Knowledge of the molecular mechanisms underlying signaling of mechanical stimuli by muscle spindles remains incomplete. In particular, the ionic conductances that sustain tonic firing during static muscle stretch are unknown. We hypothesized that tonic firing by spindle afferents depends on sodium persistent inward current (INaP) and tested for the necessary presence of the appropriate voltage-gated sodium (NaV) channels in primary sensory endings. The NaV1.6 isoform was selected for both its capacity to produce INaP and for its presence in other mechanosensors that fire tonically. The present study shows that NaV1.6 immunoreactivity (IR) is concentrated in heminodes, presumably where tonic firing is generated, and we were surprised to find NaV1.6 IR strongly expressed also in the sensory terminals, where mechanotransduction occurs. This spatial pattern of NaV1.6 IR distribution was consistent for three mammalian species (rat, cat, and mouse), as was tonic firing by primary spindle afferents. These findings meet some of the conditions needed to establish participation of INaP in tonic firing by primary sensory endings. The study was extended to two additional NaV isoforms, selected for their sensitivity to TTX, excluding TTX-resistant NaV channels, which alone are insufficient to support firing by primary spindle endings. Positive immunoreactivity was found for NaV1.1, predominantly in sensory terminals together with NaV1.6 and for NaV1.7, mainly in preterminal axons. Differential distribution in primary sensory endings suggests specialized roles for these three NaV isoforms in the process of mechanosensory signaling by muscle spindles.
NEW & NOTEWORTHY The molecular mechanisms underlying mechanosensory signaling responsible for proprioceptive functions are not completely elucidated. This study provides the first evidence that voltage-gated sodium channels (NaVs) are expressed in the spindle primary sensory ending, where NaVs are found at every site involved in transduction or encoding of muscle stretch. We propose that NaVs contribute to multiple steps in sensory signaling by muscle spindles as it does in other types of slowly adapting sensory neurons.
- sensory encoding
- muscle spindle
- voltage-gated sodium channels
The sensory neurons supplying muscle spindle receptors provide the central nervous system with information that is critical to proprioceptive function (Proske and Gandevia 2012). This information originates from ion channels engaged in mechanotransduction or action potential encoding, and significant recent advances have been made in their identification (Bewick and Banks 2015; Lin et al. 2016; Woo et al. 2015). However, knowledge remains incomplete with regard to the ion channels responsible for sustaining repetitive firing, i.e., tonic firing of spindle afferents in response to static muscle stretch. Some insight was gained from our recent discovery that tonic firing by muscle spindle group Ia afferents can be selectively blocked pharmacologically (Vincent et al. 2015). The block was achieved by two different drugs, riluzole and phenytoin, which apart from their multiple drug effects, share antagonist action on slowly inactivating Na currents, also known as Na-persistent inward currents (INaP; (Lampl et al. 1998; Schuster et al. 2012; Xie et al. 2011; Zeng et al. 2005). INaP is a plausible candidate contributor to tonic firing of muscle spindle receptors, because it participates in sustaining repetitive firing in a wide variety of neurons (Do and Bean 2003; Harvey et al. 2006; Raman et al. 1997), including the large-diameter class of dorsal root ganglia (DRG) somas that give rise to muscle spindle afferents (Baker and Bostock 1997; Xie et al. 2011). Collectively, these observations led us to hypothesize that a NaV in muscle spindle receptors contributes to the sensory encoding mechanisms that produce tonic firing. Here, we test the necessary condition that NaV channels are present in muscle spindle receptors.
