Exercise-induced adaptation of neurons in the vertebrate locomotor system

Highlights

• •Neurons in the locomotor system possess an ability to alter their excitability at cellular and molecular scales in adaptation to acute and chronic exercises.
• •Multiple mechanisms are involved in the plasticity of excitability of neurons, of which modulation of ion channels and facilitation of dendritic plasticity are particularly essential to the adaptive response during/following chronic exercise.
• •The exercise-induced modulation of ion channels is shown as an up- and/or down-regulation of transient sodium, persistent sodium, L-type calcium, delayed-rectifier potassium, and calcium-activated potassium channels.
• •Exercise also modulates receptor expression and increases nutritional factors and protein synthesis.

Abstract

Vertebrate neurons are highly dynamic cells that undergo several alterations in their functioning and physiologies in adaptation to various external stimuli. In particular, how these neurons respond to physical exercise has long been an area of active research. Studies of the vertebrate locomotor system's adaptability suggest multiple mechanisms are involved in the regulation of neuronal activity and properties during exercise. In this brief review, we highlight recent results and insights from the field with a focus on the following mechanisms: (a) alterations in neuronal excitability during acute exercise; (b) alterations in neuronal excitability after chronic exercise; (c) exercise-induced changes in neuronal membrane properties via modulation of ion channel activity; (d) exercise-enhanced dendritic plasticity; and (e) exercise-induced alterations in neuronal gene expression and protein synthesis. Our hope is to update the community with a cellular and molecular understanding of the recent mechanisms underlying the adaptability of the vertebrate locomotor system in response to both acute and chronic physical exercise.

Graphical Abstract

Exercise-induced adaptation of neurons in the vertebrate locomotor system. Chronic exercise modulates ion channels, facilitates dendritic plasticity, and increases expression of excitatory synapses and neurotrophins

1. Introduction

The vertebrate locomotor system is constantly evolving and adapting to its surrounding environment.¹ Neurons of the locomotive network display remarkable plasticity in response to increases or decreases in muscle use, upstream (brain and spinal cord) or downstream (musculoskeletal) injury/diseases, and aging.2, 3, 4 This plasticity of neurons is heavily dependent on the locomotor network from the midbrain to spinal cord. Locomotion in vertebrates is initiated in the mesencephalic locomotor region, and the precise timing and pattern of locomotor movements are controlled by central pattern generators in the spinal cord.5, 6, 7 Adaptation of neurons in the locomotor system can be classified into state-dependent (acute)^8,9 or activity-dependent (chronic) alterations of their membrane properties.^3,10, 11, 12 Extensive evidence has shown that chronic exercise leads to adaptive changes in the membrane properties and excitability of spinal motoneurons in rodents.^3,13,14 Indeed, recent studies of mouse spinal cord and midbrain neurons have found that chronic exercise induces morphological plasticity and alters their ion channel activity.^12,15,16 To fully grasp these recent observations and the underlying mechanisms of motor plasticity they are pointing to, a succinct summary of the previous research would be of great utility. Here, we first summarize the effects of acute and chronic exercise on the membrane properties of neurons in the locomotor system. We then discuss the ion channel mechanisms underlying exercise-induced changes in neuronal excitability. We further discuss the dendritic plasticity of the neurons in the spinal cord and midbrain before finally reviewing the alterations in receptor expression and protein synthesis associated with dendritic plasticity. We hope this review may solidify our understanding of and provide guidance for past and future investigations into the cellular and ionic mechanisms underlying the astounding adaptability of the locomotor system (Fig. 1 and Table 1, Table 2, Table 3).

Fig. 1.

Exercise-induced adaptation of neurons in the locomotor system. The treadmill exercise affects neurons in the CPG and MLR. The plasticity of neurons includes (A) modulated ion channels, (B) promoted dendritic plasticity, (C) increased number of excitatory synapses, and (D) increased neurotrophic expression. CPG = central pattern generator; DRG = dorsal root ganglion; MLR = mesencephalic locomotor region.

Table 1.

Electrophysiological properties in response to acute or chronic exercise.

