Current injection and receptor-mediated excitation produce similar maximal firing rates in hypoglossal motoneurons

Abstract

The maximum firing rates of motoneurons (MNs), activated in response to synaptic drive, appear to be much lower than that elicited by current injection. It could be that the decrease in input resistance associated with increased synaptic activity (but not current injection) might blunt overall changes in membrane depolarization and thereby limit spike-frequency output. To test this idea, we recorded, in the same cells, maximal firing responses to current injection and to synaptic activation. We prepared 300 μm medullary slices in neonatal rats that contained hypoglossal MNs and used whole-cell patch-clamp electrophysiology to record their maximum firing rates in response to triangular-ramp current injections and to glutamate receptor-mediated excitation. Brief pressure pulses of high-concentration glutamate led to significant depolarization, high firing rates, and temporary cessation of spiking due to spike inactivation. In the same cells, we applied current clamp protocols that approximated the time course of membrane potential change associated with glutamate application and with peak current levels large enough to cause spike inactivation. Means (SD) of maximum firing rates obtained in response to glutamate application were nearly identical to those obtained in response to ramp current injection [glutamate 47.1 ± 12.0 impulses (imp)/s, current injection 47.5 ± 11.2 imp/s], even though input resistance was 40% less during glutamate application compared with current injection. Therefore, these data suggest that the reduction in input resistance associated with receptor-mediated excitation does not, by itself, limit the maximal firing rate responses in MNs.

activation of mammalian motoneurons (MNs) by intracellular current injection elicits steady-state firing rates that increase near linearly with increasing current up to high rates, typically exceeding 50 impulses (imp)/s and occasionally above 100 imp/s (Granit et al. 1963, 1966; Jodkowski et al. 1988; Kernell 1965a; Schwindt 1973; Schwindt and Calvin 1972; Schwindt and Crill 1982). In contrast, the discharge of MNs and motor units appears to saturate at substantially lower rates (often <20–30 imp/s) in response to diverse sources of relatively steady-state synaptic inputs in cats (Alvord and Fuortes 1953; Brownstone et al. 1992; Burke 1968; Cordo and Rymer 1982; Denny-Brown 1929; Granit 1958; Granit et al. 1960; Lee et al. 2003; Prather et al. 2002; Zajac and Young 1980), monkeys (Palmer and Fetz 1985), and humans (Bailey et al. 2007; Bracchi et al. 1966; De Luca and Contessa 2012; De Luca et al. 1982; Fuglevand et al. 2015; Kiehn and Eken 1997; McGill et al. 2005; Monster and Chan 1977; Moritz et al. 2005; Mottram et al. 2009, 2014).

The cause of this disparity in firing-rate capacities of MNs activated by current injection and by synaptic drive is not known. One possibility is that the extent of depolarizing current delivered synaptically is simply less than that which can be administered directly through an intrasomatic microelectrode. Accordingly, an experimental manipulation that provides supplementary synaptic excitation above that normally delivered synaptically should lead to an increase in MN firing rate. We tested this idea recently and found that the augmentation of a descending excitatory drive to MNs with peripheral excitation (mediated by tendon vibration) did not change firing rates of motor units that had already saturated (Fuglevand et al. 2015). This result implied that intrinsic, rather than extrinsic, mechanisms are likely responsible for curbing increases in firing rate during synaptic activation of MNs.

What intrinsic mechanism(s) could act to blunt increases in firing rate when MNs are driven synaptically but not in response to current injection? One possibility is diminution of cell-input resistance associated with opening of transmitter-gated channels during synaptic drive but not occurring during current injection. Experimental findings (Berg et al. 2007, 2008; Coombs et al. 1955; Llinas and Terzuolo 1964; Paré et al. 1998; Powers and Binder 2000) and computer simulations (Barrett 1975; Binder et al. 1996; Destexhe and Paré 1999; Kuhn et al. 2004) indicate that reduction of input resistance during synaptic drive can be significant. Such a decrease in input resistance should, in turn, attenuate the effective change in membrane potential caused by a given increment in synaptic current in accordance with Ohm's law. Furthermore, a decrease in input resistance should also abbreviate the time course of postsynaptic potentials due to a decrease in membrane time constant and thereby lessen temporal summation. Collectively, decreases in membrane input resistance should undermine the efficacy by which excitatory synaptic inputs are transformed into spiking output as synaptic activity increases, which in turn, may limit the spike-frequency capacity of MNs (and other neurons as well).