Our investigation focused on NaV channels that are TTX-sensitive (TTX-S), because they, unlike TTX resistant (TTX-R) NaV channels, are necessary for the production of muscle stretch-evoked firing by muscle spindle afferents (Hunt et al. 1978). Multiple TTX-S voltage-gated Na channels qualify as potential sources of INaP. NaV1.6 stands out among them, because it produces a particularly large INaP (Chen et al. 2008; Rush et al. 2005) and is expressed by large-diameter DRG neurons, which include Ia afferents (Black et al. 2002). Additionally, NaV1.6 is present in slowly adapting mechanosensitive receptors in skin, gut, and inner hair cells, and it is necessary for tonic firing by stretch-sensitive afferents (Feng et al. 2015; Hossain et al. 2005; Lesniak et al. 2014). Although collectively, these observations point to NaV1.6, it is not the only candidate. NaV1.1 is also expressed by large-diameter DRG neurons and is known to participate in mechanosensation in the skin and the gut (Black et al. 1996; Osteen et al. 2016). NaV1.7 is also present in stretch-sensitive colorectal afferent endings (Feng et al. 2015) and is coexpressed with NaV1.6 at nodes of Ranvier in a subpopulation of small-diameter Aδ myelinated fibers in the sciatic nerve, 40% of which are known to be mechanosensitive only (Black et al. 2012; Cain et al. 2001). Here, we provide, what to our knowledge, is the first demonstration of NaV channel expression in the primary sensory endings of muscle spindle afferents. The presence of NaV1.6 at the heminodes, which are the presumptive spike-encoding regions, potentially provides the primary ending with the capacity to support repetitive firing in response to static mechanical stimulation. The similarity we find in the NaV1.6 expression by primary endings in cats, rats, and mice is consistent with the similarity that we find in the firing profiles of Ia afferents in these three species. In addition to NaV1.6 immunoreactive (IR), we found NaV IR for isoforms 1.1 and 1.7, and the differential distribution of these three suggests specialized roles for each in the process of sensory signaling by muscle spindle group Ia afferents.
The IR studies presented here were restricted to tissue recovered from normal animals, and all procedures were approved by the Georgia Institute of Technology Animal Care and Use Committee. The soleus muscles were recovered from nine deeply anesthetized (isoflurane 1.5–2.5% in 100% O2) adult animals, including one female cat (3.0 kg), five female Wistar rats (200–300 g), and four C57BL6/J mice (24–28 g). Animal use approvals by Institutional Care and Use Committees for those used in electrophysiological studies were as follows: Emory University for cats and Wright State University for rats and mice. At the end of all terminal experiments, all animals were overdosed with isoflurane (5%) and exsanguinated.
Soleus muscles were placed in 0.1 M PBS containing 4% paraformaldehyde for 40 min, and after a brief wash in a 0.1 M PBS, they were incubated in 0.1 M PBS containing 20% sucrose at 4°C overnight for cryoprotection. Sections of 60-µm thickness were cut using a Cryostat (Leica). Muscle sections were labeled in triplicate with a rabbit polyclonal antibody directed against an epitope of the rat NaV1.6 channel (ASC-009, 1:100; Alomone Laboratories). The specificity of this antibody has been previously demonstrated by complete absence of IR in NaV1.6 KO mice (Black et al. 2002; Royeck et al. 2008). A rabbit polyclonal antibody directed against an epitope of the rat NaV1.1 channel and NaV1.7 channel (ASC-001 and ASC-008, respectively; 1: 100; Alomone Laboratories) were also used. All NaV antibodies were visualized using an Alexa Fluor 488-conjugated secondary antibody (1:100, ThermoFisher Scientific). A rat monoclonal antibody against the fragment 70–89 of the classic human myelin basic protein (MBP) sequence (MAB386, 1:100; Millipore) was used to detect myelin and was visualized using an Cy5-conjugated secondary antibody (1:100, Jackson Immunoresearch Laboratories), and a chicken polyclonal against the heavy-chain (200 kDa) neurofilament protein (NF-H Aves Laboratories) was visualized using a Fluorescein-conjugated secondary antibody (1:100, Aves Laboratory). The slide mounting media (Vectashield) included DAPI in order to label cell nuclei of intrafusal muscle fibers.
Z-axis stacks of images of muscle spindle primary endings were constructed by sequentially imaging at high magnification using an Olympus FV1000 confocal microscope with a ×60 objective (NA, 1.35, oil-immersion). Stacks of images were processed and analyzed using Amaris (Bitplane) imaging software. Figures present images as flat projections of the sum of the z-axis optical slices.
