Amyotrophic lateral sclerosis (ALS) is a highly heterogeneous degenerative disease characterized by the deterioration of voluntary muscle function due to motor neuronal damage. However, there are some recent studies supporting the frequent involvement of other non-motor systems, including the visual pathways (Munte et al., 1998). The optic nerve and cells are unique parts of the central nervous system giving origin to 1.2 million axons and retinal ganglion cells (RGCs) and they are particularly vulnerable to neurodegeneration (Carelli et al., 2017). Moreover, optic neuropathies and ALS share important common pathogenic mechanisms, including increased levels of oxidative stress, mitochondrial damage and axonal transport disorders (Maresca et al., 2013). And a histopathological study demonstrates widespread astrogliosis in the subcortical white matter of occipital cortex in ALS (Kushner et al., 1991).

Since visual complaints are not common in ALS patients, relatively little is known about possible retinal alterations upon clinical examination. Optical coherence tomography (OCT) is a non-invasive imaging technique that is accurate, objective, and reproducible. It could provide high-resolution images of the retina and precisely and objectively quantify discrete neuronal layers, including the peripapillary retinal nerve fiber layer (RNFL, representing the unmyelinated axons of retinal ganglion cells), retinal ganglion cell bodies and the inner plexiform layer (GCIP, representing the health of the ganglion cell layer). This technique has been applied to detect neuroaxonal retinal abnormalities in neurodegenerative disease like Parkinson's Disease (Huang et al., 2020), progressive supranuclear palsy (Schneider et al., 2014), multiple system atrophy (Schneider et al., 2014) and Alzheimer's disease (Kirbas et al., 2013).

However, the reported changes in the retinal layers in ALS are not consistent. The first study performed in 76 ALS patients showed no significant difference from the controls for RNFL (Roth et al., 2013), while the majority of studies assessing RNFL found significant thinning within the ALS group (Ringelstein et al., 2014; Hubers et al., 2016; Mukherjee et al., 2017; Rohani et al., 2018; Rojas et al., 2019), with the superonasal quadrants being the most prominent (Mukherjee et al., 2017; Rohani et al., 2018; Rojas et al., 2019). However the most recent study conducted by our institution showed that the nasal quadrant of ALS RNFL even had a statistically thicker measurement (Liu et al., 2018). These studies included patients in different stages but did not stratify them according to their clinical phenotypes. We hypothesized that an alteration of a specific single retinal layer varied in ALS patients owing to disease heterogeneity.

In this study, we aimed to assess selective retinal nerve alterations in ALS patients in different stages and their associations with ALS progression parameters.

The retina is a highly structured tissue sharing embryonic commonality with central nervous system. It can be divided into 10 layers histologically. Photoreceptors transmit light stimulus to RGCs (Cervero et al., 2021), the neuronal cell type that projects its axon to the brain, forming the optic nerve (Carelli et al., 2004; Yu-Wai-Man et al., 2011). RGCs are the usual target in neuropathies implying an impaired mitochondrial function (Maresca et al., 2013), and this vulnerability is attributed to the unique axonal structure, characterized by a long intraretinal segment that remains unmyelinated (Carelli et al., 2004; Yu-Wai-Man et al., 2011). Since mitochondrial dysfunction is a major event in the pathology of ALS (Cozzolino and Carri, 2012), the degree of retinal thinning in ALS patients has been recently studied, with promising but discordant results. This retrospective, cross-sectional study enrolling 70 ALS patients demonstrated for the first time that (1) the RNFL is selectively targeted in specific patients; (2) when controlling for patient age and ΔFS, the RNFL formed an inverse U-shaped curve transformation, which is related to but not fully concordant with the thickness change of the RGCs; (3) The thickening of RNFL may be related to peripheral nerve motor axonal damage.

In group-wise comparisons, age related decline of RNFL thickness that has been previously reported (Parikh et al., 2007; Leung, 2012; Hondur et al., 2018) was also observed in our study. Thus age was included as a controlling variable in the relevant analysis. Due to the disequilibrium in sample size and variance between each group, we did not find statistical differences in RNFL when patients were grouped by disease duration or KCSS.

In the relevant analysis, we found that the nasal and global quadrants of the RNFL in ALS patients were significantly thickened, which is consistent with previous studies (Liu et al., 2018; Rojas et al., 2019). According to the longitudinal study conducted by Rojas et al. (2019), there is a remarkable thickness evolution from increased to thinning in both the macula and the RNFL. Notably, we demonstrated that the turning point of the alteration curve might be approximately 12 months, which has been regarded as a relatively early stage of the disease (Chio et al., 2013; Marin et al., 2016; Gille et al., 2019). In both animal models and ALS patients, neuroinflammation status was more prominent during the initial 12 months (Chen et al., 2019), and neuroprotection was indicated to be the major modulation (Chiu et al., 2008; Henkel et al., 2013; Ehrhart et al., 2015). As an extension of the brain, the same process occurs in the retina (Ramirez et al., 2017), and microglial cells intermingling between RGCs play a key role (Glass et al., 2010). Activated microglia displayed retraction of processes and enlargement of the soma, resulting in the increased thickness of RGCs. Moreover, mitochondrial dysfunction in ALS could be responsible for calcium homeostasis disruption and axonal transport impairment (De Vos and Hafezparast, 2017), which is characterized by axonal swelling due to stalling of transport (Carelli et al., 2004; Yu-Wai-Man et al., 2011). These observations suggested that in the early stage of ALS, neuroinflammatory activation of microglia and an increase in axonal caliber could produce the increased thickness observed by OCT. As the disease progresses, microglial activation switches from a neuroprotective M2 to a toxic M1 phenotype (Liao et al., 2012; Chiu et al., 2013). In this state, microglia release neurotoxic inflammatory factors that lead to retinal neuronal death (Rojas et al., 2021) (Figure 4). Additionally, Rohani et al. (2018) suggested that the thinning of the RNFL could be related to Wallerian degeneration secondary to the death of cortical neurons.