To test this possibility, we compared the maximum firing rates evoked by current injection with that elicited by glutamate receptor excitation in the same MNs. Despite substantial reduction in input resistance during glutamate activation of MNs, maximal firing rates were not different for the two methods of exciting MNs. Therefore, these results suggest that changes in input resistance alone cannot account for the limit in maximal firing rates seen during the synaptic drive of MNs.

METHODS

We used rat hypoglossal MNs (HMNs) to compare, in the same cells, firing-rate responses with current injection and glutamate application. HMNs are readily identifiable, large (20–60 μM) stellate cells that reside in the hypoglossal motor nucleus of the medulla and innervate tongue muscles. Their dendritic arbors extend primarily in the transverse plane, making them relatively less susceptible to damage associated with transverse slicing compared with spinal MNs.

Animals.

We used whole-cell patch-clamp electrophysiology and recorded from single HMNs (n = 24) in medullary slices (n = 15). All procedures were in accordance with protocols approved by the University of Arizona's Institutional Animal Care and Use Committee. Medullary slice preparations were made by the standard neuroaxis isolation procedure described previously (Pilarski et al. 2011). In brief, in vitro slices were prepared from neonatal (0- to 6-day-old) Sprague-Dawley rats, born by spontaneous birth from pregnant dams [Harlan Laboratories, Envigo (Indianapolis, IN), and Charles River Laboratories (Wilmington, MA)].

Slice preparation.

Pups were anesthetized by hypothermia and quickly decerebrated. A portion of the central nervous system, from the pons to the rostral spinal cord, was rapidly isolated in cool artificial cerebral spinal fluid (aCSF), containing (in mM) 124 NaCl, 3 KCl, 2.4 CaCl₂, 1.3 MgSO₄, 26 NaHCO₃, 1.2 NaH₂PO₄, and 30 d-glucose, with 280–300 mosM osmolarity, which was regulated at pH 7.4 and bubbled with 95% O₂ and 5% CO₂. Following further isolation and cleaning of meninges and vessels, the preparation was fixed on a putty-coated paddle, submerged in chilled aCSF, and trimmed rostrally to caudally by transversely slicing with tissue-sectioning equipment (Leica VT1000 Vibratome; Technical Products International, St. Louis, MO). We made 300 μm medullary slices that contained HMNs and then secured slices with a metal mesh in the recording chamber perfused with aCSF warmed to 27 ± 1°C.

Visualization and recordings.

HMNs were identified using infrared-differential interference contrast (IR-DIC) optics (Olympus BX50WI, IR-DIC 40 × 0.75 numerical aperture water immersion and an IR video camera, C25400-07; Hamamatsu, Schüpfen, Switzerland) and recorded using whole-cell patch-clamp electrophysiology (Axoclamp 1D system, Digidata 1320 data acquisition system; Axon Instruments, Molecular Devices, Sunnyvale, CA). Intracellular recording pipettes (3–8 MΩ; outside diameter 1.5 mm, inside diameter 0.75 mm; Sutter Instruments, Novato, CA) were filled with a recording solution containing (in mM) 135 K-gluconate, 4 KCl, 12.5 disodium phosphocreatine, 10 HEPES, 0.375 Na-GTP, and 5 ATP (Mg²⁺ salt), with 7.3 pH and 260–280 mosM osmolarity. Neurons with resting membrane potentials more positive than −50 mV were not included in the analysis. Series resistance was likely not fully compensated at high current levels, and as such, some progressive change in membrane depolarization at high current-injection levels may have been due to inadequate compensation.