Electrophysiological measurement of Ia afferent firing responses to muscle stretch-supplemented comparison of NaV1.6 staining across the three species. These data were collected, although not published, in our previous studies of the cat (Prather et al. 2011) and the rat (Vincent et al. 2016). New data were obtained from four adult mice deeply anesthetized by isoflurane inhalation throughout the entire terminal experiment, beginning with induction in a closed chamber (4–5% in 100% O2) and continuing with delivery via a tracheal cannula (1.5–2.0% in 100% O2). Subcutaneous injections of lactated Ringer solution were given to maintain adequate fluid levels and blood pressure. Respiratory rate, heart rate, oxygen saturation, and Pco2 were monitored to ensure anesthesia and overall animal health. Body temperature was recorded via a rectal probe and was maintained between 36°C and 38°C with heated water pads and a heat lamp.
The left hind limb and lumbosacral spinal cord were surgically exposed, as needed, to record the firing of individual sensory neurons in response to controlled muscle stretch (Bullinger et al. 2011). Briefly, each mouse was placed in a rigid stereotaxic frame, with legs secured and ankle and knee joints fixed at angles of ~90° and 120°, respectively. The left triceps surae muscles (medial and lateral gastrocnemius and soleus muscles) were partially freed from surrounding connective tissue, marked for their resting length measured with the leg in its fixed joint positions, and then detached from the calcaneus. The distal end of the severed Achilles tendon was tied directly to the lever of a force and length-sensing servomotor (model 305B-LR; Aurora Scientific), which provided for the application of controlled muscle stretch, while recording muscle length and force. Triceps surae nerves were freed from the surrounding tissue and were placed on a unipolar stimulating electrode, and other nerves in the left hind limb were crushed, including common peroneal, sural, and posterior tibial nerves. A laminectomy was performed from T10-S1, and the dura mater was removed to expose the spinal cord and dorsal roots. Skin flaps were tied up in the back and hind limb to create pools for mineral oil, to prevent the tissue desiccation.
Action potentials from the axons of individual muscle spindles were recorded with glass microelectrodes (filled with 2 M K+ acetate) driven into dorsal roots supported on bipolar hook electrodes. Sensory axons were randomly sampled and selected for recording when electrical stimulation of the triceps surae nerves evoked orthodromic action potentials. Sensory axons were identified as muscle spindle afferents by the pause in firing observed during electrically evoked twitch contractions of the triceps surae muscles and were further subclassified as Ia when their firing entrained to high-frequency vibration (250 Hz, 80 μm) of the muscles. Firing was then recorded in response to ramp-hold-release stretches (0.25 mm, 50-ms ramp and release, 1-s hold) from resting muscle lengths determined for ankle angle 90° and knee angle 120°. These stretch parameters were matched to those used for cat and rat by their percentage of whole muscle resting lengths; we measured ~88 mm for cat, 44 mm for rat, and 13 mm for mice. Stretch amplitudes of 2 mm, 1 mm, and 0.25 mm for cats, rats, and mice, respectively, achieved 2% strain and 40%/s strain rate for all species. Intra-axonal records of action potentials and of muscle length and force were digitized (20 kHz) and stored on a computer for later analysis using Spike2 software.
Data pooled from individuals of each species were compared statistically using one-way ANOVA (Statistica Software). The level of significance for all statistical tests was set at P < 0.05. All values are reported means ± SD.
The main objective was to identify the distribution of NaV1.6 within primary sensory endings. Results were tested for generalizability across the three mammalian species, in which muscle spindle receptors have been most thoroughly examined (Banks 2006). Study was restricted to the same muscle, the soleus, in all three species, to minimize variability introduced by muscle-specific differences (Banks et al. 2009). We concentrated analysis on 34 muscle spindle primary sensory endings (NaV1.6: rat, n = 17; mouse, n = 8; cat, n = 3; NaV1.1: rat n = 3; NaV1.7: rat n = 3) and overlooked secondary endings, which were less readily resolvable.
