 0.0258) during hyperthermia (Fig. 3A). A nonsignificant reduction in late subexcitable period (LSP) area of 3.6 ± 1.6% was observed during hyperthermia. Tibial nerve RC showed only a significant reduction in refractoriness at 2.5 ms (P = 0.0226) during hyperthermia (Fig. 3B). Comparisons between limbs in hyperthermia showed the tibial nerve to have significantly larger SNP (P = 0.0125), with no significant differences in RRP or LSP.

Fig. 3.

Fig. 3.

Open in a new tab

A–C: recovery cycle (RC) curves during temperature change. Data are expressed as group means (○) ± SE (error bars). A: RC curve of the median nerve at normal temperature (black) and in hyperthermia (gray). B: RC curve of the tibial nerve at normal temperature (black) and in hyperthermia (gray). C: RC curves during hyperthermia of the median nerve (black) and tibial nerve (gray). D: diagram of RC curve indexes with relative refractory period (RRP), supernormal period (SNP), and late subexcitable period (LSP).

Changes in Threshold Electrotonus

Depolarizing currents.

TE to small depolarizing currents in the median nerve (Fig. 4A) demonstrated significant decreases in TEd 10–20 ms (P = 0.0387), TEd peak (P = 0.0161), and TEd undershoot (P = 0.0226) alongside S2 accommodation (P = 0.0022) and accommodation half-time (P = 0.0032) during hyperthermia. The findings were less significant in the tibial nerve during hyperthermia, where only TEd 40–60 ms (P = 0.0129) and accommodation half-time (P = 0.0097) were significantly decreased (Fig. 4B). Comparisons between different nerves during hyperthermia demonstrated a significantly larger TEd 40–60 ms and S2 accommodation half-time in the hyperthermic tibial nerve (Fig. 4C).

Fig. 4.

Fig. 4.

Open in a new tab

Threshold electrotonus (TE) between temperature variables. Data are expressed as group means (○) ± SE (error bars). A: TE of the median nerve during normal (green) and hyperthermic (magenta) temperatures. B: TE of the tibial nerve during normal (gray) and hyperthermic (red) temperatures. C: TE of median (magenta) and tibial (red) nerves during hyperthermia. Top red arrow indicates TE to depolarizing current (TEd) 40–60 ms, and bottom red arrow indicates S3 at TE to hyperpolarizing current (TEh) −100%, which were found to demonstrate differential responses between the upper and lower limbs.

Hyperpolarizing currents.

TE to small hyperpolarizing subthreshold currents during hyperthermia demonstrated a significant decrease in the TEh overshoot of both median and tibial nerves compared with normothermia (P < 0.01). The accommodation to strong and long −70% (200 ms) and −100% (300 ms) hyperpolarizing currents to the median nerve showed a significant decrease in S3 for −70% (P = 0.0452) and S3 for −100% (P = 0.0065), with just significant differences in the TEh 101–140 ms slope (P = 0.0516), all of which are consistent with the graphical TE findings (Fig. 4A). Although accommodation to strong and long hyperpolarizing currents to the tibial nerve failed to demonstrate a significant change, some speculation is made to the similar graphical trend of S3 in Fig. 4B being similar to Fig. 4A of the median nerve. TE to strong and longer hyperpolarizing currents of different nerves during hyperthermia showed significantly less threshold reduction in the strongest and longest (−100%, 300 ms) hyperpolarizing current S3 phase (Fig. 4C). No significant differences were seen in TEh peaks of −70% and −100% curves.

Changes in Current-Threshold Relationship

There was no difference between median and tibial nerves at rest and no statistically significant change in either nerve on heating (Fig. 5).

Fig. 5.

Fig. 5.

Open in a new tab

Current-voltage (I/V) relationship for normal and hyperthermic temperatures. Data are expressed as means (○) ± SE (error bars). A: I/V curve in the median nerve during normal temperature (black) and during hyperthermia (gray). B: I/V curve in the tibial nerve during normal temperature (black) and during hyperthermia (gray). C: I/V curve in the hyperthermic median (black) and hyperthermic tibial (gray) nerves.