Stimulation protocol.

We recorded the maximum firing-rate responses to current injection and to receptor-mediated excitation. To stimulate the cell using current injection, we used a current clamp protocol (pClamp 10.0; Molecular Devices) that delivered 1–4 s and 400-1,200 pA triangular ramp injections of current into the soma. To ensure that maximum firing rates were achieved, peak stimulation current was adjusted to elicit firing-rate responses that increased progressively during the ascending ramp until the cell temporarily stopped spiking and then recovered spiking during the descending ramp. In classical studies of the current-frequency relation in adult MNs involving ascending rectangular current-pulse protocols, the pulse associated with such cessation of spiking typically ensues immediately after the pulse eliciting the maximum firing rate (Granit et al. 1963, Schwindt and Crill 1982). Such blockade of spiking is most likely due to sustained Na-channel inactivation during strong depolarization (Lape and Nistri 2001; Schwindt and Crill 1982). For simplicity, we will refer to this phenomenon as spike inactivation. In some experiments, we also delivered a more traditional rectangular pulse protocol involving 1-s incrementing steps of current and found that maximum steady-state firing frequencies in response to the rectangular pulses were nearly identical to that elicited by the ramp protocol.

Based on observations made during preliminary experiments, we then applied 100 mM glutamate (Sigma, St. Louis, MO), using pressure-injected puffs, into the superfusate (Picospritzer II; Parker Hannifin, Hollis, NH), 60–100 μm away from the soma to activate intensely the main excitatory receptor found on HMNs. Such potent glutamate receptor activation caused substantial depolarization; a progressive increase in firing rates, followed by a period of spike inactivation; and a subsequent return to resting membrane potential. The events occurred with a time course similar to that elicited by triangular current injection.

In separate trials, to estimate changes in input resistance associated with glutamate application, we applied brief hyperpolarizing current pulses before and during glutamate application and measured the associated change in membrane potential. Input resistance was then simply estimated as the ratio of the change in membrane potential to the magnitude of the hyperpolarizing current pulse. In the majority of cells (n = 19) for tests of input resistance, a bias current was used to hyperpolarize the membrane to avoid action-potential generation during input-resistance measurements. In five cells, instead of hyperpolarizing the membrane, TTX (10 nM; Sigma) was bath applied to the preparation during input-resistance measurements to prevent action potentials. There were no obvious differences in input-resistance measurements between the two forms of action-potential blockade, and therefore, the data obtained from these two procedures were pooled.

Data analysis.

Data were analyzed using custom scripts written with Spike2 software (Cambridge Electronic Design, Cambridge, UK). Spike waveforms were identified and used to compute instantaneous firing rates. For both forms of excitation, we calculated the maximum rate based on the mean frequency of the five spikes (4 interspike intervals) immediately preceding spike inactivation. We also measured resting membrane potential, threshold potential for spiking, and threshold current for action-potential generation, before and immediately after the two stimulation protocols were completed. Differences in maximum firing rates between current-injected and glutamate-mediated depolarization were assessed using a paired Student's t-test. Likewise, paired t-tests were used to compare resting membrane potential and spiking thresholds before and after the stimulation protocols. Furthermore, paired t-tests were used to compare input resistance before and during glutamate application. Data are presented as means ± SD, and effects are considered statistically significant for P ≤ 0.05.

RESULTS

Figure 1 shows example recordings from a single HMN activated by the two forms of excitation. In response to triangular current injection (Fig. 1A), firing rate increased progressively, until at high current levels, spikes inactivated. As the membrane then repolarized during the falling phase of the current ramp, spike production was restored. This neuron reached a peak rate of ∼43 imp/s just before spike inactivation.

Fig. 1.