Spindle Structure and Ia Afferent Innervation
Findings were entirely consistent with comprehensive descriptions compiled from earlier studies (Bewick and Banks 2015). In all preparations, muscle spindles sensory innervation was identified using a NF-H antibody for the protein known to be highly enriched at this location (Lin et al. 2016; Nahirney and Ovalle 1993). The detailed structure and complex spatial distribution of neurons innervating muscle spindles were clearly labeled by NF-H IR in rat, cat, and mouse (see Figs. 1, 2, and 3, respectively). Primary endings were identified as parent axons and branches innervating sensory terminals expressing their distinctive annulospiral form in association with intrafusal muscle fibers. The parent axons of Ia afferents branched to ultimately innervate multiple sensory terminals (e.g., Figs. 1A1a, 2A1a, and 3A1a). Labeling for MBP IR distinguished myelinated and unmyelinated portions of Ia afferents axons (e.g., Figs. 1A1b and 3A1b, and Fig. 2A1c). Sensory terminals, which are the mechanotransducer sites (Bewick and Banks 2015), and their connecting preterminal axon branches were unmyelinated. Heminodes, the alleged site of action potential initiation and encoding of mechanical stimuli (Banks et al. 1997; Quick et al. 1980), were identified in preterminal axons adjacent to the termination of axonal myelination (e.g., Figs. 1A1b and 3A1b, Fig. 2A1c and 2A1d, arrowheads).
Multiple structures within the primary endings stained for NaV1.6. The specificity of the NaV1.6 antibody was demonstrated first by the almost complete absence of NaV1.6 immunoreactivity in muscle spindles stained when the antibody was preabsorbed with the antigenic control peptide (Fig. 1E1 and E2) and by its presence at the nodes of Ranvier of axons located within the soleus muscle nerve (Fig. 1, F1 and F2), where it is known to be the predominant sodium channel (Caldwell et al. 2000).
Heminodes and preterminal axons.
The number of heminodes examined per primary ending ranged from 1 to 4. Every terminal was associated with one heminode, and in many cases, a single heminode was shared by multiple terminals. NaV1.6 IR was concentrated at the heminodes in every primary terminal (n = 34) without exception. The boxes outlined in Fig. 1A and the corresponding insets at higher magnification plainly illustrate NaV1.6 IR at the heminodes in rat primary terminals. For example, Fig. 1, A1c and A1d shows high magnification of a heminode positioned just before the preterminal axons divided to innervate different sensory endings. The same pattern was found in all three species (Fig. 1, A1c, B2a, C3a; Fig. 2, A1b, B2a, B3a; and Fig. 3, A1c). In addition, we commonly observed NaV1.6 IR in the unmyelinated preterminal axons at sites distal to the heminodes (e.g., green arrows in Fig. 1A1b, Fig. 2C, and Fig. 3A1b); however, it was significantly weaker than the one observed for NaV1.7 (see results below).
Sensory terminals and inter-nuclear spaces.
NaV1.6 IR was evident in all sensory endings sampled from all species (Fig. 1, A1b and C, Fig. 2, A1c and C, and Fig. 3A1b). We commonly observed, but did not quantify, that NaV1.6 IR staining was consistently brighter in the sensory terminals associated with intrafusal bag fibers (identified by clustering of nuclei labeled by DAPI) than with chain fibers (See Fig. 1D). NaV IR distribution within the annulospiral sensory terminal was uneven. This pattern of expression is likely due to an uneven antibody penetration occurring during the immunostaining of a large and complex structure, such as the sensory terminal. In other less complicated structures such as the heminode, NaV IR was invariably robust in all primary endings studied. Although the exact structure labeled by NaV1.6 IR within the sensory terminals was not identified in the present study, we determined that it did not completely colocalize with NF-H (Fig. 1, A1a and B; Fig. 2, A1a and B, and Fig. 3, A1a). In all primary endings sampled, we also observed strong NaV1.6 IR between the nuclei of intrafusal muscle fibers, particularly, in juxtaequatorial regions (arrowheads in Fig. 1D).
Ia afferent firing properties in the cat, rat and mouse.
The firing responses of Ia afferents to passive-muscle stretch were compared across species using unpublished data collected from earlier studies in this laboratory, including Prather et al. (2011) and Vincent et al. (2016). Data were selected from a small set of ramp-hold-release stretches having the same strain (2%) and strain rates (40%/s) in each species (see methods). Fig. 4 illustrates representative firing responses. Overall, firing profiles were qualitatively similar, exhibiting a high-frequency initial burst of three or four spikes followed by firing in dynamic stretch that adapted to lower rates during static stretch. Firing properties are compared for pooled samples of Ia afferents in each of the three species in Table 1. Tonic firing extended virtually throughout the full 1 s of the stretch hold phase in all species, corroborating earlier in vivo studies of cat and rat, as well as findings for mouse Ia afferents, which have been studied systematically in vitro (Wilkinson et al. 2012) and to a limited extent in vivo (Nakanishi and Whelan 2012). Remaining parameters exhibit quantitative differences, which in some comparisons, reach statistical significance between species. Measures of dynamic sensitivity (dynamic firing rate and dynamic index) and tonic firing rates were all significantly greater in the mouse. We attribute these differences to species variations, since the responses were studied under similar experimental conditions and in response to muscle stretch applied with identical strain and strain rates in all species. Additional study will be required to determine whether these differences in firing behavior relate to possible differences in levels of NaV expression.