DISCUSSION

Axonal excitability studies in lower limb motor axons have shown that hyperthermia has marked effects on RC superexcitability and accommodation to small depolarizing currents. Also (and previously unreported), we demonstrate late-onset differences in the response to strong hyperpolarizing currents between the upper and lower limbs during hyperthermia. This demonstrates differential HCN channel behavior between the upper and lower limbs that, as recent evidence suggests, may render the lower limbs more susceptible to dysfunction in disease states and could underlie the pathophysiological mechanisms responsible for the symptomatology seen in LDPN.

Effects of Hyperthermia on Stimulus Response, Strength Duration, and Recovery Cycle Between Limbs

Hyperthermia demonstrated opposing shifts of the stimulus response compared with normothermia; the median nerve necessitated an increased stimulus to elicit a peak response, whereas the tibial nerve reduced the stimulus required. There was nothing to suggest temperature-related changes at the neuro-muscular junction or directly to muscle fibers; maximum amplitude and morphology of the CMAP remained relatively stable, and there was no evidence of conduction block. Therefore this change can be taken to show that the median nerve axons at the wrist were more hyperpolarized and therefore required a higher stimulus current to exceed the axon threshold (Henderson et al. 2006). This is supported by the reduced SD slope, alongside a slight reduction in the raw peak CMAP amplitude, constituting a level of local axon hyperpolarization (Krishnan et al. 2009). Such changes indicate that the tibial nerve may be better able to counteract chronic hyperpolarizing changes compared with the median nerve.

Hyperthermia in both the upper and lower limb axons from resting temperature demonstrated reductions in refractoriness, with reduction in RRP also seen in the upper limb. This suggests an acceleration of fast K+ channel kinetics in the node and paranode caused by hyperthermia of the axon, as the opposite effect is seen during cooling (Kiernan et al. 2001). Comparisons between median and tibial nerves during hyperthermia revealed no significant differences. This suggests that fast K+ kinetics are not significantly different between the upper and lower limbs and is consistent with previous findings in the peroneal nerve (Kuwabara et al. 2000). Overall, this leftward shift of the RC curve again indicates hyperpolarization of the axon membrane (David et al. 1995).

Increased Nodal Na+ Driving Current in Tibial Motor Axons

A significant reduction of median nerve supernormality (SNP) was seen in hyperthermia. This was surprisingly not seen in the tibial nerve despite a larger temperature change from rest, which would have been predicted (Burke et al. 1999). The SNP originates from a depolarizing afterpotential through spontaneous charge flow into the node from current stored in the internode. This makes the SNP very sensitive to internodal resistance properties (David et al. 1995; Kiernan and Bostock 2000). Although the axon hyperpolarization could create differences in the SNP, this would be a result of extra-axonal K+ accumulation due to activation of the Na+-K+ pump, which would cause significant differences in RRP between limbs, a finding not seen previously (Kuwabara et al. 2001a). It could be possible that the larger and resilient SNP of the tibial nerve may reflect greater capacitance in the internode creating a larger SNP; however, this is physiologically implausible. A longer internodal region should proportionally increase conduction velocity, but tibial nerve velocity is slower than the median nerve (Kimura 2013; Simpson et al. 2013). In addition, differences in axon diameter would cause disparity in the RRP, which was not seen between nerves (Kiernan et al. 2000). It is therefore suggested that these findings are attributed to an increased nodal Na+ driving current to overcome hyperpolarization and incur a larger SNP as seen in the tibial nerve. An alternate hypothesis would be an increased myelin thickness, although this is not supported by histological findings (Debanne et al. 2011).

Decreased Slow K+ Conductance in Tibial Nerve Axons

During hyperthermia the TEd 40–60 ms, representing the S2 phase, was significantly increased in the tibial nerve compared with the median nerve. The S2 period reflects a decrease in excitability of the axon, physiologically attributed to the activation of slow nodal K+ channels (Kiernan et al. 2000). This was supported by a significant increase in the accommodation half-time of tibial axons, indicating a longer accommodation of slow K+ channels to restore membrane potential (Bostock and Baker 1988). The TEd overshoot phase, attributed to deactivation of slow K+ channels, was not significantly different between upper and lower limbs in hyperthermia. This finding indicates that although the kinetics of the slow K+ channels are similar between upper and lower limbs, the slow K+ channel expression is reduced in the lower limb. This difference in physiology gives one mechanism that may render the lower limbs more susceptible to dysfunction in disease conditions (see also Kuwabara et al. 2000, 2001a).