Example firing responses in a hypoglossal motoneuron to current injection (A) and glutamate receptor-mediated excitation (B). Firing rate (top), membrane potential (middle), and stimulus protocol (bottom) involving triangular ramp current injection (A) or glutamate (100 mM) application (arrow) to superfusate (B). Dashed, horizontal lines indicate maximum firing rates for 5 spikes immediately before spike inactivation. C: means (SD) of maximum firing rates elicited by current injection and glutamate receptor-mediated excitation. There was no significant difference in firing rates between the 2 conditions. imp/s, impulses per second.

The response was similar for receptor-mediated excitation, as shown in Fig. 1B. Membrane potential depolarized vigorously and was associated with progressively higher spike rates up until spikes inactivated. The maximal firing rate in this case was ∼47 imp/s. As glutamate was washed away in the perfusate, glutamate concentration dwindled, leading to repolarization of membrane potential and temporary restoration of spiking.

The time courses and trajectories of membrane potential were roughly similar for current injection and glutamate application. However, once the spikes failed during current injection, membrane potential (Fig. 1A) appeared to depolarize, more or less in proportion to injected current. This may, in part, have been due to inadequate compensation for electrode resistance at high levels of injected current. On the other hand, with glutamate application (Fig. 1B), membrane potential tended to level off once spikes failed. This may reflect activation of potent inward currents at highly depolarized levels (Hamm et al. 2010; Lee and Heckman 1998; Li and Bennett 2007). Nevertheless, for both cases (current injection and glutamate application), the upper limit of firing rate appeared associated with spike failure, resulting from sustained, strong depolarization, contributing to Na-channel inactivation.

Overall, maximal firing rates did not differ between the two methods of excitation used to activate the 24 cells (Fig. 1C; current injection 47.5 ± 11.2 imp/s and glutamate 47.1 ± 12.0 imp/s). These maximal firing rates are similar to those reported previously for neonatal HMN in response to current injection (Pilarski et al. 2011). There was no significant difference in the average rate of rise of membrane potential during the prespiking phase for the two conditions (current injection 95.8 ± 26.5 mV/s and glutamate 118.3 ± 91.0 mV/s; P = 0.786). Furthermore, there were no significant changes (P > 0.05) in basic cell properties before vs. after the two stimulation protocols, including resting membrane potential (before: −60.3 ± 11.6 mV, after: −61.5 ± 3.9 mV), resting input resistance (before: 100.2 ± 33.7 MΩ, after: 98.2 ± 23.6 MΩ), threshold potential for spiking (before: −35.5 ± 18.2 mV, after: −32.5 ± 12.8 mV), and threshold current for action-potential generation (before: 520 ± 515 pA, after: 537 ± 418 pA). The absence of change in these properties suggests that the neurons remained reasonably healthy despite the intensity of the intervening activation protocols.

Even though maximum firing responses were not different when generated by current injection and glutamate application, we anticipated that input resistance would decrease during receptor-mediated excitation. Figure 2A depicts an example of such a recording, during which hyperpolarizing pulses (−50 pA) were applied. In this case, TTX was applied to the bath to prevent elicitation of action potentials. Before the glutamate puff (Fig. 2A), hyperpolarizing current pulses evoked relatively large changes in membrane potential (∼4.5 mV), consistent with a relatively high input resistance (∼90 MΩ). The same magnitude of current pulses applied immediately following the glutamate puff, however, caused smaller deflections in membrane potential (∼2.5 mV), indicating a reduction in input resistance (to ∼50 MΩ) during glutamate application. On average (Fig. 2B), input resistance during glutamate receptor activation was 40% less than that before glutamate application (60.4 ± 20.2 MΩ vs. 102.2 ± 26.7 MΩ; P < 0.05). Taken together, our results suggest that despite a substantial reduction in membrane input resistance, the maximal firing rates elicited by glutamate receptor-mediated excitation of HMNs were not different from that evoked by current injection.

Fig. 2.

