**2.2 Histopathological studies in ALS and neuroinflammation in retinal tissue**

(**Table 1**, **Figure 1**) Neuroinflammation is a pathophysiological mechanism, which involves the activation of astroglial and microglial cells, and it occurs in many neurodegenerative diseases, such as Parkinson's disease, Alzheimer's disease, ALS, and glaucoma [62]. Microglial cells are the macrophages of the CNS and have the ability to respond to injury by becoming activated; they can proliferate, migrate, and change shape, acquiring an amoeboid appearance in the most active state [62, 71]. On the one hand, in an attempt to protect against damage, microglial cells can secrete proinflammatory molecules, such as interferon γ or interleukin (IL)-1β [62].

#### **Figure 1.**

 *Summary of retinal changes in amyotrophic lateral sclerosis. (A) Healthy retina. (B) Retinal changes in ALS. Most of retinal changes are mainly detected in the inner layers: (i) ganglion cell loss; (ii) activated microglia in outer plexiform layer (OPL) and inner layer complex (ILC) (constituting an inner plexiform layer and a nerve fiber–ganglion cell layer); (iii) p62-positive and TDP43-negative protein aggregates mostly in the inner nuclear layer with some in the ganglion cell layer (GCL); (iv) ubiquilin 2-positive inclusions mostly in the inner plexiform layer (IPL), with a smaller amount in the outer plexiform layer (OPL) and in the GCL; (v) periodic acid Schiff (PAS)-positive spheroids in the retinal nerve fiber layer (RNFL) and phosphorylated form of neurofilament (P-NF)-positive spheroids in the in the peripheral and peripapillary RNFL; (vi) non-phosphorylated form of neurofilament (NP-NF)-positive spheroids in the RNFL; (vii) hSOD1-positive vacuoles in in the IPL, with hardly any in the GCL and inner nuclear layer (INL).* 

Nonetheless, uncontrolled activation of the M1 phenotype can lead to a state of chronic inflammation, which can induce neuronal death. On the other hand, microglial cells can also secrete anti-inflammatory molecules, such as IL-10 and the enzyme arginase 1 (Arg1), in order to control inflammation, repair tissue, and improve neuronal survival [ 62 , 71 – 73 ]. Consequently, activated microglia can acquire two different activation phenotypes: an M1 or proinflammatory phenotype vs. an M2 or anti-inflammatory phenotype, both of which can be influenced by molecules derived from surrounding cells such as astrocytes [ 62 , 74 ]. Astrocytes are glial cells of ectodermal origin that perform numerous functions for neuronal survival [ 75 ], such as maintenance of the volume and composition of the extracellular space, maintenance of the blood-brain barrier, and regulation of synaptic transmission [ 76 ], as well as metabolic maintenance and neuronal survival [ 77 , 78 ]. When astrocytes are damaged and consequently activated, "astrogliosis" occurs [ 79 ]. If this astrogliosis is severe, a glial scar may form [ 75 ]. Reactive astrocytes can interact with microglia and neurons and can impair the function of neurons after an injury [ 80 ].

 Astrocyte activation [ 81 ], microglial activation [ 82 ], and the appearance of lymphocytes [ 83 ] have been found in animal models of ALS (with SOD1 mutations) and in ALS patients. In ALS there are reactive microglia and astrocytes, which can result in motor neuron injury and subsequent death [ 73 , 74 , 81 ]. The SOD1G93A mouse model is one of the most suitable and widely used for preclinical studies in ALS, attributable to the animals having an analogous phenotype to patients. These animals develop limb paralysis due to the loss of motor neurons in the spinal cord, with a reduced lifetime of 150 days [ 84 ]. Microglial activation also occurs in ALS, as observed in SOD1-mutated mice and in spinal cord samples from ALS patients, which could exacerbate neuronal damage [ 73 , 74 , 81 , 85 ]. In fact, it has been shown that exogenous extracellular mutation of SOD1G93A is not directly toxic to motor neurons, but requires microglial activation for toxicity in primary motor neuron and glia cultures [ 86 ]. Furthermore, in SOD1 transgenic mice, activated astrocytes and microglia have been shown to contribute to disease progression but not to disease

#### *Retinal Disorders in Humans and Experimental ALS Models DOI: http://dx.doi.org/10.5772/intechopen.107052*

onset [87–89]. In ALS, microglial activation and proliferation have been observed in areas of significant motor neuron loss, such as the motor cortex, brainstem motor nucleus, corticospinal tract, and ventral horn of the spinal cord [90–94], as well as in areas with mild degeneration [95]. Precisely, in postmortem spinal cord analysis of patients with advanced stages of ALS, reactive astrocytes were found in the dorsal and ventral horn of the spinal cord [96] and in the gray [97] and white matter [34] of the cerebral cortex. Similarly, reactive microglia were found in the motor nucleus of the brainstem, motor cortex, corticospinal tract, and ventral horn of the spinal cord [91]. Reactive microglia were also observed in vivo using the PET imaging technique C-PK11195, finding a close relationship between microglial activation and upper motor neuron damage, but not lower motor neuron injury [98]. Moreover, in the SOD1 model, it was confirmed that overexpression of the SOD1 mutation in glial cells contributes to motor neuron damage, and that the degree of neuronal injury depends on the degree of glial cell pathology [99]. Microglial cells of SOD1-mutated mice suffer different degrees of morphological changes from resting to macrophagic amoeboid forms [91]. Lastly, symptomatic SOD1 transgenic mice also have increased numbers of microglial cells, mainly due to the proliferation of resident microglia [100].

