**3. Motor neuron functional identities**

#### **3.1 Motor neuron functional diversity: alpha, beta and gamma motor neurons**

The mammalian spinal cord displays a diverse population of motor neurons which is correlated with the heterogeneity of the muscle fiber types they innervate (**Figure 1**). Motor neurons within the mammalian spinal cord can be categorized into functionally diverse classes and subtypes based on several properties: size and morphology, electrical properties, molecular marker expression and function. Based on these characteristics, motor neurons can be divided into three main types, alpha motor neurons (α-MNs), beta motor neurons (β-MNs) and gamma motor neurons (γ-MNs) and several subtypes. Somatic α-MNs exclusively innervate extrafusal muscle fibers that regulate muscle force and movement. While gamma motor neurons innervate the smaller intrafusal fibers located within the muscle spindle proprioceptive organs, regulate the sensory information reported during muscle stretch from the muscle spindle, and thereby contribute to motor control. And beta motor neuron, innervate both extrafusal and intrafusal fibers, and may contribute to both muscle contraction and sensory information gathered from muscle spindles during muscle stretch, although their exact functional contribution remains unknown.

#### **3.2 Alpha motor neurons**

Motor neurons and the muscle fibers their axons innervate display a spectrum of properties that are used to categorize them into subtypes. For instance, there is exquisite matching of muscle contractile properties such as isometric twitch speed, maximum force and endurance and alpha motor neuron properties like size and morphology, excitability and firing pattern, which together allow them to contract synchronously as a "motor unit" to drive muscle contraction [49, 50]. Alpha motor units can be classified into three subtypes: (1) slow-twitch fatigue resistant (S), fasttwitch fatigue-resistant (FR), and fast-twitch fatigable (FF) [51]. Each of these motor units possesses a spectrum of properties: S-type are made of type I muscle fibers that contract slow, develop small quantity of force and are very fatigue-resistant, FR-type are comprised of type IIA fibers that contract faster, develop more force and are less fatigue-resistant than the S-type, while the FF-type are made of type IIB fibers that contract the fastest, develop the highest force and are highly fatigable [50, 51].

Alpha motor neurons, which comprise of one-third of motor neurons in a given motor pool, can be identified based on their size and morphology, excitability, biophysical properties, and the expression of molecular markers. Alpha motor neurons are sequentially activated depending on two linked properties of their cell membrane, that is their soma size and morphology (dendritic arborization) and intrinsic properties (quantity and diversity of ion channels) [52, 53]. For example, as a motor pool receives information from descending inputs for eliciting the contraction of the muscle it innervates, S-type motor neurons, possessing smaller soma sizes and more simple dendritic arborization, are activated first because: (1) they possess a higher input resistance, meaning a larger change in cell membrane voltage over constant current injection, and

(2) a lower threshold due to cell membrane Na+ ion channel sensitivity. Thus, S-type motor neurons fire action potentials with lower synaptic current when compared to the larger FF-type motor neurons. Thus, this sequential activation from S-type motor neuron to FF-type motor neuron is thought to follow Henneman's "size principle" and has implications for motor unit recruitment [54, 55]. Moreover, FR-type motor neurons and motor units have intermediate characteristics between S- and FF-subtypes. Thus, activities that require sustained muscle contraction (standing or walking) results in the recruitment of S-type motor neurons that activate slow motor units, while activities that require potent bursts of muscle contraction (running or jumping) initiates FF-type motor neurons recruitment and subsequent activation of fast motor units, elegantly matching motor neuron morphology and electrical properties with motor unit size required for specific movement tasks [56].

Since motor neurons fire repetitive action potentials, their firing rate can be used to classify them. The firing rate is shaped by persistent inward currents (PICs) generated by voltage-gated Na+ and Ca2+ currents, which are prolonged on dendrites of S-type motor neurons than FF-type motor neurons [57]. PICs amplify and limit the modulation of firing rate, making the S-type motor neurons highly excitable with an initial steep firing rate and subsequent saturation [57]. Another property that determines motor neuron firing rate is the after-hyperpolarization (AHP) phase after the action potential, which is generated by Ca2+-dependent K+ currents. The influx of Ca2+ during the firing of action potentials and the intracellular diffusion, pumping and interaction with proteins regulates AHP-decay times [58–60]. Thus, FF-type motor neurons have a shorter AHP-decay time, therefore a higher maximum firing frequency than S-type motor neurons, which ultimately matches the contractile frequency of the muscle fiber type these motor neurons innervate [3, 61, 62].