Species	Acute/chronic	Intensity	Duration	Neuron	Parameter	Change	Reference
Cat	Acute	/	/	Motoneuron	Rin	↓	Shefchyk and Jordan (1985)20 Gosgnach et al. (2000)21
Cat	Acute	/	/	Motoneuron	AHP amplitude	↓	Brownstone et al. (1992)22
Cat	Acute	/	/	Motoneuron	Vth	↓	Krawitz et al. (2001)8 Brownstone et al. (1992)22 Fedirchuk et al. (1998)24 Power et al. (2010)25 Brownstone et al. (1994)27
Rat	Acute	/	/	Motoneuron	Vth	↓	Gilmore et al. (2004)26
Rat	Acute	/	/	Motoneuron	AHP amplitude	↓	Schmidt (1994)28
Rat	Chronic	5 days/week, 2 h/day, 30 m/min, at a 10% grade	16 weeks	Motoneuron	RMP, spike trigger level, and spike rise time	↓	Beaumont and Gardiner (2003)10
Rat	Chronic	Spontaneous	12 weeks	Motoneuron	RMP and spike trigger level	↓	Beaumont and Gardiner (2002)11
					AHP amplitude	↑
Rat	Chronic	Spontaneous	16–20 weeks	Motoneuron	F–I curve	←	Macdonell et al. (2012)43
Rat	Chronic	Voluntary weight-lifting training	5 weeks	Motoneuron	Spike rise time and rheobase	↓	Krutki et al. (2017)44
					F–I slope	↑
Rat	Chronic	Compensatory overload	5 weeks or 12 weeks	Motoneuron	AHP amplitude and Rin	↑	Krutki et al. (2015)45
					Spike rise time and rheobase	↓
Rat	Chronic	Transcutaneous trans-spinal direct current stimulation (100 μA, 15 min), 5 days/week	5 weeks	Motoneuron	Rin, maximum steady-state firing frequency, and F–I slope	↑	Bączyk et al. (2020)46
					Vth for spike generation and doublet threshold	↓
Rat	Chronic	Hindlimb unweighting	2 weeks	Motoneuron	Rheobase and Vth	↑	Cormery et al. (2005)47
					AHP amplitude	↓
					F–I curve	→
Mouse	Chronic	10–13 m/min for 60 min	3 weeks	Lamina X interneuron	Vth and rheobase	↓	Chen et al. (2019)12
				Dorsal horn interneuron	AP amplitude	↑
Mouse	Chronic	10–13 m/min for 60 min	3 weeks	Serotonergic neuron of the midbrain	Vth	↓	Ge and Dai (2020)16
					AP amplitude and firing frequency	↑

Note: ↑ means increased; ↓ means decreased; ← means leftward shift; → means rightward shift.

Abbreviations: AHP = afterhyperpolarization; AP = action potential; F–I = frequency–current; Rin = input resistance; RMP = resting membrane potential; Vth = voltage threshold.

Table 2.

Ion channel in response to exercise training in rodents.

Species	Exercise protocol	Intensity	Duration	Neuron affected	Ion channel modulated	Change in channel conductance, kinetics, or protein	Reference
Mouse	Treadmill	10–13 m/min, 60 min/day, 6 days/week	3 weeks	Lamina X interneuron	Persistent sodium channel	↑	Chen and Dai (2022)15
Mouse	Treadmill	10–13 m/min, 60 min/day, 6 days/week	3 weeks	Midbrain 5-HT neuron	Persistent sodium channel	↑	Ge and Dai (2020)16
Mouse	Treadmill	10–13 m/min, 60 min/day, 6 days/week	3 weeks	Lamina X interneuron	L-type calcium channel	↑	Chen and Dai (2022)15
Mouse	Treadmill	10–13 m/min, 60 min/day, 6 days/week	3 weeks	Midbrain 5-HT neuron	L-type calcium channel	↑	Ge and Dai (2020)16
Rat	Exercise wheels	Spontaneous	12 weeks	Motoneuron	K(Ca)	↑	Beaumont and Gardiner (2002)11
Rat	Bilateral tenotomy	Compensatory overload	5 or 12 weeks	Motoneuron	K(Ca)	↑	Krutki et al. (2015)45
Rat	Treadmill	27 m/min at a 10° incline for 60 min, twice per day, 5 days/week	16 weeks	Motoneuron	K(Ca)	↓	Woodrow et al. (2013)92
Rat	Treadmill	20 m/min, ∼50%–55%VO2max, 60 min/day, 3 or 5 days/week	12 weeks	Cerebral artery smooth muscle cell	K(Ca)	↑	Li et al. (2013)93

Note: ↑ means increased; ↓ means decreased.

Abbreviations: 5-HT = serotonin; K(Ca) = calcium-activated potassium channel; VO_2max = maximal oxygen uptake.

Table 3.

Morphology, receptors, or neurotrophins of neurons in response to exercise training in rodents.