NaV1.1 and NaV1.7 IR
Heminodes and preterminal axons.
NaV1.1 IR was completely absent at the heminodes and preterminal axons in all primary endings in which it was examined. The box outlined in Fig. 5A and the corresponding inset at higher magnification clearly show the lack of NaV1.1 IR at the termination of axon myelination (Fig. 5, A1b–A1d, B2a, and C2a, white arrowheads), where the heminode normally is positioned and in the preterminal axons innervating the different terminals (green arrows in Fig. 5, A1a). NaV1.7 IR on the other hand was consistently present in the preterminal axons in every primary ending examined (Fig. 6, A1a and A1b, B and C). NaV1.7 IR extended from the end of the axonal myelination, i.e., the heminode area, and along the preterminal axons all the way to the sensory terminals (arrowhead in Fig. 6A1b and green arrows in C). Figures 5 and 6 show at higher magnification that NaV1.7 IR is constrained within the boundaries of the preterminal axon labeled with NF-H (Fig. 6A1c and A1d; B2a and B2b, and C3a and C3b, green arrows). In all primary endings analyzed, the diameter of the preterminal axon stained with NF-H and NaV1.7 at the end of the axonal myelination ranged between 1.1 and 1.3 µm (see Fig. 6A1d, arrowhead). This means that at the heminode, NaV1.7 channels in the axon are likely to be flanked by NaV1.6 channels (cf. Fig. 1A1d and Fig. 6A1D).
Sensory terminals and internuclear spaces.
NaV1.1 and NaV1.7 IR were present in all of the sensory endings examined (Figs. 5 and 6, A1b and C). Although we did not quantify NaV1.1 IR, it appeared consistently brighter than NaV1.7 IR in all the sensory terminals associated with intrafusal bag fibers (cf. Fig. 5, A1b and C with Fig. 6, A1b and C). In addition, NaV1.1 IR, like NaV1.6 IR, was clearly seen between the nuclei located in the juxtaequatorial regions of intrafusal muscle fibers (arrowheads in Fig. 5D), and contrary to NaV1.6 and NaV1.7 IR, the IR of this channel was consistently observed in the sensory terminals associated with chain intrafusal fibers (Fig. 5D). Our findings follow those showing the targeting of Na channels to specific regions in sensory axons of other vertebrates (Waxman et al. 1972).
NaV1.6 IR observed in the preterminal axons and in sensory terminals might have been a product of cross IR with NaV1.1 and/or NaV1.7 antigens. To test this possibility, we preabsorbed the NaV1.6 antibody with the antigenic control peptides for NaV1.1 and NaV1.7 and stained a soleus muscle using the same protocol described in methods. NaV1.6 IR was clearly present and distributed in the patterns described above using the regular protocol (Fig. 7, A1a and A1d, arrowheads), thereby excluding the possibility that NaV1.6 cross-reacted with the other NaV channels analyzed.
Results presented here provide the first evidence of TTX-S NaV channel expression in muscle spindle primary endings. Study focused on NaV1.6 IR, which we found in cat, rat, and mouse at heminodes, preterminal axons, and sensory terminals, i.e., at every site involved in transduction or encoding of muscle stretch. Its presence, particularly at the spike-encoding region establishes NaV1.6 as a candidate source of INaP and possibly also tonic firing of the primary ending. Consistent with this notion, we find uniformity both in NaV1.6 expression and in the occurrence of tonic firing for all three species. Exploration of additional NaV isoforms in the rat revealed isoforms 1.1 and 1.7 differentially distributed at sites within the primary ending. The NaV expression patterns that we observed suggest a prominent role for these three NaV isoforms in multiple steps of mechanosensory signaling (see Bewick and Banks 2015), and below, we propose how the known biophysical properties of NaV channels might contribute to various steps in sensory signaling by mammalian muscle spindles.