Alpha motor neuron subtypes can also be distinguished by the expression of a subset of genes. For instance, studies have shown that FF-type motor neurons express: Calcitonin gene-related peptide (CGRP)/calca, Chondrolectin, Matrix metallopeptidase 9 (MMP-9), and Delta-like homolog 1 (Dlk1) [63–67]. While a synaptic vesicle protein, SV2a is expressed postnatally in presynaptic terminals of S-type motor neurons that innervate type I and small type IIA muscle fibers [68]. Moreover, a study in rat showed that Ca2+-activated K+ (SK) channels are expressed in S-type alpha motor neurons and the electrophysiological recordings of these SK3<sup>+</sup> motor neurons showed medium size AHP-duration, which seems to be in the range of S-type alpha motor neurons [69]. Other putative markers that are expressed in all three alpha motor neuron subtypes (FF, FR and S) are Hb9::GFP, NeuN, Osteopontin, and Na<sup>+</sup> /K+ ATPase (Atp1a1 and Atp1a3) (both alpha 1 and 3 isoforms in FF- and FR-subtypes), while UCHL1::eGFP and Na<sup>+</sup> /K+ ATPase (alpha 1 isoform only) are expressed in S-type [70–77].

#### **3.3 Beta motor neurons**

Another class of motor neuron, the beta motor neurons, were thought to exist only in lower vertebrates such as reptiles, amphibians and birds but are also shown to exist in mammals and comprise one-third of all motor units and innervate threefourths of all muscle spindles [3, 78, 79]. Despite their abundance in quantity, limited anatomical and functional characterization suggests that beta motor neurons possess intermediate properties between alpha and gamma motor neurons and may play important roles in regulating motor behavior. Unlike alpha and gamma motor neurons which exclusively innervate extrafusal and intrafusal fibers, respectively, beta

#### *From Motor Neuron Specification to Function: Filling in the Gaps DOI: http://dx.doi.org/10.5772/intechopen.114298*

motor neurons innervate both extrafusal and intrafusal fibers, and thus, regulate both muscle contraction and sensory information from the muscle spindle [78]. Based on their hybrid innervation pattern, beta motor neurons can be subdivided into two subtypes: static and dynamic. Static beta motor neurons innervate type IIa extrafusal fibers and bag2 intrafusal fibers, while dynamic motor neurons innervate type I extrafusal fibers and bag1 intrafusal fibers [3]. Thus, the functional role of beta motor neurons is currently unresolved, however, based on their anatomical properties, they seem to play a role in movement and maintenance of posture [3]. Moreover, studies that aim to identify molecular marker expression and electrophysiological properties of beta motor neurons are especially needed in testing what type of functional role they may play in movement and movement control.

#### **3.4 Gamma motor neurons**

Like the alpha motor neurons, gamma motor neuron subtypes can be classified based on their distinct patterns of morphology, connectivity, electrical and functional properties. In mammals, gamma motor neurons represent about one-third of the motor neurons in a given spinal motor pool and are distinguished by their singular role of regulating muscle spindle sensitivity and motor control. Gamma motor rely on GDNF secreted from muscle spindle sensory receptors buried deep within skeletal muscle for their postnatal survival, although it is not known how they survive before maturation [70]. Gamma motor neurons innervate the intrafusal fibers within the muscle spindle and receive presynaptic input from sensory neurons within the muscle spindles. Thus, gamma motor neurons enable continuous flow of sensory information about muscle length and ensure fluid muscular action during muscular contractions [6, 80]. Gamma motor neurons can be subdivided into two types based on their intrafusal fiber innervation patterns and their role in regulating muscle spindle sensory information. Dynamic gamma motor neurons innervate bag1 intrafusal fibers and are thought to regulate muscle length during locomotion, while static gamma motor neurons innervate bag2 and nuclear chain fibers and are thought to be involved in body posture [81].

Gamma motor neurons possess unique biophysical properties which were identified based on intracellular and patch clamp recordings in different animal models. Early on, intracellular recordings in the ventral spinal cord of the cat identified that gamma motor neurons had slower conduction velocity (since they have small axon diameter) compared to alpha motor neurons [82–84]. Moreover, studies showed that gamma motor neurons possess lower discharge threshold, higher discharge rates and lower membrane input resistance when compared to alpha motor neurons [85]. Furthermore, gamma motor neuron subtypes display unique firing properties: dynamic gamma motor neurons increase the discharge rate of primary sensory afferents when muscle is stretched, while static gamma motor neurons seem to have no effect on primary sensory afferent firing [86]. Studies in mice have also characterized gamma motor neuron electrophysiological properties. Immature gamma motor neurons in young mice (P0-P6) expressing GFP under serotonin receptor 1d (5-ht1d) promoter were patch clamped and revealed that gamma motor neuron electrical properties of rheobase current, input resistance and AHP-decay time ranged between FF-type and S-type alpha motor neurons [87]. In a recent study, more mature gamma motor neurons from mice (P20–22) labeled with high levels of Fluorogold (FG) (retrograde tracer) showed low rheobase current, high firing frequencies and gain when compared to alpha motor neurons [88]. This study matches some of the gamma motor neuron signature properties observed in the cat.