Change in input resistance during glutamate application. A: example changes in membrane potential to hyperpolarizing current pulses (50 pA), before and following glutamate application (arrow). TTX was applied to bath to prevent action potentials. Input resistance was estimated by measuring the change in membrane potential, divided by current pulse amplitude. B: means (SD) of input resistance, before and immediately after glutamate application. Overall, input resistance decreased by 40% during receptor-mediated excitation compared with control (*P < 0.05).

DISCUSSION

Our results show that the maximal firing rates of HMNs evoked by somatic current injection and that elicited in response to glutamate receptor activation are not different from one another. Indeed, the extent of depolarization achieved in response to application of high-concentration glutamate was sufficient to induce spike inactivation associated with the upper limit of firing rate, despite a substantial reduction in input resistance that would tend to counteract changes in membrane potential. As such, the conductance state alone cannot account for the marked differences in maximal firing rates seen in MNs activated by current injection and that reported for MNs and motor units activated by synaptic input.

Nevertheless, a high conductance state associated with intense synaptic activity would not seem to be without impact on the integrative capabilities of neurons (Chance et al. 2002; Destexhe et al. 2003; Mitchell and Silver 2003; Powers and Binder 2001). For example, background synaptic activity is known to attenuate postsynaptic potentials due to reduced input resistance when the membrane potential is below threshold (Coombs et al. 1955; Kuhn et al. 2004; Llinas and Terzuolo 1964). Likewise, when the membrane potential is above threshold, one might expect current spike-frequency gain to diminish progressively with increased synaptic activity because of the drop in input resistance. Such a reduction in gain would mean that disproportionately more current would need to be delivered to achieve a given increment in firing rate under conditions of high synaptic activity compared with low activity (or during current injection). Surprisingly, however, previous experimental work has shown that current-frequency gain changes little in the presence of intense synaptic activity compared with that without synaptic input (Granit et al. 1966; Kernell 1965b, 1969; Schwindt and Calvin 1973). It may be that the magnitude of the active conductances associated with above-threshold membrane potential and repetitive spiking is simply so much larger than synaptic conductances that it is the dominant factor shaping the current-frequency relation (Kernell 1969; Kernell and Sjöholm 1973; Schwindt and Calvin 1973).

One important concern with the present study is that neonatal MNs (as tested here) are in a developmental stage, and their properties can differ markedly from that of more mature MNs (Berger 2000; Berger et al. 1996; Cameron and Núñez-Abades 2000; Greer and Funk 2005; Rekling et al. 2000). These include, among others, differences in physical dimensions (neonatal MNs are smaller with less-elaborate dendrites) (Núñez-Abades and Cameron 1995), passive electrical properties (neonatal MNs have higher input resistance) (Cameron et al. 2000; Durand et al. 2015; Fulton and Walton 1986; Haddad et al. 1990; Viana et al. 1994), active properties (neonatal MNs have lower densities of fast Na⁺ and high voltage-activated Ca²⁺ channels) (Carlin et al. 2000; Garcia et al. 1998; Jiang et al. 1999; Umemiya and Berger 1994), and repetitive firing responses (neonatal MNs have higher current-frequency gain and lower maximal firing rates) (Fulton and Walton 1986; Viana et al. 1995). Therefore, it seems reasonable to question whether neonatal MNs provide an apt model of firing-rate saturation observed in adult MNs. In the present study, we compared maximal firing rates with both current injection and receptor-mediated excitation in the same neurons under otherwise identical experimental conditions. As such, this allowed us to conclude that maximal firing rates are not different under these two forms of excitation for neonatal MNs, despite substantial, additional reduction in input resistance during glutamate application. Whereas it seems likely that these results will pertain to adult MNs, this awaits confirmation.