Positive immunostaining for NaV1.6 was found at all heminodes in every spindle primary ending sampled in this study. This finding follows demonstrations of NaV1.6 clustered in the heminodes at two other types of mechanosensory afferents, one that innervates outer hair cells in the cochlea (Hossain et al. 2005) and another that supplies Merkel touch receptors in the skin (Lesniak et al. 2014). Heminodes are thought to be sites where receptor currents representing features of their mechanical stimuli are encoded in the firing rates and patterns of action potential trains (Bewick and Banks 2015; Loewenstein and Rathkamp 1958; Quick et al. 1980). These encoding functions appear similar to those occurring at the axon initial segments, where synaptic currents are integrated and translated into action potential firing (Clark et al. 2009) and where NaV1.6 is also found concentrated in a number of neuron types (Brocard et al. 2016; Hu et al. 2009; Osorio et al. 2010; Royeck et al. 2008). The presumptive encoding function of heminodes appears well served by the three voltage-gated currents mediated by NaV1.6. The transient Na current (INaT) is large, but brief, initiated by small depolarization, and, thereby, well suited to facilitate action potential initiation (Royeck et al. 2008). A resurgent current (INaR) produced by NaV1.6 repriming kinetics is present in some large-diameter DRG neurons (Cummins et al. 2005), has the potential to support firing at high firing rates (Khaliq et al. 2003; Mercer et al. 2007) and can aid in sustaining repetitive firing, as has been shown for Purkinje cells (Raman et al. 1997). Both INaT and INaR are recognized for their capacity to support action potential generation at axon initial segments, where NaV1.6 is shown to aggregate (Hu et al. 2009; Osorio et al. 2010; Royeck et al. 2008). A third current mediated by NaV1.6, INaP, also activates near resting membrane potential threshold, is larger in comparison with some other NaV isoforms (Chen et al. 2008), and persists >600 ms (Crill 1996; Raman et al. 1997). At the heminodes, INaP might support repetitive firing and adjust firing rate in relation to the magnitude of arriving receptor potentials, as it does for synaptic currents in other neurons (e.g., Hultborn et al. 2003). A critical role for NaV1.6 in supporting tonic firing was recently assigned to the sensory endings of stretch-sensitive colorectal afferents (Feng et al. 2015). For spindle primary endings, the presence of NaV1.6 IR aggregations at heminodes qualifies it to mediate inward Na currents, which in other systems described above support both initiation and repetitive occurrence of action potentials.
The robust immunostaining that we found for NaV1.6 in the annulospiral sensory terminals of the spindle primary endings was unexpected. The capacity for NaV1.6 to generate action potentials seems misplaced at these terminals, which are not generally thought to be sites of action potential generation (Banks et al. 1997; Bewick and Banks 2015; Quick et al. 1980). The spindle terminals are instead the probable sites of mechanotransduction, where muscle stretch activates aggregations of mechanically gated ion channels to produce receptor currents (Bewick and Banks 2015). In the absence of evidence for mechanical gating, NaV1.6 is not expected to contribute in mechanotransduction. Nonetheless, we recognize a potential role for NaV1.6 in the sensory terminals that could be critical for sensory signaling. NaV1.6 INaP might assist in conveying the receptor potential from its origin in the unmyelinated sensory terminal to the site of action potential generation in the heminodes. The electrotonic pathway from the sensory terminal through the preterminal axons might result in considerable decrement in receptor current before its arrival at the heminodes. Although the absence of information on the input conductance of spindle sensory terminals precludes estimating the magnitude of current loss, some fraction of the receptor is expected to be lost to electrotonic decay in the terminal’s expansive membrane area of ~17,500 to 18,500 µm2 measured in the adult cat (Banks et al. 1982), and to conductance shunting occurring when mechanically gated channels are activated. INaP has the potential to compensate for electrotonic reduction of the receptor current, just as it is proposed to do for synaptic currents that, without amplification, may be substantially reduced as they spread through the dendritic arbors of neurons (Crill 1996), and therefore, less effective in initiating or modulating neuron firing (Binder 2002; Hultborn et al. 2003). In the sensory terminals of spindle primary endings, NaV1.6 is well positioned through its close proximity to mechanotransducers, to be activated by receptor potentials and to generate INaP that boosts, i.e., amplifies, the receptor current. The amplification might apply over the full time course of the receptor current, owing to the persistence of INaP. In this way, the receptor current’s representation of all phases of a discrete muscle stretch, e.g., early dynamic and later static phases would be preserved. We suggest further that NaV1.6 concentration in segments of preterminal axons might provide additional amplification stations for receptor current along its electrotonic path to the heminodes. Alternatively, NaV1.6 might support action potentials in the preterminal axon (Ito and Ito 1976), as it does in unmyelinated axons of other neurons (Black et al. 2002).
