Role of Glial Cells in Driving Disease Progression
There is increasing evidence for a role of the neighboring non-neuronal cells in ALS. Under normal conditions, glial cells provide nutritional and trophic support to MNs, but in ALS, they appear to exacerbate neurodegeneration in a non-cell autonomous fashion. These cells include microglia, astrocytes, oligodendrocytes and Schwann cells. Limiting the expression of mutant SOD1 to MNs only does not lead to neurodegeneration in mice (Pramatarova et al., 2001; Lino et al., 2002), and chimeric mouse studies have established that the presence of mutant SOD1G93A in glial cells induces neurodegeneration and MN loss (Papadeas et al., 2011). Both microglia and astrocytes appear to enhance disease progression by inducing neuroinflammation, whereas oligodendrocytes drive disease initiation. Non-neuronal cells may also be involved in the spread of pathological proteins in ALS (Thomas et al., 2017; Porta et al., 2018). However, whilst misfolded proteins released by MNs can be taken up by glial cells, they may be less toxic to these cells than to MNs (Benkler et al., 2018).

Microglia
Microglia are the main immune cells of the CNS (Fujita and Kitamura, 1975; Hickey and Kimura, 1988; Lawson et al., 1990). In ALS patients, activated microglia increase in CNS regions that are susceptible to neurodegeneration (Kawamata et al., 1992) and in SOD1G93A mice, enhanced microglial reactivity precedes nerve denervation at the NMJ (Alexianu et al., 2001; Saxena et al., 2009). Microglia exist in both resting and activated states [reviewed in Perry and Holmes (2014)] and in ALS, activated microglia display two distinct phenotypes. The neuroprotective M2 phenotype promotes tissue repair and supports MN survival by releasing neuroprotective factors, and the toxic M1 phenotype produces cytokines, enhances inflammation, and induces cell death (Liao et al., 2012). Studies in mutant SOD1 mice reveal that the numbers of microglia increase during disease progression, but they vary between the neuroprotective M2 and toxic M1 phenotypes (Liao et al., 2012; Chiu et al., 2013). In lumbar spinal cords of pre-symptomatic SOD1G93A mice, the anti-inflammatory M2 microglia predominate (Gravel et al., 2016), whereas at disease onset and during progression, the proinflammatory M1 type is more common (Beers et al., 2011). Microglial-specific ablation of mutant SOD1G37R in mice does not affect disease initiation, but it significantly slows disease progression (Boillée et al., 2006b), indicating that microglia enhance the progression, but not the onset, of disease in transgenic mutant SOD1 mice. However, contradictory findings were obtained in the TDP-43 rNLS8 model, where microglia were neuroprotective and not neurotoxic (Spiller et al., 2018). Interestingly, knockdown of C9orf72 in mice alters microglial function and induces age-related neuroinflammation, but not neurodegeneration (Lall and Baloh, 2017). Further investigations are required to examine the role of microglia in other ALS disease models, and to determine whether MN subtypes display different vulnerabilities to microglia-mediated protective and/or toxicity in ALS.

Astrocytes
Astrocytes perform multiple homeostatic functions in the CNS; they regulate the plasticity of synapses and synthesis of neurotransmitters (Ullian et al., 2004; Volterra and Meldolesi, 2005; Sloan and Barres, 2014), they maintain the blood brain barrier, and they provide neurotrophic support to MNs by releasing glial-derived neurotrophic factor (GDNF) and transforming growth factor β1 (TGF-β1) amongst others. Like microglia, during the neurodegenerative process, astrocytes can exist in two states, either reactive or activated, and activated astrocytes lose their neuroprotective functions and become neurotoxic during disease (Yamanaka et al., 2008; Ilieva et al., 2009; Valori et al., 2014; Das and Svendsen, 2015). Also, like microglia, astrocytes are implicated in the progression rather than onset of ALS. Deletion of SOD1 from astrocytes slowed disease progression, but not disease onset, in SOD1G93A mice (Yamanaka et al., 2008; Wang L. et al., 2011), whereas deletion of mutant SOD1 from MNs did delay onset (Boillée et al., 2006a; Wang L. et al., 2009). Furthermore, gene expression changes in MNs, astrocytes and oligodendrocytes start just before disease onset in SOD1G37R mice, but these alterations are first observed in MNs (Sun et al., 2015). Recently, two different subsets of reactive astrocytes were described in the adult CNS, A1 and A2 (Liddelow et al., 2017; Clarke et al., 2018; Miller, 2018) and the A1 reactive astrocytes were associated with the death of both neurons and oligodendrocytes (Liddelow et al., 2017).
There is increasing evidence that astrocytes mediate MN degeneration via the release of neurotoxic factors. Soluble toxic compounds produced by astrocytes expressing mutant SOD1 trigger the selective loss of spinal MNs (Nagai et al., 2007), but not spinal GABAergic neurons, consistent with the specific vulnerability of these cells in ALS (Nagai et al., 2007). Astrocytes in the ventral spinal cord can be distinguished from astrocytes in the dorsal spinal cord by expression of semaphorin A3 (Sema3a), which is implicated in the specific vulnerability to FF-MNs in ALS (see section “Neuroprotective and Neurotoxic Factor Expression in MN Subpopulations” below). Furthermore, astrocytes are also implicated in MN loss and disease progression by mediating AMPA receptor-induced excitotoxicity via EAAT2/GLT-1, as discussed below (section “Neuronal Excitability”). Expression of mutant TDP-43M337V in rat astrocytes led to down-regulation of neurotrophic genes, up-regulation of neurotoxic genes and progressive MN degeneration (Tong et al., 2013; Huang et al., 2014). Conditioned medium from primary astrocyte cultures of SOD1G86R and TDP-43A315T mice also induces MN death through activation of sodium channels and nitro-oxidative stress (Rojas et al., 2014). Furthermore, astrocytes expressing mutant FUSR521G trigger MN death by secreting pro-inflammatory tumor necrosis factor (TNF)-α (Kia et al., 2018). SOD1G93A aggregates in astrocytes appear in late disease stages, selectively in regions with extensive neuronal degeneration and prominent astrogliosis (Jaarsma et al., 2008). This raises the possibility that astroglial aggregate formation is triggered by MN degeneration, implying that disease may spread from neurons to glia (Jaarsma et al., 2008; Sun et al., 2015).
Together these studies suggest the involvement of astrocytes in the selective degeneration of MNs in ALS. Under normal conditions, astrocytes may be able to cope with the expression of low levels of misfolded proteins, but, during cell stress or in the context of MN degeneration, they become more vulnerable, and release factors toxic to MNs, thus producing a vicious cycle. However, the relative resistance of neuronal populations surrounded by reactive astrocytes indicates that the vulnerability of MNs is also determined by cell-autonomous components, such as their genetic background and transcriptional/translational profiles (Boillée et al., 2006a; Sun et al., 2015).