#### **3.5 Mechanisms underlying motor neuron functional diversification**

Differences in the connectivity and physiological properties of motor neurons were first reported in the 1940s and 1950s [84, 89–91], and the significance of these differences have been recognized soon afterwards [6, 84, 92]. Beta and gamma motor neurons were described a few years later [84, 93, 94], and their discovery together significantly broadened our understanding of the neuromuscular bases of movement control and have become canonical topics of neurobiology and neurophysiology textbooks. It therefore seems astonishing that it took over 60 years until the first mechanisms promoting motor neuron functional diversification were discovered [67, 88, 95, 96]. This is likely explained by the difficulties in discovering molecular markers for functional motor neuron types in tetrapod vertebrates due to their general lack of correlation with fixed anatomical features.

Since the first molecular markers for functional motor neuron types were reported, it then took another couple of years until first insights into the mechanisms promoting the diversification of alpha motor neurons into fast and slow types were reported. Herein, the type I transmembrane protein and non-canonical Notch ligand Delta-like homolog 1 (DLK1) was shown to be both necessary and sufficient to promote fast alpha motor neuron gene expression and biophysical signatures required for peak force execution [67]. DLK1 appears to operate in part through the activation of regulatory ion channel subunits including *Kcng4*/KV6.3/4, apparently reflecting its requirement for both specification as well as maturation of fast alpha motor neurons (**Figure 2**) [67]. The transition between initial specification and maturation of

#### **Figure 2.**

*Schematic summarizing mechanisms identified so far promoting motor neuron functional diversification. While mechanisms promote the acquisition of motor neuron positional identities (not shown) and the organization into different motor pools, parallel or subsequent mechanisms promote the diversification of motor neurons into different alpha, beta and gamma motor neuron types and subtypes. Cell-autonomous mechanisms have been identified underlying the specification and/or maturation alpha and gamma motor neuron types and subtypes, while the maintenance of some alpha motor neuron properties appear to rely on non-cell-autonomous signals by ventral horn astrocytes. Mechanisms underlying the specification of slow alpha motor neurons and beta motor neurons remain to be identified.*

#### *From Motor Neuron Specification to Function: Filling in the Gaps DOI: http://dx.doi.org/10.5772/intechopen.114298*

motor neurons is further reflected by global shifts in gene expression and chromatin accessibility for different transcription factor classes [97].

More recent work identified two transcription factors PRDM16 and MECOM as early determinants of primary and fast secondary motor neurons in zebrafish, which are likely homologous to tetrapod fast alpha motor neurons [96]. Both transcription factors broadly, but not completely, regulate fast motor neuron gene expression, raising the question of how these molecules are linked to the actions of DLK1 and Notch signaling (**Figure 2**) [96]. The incorporation of neurons into functional circuitries entails not only their specification but also their maturation including the acquisition of specific membrane electrical and firing properties [98]. An intriguing non-cellautonomous mechanism was found to promote and maintain the mature state of fast alpha motor neuron properties [95]. The potassium channel subunit KIR4.1 expressed by ventral horn astrocytes appears to be required for maintaining fast alpha motor neuron soma size and function through the paracrine activation of mTOR signaling (**Figure 2**) [95]. Again, it remains to be determined how these cell-autonomous (DLK1, PRDM16, MECOM) and non-cell-autonomous mechanisms intersect during alpha motor neuron diversification and maturation (**Figure 2**).

#### **4. Phylogenetic considerations**

The spinal motor systems of fish and tetrapod vertebrates differ in key aspects that reflect adaptations specific to swimming and terrestrial locomotion. For instance, fish generally lack muscle spindles and consequently lack gamma or beta motor neurons regulating muscle spindle proprioception [99]. The appearance of muscle spindle and fusimotor systems providing and regulating muscle proprioception obviously represent adaptations to terrestrial locomotion. Another difference between fish and tetrapods is the spatial arrangement of the force-generating alpha motor neurons in the spinal cord. In zebrafish, these neurons are organized into distinct motor columns presynaptically connected to dedicated interneuron modules and postsynaptically to muscle groups containing either slow, intermediate or fast muscle fiber types [100–103]. This arrangement allows rapid transitions from slow undulating swimming to fast escape movements and is apparently adapted to pelagic locomotion. Zebrafish larvae initially possess only fast muscle fibers and primary motor neurons facilitating escape swimming movements, while intermediate and slow muscle fibers and matching secondary motor neuron subtypes are generated later during the transition to adulthood.