Our IR studies also revealed the presence of two other NaV isoforms in the primary endings of muscle spindles of the rat soleus, the NaV1.1 and NaV1.7 isoforms. We selected these two isoforms because, on the basis of the available evidence, the presence of the others at this location is unlikely. NaV1.2 is mainly expressed in the central nervous system, NaV1.3 is primarily expressed in the central neurons during embryonic and postnatal development, NaV1.4 primarily in skeletal muscle, and NaV1.5 in cardiac muscle (Beckh 1990; Catterall et al. 2005; Goldin 2001; Rush et al. 2007). The other two NaV isoforms, NaV1.8 and NaV1.9, are found mainly in small sensory neurons. Functionally, it seems unlikely that repetitive firing by spindle afferents requires either NaV1.8 or NaV1.9, because both isoforms are resistant to tetrodotoxin (TTX-R; Catterall et al. 2005), which is known to completely block the electrical response of muscle spindles to stretch (Hunt et al. 1978). We acknowledge that although unnecessary for firing, TTX-R NaV channels may play a role, especially since NaV1.8 expression has been shown in dissociated-cultured, large-diameter neurons in DRG, as well as in a large group of neurons with myelinated A fibers (Ramachandra et al. 2012; Shields et al. 2012).
As mentioned earlier, NaV IR for both 1.1 and 1.7 isoforms was clearly observed in the primary endings of Ia afferents of the rat soleus. Their distribution, however, varied significantly. Strong NaV1.1 IR was observed at the sensory terminals of the primary endings of bag fibers, whereas NaV1.7 IR was the weaker of the two isoforms at this location. NaV1.1 IR predominated in the sensory terminal of chain fibers. NaV1.7 IR, on the other hand, was very strong in the presynaptic axon extending from the heminode areas to the sensory terminals where NaV1.1 IR was completely absent. In the sensory terminals NaV1.1 is likely to be another important source of INaP. Purkinje neurons from NaV1.1 KO mice, have significantly reduced INaP and resurgent inward currents compared with neurons from wild-type animals (Kalume et al. 2007). Therefore, and as discussed earlier for NaV1.6, NaV1.1 is also well positioned to boost the receptor currents produced by mechanotransducers, especially in intrafusal chain fibers, where it predominates. It is interesting that deletion of NaV1.1 or NaV1.6 renders mice ataxic, significantly affecting limb coordination and motor reflexes (Kalume et al. 2007; Raman et al. 1997). The motor disorders were associated with the loss of INaP and INaR in Purkinje neurons. Our findings suggest that ataxia in these KO mice might also have resulted from the deletion of NaV1.1 and NaV1.6 from muscle spindle primary endings.
NaV1.7 was clearly the predominant isoform in the preterminal axons. This finding adds to what others have reported in stretch-sensitive colorectal afferent endings (Feng et al. 2015). In addition to its presence in mechanosensitive receptors in the gut, coexpression of NaV1.7 and NaV1.6 has been shown at nodes of Ranvier in a subpopulation of small-diameter Aδ myelinated fibers in sciatic nerve (Black et al. 2012), 40% of which are mechanosensitive only (Cain et al. 2001). Thus, the available evidence, together with our results, points to a probable, but unrecognized role, of this channel in the process of mechanosensory signaling. NaV1.7 is known to produce ramp currents (Cummins et al. 1998) and, via this process, has the capacity to amplify subthreshold potentials, and to decrease the threshold for action potential initiation (Dib-Hajj et al. 2013; Rush et al. 2007).