Differences in Hyperpolarizing Threshold Electrotonus: Evidence for Different HCN Channel Expression of Tibial Nerve

A significant reduction in TEh overshoot was seen in both axons during hyperthermia compared with normothermia. As TEh is proportionate to deactivation of slow K+ channels and deactivation of IH, these findings would support the findings in TEd of less expressed slow K+ channels in the lower limb.

Perhaps most novel of the findings was a significantly increased S3 period to the strongest and longest (−100% for 300 ms) hyperpolarizing currents of the tibial nerve compared with the median nerve in hyperthermia. This demonstrates greater accommodation in the tibial nerve during hyperthermia, which implies a greater inward-rectifying IH in the tibial nerve relative to the median nerve. This study was unable to directly compare differences of IH at rest because of differences in resting temperatures of the two nerves. Previous studies found no significant differences between S3 phases (Kuwabara et al. 2001a). It is therefore implied that when subjected to hyperthermic stress the lower limb tibial nerve exhibits greater accommodation to hyperpolarization through higher IH, probably due to increased conductance of individual channels, although this has not been studied in vitro (Tomlinson et al. 2010). We therefore speculated that HCN channels are expressed differently in the lower limb or the HCN channels themselves have different properties that permit greater IH conductance.

Although different HCN channel expression is likely to be the rationale of these findings, the graphical changes indicate a mechanism unanticipated by the original hypothesis. It was surprising that the most significant change was mostly notable after 200 ms of TEh −100% plots. This is rather atypical for previous studies of HCN excitability in the median nerve because of its late onset and demonstrates why these changes are not reflected significantly within the I/V curve due to its shorter stimulus duration (Howells et al. 2012; Tomlinson et al. 2010). Of the four HCN isoforms both HCN1 and HCN2 are known to be distributed in human peripheral nerve, but HCN3 is of uncertain distribution and HCN4 primarily distributed in the central nervous system (Doan et al. 2004). Although HCN1 was originally hypothesized to be expressed differently in the lower limb, these findings allow the theory that a slower HCN isoform is more likely to be responsible for these late changes because of its time of activation. It has been documented in cellular studies that the activation times for HCN1–4 are 30 ms, 184 ms, 265 ms, and 461 ms, respectively, which therefore makes HCN2 and HCN3 viable candidates and also supports the theory that slower HCN isoforms underpin the late strong hyperpolarization as postulated by Nodera and Rutkove (2012) and Howells et al. (2012). As no data exist on HCN isoform location in peripheral axons of humans in vitro, it may be that our observations represent the first indication of the existence of slower HCN isoforms in human peripheral nerves.

Dysfunctional HCN Channels: Potential Implications for LDPN

HCN channel dysfunction has been demonstrated in experimentally induced diabetic wild-type mice and rat nodose ganglion cells (Shimatani et al. 2015; Tu et al. 2010). There has been little study of HCN channel function, through additional hyperpolarizing protocols, in early or established acquired neuropathy in humans in vivo, research of which now appears necessary to establish our speculations of whether this mechanism has a substantial contribution to the development to the distally generated symptoms in actual patients with LDPN. Studies in sensory axons may further elucidate a mechanism of slow HCN channel isoforms, especially given the higher expression of HCN channels in sensory nerve axons (Howells et al. 2012).