Bearing in mind all of the above, both the microglia and the astrocytes play an important dual role in the progression of the ALS. Nevertheless, most studies about the involvement of microglia in ALS have been conducted in the motor cortex, brainstem motor nucleus, corticospinal tract, and ventral horn of the spinal cord. To our knowledge, there are only three studies that investigated the glial cells of the retina in relation to ALS [101–103].

In the first one, a mouse model of ALS devoid of RAN-binding protein 2 (Ranbp2), microglial activation was confirmed. Ranbp2 is a protein, which plays an important role in nucleocytoplasmic transport and whose regulation is affected in both sporadic and familiar ALS [104]. In this ALS mouse model, there was microglial activation with an increase in the number of microglial cells surrounding retinal ganglion cells (RGCs), as well as a noteworthy increase in amoeboid forms relative to controls. In addition, there was an increase in metalloproteinases in RGCs, and both hypertrophy in RGCs and axonopathy in the optic nerve were found [102].

In the second model, a TG mouse model of ALS SOD1 (SOD1G93A), there was a vacuolization, with hSOD1-positive vacuoles placed in the dendrites of excitatory retinal neurons, which were detected principally in the inner plexiform layer (IPL) and hardly in the GCL and INL; however, no signs of activation of either the astroglia or the microglia of the retina were shown compared with to the wild-type mice [101]. However, the authors did not rule out the possibility that the microglia were undergoing functional changes (in cytokines) related to the inflammatory process. Nevertheless, neuronal changes observed in this SOD1G93A ALS model in the brain at 50 days of age were followed by microglial morphological changes at 60 days [105–107]. Therefore, the authors concluded that, if there is an inflammatory process in the retina, microglia would be in a different, less reactive or even neuroprotective phenotype [101].

Lastly, the third transgenic murine SOD1G93A model of ALS in an advanced stage of the disease (120 days) showed a loss of the number of Brn3a+ RGCs and a microglial activation in retinal tissue [103]. Signs of microglial activation were found in different retinal sectors (superior, inferior, nasal, and temporal) of different retinal layers: outer plexiform layer (OPL) and inner layer complex (ILC) (constituted by an inner plexiform layer and a nerve fiber–ganglion cell layer). In addition, the microglial activation in this SOD1G93A model of ALS showed a cell thickening in the area occupied by each microglial cell, a significant increase in the area of microglial

arborization with hyper-ramifications in the inferior sector of the OPL, retractions of cell processes, and migration and clustering of cells in some areas of the retina, but no increase in the number of microglial cells [103]. Moreover, phenotypic analysis of the microglia showed an M1 phenotype or proinflammatory state of microglia, as the cells were intensely labeled with anti-IFNγ and anti-IL-1β but did not stain with the characteristic M2 markers (anti-arginase 1 and anti-IL-10) [103]. The significant decrease in the total number of Brn3a<sup>+</sup> RGCs at 120 days of illness would be consistent with the damage observed in the RGCs of the ALS models discussed above [37, 101, 102], as well as with the thinning of the peripapillary retinal nerve fiber layer (pRNFL), observed by OCT, in ALS patients compared with controls [36–44]. Consequently, these data would support that, in ALS, not only are motor neurons affected but also RGC loss occurs, considering this disease as a multisystemic disease [103].

In none of the abovementioned models were changes in the outer segments of the photoreceptors found. This could indicate that neither this layer of the retina nor the outer blood-retinal barrier (BRB) would be compromised in these animals. Because, when the outer BRB is disrupted, as in a glaucoma model of laser-induced ocular hypertension, there are morphological changes and an increase in the number of microglial cells in the photoreceptor outer segment layer [108–112]. Moreover, no changes in the number of microglial cells were found in either the OPL or the ILC [101, 103]; however, the group of Rojas et al. described signs of microglial activation [103]. This difference in results in the same experimental model could be due to the fact that Ringer et al. [101] used retinal sections, while Rojas et al. [103] used retinal whole mounts.