Oligodendrocytes and Schwann Cells
The two glial cell types responsible for myelination of axons have also been investigated in the context of ALS. Oligodendrocytes myelinate axons in the CNS whereas Schwann cells are responsible for myelination in the peripheral nervous system (PNS). Whilst they perform similar functions, there are also important differences between these two cell types. Schwann cells form a single myelin sheath around one single axon, whereas oligodendrocytes myelinate many different axons. Furthermore, there are differences in the protein composition of CNS and PNS myelin.
In ALS, TDP-43 pathology has been detected in oligodendrocytes in the motor cortex and spinal cord of both SALS and FALS patients (Arai et al., 2006; Mackenzie et al., 2007; Tan et al., 2007; Zhang et al., 2008; Seilhean et al., 2009; Murray et al., 2011; Philips et al., 2013). In addition, FUS forms cytoplasmic aggregates in oligodendrocytes from ALS patients bearing FUSR521C or FUSP525L mutations (Mackenzie et al., 2011). Degeneration of oligodendrocytes and their precursors was also linked with axon demyelination in both SALS and FALS patients (Kang et al., 2013). In SOD1G93A mice, oligodendrocyte loss in the spinal cord occurs before symptoms appear and importantly, before MN loss, implying that oligodendrocytes are associated with disease onset. This MN loss increases with disease progression, resulting in MNs with only partially myelinated axons in SOD1G93A mice and SOD1G93A rats (Niebroj-Dobosz et al., 2007; Kang et al., 2013; Philips et al., 2013). Whilst the proliferation of oligodendrocyte precursors may compensate for this loss, newly synthetized oligodendrocytes failed to mature and remain dysfunctional in SOD1G93A mice (Magnus et al., 2008; Philips et al., 2013). Recently, SOD1G85R was able to transfer from MNs to nearby oligodendrocytes (Thomas et al., 2017). The selective removal of mutant SOD1 from NG2+ oligodendrocyte progenitors, but not mature oligodendrocytes in SOD1G37R mice, leads to delayed disease onset and prolonged survival (Kang et al., 2013), further suggesting that mutant SOD1-induced oligodendrocyte defects are detrimental to MNs in ALS.
Schwann cells are required for the long-term maintenance of synapses at the NMJ (Reynolds and Woolf, 1992; Son and Thompson, 1995; Reddy et al., 2003). Early studies demonstrated that myelin is altered along peripheral nerves in ALS patients, implying that Schwann cells are involved in disease (Perrie et al., 1993). However, unlike the other glial cell types, more recent studies on the role of Schwann cells in ALS have reached conflicting conclusions. Knockdown of SOD1G37R within Schwann cells significantly accelerates disease progression, concomitant with a specific reduction in insulin-like growth factor (IGF-I), which is protective to MNs (see section “Neuroprotective and Neurotoxic Factor Expression in MN Subpopulations” below) (Lobsiger et al., 2009). This surprising finding, implying that SOD1G37R is protective in Schwann cells, could be linked to the dismutase activity of SOD1. Whereas SOD1G37R retains its enzymatic activity, SOD1G85R does not, and similar experiments performed in SOD1G85R mice resulted in opposite findings; Schwann cell specific knock-down of SOD1G85R delayed disease onset and extended survival (Wang et al., 2012). Furthermore, TGF-β1 produced by Schwann cells promotes synaptogenesis by increasing nerve-muscle contacts (Feng and Ko, 2008), in contrast to TGF-β1 expression in astrocytes which accelerates disease progression in SOD1 mice (Endo et al., 2015). Hence, the role of Schwann cells in ALS remains unclear.