Species	Exercise protocol	Intensity	Duration	Neuron affected	Functional modulated	Change in morphology, receptors, or neurotrophins	Reference
Rat	Treadmill	30 m/min, at a 5% grade, 2 h/day, 5 days/week	16 weeks	Lumbar motoneuron	Calcitonin gene-related peptide	↑	Gharakhanlou et al. (1999)103
Mouse	Voluntary wheel running or treadmill	60% of their maximum speed and then increased to 70%, 80%, and 95% of their maximum speed for 2-min intervals, the treadmill speed was increased by 0.8 m/min until the animal was fatigued	5 months	Gastroc muscle	SERCA1a and SERCA2a	↑	Morissette et al. (2014)104
Rat	Treadmill	2 h/day, 30 m/min, 5 days/week	10 weeks	Motoneurons innervating the slow-twitch soleus muscles	Soma area and total SDH activity	↑	Nakano et al. (1997)105
Rat	Treadmill	2 h/day, 30 m/min, at a 10% grade	16 weeks	Motoneuron	Membrane capacitance	↑	Beaumont and Gardiner (2003)10
Mouse	Treadmill	1 h/day, 14 m/min, at a 25° incline	8 weeks	Hippocampus and motor cortex	Complexity of dendritic branches and synaptic plasticity-related protein expression	↑	Feng et al. (2021)106
Mouse	Treadmill	5 m/min for 5 min, 10 m/min for 35 min, 12 m/min for 5 min, 15 m/min for 5 min, 18 m/min for 5 min, cool-down 5 min, 60 min/day, 6 day/week	3 weeks	Brainstem 5-HT neuron	The number of branch points, dendritic branches and length	↑	Ge and Dai (2020)16
Mouse	Treadmill	5 m/min for 5 min, 10 m/min for 35 min, 12 m/min for 5 min, 15 m/min for 5 min, 18 m/min for 5 min, cool-down 5 min, 60 min/day, 6 days/week	3 weeks	Lamina X interneuron	The number of primary dendrites and total dendritic length	↑	Chen and Dai (2022)15
Rat	Treadmill or voluntary wheel	Treadmill: 27 m/min at a 10° incline for 60 min, twice per day, 5 days/week	16–20 weeks	Motoneuron	5-HT1A, GABAAa2, and for SK2 from 16-week-trained, mGluR1 in the voluntary wheel-trained	↓	Woodrow et al. (2013)92
Rat (spinal transection)	Passive cycling	30–50 rpm, 1 h/day	3 months	Sacrocaudal motoneuron	5-HT2A, 5-HT7, and KCC2 expression	↑	Chopek et al. (2015)107
Mouse	Treadmill	25 m/min and 28 m/min, 60 min/day	2 months	Neuromuscular junction	The synapse size of NMJs	↑	Andonian and Fahim (1987)109
Rat	Treadmill	6 m/min for 8–12 min at the onset of training, at 15 m/rain for 20–24 min midway in the period, and at 22 m/min for 30–35 min during the last week, 5 days/week	6 weeks	Neuromuscular junction	The synapse size of NMJs	↑	Waerhaug et al. (1992)110
Rat	Treadmill	24 m/min, 20 min/day, 5 days/week	12 weeks	Neuromuscular junction	The synapse size of NMJs	↑	Deschenes et al. (1993)111
Mouse	Treadmill	28 m/min, 60 min/day, 5 days/week	12 weeks	Neuromuscular junction	Nerve terminals	↑	Fahim (1997)113
Rat	Treadmill or voluntary wheels	27 m/min at a 10% incline for 60 min, twice per day, 5 days/week	16–18 weeks	Dorsal root ganglion	MOR	↑	Paddock et al. (2018)114
					5-HT1A, tyrosine-related kinase receptor A and B, and DOR mRNA levels	↓
SCI adult rat	Treadmill	5–12 m/min for 15 min, twice a day, 5 days/week	6 weeks	Lumbar motoneuron	Total neurite length	↑	Wang et al. (2015)118
Guinea pig	Treadmill	12 days: 20 min/day in two 10 min periods; 5 days: 30 min/day in two 15 min periods; 2 days: 90 min/day in three 30 min periods; 10 days: 135 min/day in three 45 min periods	29 days	Motoneuron	Glucose-6-phosphate dehydrogenase and protein synthesis	↑	Edstrom (1957)119

Note: ↑ means increased; ↓ means decreased.

Abbreviations: 5-HT = serotonin; DOR = delta-type opioid receptor; GABA_Aa2 = GABA A receptor, subunit alpha 2; KCC2 = potassium-chloride cotransporter; mGluR1= glutamate receptor metabotropic 1; MOR = opioid receptor μ subunit; NMJs = neuromuscular junctions; rpm = revolutions per minute; SDH = succinate dehydrogenase; SCI = spinal cord injury; SERCA = sarcoplasmic reticulum calcium-ATPase; SK2 = small conductance calcium-activated potassium channel 2.