As mentioned above, microglial cells have two distinct phenotypic states that can exert neurotoxic or neuroprotective responses depending on the physiological conditions in which they are found. During ALS progression, activated microglia represent a continuum between the neuroprotective M2 phenotype and the neurotoxic M1 phenotype [113]. In SOD1 ALS animal models, in early stages of the illness, microglia in the lumbar spinal cord expressed markers related to the M2 neuroprotective phenotype (Ym1 and CD206); however, in the late stages of the disease, microglia in the lumbar spinal cord expressed markers related to the M1 neurotoxic phenotype (high levels of NADPH oxidase 2 (NOX2)) [74], suggesting that there is a polarization from a neuroprotective phenotype to a cytotoxic phenotype that induces motor neuron damage. In the retina, there is only one study that analyzed whether microglia are in an M1 or M2 activation phenotype [103]. The results of this study showed that, in 120-day-old SOD1G93A mice, microglia were strongly labeled with antibodies against M1 inflammatory cytokines (IFNγ and IL-1β), but not with those against M2 anti-inflammatory cytokines (arginase-1 and IL-10), suggesting that at an advanced stage of the disease retinal microglial cells are in an M1 activation phenotype or in a pro-inflammatory state that could be neurotoxic to RGCs, as demonstrated by the loss of these neurons. These results are consistent with the findings in spinal cords of the same animal model, where microglia in an advanced stage of the disease showed a neurotoxic M1 phenotype, demonstrating the dual role (neuroprotective/neurodegenerative) of microglial cells during the ALS process [74]. Therapeutic approaches that target microglia polarization and result in the induction of the M2 phenotype are promising strategies to ameliorate local neurodegeneration and clinical outcome of the disease [114].

#### **2.3 Histopathological studies in ALS and retinal spheroids and axon pathology**

(**Table 1**, **Figure 1**) Alterations in axonal transport (retrograde and anterograde) are a hallmark of ALS, being impaired both in ALS patients and in mutant SOD1

#### *Retinal Disorders in Humans and Experimental ALS Models DOI: http://dx.doi.org/10.5772/intechopen.107052*

mice. In the spinal motor neuron axons, an accumulation of altered mitochondria, neurofilaments, and autophagosomes [12, 58] was demonstrated. On the one hand, mutated dyneins in ALS mice cause this accumulation in the axons of mitochondria and autophagosomes [58]. On the other hand, altered autophagosomes do not eliminate either altered mitochondria or dilated endoplasmic reticules, which accumulate in the axons of motor neurons and cause them to malfunction [12].

There is only one study that focused on this important pathological mechanism in the retina [115]. This study analyzed retinal sections of postmortem eyes from ALS patients with periodic acid Schiff (PAS) and phosphorylated (P-NF) and nonphosphorylated (NP-NF) forms of neurofilament (NF), compared with age-matched controls. Three kinds of spheroids were revealed. First, PAS-positive spheroids with a diameter bigger than 9.07 μm in the retinal nerve fiber layer (RNFL) were observed in most ALS patients (but only in half of controls), most commonly in the pRNFL and the peripheral RNFL, but rarely in the central RNFL in patients with ALS. The density of PAS-positive spheroids was significantly greater in the pRNFL. Second, P-NF-positive spheroids ranging from 8 to 15 μm in diameter were observed in the peripheral and pRNFL only in ALS patients. Additionally, ALS patients showed a stronger P-NF signal intensity in the RNFL in the peripheral, central, and peripapillary regions. Third, NP-NF spheroids ranging from 7 to 10 μm in diameter were observed in the RNFL in some of ALS patients (but not in controls). In addition, in most of the ALS patients, the NP-NF signal was increased in the RNFL and IPL. Nevertheless, there was no significant correlation of these retinal spheroids and axon pathology with the clinical characteristics of the ALS patients (age at death, gender, disease duration, mode of disease onset, revised ALS functional rating scale, and rate of disease progression) [115].

Consequently, patients with ALS show not only hallmark findings in spinal cord motor neurons pointing to disrupted axon transport [116–121] but also retinal spheroids and axon pathology as a shared pathogenesis [115]. Transgenic mice with dysfunctional microtubule-associated motor proteins also display such findings [122–124].

#### **2.4 Retinal vessel pathology and ALS**

(**Table 1**, **Figure 1**) Retinal vessels are a reflection of small blood vessels in the brain [125]. Parallel vessel pathology in the retinal and cerebral small blood vessels has been demonstrated in many systemic diseases such as coronary heart disease [126] or stroke [127], as well as in some neurodegenerative diseases such as Alzheimer's disease [128, 129] (even in subjects at high genetic risk of developing Alzheimer's disease [130]) and Parkinson's disease [131].

Some ALS-induced changes have also been described in small blood vessels of the brain, which include a loss of pericytes, endothelial cell degeneration, capillary leakage, downregulation of tight junction proteins, and microhemorrhages in patients with ALS [132, 133]. Moreover, alterations of the structure of small blood vessels of the skin and muscles in ALS patients have been described [134, 135].

There was only one study that analyzed retinal vessel pathology in ALS patients with Spectralis OCT but not with angio-OCT. This study described a thicker outer wall of retinal vessels in ALS patients compared with controls, which may be related to the findings in small blood vessels in skin and muscle biopsies. There were neither significant differences in the vessel diameters between ALS patients with spinal onset and bulbar onset, nor a correlation between the vessel measurements and clinical parameters (disease duration and ALS Functional Rating Scale—Revised (ALSFRS-R)) [136].