If changes in input resistance cannot readily account for the differences typically observed in maximal firing rates for MNs activated by current injection and in response to synaptic activity, then what can? One possibility is that simply related to species differences. Current-injection studies primarily have involved cat MNs, whereas responses to synaptic activation have predominantly been reported for human motor units during voluntary contraction. As such, it could be that human MNs lack the capacity to generate action potentials at high rates. However, there is substantial evidence indicating marked firing-rate saturation in cat MNs during synaptic activation (Alvord and Fuortes 1953; Bracchi et al. 1966; Brownstone et al. 1992; Burke 1968; Cordo and Rymer 1982; Denny-Brown 1929; Granit 1958; Granit et al. 1960; Kernell and Sjöholm 1975; Lee et al. 2003; Prather et al. 2002; Tansey and Botterman 1996; Zajac and Young 1980). Most importantly, Lee et al. (2003) demonstrated considerable firing-rate saturation in spinal MNs of decerebrate cats in response to excitatory synaptic inputs mediated by muscle stretch, yet little saturation (and substantially higher discharge rates) in the same MNs when activated by somatic current injection. Lee et al. (2003) suggested that the lower firing rates during muscle stretch might have been due to activation of synaptic inhibition together with excitation. In any event, differences in firing-rate responses to current injection and synaptic activation cannot be readily attributed to differences in species ordinarily used for the two types of experiments.

Another possibility is that the full extent of depolarizing current that can be marshaled by synaptic inputs to act upon MNs during voluntary activity may simply be less than that needed to drive MNs to discharge at high rates (Binder 2003; Binder et al. 1996; Powers and Binder 2001). In addition, synaptic inhibition directed at MNs might progressively increase as the strength of muscle contraction increases (Denny-Brown 1929; Granit 1958; Granit et al. 1960; Heckman and Binder 1993; Powers et al. 2012; Tansey and Botterman 1996) and thereby limit MN firing rates under natural circumstances. Recently, however, we have shown that providing supplemental excitatory synaptic input to human motor units whose firing rates had already saturated at relatively low rates in response to voluntary drive had practically no effect on firing rate (Fuglevand et al. 2015). Consequently, at least in that case, it would appear that limits in firing rate were not due to inadequate excitation but rather to some intrinsic factor(s) limiting spike-frequency output.

An additional intrinsic mechanism that might partially underlie firing-rate saturation during synaptic activity, paradoxically, is activation of persistent inward currents (PICs) (Fuglevand et al. 2015; Heckman et al. 2008; Hornby et al. 2002; Kiehn and Eken 1997; Powers and Binder 2001; Taylor and Enoka 2004). PICs activate near recruitment threshold (Bennett et al. 1998) and provide a potent source of depolarizing current to MNs (Bennett et al. 1998; Lee and Heckman 1998, 2000; Lee et al. 2003). Because PICs primarily originate in the dendrites, the relatively high-input resistance associated with thin dendrites might cause dendritic membrane potential to shift abruptly into a highly depolarized state and thereby markedly diminish synaptic driving potential (Cushing et al. 2005; Powers and Binder 2001). This, in turn, could make dendrites relatively unresponsive to additional synaptic input, clamp membrane potential, and spiking output at relatively fixed levels (Fuglevand et al. 2015; Lee et al. 2003; Powers and Binder 2001). Because the L-type calcium channels that are thought to be key mediators of PICs (Carlin et al. 2000; Hounsgaard and Kiehn 1989) in mature MNs are relatively undeveloped in neonatal MNs (Jiang et al. 1999), PICs were unlikely to have been engaged in the present study. Future experiments that compare firing-rate capacity in adult MNs with PICs enabled and disabled should help illuminate the role that PICs play in limiting firing rate output in MNs.

GRANTS

Funding for these experiments was provided by the National Institutes of Health (R01-DC769201 to R. F. Fregosi and R01-NS079147 to A. J. Fuglevand).

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

Author contributions: H.E.W., R.F.F., and A.J.F. conception and design of research; H.E.W. performed experiments; H.E.W. and A.J.F. analyzed data; H.E.W., R.F.F., and A.J.F. interpreted results of experiments; H.E.W. and A.J.F. prepared figures; H.E.W. and A.J.F. drafted manuscript; H.E.W., R.F.F., and A.J.F. edited and revised manuscript; H.E.W., R.F.F., and A.J.F. approved final version of manuscript.