The modular architecture of motor neuron subtypes in fish contrasts with that found in tetrapods [102], particularly in mammals, in which most motor pools comprise a mosaic of motor neuron types, including the different alpha motor neuron types, gamma motor neurons and small numbers of beta motor neurons [3]. Exceptions can be found in motor pools connecting to specialized muscles, such as predominantly slow antigravity muscles like the *m. soleus* of the calf or muscles involved in explosive force generation [104], such as the *m. rectus femoris* of the thigh, which are enriched in slow or fast alpha motor neurons, respectively [67]. Moreover, some muscles in birds and non-avian reptiles display regionalized abundances of slow or fast muscle fibers, which are reflected by similar spatial separation of alpha motor neuron types in the corresponding motor pools. Moreover, particularly in primates, motor pools supplying distal forelimb muscles involved in dexterous movements tend to be enriched in gamma motor neurons. Nevertheless, in mammals, and tetrapod

vertebrates in general, motor neuron types within the motor pools typically intermingle and do not spatially segregate, as do the fiber types in the skeletal muscles.

While a degree of conservation of the mechanisms underlying the specification of alpha motor neuron types and the axial motor neurons of fish are expected, it is likely that the transition to mosaic organization of motor pools in tetrapods will be reflected by differences in the mechanisms of specification and functional specializations of alpha motor neuron types. In mouse, for instance, DLK1 promotes a gene expression signature specific for fast alpha motor neurons [67], which, however, shows little overlap with that promoted by PRDM16 and MECOM in zebrafish primary and fast secondary motor neurons [96]. While these discrepancies may in part stem from differences in the transcriptome profiling methodologies used by both studies, it is likely that the spatial and functional reorganization of functional motor neuron types in tetrapods reflects an underlying reorganization of genetic circuitries promoting alpha motor neuron diversification.

A similar difference concerns the molecular profile of slow alpha and gamma motor neurons, with slow secondary motor neurons in zebrafish apparently expressing ERR2. In mouse and chick, high levels of ERR2 (together with its paralogue ERR3) mark gamma motor neurons, while lower levels are initially also expressed by alpha motor neurons [65, 71, 88, 96]. It is therefore tempting to speculate that tetrapod fusimotor (beta and gamma) motor neurons evolved from slow motor neurons with a genetic program that enhanced certain sets of preconfigured properties, such as low firing thresholds and small soma sizes. Intriguingly, while fish trunk muscles lack muscle spindles [99], the jaw muscles of some fishes possess muscle-spindle like structures likely supporting rapid prey capture and manipulation movements [105]. It will be highly interesting to test whether such spindle-like structures would receive innervation from motor neurons and how such hypothetical 'ancestral' beta motor neurons would differ in their gene expression profile from axial motor neurons.

#### **5. Conclusions**

The two aspects motor neuron diversity discussed in this chapter raise the question of whether both are coordinated and if yes, how? At first glance, both types of motor neuron identity, positional and functional, appear to be established independently form each other. In tetrapod vertebrates, a typical motor pool contains a complement of alpha, beta and gamma motor neuron types and subtypes. Developing motor neurons eventually coalesce into motor pools under the influence of both cell-intrinsic and axon target-induced mechanisms, which in addition to positioning motor neuron somas shape dendritic arbors and presynaptic connectivity. Thus, subsets of motor neurons must somehow acquire distinct sets of properties overlayed on these motor pool-specific features. Moreover, some motor pools are enriched in certain functional motor neuron types mirroring the enrichment in certain fiber types in the muscle they supply. This suggests at least some degree of coordination or crosstalk between motor neuron and muscle fiber type diversification, which is supported by the mutual influence motor neuron and muscle fiber types have on each other. However, this mutual influence only seems to work to some degree, as the conversion of motor neurons from one type to another does not lead to a complete conversion to the corresponding muscle fiber type and *vice versa*. Moreover, there are several lines of evidence that the initial specification of both motor neuron and muscle fiber types occurs in a cell-autonomous fashion but that they to some degree (yet not completely) remain

### *From Motor Neuron Specification to Function: Filling in the Gaps DOI: http://dx.doi.org/10.5772/intechopen.114298*

sculptable throughout later life. This initial cell-autonomy suggests that there may be some intersection between the interpretation of positional signal by motor neurons and the eventual generation of certain ratios of motor neuron types, such as the relative abundance of slow motor neuron and muscle fiber types in the *soleus* motor pool and muscle, respectively. There are therefore many open questions remaining, not only regarding the functional diversification of motor neurons but also regarding to what degree and how the underlying mechanisms are coordinated with those promoting motor neuron positional identities.