We speculate that a difference in HCN channel expression in the lower limbs seen in these findings would have direct implications for the symptomatology of acquired neuropathies. Previous studies in diabetic neuropathy have demonstrated subtle inward rectification, which indicates altered IH (Horn et al. 1996). Limb paresthesia and neuropathic pain can be the most common symptoms in acquired neuropathy, which is a pertinent feature because of the mechanism of paresthesia being a result of ectopic impulse activity where axons activate and discharge asynchronously or spontaneously (Dyck et al. 1985; Mogyoros et al. 2000). Where repetitive firing may occur, such as in trains of impulses, axons are known to hyperpolarize (Bostock and Bergmans 1994; Kiernan et al. 1997). Therefore, an inability to reverse a hyperpolarized axon state effectively may lead to the development of ectopic impulse activity and worsen paresthesia by reducing the safety margin for conduction (Kiernan et al. 1997). Relative overactivity of mechanisms of hyperpolarization during pathological stress would therefore render a nerve more excitable and liable to generate action potentials. This is a known mechanism of adenosine 3′,5′-cyclic monophosphate (CAMP)- and temperature-dependent tachycardia and may also explain the worsening of neuropathic symptoms when hot. HCN2 channels also play an important contribution to the generation of neuropathic pain seen in mouse models of acquired neuropathy (Jiang et al. 2008; Tsantoulas et al. 2017), whereas their inhibition with ivabradine and lamotrigine, an antianginal and antiepileptic medication, have HCN blocking properties and are interestingly both clinically effective in neuropathic pain relief (Eisenberg et al. 2001; Young et al. 2014). Both of these medications drive functional kinetic differences in HCN channel function, and thus future research using the methods described in this report may help elucidate their mechanism of action in future in vivo studies of patients with LDPN.

Why Would Late HCN Channels Be More Expressed in the Lower Limbs?

From a physiological perspective, the justification behind differential HCN channel expression is uncertain. According to the Henneman size principle, slower motor fibers are recruited earlier; thus it is possible that the changes in IH may reflect differences in slow and fast axons. As AH had a higher stimulus requirement to elicit a peak response, it is feasible that more “slow” fibers were recruited, contributing to the findings (Kudina and Andreeva 2014; Lorenz and Jones 2014; Mendell 2005). It is, however, our theory that the lower limbs require more stable tetanic contraction to maintain weight bearing and posture control; therefore their greater firing rates and need to counter excessive hyperpolarization is managed through this proposed upregulation or higher expression of HCN channels (Johns and Fuglevand 2011). This would be supported by the fatigue resistance of the AH muscle, meaning that a slower HCN isoform may be in higher concentration to limit hyperpolarization and permit more continuous firing of axons (Kelly et al. 2013). Such hypotheses could be supported by similar experimental methodology in quadrupeds. This physiological compensation for more stable firing is required for postural control but therefore may render the lower limbs more susceptible in disease states.

Conclusions

This study provides the first known data demonstrating the differences in axonal excitability of the lower limbs during hyperthermia. It was found that there is most likely greater inward nodal Na+ driving current and poorer expression of slow K+ conductance in the lower limbs. We have discovered novel findings of differential accommodation to strong and long (−100%, 300 ms) hyperpolarizing currents during hyperthermia between the upper and lower limbs that demonstrated greater IH conductance in the lower limb tibial nerve. The late onset of this change is speculated to be a result of a slower HCN channel isoform, possibly HCN2 or HCN3. The findings of this study suggest that the differential expression of HCN channels between upper and lower limbs may contribute to the patterns of symptoms in some diseases as a potentially new liable site for dysfunction in acquired neuropathy. Further studies are necessary to confirm this last speculation, particularly with investigation of sensory axons and in diabetic patients without neuropathy, to identify whether this technique may be of use in early diagnosis of neuropathy where traditional nerve conduction studies lack sensitivity. Positive findings may prompt later experimental models beneficial to establish a pharmacological role of HCN channels in the early diagnosis of LDPN and management of disease progression and pain relief. The ability to imply ion channel function and dysfunction in humans in vivo with the methods of nerve excitability should also help take research directly to the bedside and reduce the number of animal studies needed in the development of new treatments.

GRANTS

This research was financially supported by Health Education England through the National School of Healthcare Science Scientist Training Programme funding stream and was registered on the National Institute for Health Research portfolio (Ref. 193691).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

O.R.M., P.L.F., and C.E.G.M. conceived and designed research; O.R.M. performed experiments; O.R.M. and C.E.G.M. analyzed data; O.R.M. and C.E.G.M. interpreted results of experiments; O.R.M. prepared figures; O.R.M., P.L.F., and C.E.G.M. drafted manuscript; O.R.M., P.L.F., and C.E.G.M. edited and revised manuscript; O.R.M., P.L.F., and C.E.G.M. approved final version of manuscript.