In this review, the membrane properties used to describe the neuronal excitability included rheobase (the minimum current required to evoke a spike), voltage threshold (Vth, the lowest membrane potential for spike generation), resting membrane potential (RMP), input resistance (Rin, calculated by membrane potential deflection divided by negatively injected current), frequency and current relationship (F–I relationship, established by firing frequency vs. step current), action potential (AP), and afterhyperpolarization (AHP, hyperpolarization of membrane potential after an AP). The ion channels focused on in this review included: (a) transient sodium channels, which determine the threshold of AP generation, (b) persistent sodium channels, which generate persistent inward currents (PICs) important for motoneuron excitability, (c) L-type calcium channels, which generate PICs regulating rhythm generation in the locomotor system, (d) delayed-rectifier potassium (K(DR)) channels, which modulate Vth, and (e) calcium-activated potassium channels (K(Ca)), which mediate AHP.

2. Acute exercise increases neuronal excitability

Our present understanding of the relationship between acute exercise and neuronal excitability has emerged from comprehensive studies of fictive locomotion in different animal models. Fictive locomotion refers to the generation of neural activity patterns in the nervous system resembling those patterns observed during actual locomotion, but without actual movement of the limbs or body; hence, the resulting locomotion is considered not real but fictive.¹⁷ Studies of fictive locomotion in cats have demonstrated that acute exercise leads to enhanced neuronal excitability. For instance, electrical stimulation of mesencephalic locomotor regions of decerebrated cats generates locomotion¹⁷ via the increase of the excitability of their spinal motoneurons.^8,18, 19, 20 Specifically, their lumbar motoneurons undergo a reduction in Rin^20,21 followed by a decrease in amplitude of AHP during similar interspike intervals.²² Additional studies of fictive locomotor outputs in cat motoneurons have also revealed significant changes in their F–I relationship, a pattern of nonlinear voltage-dependent excitation, and perhaps most significantly, a hyperpolarization of their Vth required for AP generation.^8,9,22, 23, 24, 25, 26, 27 The varying changes in F–I relationship during fictive locomotion are shown as the left-shift of F–I relationship with alteration of F–I slopes during the excitatory phase of the locomotor drive potentials and the right-shift of F–I relationship with changes in F–I slopes in the inhibitory phase of the locomotor drive potentials. In addition, the disappearance of F–I relationships in the excitatory phase of the locomotor drive potentials is also observed in some cells, where the motoneuron discharge rates are saturated without changing with the injected currents.^22,24 Similar alterations in neuronal membrane properties indicating acute exercise-enhanced neuronal excitability have also been reported in other species. Studies of fictive locomotion in decerebrated neonatal rats likewise demonstrate that their spinal motoneurons undergo decreases in amplitude of AHP and a hyperpolarization of their Vth during locomotor-like activity.^9,26,28 Increases in neuronal excitability experienced during acute exercise ultimately dissipate back into a steady state once the locomotor state terminates such that the cycle may repeat again. Mechanism studies suggest that ion channels are responsible for regulating this cycle. Experimental and modeling studies indicate that the hyperpolarization of Vth can be mediated by the upregulation of transient sodium channels^23,29,30 or the downregulation of K(DR) channels.³¹ Furthermore, excitation of motoneurons is mediated by the activation of PICs generated by L-type calcium or persistent sodium channels, which are required for repetitive discharge of spinal motoneurons32, 33, 34, 35 and amplification of synaptic inputs from excitatory reflexes.³⁶ It is noted that Vth and PICs can be modulated by monoamines during acute exercise. It has been shown previously that serotonergic (5-HT) receptors (5-HT_1A, 5-HT_2A, and 5-HT₇) are co-expressed in the spinal neurons, which are activated by electrical stimulation of the mesencephalic locomotor region in cats, suggesting that monoaminergic fibers contact neurons involved in generating locomotion.³⁷ Furthermore, previous studies have reported that serotonin hyperpolarizes Vth³⁸ and enhances PICs in the spinal neurons of rodents.³⁹ These studies suggest that monoamines such as serotonin can play an essential role in modulating neuronal excitability during acute exercise.

Although it is impossible to directly isolate motoneurons from the human spinal cord for recording, researchers have used non-invasive measurement techniques to study the activity of motoneuron pools during different motor tasks (acute exercise). In several studies of human motoneurons, they have investigated the task-dependence of spinal motoneuron excitability during arm cycling by comparing the responses evoked during the cycling to those evoked during position- and intensity-matched isometric contractions.40, 41, 42 Their findings suggest that the task-dependency in spinal motoneuron excitability is present but depends on the arm position during arm cycling.^14,40 This task dependency in human spinal motoneurons during acute exercise has some similarity to the state-dependent property in spinal motoneurons during fictive locomotion in non-human vertebrates.

