Abstract

Introduction

Reactive oxygen species were traditionally considered to be detrimental to cell integrity and held responsible for a variety of damages to cellular structures, ultimately resulting either in premature cell death by apoptosis or in cancerogenesis (Lushchak, 2014). Over the course of years of intensive research, a more differentiated view on the role of reactive oxygen species both in health and disease was developed and continues to be refined (LeVine, 1992; Auten and Davis, 2009; Forman et al., 2010; Fang, 2011; Szumiel, 2011).

Among the enzymes regulating reactive oxygen species, superoxide dismutases (SODs) play an important role by facilitating the transformation of the superoxide anion (O₂^•−) into hydrogen peroxide (H₂O₂), which is then further processed by a variety of enzymes (Fridovich, 1997). Humans express three distinct SODs, namely SOD1–3. While SOD2 is localized within the mitochondria and SOD3 is located extracellularly, SOD1 is mainly found in the cytoplasm (Fukai and Ushio-Fukai, 2011). However, minor amounts are also localized in the mitochondrial intermembrane space and the nucleus of eukaryotic cells (Higgins et al., 2002; Chung, 2017).

Previously, variants of SOD1 have been implicated in the pathogenesis of familial amyotrophic lateral sclerosis (ALS), a debilitating neurological disorder characterized by the progressive degeneration of motor neurons. SOD1 variants are causative for a significant proportion of familial ALS cases, ranging from about 13% in western to up to 30% in Asian populations (Kaur et al., 2016; Zou et al., 2017). While initially believed to exert mainly gain-of-function effects on the enzyme’s activity, it has become clear that the described variants result in a variety of effects, some of which are suggestive of a prion-like pathomechanism (Vijayvergiya et al., 2005) or mitochondrial dysfunction (Magrane et al., 2009). These emerging new understandings of the pathomechanism of familial ALS and indeed also sporadic ALS (Alexander et al., 2002) continue to give insights into the origins of this debilitating disorder. Interestingly, the overwhelming majority of SOD1 variants associated with familial ALS show an autosomal-dominant inheritance pattern with homozygous variants being a rare exception (Andersen et al., 1995, 1996, 1998). Heterozygous SOD1 variants are therefore well established as a cause of familial ALS. These variants result in varying levels of SOD1 enzyme activity, ranging from severely reduced to levels above those observed in wild-type SOD1 (Borchelt et al., 1994; Keskin et al., 2017). Notably, the observed enzyme activity shows no association with clinical severity (Cleveland et al., 1995).

There is a growing number of Sod1-deficient mouse models. While many were created using ALS-causing mutations observed in humans in order to study the associated phenotype (Dal Canto and Gurney, 1995; Tu et al., 1996), complete knockouts have also been created. Interestingly, no motor neuron disease has been observed in these mice (Reaume et al., 1996). However, they show extensive muscle involvement consisting of progressive motorneuronopathy with axonal denervation (Frey et al., 2000; Hegedus et al., 2007) resulting in secondary muscle pathology (Muller et al., 2006). This phenotype is associated with increased oxidative stress secondary to mitochondrial dysfunction (Fischer et al., 2012).

In addition to neuronal manifestations, these animals exhibit extraneuronal phenotypes with a shortened lifespan, hepatocellular carcinoma (Elchuri et al., 2005) and altered hepatic energy metabolism (Wang et al., 2012) as well as exhibiting endocrinological abnormalities (Matzuk et al., 1998). For an extensive review of Sod1-deficient murine models, see Saccon et al. (2013).

In this study, we report on a homozygous loss-of-function SOD1 variant identified in a patient with a debilitating neurological phenotype. The variant leads to SOD1 activity levels below measurable ranges and is associated with a phenotype marked by hyperekplexia, ataxia and muscular hypotonia in addition to severe psychomotor retardation.

Materials and methods

Results

Case report and radiography studies

The index patient is the third child of consanguineous parents (first degree cousins) of Afghan origin (Fig. 1). During pregnancy, polyhydramnios was noted. The patient was born at 39 weeks of pregnancy via emergency Caesarean section due to a maternal indication.

Open in a separate window

Figure 1

Pedigree of the index patient. His parents are first degree cousins. No other affected individuals were reported. Notably, there are no known cases of (familial) ALS in the kinship, while many of the reported individuals are at or above the mean age of onset observed in SOD1-related familial ALS.