ACKNOWLEDGMENTS

O. R. Marmoy thanks participants for their kind donation of time.

REFERENCES

  1. Arnold R, Kwai NC, Krishnan AV. Mechanisms of axonal dysfunction in diabetic and uraemic neuropathies. Clin Neurophysiol 124: 2079–2090, 2013. doi: 10.1016/j.clinph.2013.04.012. [DOI] [PubMed] [Google Scholar]
  2. Bostock H, Baker M. Evidence for two types of potassium channel in human motor axons in vivo. Brain Res 462: 354–358, 1988. doi: 10.1016/0006-8993(88)90564-1. [DOI] [PubMed] [Google Scholar]
  3. Bostock H, Bergmans J. Post-tetanic excitability changes and ectopic discharges in a human motor axon. Brain 117: 913–928, 1994. [DOI] [PubMed] [Google Scholar]
  4. Bostock H, Cikurel K, Burke D. Threshold tracking techniques in the study of human peripheral nerve. Muscle Nerve 21: 137–158, 1998. doi:. [DOI] [PubMed] [Google Scholar]
  5. Boulton AJ. Management of diabetic peripheral neuropathy. Clin Diabetes 23: 9–15, 2005. doi: 10.2337/diaclin.23.1.9. [DOI] [Google Scholar]
  6. Burke D, Mogyoros I, Vagg R, Kiernan MC. Temperature dependence of excitability indices of human cutaneous afferents. Muscle Nerve 22: 51–60, 1999. doi:. [DOI] [PubMed] [Google Scholar]
  7. David G, Modney B, Scappaticci KA, Barrett JN, Barrett EF. Electrical and morphological factors influencing the depolarizing after-potential in rat and lizard myelinated axons. J Physiol 489: 141–157, 1995. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Debanne D, Campanac E, Bialowas A, Carlier E, Alcaraz G. Axon physiology. Physiol Rev 91: 555–602, 2011. doi: 10.1152/physrev.00048.2009. [DOI] [PubMed] [Google Scholar]
  9. Doan TN, Stephans K, Ramirez AN, Glazebrook PA, Andresen MC, Kunze DL. Differential distribution and function of hyperpolarization-activated channels in sensory neurons and mechanosensitive fibers. J Neurosci 24: 3335–3343, 2004. doi: 10.1523/JNEUROSCI.5156-03.2004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Dyck PJ, Karnes JL, Daube J, O’Brien P, Service FJ. Clinical and neuropathological criteria for the diagnosis and staging of diabetic polyneuropathy. Brain 108: 861–880, 1985. doi: 10.1093/brain/108.4.861. [DOI] [PubMed] [Google Scholar]
  11. Eisenberg E, Lurie Y, Braker C, Daoud D, Ishay A. Lamotrigine reduces painful diabetic neuropathy: a randomized, controlled study. Neurology 57: 505–509, 2001. [DOI] [PubMed] [Google Scholar]
  12. Emery EC, Young GT, McNaughton PA. HCN2 ion channels: an emerging role as the pacemakers of pain. Trends Pharmacol Sci 33: 456–463, 2012. doi: 10.1016/j.tips.2012.04.004. [DOI] [PubMed] [Google Scholar]
  13. Henderson RD, Ridall GR, Pettitt AN, McCombe PA, Daube JR. The stimulus-response curve and motor unit variability in normal subjects and subjects with amyotrophic lateral sclerosis. Muscle Nerve 34: 34–43, 2006. doi: 10.1002/mus.20561. [DOI] [PubMed] [Google Scholar]
  14. Horn S, Quasthoff S, Grafe P, Bostock H, Renner R, Schrank B. Abnormal axonal inward rectification in diabetic neuropathy. Muscle Nerve 19: 1268–1275, 1996. doi: 10.1002/mus.880191002. [DOI] [PubMed] [Google Scholar]
  15. Howells J, Czesnik D, Trevillion L, Burke D. Excitability and the safety margin in human axons during hyperthermia. J Physiol 591: 3063–3080, 2013. doi: 10.1113/jphysiol.2012.249060. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Howells J, Trevillion L, Bostock H, Burke D. The voltage dependence of Ih in human myelinated axons. J Physiol 590: 1625–1640, 2012. doi: 10.1113/jphysiol.2011.225573. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Jiang YQ, Sun Q, Tu HY, Wan Y. Characteristics of HCN channels and their participation in neuropathic pain. Neurochem Res 33: 1979–1989, 2008. doi: 10.1007/s11064-008-9717-6. [DOI] [PubMed] [Google Scholar]