NaV expression in the muscle spindle was not restricted to primary endings. In a mouse spindle, we were confident in our positive identification of one secondary ending, in which we found clear expression of NaV1.6 in the heminode and terminal endings. This isolated observation may indicate that the roles we propose for NaV1.6 extend to secondary endings. In addition, NaV1.6 and NaV1.1 were observed between nuclei in the juxta regions of bag and chain intrafusal muscle fibers. The intranuclear space is also found to express the provisional mechanotransducing channel, α-ENaC (see Simon et al. 2010). The function of Na channel expression in intranuclear sites at the moment is unknown.
Physiological data demonstrate qualitative similarity among these three species. To our knowledge, this comparison has not been reported. With respect to tonic firing, it establishes for the first time that Ia afferents in these species are indistinguishable. The differences in other parameters are interesting, and we attribute them to species differences, since the responses were obtained using strain and strain rates that were identical in all species.
The distribution described above for NaV isoforms is summarized in Fig. 8. This figure is an adapted version of the model advanced by Bewick and Banks (2015), which represented knowledge accumulated up to that time about the molecular mechanisms of primary ending excitability. To this model, we add new details about NaV channel presence in sensory regions involved in encoding by muscle spindles. In the terminals, close proximity to mechanotransducers positions NaV isoforms, primarily 1.1 and 1.6, to detect and amplify receptor potential current. In the preterminal axons, the strong presence of NaV1.7 may serve to boost the receptor current along its electrotonic path to the heminode (Cummins et al. 1998; Rush et al. 2007). Lastly, at the heminode, the presumed site of action potential generation (Bewick and Banks 2015; Loewenstein and Rathkamp 1958; Quick et al. 1980), NaV1.7 may act as a “threshold channel” (Dib-Hajj et al. 2013), while NaV1.6 takes primary responsibility for action potential initiation, as well as for encoding receptor potential features into firing rate modulation and tonic firing.
The discussion above attributes particular importance to INaP. This focus originated from our interest in identifying mechanisms that yield the tonic firing that characterizes spindle afferents, i.e., sustained repetitive firing in response to steady depolarization produced by static muscle stretch (Hunt et al. 1978). This mechanism is unknown and is not expressly indicated in ion-channel models of primary spindle endings (cf. Bewick and Banks 2015). The potential importance of INaP to this mechanism is suggested by our recent findings that the drugs riluzole and phenytoin, which share only blockade of INaP among their many effects (Lampl et al. 1998; Urbani and Belluzzi 2000; Xie et al. 2011; Zeng et al. 2005), result in substantial reduction of primary spindle afferent firing in vivo when administered acutely to adult rats (Vincent et al. 2015). We also reported that, as a predicted expression of persistent inward current (Harvey et al. 2006), the threshold for repetitive firing increases for ramp stretch at slow velocity (Vincent et al. 2016). These findings suggested to us that channels mediating INaP are expressed peripherally by muscle spindle afferents, which belong to the group of large-diameter DRG cells that respond to intrasomatic current injection by producing INaP and repetitive firing, both of which are blocked by riluzole (Xie et al. 2011). If INaP does operate in spindle endings, possibly mediated by NaV1.6 and NaV1.1, then considerations presented earlier in the discussion indicate that it might support tonic firing by amplifying sustained depolarization of the receptor potential and/or by supporting repetitive firing at heminodes. Impairment of either process has the potential to produce the reduction of tonic firing that we observe with antiepileptic drugs, as well as with chronic treatment of rats with the anticancer agent oxaliplatin (Bullinger et al. 2011; Vincent et al. 2015, 2016). Therefore, we hypothesize that these agents reduce tonic firing by impairing the INaP mediated by either NaV1.6, NaV1.1, or both in spindle primary endings.
This work was supported, in part, by funding from National Institute of Neurological Disorders and Stroke Grant P01NS-057228.
No conflicts of interest, financial or otherwise, are declared by the authors.
T.C.C. and D.I.C. wrote the manuscript; D.I.C. performed immunohistochemistry and confocal imaging; J.A.V performed electrophysiological recordings.
We gratefully acknowledge the consultations and critical review of this manuscript by Dr. Robert Banks and Dr. Guy Bewick. We also thank Dr. Francisco Alvarez and Dr. Martin Pinter for facilitating some of the equipment that allowed this study to be possible.
- Copyright © 2017 the American Physiological Society