The canonical view is that RGC degeneration may affect the soma after axonal involvement in a retrograde march (Carelli et al., 2004). In our study, GCIP was in accordance with the RNFL thickness in each group and it thickened significantly along with the RNFL in the initial 12 months. Nevertheless, we did not detect a decline in GCIP thickness similar to that of the RNFL one year later, indicating that the earlier pathological events may involve axons of nerve fibers preceding RGCs loss, which suggests a “dying-back” degeneration pattern (Dadon-Nachum et al., 2011) in the retina of ALS.

To detect the association between early-stage axonal degeneration of retinal nerve fibers and peripheral motor neurons, we assessed the CMAP amplitude of the distal motor nerve and calculated the CMAP decline score. In the T-RNFL group, the CMAP score was higher although not significant. Notably, when controlling for ΔFS and the site of onset, the thickening of the RNFL was significantly correlated with the decline in APB-CMAP within 12 months. From our previous research, the discrepancy in the CMAP was obvious at 12 months (Yu et al., 2021), and the decline in the CMAP at the distal median nerve within 12 months can reflect peripheral nerve axonal injury and indicate disease prognosis outside the onset site (Eisen and Kuwabara, 2012; Imai et al., 2020; Yu et al., 2021). In our study, we also found that patients in the T-RNFL group within 12 months progressed faster than others. Dibaj et al. (2011) found that during the course of microglia-mediated neuroinflammation in the central nervous system, degenerated lower motor axons were also surrounded by phagocytic phenotype macrophages. Therefore, early stage RNFL thickening might be related to the axonal inflammatory response and axonopathy. In the I-RNFL group, RGCs and peripheral motor neurons were evidently lost, and deep tendon reflexes were hard to induce in the clinically weak and wasted muscles. In this scenario, we found that the UMN score declined significantly with the thinning of the RNFL.

Our study showed that there was significant heterogeneity in RNFL changes among different ALS patients. This type of selective alteration may be related to the patients' genotype, metabolism in vivo, environment in vitro and the expression level and spread rate of pathogenetic proteins. RNFL thinning in the early stage may indicate a significantly faster deterioration (Table 5); however, patients whose RNFL was still unaffected after one year may have a relatively slower rate of progression (Table 4). The selective alteration of the RNFL, especially in the early stage, could be a new non-invasive biomarker to predict disease progression.

With the burgeoning OCT analysis in ALS patients, our findings about the retinal nerve fibers may represent a unique susceptibility of the RNFL to ALS pathology. However, there are several limitations to this study. Although we controlled for age and ΔFS to minimize the heterogeneity of the disease in correlation analysis, this is still a retrospective, cross-sectional study. A follow-up evaluation will be more convincing to prove the alteration trend. Second, although previous studies have detected alterations of the early components of the visual evoked potentials (VEPs) (Munte et al., 1998), we did not record VEPs in our cohorts, since this examination is time-consuming and requires cooperation from the patient. Third, the relationship between RNFL thickness and magnetic resonance imaging (MRI) measurements of optic nerve and brain structure volumes was not detected in our study. Since previous studies have reported the association between RNFL thickness and the volumes of brain regions vulnerable to aging or neurodegeneration (Shi et al., 2020), more imaging parameters are expected to be included in future studies. Further, we did not take macular layers such as the inner nuclear layer, outer plexiform layer or outer nuclear layer into account. Kim et al. (2019) found progressive outer retinal thinning in frontotemporal degeneration (FTD) patients. Given that up to 13% of ALS cases are associated with coincident FTD (Phukan et al., 2012), it is noteworthy to analyze the cognitive impairment and outer retinal thickness in this cohort in the future.

Visual symptoms are not a prominent complaint in most patients with ALS and can be easily concealed by motor system involvement. Our data support previously reported clinical and neurophysiological studies on visual disturbances in ALS (Rojas et al., 2019) and indicate a neurodegenerative process involving the retina. According to the four pTDP-43 pathology propagation stages (Brettschneider et al., 2013), occipital cortex and retinal axons rarely showed pTDP-43 inclusions in early stage, thus RNFL thinning in the early stage may indicate a faster propagation and deterioration. This inverse U-shaped curve transformation might be due to the initial priming and then inflammatory reaction of microglia intermingling between retinal nerve axons, and the RNFL thickening may be related to the early axonal inflammatory response. In this scenario, retinal changes emerge as a complementary window to approach the disease and can be used to assess disease progression in ALS patients.

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author.

The studies involving human participants were reviewed and approved by the Ethical Committee of Peking University Third Hospital (approval No. M2019388). The patients/participants provided their written informed consent to participate in this study.

YiZ and DF designed the research. YiZ wrote the main manuscript text. YiZ, XL, and JF performed data analysis and interpreted the findings. YuZ and SZ contributed to the recruitment and evaluation of patients. XY contributed to the acquisition of imaging data. All of the authors read the draft, made contributions, and approved the final manuscript.

This study was supported by the National Natural Science Foundation of China (81873784 and 82071426) and Clinical Cohort Construction Program of Peking University Third Hospital (BYSYDL2019002).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.