3. Chronic exercise enhances neuronal excitability

Through several studies in rats, it has been observed that the electrophysiological properties of spinal motoneurons vary in response to several different forms of chronic exercise. A regimen of “forced exercise” subjecting rats to 2 h of treadmill training per day for up to 16 weeks caused their spinal motoneurons to experience significant hyperpolarization of RMP and Vth, reductions in spike rise time, and decreases in F–I slope.¹⁰ Rats given access to an exercise wheel for 12–20 weeks as a form of “voluntary exercise” caused their spinal motoneurons to experience a similar hyperpolarization of RMP and Vth and lowering of F–I slopes in addition to increases in AHP amplitude and a leftward shift of F–I curves.^11,43 Rats that underwent a “resistance-type exercise” regimen for 1 h per day, 5 days per week for 5 weeks caused their spinal motoneurons to experience a decrease in spike rise time and rheobase and an increase in F–I slope.⁴⁴ Furthermore, rats subjected to compensatory muscle overload through tenotomy of synergistic muscles for up to 12 weeks experienced an increase in AHP amplitude and Rin and a decrease in spike rise time and rheobase.⁴⁵ Rats that underwent repeated sessions of transcutaneous, trans-spinal direct current stimulation for 5-week saw increased Rin, decreased Vth for spike generation and doublet threshold, as well as increased maximum steady-state firing frequency and F–I slope in their spinal motoneurons.⁴⁶ In contrast, 2 weeks of hindlimb un-weighting resulted in decreased excitability coinciding with increased rheobase, depolarized Vth, reduced AHP amplitude, and rightward-shifting of the F–I curve.⁴⁷ Altogether, these studies suggest chronic exercise, in its various forms, increases rodent spinal motoneuron excitability by modulating its RMP, Vth, rheobase, firing frequency, and AHP. Chronic exercise increases the gain of the F–I relationship, which physiologically translates into the motoneuron becoming more sensitive to its synaptic inputs, explaining their increased firing frequencies and excitability. This adaptation allows motoneurons to maintain output more efficiently and optimize force production for skeletal muscle movement.^3,48 Overall, these adaptive changes in electrophysiological properties observed in rodents indicate that spinal motoneurons become more excitable and fatigue resistant after chronic exercise. It is worth noting that while the majority of studies have demonstrated that chronic exercise increases motoneuron excitability, recent research has revealed that a 5-week endurance training program on a treadmill led to notable decreases in the Rin values and excitability of fast motoneurons.⁴⁹

Spinal interneurons also undergo adaptive changes with chronic exercise. A recent study¹² in mice (postnatal 42–50 days) reported that 3-week treadmill training increased neuronal excitability with hyperpolarization of Vth and reduction of rheobase in ventromedial and laminar X interneurons, accompanied by a significant increase of AP amplitude in dorsal horn interneurons. These effects depended on the anatomic distribution of the spinal interneurons. Furthermore, another study¹⁶ demonstrated that 3-week treadmill training hyperpolarized Vth, increased AP amplitude, and enhanced firing frequency in 5-HT neurons of the midbrain in mice (postnatal 42–50 days). These findings suggest that exercise-induced plasticity of neuronal excitability is achieved not only in the spinal cord but also in the midbrain, although there are a few examples where exercise does not alter the basic neuronal electrophysiological properties.⁵⁰

Since direct recording of human motoneuron membrane properties is not feasible, much of our understanding of how these neurons adapt in response to exercise comes from indirect measurements using surface or intramuscular electromyography.⁵¹ Human models involving reduced or disused physical activity, such as limb immobilization,⁵² bedrest,⁵³ and spinal cord injury⁵⁴ are also utilized to investigate the plasticity of a motor unit (MU, a motoneuron with all of the skeletal muscle fibers it innervates). Research on the impact of chronic exercise on MUs reveals that lifelong high-intensity physical activity may preserve motor unit function with advancing age.^4,55,56 Moreover, both strength training and endurance training have been shown to enhance the output of the motoneuron pool, suggesting additional recruitment of MUs during exercise.⁵⁷ In fact, a significant increase in the discharge rates of MUs is observed in the recording of a single motor unit using intramuscular electromyography during maximal effort.^58,59 Collectively, these findings indicate that exercise interventions increase the output of single MUs, which in turn increases muscle strength.

4. Exercise modulates ion channels

Ion channel activity is a significant regulator of the spinal motor system's adaptability to acute or chronic exercise. Early modeling studies suggest that the transient sodium, persistent sodium, and K(DR) channels play a dominant role in regulating Vth in cat lumbar motoneurons during fictive locomotion.³⁰ Indeed, numerous other studies have confirmed various channels are involved in regulating neuronal plasticity and excitability during acute or after chronic exercise.^3,23,30 These channels primarily include (a) transient sodium channels, (b) persistent sodium channels, (c) L-type calcium channels, (d) K(DR) channels, and (e) K(Ca) channels.^23,60, 61, 62