  18. Johns RK, Fuglevand AJ. Number of motor units in human abductor hallucis. Muscle Nerve 43: 895–896, 2011. doi: 10.1002/mus.22071. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Kelly LA, Racinais S, Cresswell AG. Discharge properties of abductor hallucis before, during, and after an isometric fatigue task. J Neurophysiol 110: 891–898, 2013. doi: 10.1152/jn.00944.2012. [DOI] [PubMed] [Google Scholar]
  20. Kiernan MC, Bostock H. Effects of membrane polarization and ischaemia on the excitability properties of human motor axons. Brain 123: 2542–2551, 2000. [DOI] [PubMed] [Google Scholar]
  21. Kiernan MC, Burke D, Andersen KV, Bostock H. Multiple measures of axonal excitability: a new approach in clinical testing. Muscle Nerve 23: 399–409, 2000. doi:. [DOI] [PubMed] [Google Scholar]
  22. Kiernan MC, Cikurel K, Bostock H. Effects of temperature on the excitability properties of human motor axons. Brain 124: 816–825, 2001. doi: 10.1093/brain/124.4.816. [DOI] [PubMed] [Google Scholar]
  23. Kiernan MC, Mogyoros I, Hales JP, Gracies JM, Burke D. Excitability changes in human cutaneous afferents induced by prolonged repetitive axonal activity. J Physiol 500: 255–264, 1997. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Kimura J. Electrodiagnosis in Diseases of Nerve and Muscle: Principles and Practice (4th ed). Oxford: Oxford Univ. Press, 2013. [Google Scholar]
  25. Krishnan AV, Lin CS, Kiernan MC. Nerve excitability properties in lower-limb motor axons: evidence for a length-dependent gradient. Muscle Nerve 29: 645–655, 2004. doi: 10.1002/mus.20013. [DOI] [PubMed] [Google Scholar]
  26. Krishnan AV, Lin CS, Park SB, Kiernan MC. Axonal ion channels from bench to bedside: a translational neuroscience perspective. Prog Neurobiol 89: 288–313, 2009. doi: 10.1016/j.pneurobio.2009.08.002. [DOI] [PubMed] [Google Scholar]
  27. Kudina LP, Andreeva RE. Excitability properties of single human motor axons: are all axons identical? Front Cell Neurosci 8: 85, 2014. doi: 10.3389/fncel.2014.00085. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Kuwabara S, Cappelen-Smith C, Lin CS, Mogyoros I, Burke D. Differences in accommodative properties of median and peroneal motor axons. J Neurol Neurosurg Psychiatry 70: 372–376, 2001a. doi: 10.1136/jnnp.70.3.372. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Kuwabara S, Cappelen-Smith C, Lin CS, Mogyoros I, Burke D. Effects of voluntary activity on the excitability of motor axons in the peroneal nerve. Muscle Nerve 25: 176–184, 2002. doi: 10.1002/mus.10030. [DOI] [PubMed] [Google Scholar]
  30. Kuwabara S, Cappelen-Smith C, Lin CS, Mogyoros I, Bostock H, Burke D. Excitability properties of median and peroneal motor axons. Muscle Nerve 23: 1365–1373, 2000. doi:. [DOI] [PubMed] [Google Scholar]
  31. Kuwabara S, Lin CS, Mogyoros I, Cappelen-Smith C, Burke D. Voluntary contraction impairs the refractory period of transmission in healthy human axons. J Physiol 531: 265–275, 2001b. doi: 10.1111/j.1469-7793.2001.0265j.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Lorenz C, Jones KE. IH activity is increased in populations of slow versus fast motor axons of the rat. Front Hum Neurosci 8: 766, 2014. doi: 10.3389/fnhum.2014.00766. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Low PA, Lagerlund TD, McManis PG. Nerve blood flow and oxygen delivery in normal, diabetic and ischemic neuropathy. Int Rev Neurobiol 31: 355–438, 1989. [DOI] [PubMed] [Google Scholar]