The family reported normal development until the age of 9 months when the patient was able to crawl. Shortly afterwards, progressive psychomotor decline marked by loss of motor abilities and progressive ataxia began.

Upon initial presentation at our tertiary care centre at the age of 6 years, he showed a significant combined developmental delay as he was not able to sit or stand unsupported. Additionally, he presented with dysmorphic features such as low set, posteriorly rotated ears (Fig. 2A) and overlapping toes (T2, 3 on the left and T3, 2 on the right) (Fig. 2B).

Open in a separate window

Figure 2

Phenotype of the index patient at 6 years of age. Dysmorphic features such as low set, posteriorly rotated ears and overlapping toes are present. A clinical examination demonstrating hyperekplexia-like symptoms is available online.

Clinical examination revealed an increased muscle tone of both upper and lower limbs with persistently bended arms in pronated position whereas muscular mass and distribution appeared normal. Corresponding to the spasticity he demonstrated hyper-reflexive brachioradial reflexes and more pronounced hyperreflexia in patellar reflexes together with enlarged reflex zones. Furthermore, we observed pyramidal path symptoms such as positive Babinski sign and bilateral exhaustible clonus of the feet (Supplementary Video 1). In addition, we stated pronounced symptomatic hyperekplexia with persistent glabellar tap sign as well as an incomplete Moro reflex consisting of the initial abduction, while the typical subsequent adduction reaction was lacking.

Truncal and proximal muscular hypotonia leading to the inability to sit or stand was noted. Therefore, it was not possible to examine for truncal ataxia. The patient showed no nystagmus or oculomotor dysfunction, but slight tremor of both hands and fingers upon movement. Cognitive functions were impaired; however, he was able to understand and/or implement simple correlations but not complexes requirements (e.g. directed pointing). Therefore, detailed assessment of these symptoms remains unanswered. Only non-verbal communication is possible.

Cranial MRI was performed at the age of 2 and again at 6 years. The initial MRI was inconclusive. At the age of 6 years, mild cerebellar atrophy with discreetly enlarged interfoliar spaces in the region of the anterior vermis was diagnosed (Fig. 3).

Open in a separate window

Figure 3

T₂-weighted cranial MRI of the index patient at the age of 6 years. In general, mild cerebellar atrophy is present. The cerebellar interfoliar spaces at the anterior vermis are discreetly enlarged (indicated by arrows) as are the cerebellar hemispheres in the lateral view (D).

Extensive metabolic screening, including serum organic acids, serum acylcarnitine profile, urinary oligosaccharides and amino acids, purines/pyrimidines, and serum lactate analysis was performed but results were inconclusive.

Evaluation of antioxidant vitamins and trace elements revealed consistently low levels of blood manganese between 3.1 and 4 ng/ml (reference: 7–11 ng/ml) and zinc [672 µg/l (reference: 750–1400 µg/l)].

Heterozygous carriers identified within this family did not show any overt neurological phenotype. There were no cases of ALS reported in this family.

Superoxide dismutase functional assay

The activity of SOD1 of the index patient was below the detection level of 2 units/mg, whereas SOD2 activity was within normal ranges. Heterozygous carriers of the identified variant had an approximately halved SOD1 activity of 12 and 11 units/mg when compared to wild-type controls (55 and 51% of wild-type SOD1 activity, respectively).

In contrast, the activities measured in three healthy control subjects were all within reference ranges (Fig. 4).

Open in a separate window

Figure 4

SOD1 activity measurement of the patient and heterozygous carriers compared to wild-type controls. Measurements were performed on isolated blood leucocytes. The index patient had no detectable SOD1 activity. Heterozygous carriers showed an approximately halved SOD1 activity of 12 and 11 units/mg protein, corresponding to 55% and 51% of the activity measured in wild-type (WT) controls, respectively.

Discussion

In this study, we describe a patient carrying a truncating homozygous SOD1 variant resulting in the total absence of SOD1 activity. Human SOD1 deficiency has, to our knowledge, never been described before.

The link between SOD1 variants and familial ALS has been firmly established. Indeed, research on SOD1-related ALS has been an important way of delineating the pathogenesis of this debilitating neurodegenerative disorder.

