Introduction
Amyotrophic lateral sclerosis (ALS) is a late-onset, progressive and fatal neurodegenerative disease which primarily affects motor neurons (MNs) of the motor cortex of the brain, brainstem motor nuclei and anterior horn of the spinal cord (Kiernan et al., 2011; Renton et al., 2014; Al Sultan et al., 2016; Taylor et al., 2016). ALS commonly begins in late-adulthood, when patients first experience focal symptoms, such as weakness in the limb or bulbar muscles, as well as widespread fasciculations. The disease then usually progresses in an organized way to adjacent areas of the central nervous system (CNS), and consequently symptoms appear in other regions of the body. Several clinical subsets of ALS can be distinguished by the anatomical location first affected (Renton et al., 2014; Taylor et al., 2016). This includes bulbar onset, where symptoms first appear in the muscles controlling speech, mastication and swallowing; and limb onset, where symptoms present initially in the upper (arm or hand) or lower limbs (leg or foot). Bulbar onset patients face a much worse prognosis than those with spinal onset ALS, where the average survival time following diagnosis is less than 2 years. However, in patients with the much rarer respiratory onset form (3–5%), the prognosis is even worse as the survival time following diagnosis is only 1.4 years (Swinnen and Robberecht, 2014). At disease end stage, only support and palliation are available, and patients usually die from respiratory failure, typically 3–5 years after diagnosis (Taylor et al., 2016). There are currently few effective treatments. Hence there is an urgent need to understand the underlying causes and risk factors for ALS to discover new therapeutic targets.
Neurons have complex and extended morphologies compared to other cell types, and within the CNS, neurons can vary greatly in their properties. MNs are unique cells amongst neurons because they are large, even by neuronal standards, with very long axons, up to 1 m in length in an adult human. MNs can be distinguished into two main categories according to their location in the CNS: upper MNs (UMNs) located in the cortex, and lower MNs (LMNs) located in the brainstem and spinal cord. The spinal MNs comprise both visceral MNs of the thoracic and sacral regions, which control autonomic functions, and somatic MNs, which regulate the contraction of skeletal muscles and thus control movement. The diversity of MNs reflects the variety of targets they innervate, including a wide range of muscle fiber types. UMNs and LMNs differ in the location of their cell bodies, the neurotransmitters released, their targeting and symptoms resulting from their injury.
It is unknown why MNs are specifically targeted in ALS and remarkably, MNs are not equally affected (Rochat et al., 2016; Nijssen et al., 2017). Whilst both UMNs and LMNs are involved, some LMN subtypes are relatively resistant to neurodegeneration. Spinal cord and hypoglossal MNs are amongst the first to degenerate, hence the ability to speak, breath and move is lost early in disease course. As ALS progresses, specific MN subtypes then preferentially deteriorate. However, some MNs are spared until disease end stage, such as oculomotor neurons and Onuf’s nuclei MNs, and as a result, patients retain normal visual, sexual and bladder function throughout the disease course. The resistant MNs differ significantly from the vulnerable MNs anatomically and functionally, and they possess distinct transcriptomes, metabolic and developmental profiles. Surprisingly, there are also differences in vulnerability amongst spinal MNs, because those that are part of the faster motor units degenerate before those in the slower motor units (Frey et al., 2000; Pun et al., 2006; Hegedus et al., 2007; Hadzipasic et al., 2014; Sharma et al., 2016; Spiller et al., 2016a), thus adding further complexity to the question of MN vulnerability.
ALS shares clinical and pathological features with frontotemporal dementia (FTD), a type of dementia that involves impaired judgment and executive skills. In FTD, the loss of cortical MNs is accompanied by loss of neurons in the frontal and temporal cortices, which correlates clinically with the symptoms of FTD (Neumann et al., 2006; Burrell et al., 2016). The relationship between ALS and FTD has been confirmed through genetic studies, and these two conditions are now considered to be at opposite ends of the same disease continuum (Taylor et al., 2016; Shahheydari et al., 2017). Hence, while ALS was historically judged as a disorder affecting the motor system only, it is now recognized that non-motor features are present (Fang et al., 2017). A wealth of evidence also demonstrates that ALS is a heterogeneous disorder. The clinical symptoms, including the proportion of UMN and LMN signs, age of onset, disease duration, and association with other conditions, are major features contributing to its highly variable phenotypes. As well as the development of FTD (Strong and Yang, 2011), ALS can also involve cognitive impairment in up to 50% of patients (Tsermentseli et al., 2012), the autonomic nervous system (Piccione et al., 2015), supranuclear gaze systems (van der Graaff et al., 2009; Donaghy et al., 2011), and extrapyramidal motor signs (Pradat et al., 2002). Sensory, olfactory and visual dysfunction have also been described in some patients (Bede et al., 2016). In addition, there are also other conditions affecting MNs that share similarities, but also striking differences, to ALS. In particular, primary lateral sclerosis (PLS) affects UMNs but it progresses much slowly than ALS. It also has a significantly lower mortality rate (Tartaglia et al., 2007), consistent with the relative resistant of LMNs in ALS.
One of the main pathological characteristics of ALS is the presence of insoluble protein inclusions in the soma of MNs. TAR DNA binding protein-43 (TDP-43) is the major component of these inclusions (Arai et al., 2006; Neumann et al., 2006) in almost all (∼97%) ALS patients and ∼50% FTD patients (Arai et al., 2006; Neumann et al., 2006; Mackenzie et al., 2007; Scotter et al., 2015; Le et al., 2016). Loss of TDP-43 from the nucleus is evident in MNs from ALS/FTD patient tissues, concomitant with the formation of TDP-43 inclusions in the cytoplasm of both MNs and glia. Neuropathological studies have also revealed that the clinical course of ALS reflects the presence of TDP-43 pathology, from its deposition at an initial site of onset, to its spread to contiguous regions of the CNS (Brettschneider et al., 2013). Mutations in TDP-43 are also present in 5% of familial forms of ALS (Sreedharan et al., 2008). In the genetic types of ALS, it remains unclear why MNs are specifically affected when the mutant proteins are ubiquitously expressed. Males are affected more by ALS than females, and ethnic populations show differences in the incidence rates of ALS, further highlighting the contribution of genetics to ALS.
Whilst our understanding of the etiology of ALS has increased significantly in recent years, major gaps in our knowledge remain. In this review, we address several unanswered questions regarding the unique susceptibility of specific types of MNs in ALS: Why does neurodegeneration spread throughout specific neural networks? How can ubiquitously expressed genes be selectively toxic to MNs? Why are some MN subtypes more vulnerable to degeneration than others? We also discuss the role of the neuronal network and the specific cellular microenvironment in driving cell-to-cell disease progression, plus the importance of genetics in influencing susceptibility of specific neuronal subpopulations. Finally, we discuss the role of aging as a potential risk factor for the susceptibility of specific MN subtypes. A thorough comprehension of why specific cell types degenerate is imperative to our understanding of ALS because it provides important clues as to what initiates neurodegeneration, and how this knowledge may be harnessed therapeutically.