  34. Mendell LM. The size principle: a rule describing the recruitment of motoneurons. J Neurophysiol 93: 3024–3026, 2005. doi: 10.1152/classicessays.00025.2005. [DOI] [PubMed] [Google Scholar]
  35. Mogyoros I, Bostock H, Burke D. Mechanisms of paresthesias arising from healthy axons. Muscle Nerve 23: 310–320, 2000. doi:. [DOI] [PubMed] [Google Scholar]
  36. Moore CE, Cockerill HJ, Marmoy OR. Temperature dependent changes in HCN channel current (Ih) in human motor axons of different threshold. Implications for generation of febrile seizures (Abstract). Clin Neurophysiol 127: e58, 2016. doi: 10.1016/j.clinph.2015.11.191. [DOI] [Google Scholar]
  37. Nodera H, Rutkove SB. Accommodation to hyperpolarizing currents: differences between motor and sensory nerves in mice. Neurosci Lett 518: 111–116, 2012. doi: 10.1016/j.neulet.2012.04.065. [DOI] [PubMed] [Google Scholar]
  38. Quasthoff S. The role of axonal ion conductances in diabetic neuropathy: a review. Muscle Nerve 21: 1246–1255, 1996. [DOI] [PubMed] [Google Scholar]
  39. Said G. Diabetic neuropathy—a review. Nat Clin Pract Neurol 3: 331–340, 2007. doi: 10.1038/ncpneuro0504. [DOI] [PubMed] [Google Scholar]
  40. Shimatani Y, Nodera H, Osaki Y, Banzrai C, Takayasu K, Endo S, Shibuta Y, Kaji R. Upregulation of axonal HCN current by methylglyoxal: potential association with diabetic polyneuropathy. Clin Neurophysiol 126: 2226–2232, 2015. doi: 10.1016/j.clinph.2015.02.058. [DOI] [PubMed] [Google Scholar]
  41. Simpson AH, Gillingwater TH, Anderson H, Cottrell D, Sherman DL, Ribchester RR, Brophy PJ. Effect of limb lengthening on internodal length and conduction velocity of peripheral nerve. J Neurosci 33: 4536–4539, 2013. doi: 10.1523/JNEUROSCI.4176-12.2013. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Tavaoli M, Mojaddidi M, Fadavi H, Malik RA. Pathophysiology and treatment of painful diabetic neuropathy. Curr Pain Headache Rep 12: 192–197, 2008. [DOI] [PubMed] [Google Scholar]
  43. Tomlinson SE, Burke D, Hanna M, Koltzenburg M, Bostock H. In vivo assessment of HCN channel current (Ih) in human motor axons. Muscle Nerve 41: 247–256, 2010. doi: 10.1002/mus.21482. [DOI] [PubMed] [Google Scholar]
  44. Trevillion L, Howells J, Bostock H, Burke D. Properties of low-threshold motor axons in the human median nerve. J Physiol 588: 2503–2515, 2010. doi: 10.1113/jphysiol.2010.190884. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Tsantoulas C, Laínez S, Wong S, Mehta I, Vilar B, McNaughton PA. Hyperpolarization-activated cyclic nucleotide-gated 2 (HCN2) ion channels drive pain in mouse models of diabetic neuropathy. Sci Transl Med 9: eaam6072, 2017. doi: 10.1126/scitranslmed.aam6072. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Tu H, Zhang L, Tran TP, Muelleman RL, Li YL. Diabetes alters protein expression of hyperpolarization-activated cyclic nucleotide-gated channel subunits in rat nodose ganglion cells. Neuroscience 165: 39–52, 2010. doi: 10.1016/j.neuroscience.2009.10.002. [DOI] [PubMed] [Google Scholar]
  47. Young GT, Emery EC, Mooney ER, Tsantoulas C, McNaughton PA. Inflammatory and neuropathic pain are rapidly suppressed by peripheral block of hyperpolarisation-activated cyclic nucleotide-gated ion channels. Pain 155: 1708–1719, 2014. doi: 10.1016/j.pain.2014.05.021. [DOI] [PubMed] [Google Scholar]

Articles from Journal of Neurophysiology are provided here courtesy of American Physiological Society

RESOURCES