4.1. Transient sodium channels

Transient sodium channels determine the threshold for AP generation and make essential contributions to the regulation of neuronal output.^63,64 Recent studies have found that 3 weeks of treadmill exercise increased neuronal excitability and caused hyperpolarization of Vth in mouse spinal interneurons and midbrain 5-HT neurons.^12,15,16 In fact, transient sodium channels play a dominant role in modulating Vth and are one of the main regulators in the recruitment order of MUs in mammals.^65,66 Modeling studies suggest increasing transient sodium channel conductance and/or hyperpolarizing channel state variables (activation m and inactivation h) hyperpolarize Vth in cat lumbar moroneurons.³⁰ Furthermore, in vitro experiments in spinal neurons of neonatal rodents have shown that Vth can be hyperpolarized by the voltage-gated sodium channel 1.1 (Nav1.1) agonist veratridine.²⁹ However, previous reports indicate that activation of protein kinase C depolarizes Vth.⁶⁷ In addition, the α-subunit of sodium channels can be rapidly phosphorylated by protein kinase C at different membrane potentials, suggesting an element of state-dependent regulation of transient sodium channels.68, 69, 70 Furthermore, gene expression of the α-subunit of the transient sodium channel Nav1.6 is significantly upregulated after just 5 days of treadmill training.³ Altogether, these results indicate that modulation of transient sodium channels could be a potential mechanism underlying Vth hyperpolarization during acute or chronic exercise. During fictive locomotion, the F–I relationship of spinal motoneurons is highly varied,^9,22,23 but the mechanisms underlying these variations are still unknown. Simulation results suggest multi-channel modulations are involved in motor output during locomotion and further show that enhancement of transient sodium channel conductance and/or hyperpolarization of channel state variables (m and h) increase the F–I slope and shift the curve to the left. These results are consistent with experimental observations²³ and conclusively demonstrate that transient sodium channels substantially contribute to neuronal plasticity and excitability in response to exercise.

4.2. Persistent sodium channels

Persistent sodium channels contribute to the generation of PICs^33,71, 72, 73, 74, 75 and play an important role in the plasticity of spinal interneurons and brainstem 5-HT neurons.^15,16 Previous studies have reported that persistent sodium channels regulate Vth and repetitive firing in spinal motoneurons,³⁶ interneurons,⁷⁶ hippocampal neurons,⁷⁷ and brainstem 5-HT neurons.⁷⁸ In addition, persistent sodium channels are critical in regulating pacemakers and locomotion speed.⁷⁹ Recent studies have discovered that 3 weeks of treadmill training increases neuronal excitability through the enhancement of persistent sodium channels in spinal interneurons¹⁵ and midbrain 5-HT neurons in mice.¹⁶ Three weeks of treadmill training significantly hyperpolarizes the activation voltage of persistent sodium channels, suggesting they play a key role in regulating spike initiation and excitability. Modeling studies further confirm that PICs generated by persistent sodium channels are involved in the recruitment of spinal motoneuron pools and force generation by skeletal muscle.^48,66 Studies of their F–I relationship suggest persistent sodium channels also play a role in regulating output from spinal neurons during locomotion.^23,78,80 More recent studies suggest persistent sodium channels may contribute to the hour-long effects of trans-spinal stimulation combined with epidural polarization, suggesting they also play a role in promoting rehabilitation.^81,82 Overall, persistent sodium channels help to facilitate the adaptation of the locomotor system in response to exercise.

4.3. L-type calcium channels

L-type calcium channels, mediated by voltage-gated calcium channel 1.3 (Cav1.3) are another generators of PICs,^33,71, 72, 73 and they play an important role in spinal locomotor network rhythm generation.^61,83 In studies of human MU recruitment, PICs mediated by L-type calcium channels have been shown to reduce recruitment.⁸⁴ Modeling studies also emphasize the role of L-type calcium channels in regulating motoneuron excitability, recruitment, and motor output during locomotion.²³ L-type calcium channels have also been shown to amplify synaptic excitatory input, enhance firing frequency,⁷³ and maintain continuous firing in spinal neurons.^75,78,80 Recent studies have revealed that 3 weeks of treadmill exercise enhances L-type calcium channel activity in laminar X interneurons and 5-HT neurons of the dorsal raphe nucleus.^12,16 Further studies have confirmed L-type calcium channels mainly contribute to the regulation of repetitive firing, the facilitation of spike initiation (along with persistent sodium channels), and exercise-prolonged hysteresis of firing after chronic exercise (along with persistent sodium channels).^12,16 Taken together, L-type calcium channels are crucial modulators of neuronal function, motor control, and the adaptive responses of the locomotor system to exercise intervention.

4.4. K(DR) channels

K(DR) channels play a major role in modulating Vth in spinal motoneurons.³⁰ They also regulate diverse membrane properties, including AP amplitude and duration, firing frequency, RMP, and neurotransmitter release.85, 86, 87 Modeling studies suggest K(DR) channels contribute mainly to hyperpolarization of Vth during fictive locomotion.^30,48 Modeling a 70% reduction of conductance of K(DR) channels in the initial segment of motoneurons results in a 3- to 5-mV hyperpolarization of Vth, comparable to the amount observed during locomotion.³⁰ This prediction was verified experimentally in mouse spinal motoneurons using the KCNQ/Kv7 channel antagonist XE-991.³¹ Simulation results further show downregulation of K(DR) channels increases the F–I slope in a manner consistent with experimental observations during locomotion.²³ These results in sum demonstrate that K(DR) channels are one of the major modulators of the locomotor system's adaptability to exercise.

4.5. K(Ca) channels

Two types of K(Ca) channels mediate AHP, small conductance K(Ca) and big conductance K(Ca). Small conductance K(Ca) channels are activated by cytosolic calcium in response to calcium influx via voltage-gated calcium channels during AP generation.^62,88,89 K(Ca) channels control many physiological processes, from the firing properties of neurons to the control of neurotransmitter release.⁹⁰ During repetitive firing, the inter-spike interval is dependent on the amplitude and duration of the AHP.^60,91 Modeling studies have shown that reducing K(Ca) channel conductance by 30%–50% leads to increases in the F–I slope,²³ the number of recruited motoneurons,⁴⁸ and the force production of skeletal muscles.⁶⁶ Therefore, modulation of K(Ca) channels regulates discharge rate and neuronal excitability. The effect of exercise intervention on the amplitude of AHP mediated by K(Ca) channels is inconsistent. Experimental results show the amplitude of AHP is reduced in cat and rat spinal motoneurons during fictive locomotion,^9,22,25 while results from genetic studies show gene expression of K(Ca) channels decreases in rat spinal motoneurons after chronic exercise.⁹² However, the amplitude of AHP significantly increases in fast motoneurons of rats subjected to chronic compensatory muscle overload for 5 weeks or 12 weeks⁴⁵ and in motoneurons of rats subjected to spontaneous exercise wheels for 12 weeks.¹¹ Increased AHP is also observed in cerebral artery smooth muscle cells after chronic exercise.⁹³ Furthermore, the amplitude of AHP does not undergo significant changes following 3 weeks of exercise.^15,16 The mechanism underlying this paradoxical increase in AHP amplitude remains an open question. A possible interpretation of this paradox is that an increase in AHP amplitude following training is used by the motor system to reduce PIC activation and thus firing rates to offset fatigue. A further study is required to test this hypothesis. Overall, K(Ca) channels play a certain role in regulating neuronal excitability and motor output during exercise; however, additional research is needed to elucidate the mechanisms and factors contributing to the contradictory increases in AHP amplitude observed under different exercise conditions.

5. Exercise facilitates morphological and functional plasticity of neurons

A key feature of the central nervous system is the ability of its neurons to alter their morphology and connectivity in response to sensory experience and other changes in the environment. Neurons with distinct dendritic morphologies have different functions in neural circuits, as they will make unique connections with other neurons.⁹⁴ Voluntary physical exercise is one of the most studied activities that has been shown to positively influence adult neuroplasticity.⁹⁵ Neuroplasticity is an umbrella term that includes all the functional and structural changes occurring within a neural circuit, and it involves various mechanisms, such as synaptic plasticity, dendritic remodeling, and neurogenesis.96, 97, 98 Neurons are typically comprised of a cell body, an axon through which they transmit information to other neurons, and a dendritic arbor where input from other neurons is primarily received.⁹⁹ Neurotransmitter receptors are largely restricted to the surface of dendritic spines, which are small membranous protrusions whose structural geometry correlates with the strength of synaptic connections.¹⁰⁰ Neuronal plasticity is correlated to the density and morphology of the dendritic spines given they are the main sites of synaptic input for neurons.¹⁰¹

Exercise can modulate neuronal excitability by regulating various ion channels, which can in turn alter neuronal morphology. Several studies have shown exercise causes the regeneration of spinal cord neurons by increasing the length of neuronal dendrites, causing the synthesis of nutritional factors inside neurons, and increasing the density of neuronal synapses.^3,102, 103, 104 Significant adaptations in neural morphology, including changes in soma size and dendritic number, as well as herald changes in functional properties.³ A previous study demonstrated endurance training increases the average soma diameter of rat soleus motoneurons.¹⁰⁵ Additional research has shown chronic exercise leads to an increase in motoneuron membrane capacitance, an index of cell size, suggesting motoneuron size increases following exercise.¹⁰ Exercise promotes dendritic spine formation and enhances motor learning in the mouse hippocampus and motor cortex.¹⁰⁶ Recent research has revealed 3 weeks of treadmill exercise facilitates dendritic plasticity in midbrain 5-HT neurons of juvenile mice as evidenced by an increase in the number of dendritic length, branches, and branch points within a range of 50–200 μm from the soma.¹⁶ These results matched those of similar studies in spinal lamina X neurons.¹⁵ Past studies have also confirmed that the distribution of the Cav1.3 is concentrated in neuronal dendrites.^33,73 In fact, exercise can increase PICs mediated by Cav1.3.^15,16 Altogether, these results illustrate a model for the cellular and ionic basis of exercise-induced dendritic plasticity.

6. Exercise modulates receptor expression

Exercise also alters synaptic density and structural formation. Previous studies have shown that passive cycling can increase synaptic density in rat lumbar motoneurons by increasing the expression of 5-HT_2A, 5-HT₇, and potassium-chloride cotransporter 2 receptors in extensor motoneurons.^92,107 Conversely, exercise reduces the expression of 5HT_1A receptor and γ aminobutyric acid receptor subunit α 2.⁹² Furthermore, resistance training increases the number of excitatory synapses on spinal motoneurons while inhibitory synapses remain unaffected, suggesting motoneurons receive greater excitatory synaptic input from chronic exercise.¹⁰⁸ This change in neurotransmitter receptor densities suggest exercise enhances spinal motoneuron dendritic plasticity. The neuromuscular junction is a specialized synapse formed between motoneurons and skeletal muscle fibers. Early studies have shown endurance training increases the synapse size of neuromuscular junctions in adult mice¹⁰⁹ and rats.^110,111 Exercise also induced hypertrophy of neuromuscular junctions of the extensor digitorum longus and gluteus maximus in adult mice and rats.^112,113

Chronic exercise also affects receptor expression in neurons involved in sensory signal transmission of the dorsal root ganglion. These changes include an increase in opioid receptor μ subunit mRNA levels in conjunction with decreases in 5-HT_1A, tyrosine-related kinase receptors A and B, and delta-type opioid receptor mRNA levels.¹¹⁴ Furthermore, exercise also promotes morphological and functional plasticity in astrocytes.¹¹⁵ Recent studies have demonstrated that exercise induces various changes in astrocytes, including increased proliferation, improved maintenance of basal levels of catecholamine, enhanced glutamate uptake, and increased trophic factor release.¹¹⁵

7. Exercise increases nutritional factors and protein synthesis

Neurotrophins are a group of soluble growth factors that can modulate neuronal synaptic function, plasticity, and survival.¹¹⁶ Brain-derived neurotrophic factor plays a significant role in promoting dendritic plasticity,¹¹⁷ and their increased expression is associated with changes in neuronal morphology. Treadmill training has been shown to significantly increase brain-derived neurotrophic factor expression in rat lumbar motoneurons in conjunction with an increase in total neurite length.¹¹⁸ It has been known since the 1950s that motoneurons of treadmill-trained rats exhibit increased staining intensity for glucose-6-phosphate dehydrogenase, which suggests heightened synthesis of that protein.¹¹⁹ Moreover, a study in mice has reported moderate-intensity treadmill exercise leads to increased expression of several factors in the motor cortex of a crushed spinal cord injury, including brain-derived neurotrophic factor, insulin-like growth factor 1, phosphorylated ribosomal S6 protein, and protein kinase B.¹²⁰ These studies support the notion that chronic exercise can induce alterations in nutritional factors and protein synthesis that ultimately facilitate neuronal plasticity.

8. Conclusion

Vertebrate neurons possess an astounding ability to alter themselves at cellular and molecular scales in adaptation to their surroundings. Nowhere is this truer than in the case of the neurons composing the locomotor system, which display remarkable plasticity in response to increases or decreases in muscle use, upstream or downstream motor injury, and aging. Research in various animal models from the past up to the present has shown that exercise induces significant alterations in the function and physiology of these neurons. Multiple mechanisms are involved in the plasticity and excitability of neurons, and the modulation of ion channels and facilitation of dendritic plasticity are of particular importance to their adaptive response during/following chronic exercise. Exercise also modulates the gene expression of receptors and the protein synthesis of neurotrophins, leading to an increase in neuronal excitability and changes in neuronal morphology that extended dendritic growth and plasticity. Overall, we have reviewed a diverse array of quality studies that use multidisciplinary approaches and various animal models to understand the adaptations incurred on the vertebrate locomotor system due to exercise. This investigation has significantly improved our understanding of the subject while raising questions and hinting at directions for future research in this field. For instance, work remains to be done reconciling the observations that exercise seemingly reduces AHP amplitude mediated by K(Ca) channels in cat and rat motoneurons during fictive locomotion but increases it in rats subjected to chronic exercise. Furthermore, our understanding of the human locomotor system is greatly limited by the fact that most of these studies occurred in animal models. Moving forward, there is a need for the development of new methods and experimental designs that will allow us to recapitulate the findings we have thus far amassed in cats, rats, and mice, in humans.