A vast number of variants in SOD1 have been identified in familial ALS but despite extensive research, the exact pathomechanism by which these alterations cause familial ALS remains elusive. Recently, it has become clear that there are probably several ways in which SOD1 variants exert their causative role in the disorder. It has been firmly established that a vast number of variants leads to a gain-of-function of the gene product SOD1 (Bruijn et al., 1998; Allen et al., 2003). Furthermore, SOD1 aggregation (Jonsson et al., 2008) and mitochondrial dysfunction (De Vos et al., 2007) have been found to play a major role in the disorder.

The role of loss-of-function mutations is somewhat disputed. While this concept was initially believed to be the underlying cause of familial ALS (Deng et al., 1993; Rosen et al., 1993), mounting evidence against this notion has been derived from animal models (Ratovitski et al., 1999; Boillee et al., 2006). In addition, various human SOD1 variants that retain the full dismutase enzyme activity have been described in familial ALS (Borchelt et al., 1994; Hayward et al., 2002). Therapeutic trials of antisense oligonucleotides in murine models carrying human SOD1 variants support this notion (McCampbell et al., 2018).

SOD1 loss-of-function is a new disease with recessive mode of inheritance and heterozygous carriers not affected by the disorder. Although the eldest investigated heterozygous carriers (35 and 44 years, respectively) could still develop symptoms of familial ALS, there have to be several older heterozygous carriers in earlier generations of the family without any symptoms of familial ALS (Fig. 1) (Bali et al., 2017), showing that a reduction in enzyme activity—as in heterozygous carriers—does not lead to familial ALS whereas complete loss-of-function leads to a different disease.

The disorder observed in our patient differs significantly from motor neuron disease. While it shares some similarities, such as signs of muscular denervation observed indirectly in electromyography and muscle ultrasound, the vast majority of symptoms have not been described in ALS before. The age of onset in early childhood is atypical, as is the phenotype indicative of first motor neuron affection. The hyperekplexia-like presentation is a hallmark of the syndrome, which might facilitate diagnosis in the future.

The phenotype bears distinct similarities to the presentation of total murine Sod1 knockouts (Matzuk et al., 1998; Elchuri et al., 2005; Wang et al., 2012; Sakellariou et al., 2018) in which similar features such as tremor, reduced muscle mass, and motor axonopathy (Shefner et al., 1999) were noted.

The total absence of any measurable SOD1 activity suggests a major role of oxidative stress or dysregulation in the pathogenesis of the disease. This is in part supported by the finding of hypomanganesaemia in the index patient, as manganese is a known antioxidant, either as part of SOD2 or independently (Coassin et al., 1992; Aguirre and Culotta, 2012).

Given the severe and debilitating nature of the disorder, the need for therapeutic options is evident. Based on SOD1’s role in metabolism of reactive oxygen species, antioxidant therapy, e.g. by external supplementation, represents an intriguing approach. Indeed, the antioxidant N-acetylcystein was shown to have positive effects on anaemia and autoantibody generation in Sod1^−/− mice (Iuchi et al., 2007).

In addition to antioxidant compounds, SOD mimetics represent another promising therapeutic approach. These agents have been suggested as treatment options in the context of Parkinson’s disease (Filograna et al., 2016) and radiation damage (Anderson et al., 2018), among others. In a phase 1b/2a trial in patients undergoing radiation therapy for oral carcinoma, the mimetic GC4419 showed acceptable safety (Anderson et al., 2018), making it a candidate for application in SOD deficiency in the future.

The findings presented in this study not only shed light on the pathomechanism of SOD1-related ALS but are also of high relevance for the intensely investigated therapeutic strategy of SOD1 silencing. Previous studies in various ALS disease models have established the general efficacy of silencing using antisense oligonucleotides, shRNA, miRNA, and other compounds (for a review see van Zundert and Brown, 2017). Due to the lack of ALS in SOD1 knockout mice, safety of silencing of both wild-type and mutant SOD1 is assumed (van Zundert and Brown, 2017).

Given the severe neurological manifestation of SOD1 deficiency in our patient, the possibility of adverse effects following SOD1 silencing must be considered. From a pathomechanistic point of view, silencing of the principal cytoplasmic superoxide dismutase might result in increased oxidative damage as well as altered redox signalling.

Findings from animal as well as early clinical studies did not indicate any adverse events attributable to SOD1 deficiency (Miller et al., 2013; Thomsen et al., 2014; Stoica et al., 2016; Borel et al., 2018). This can in part be explained by the incomplete representation of human (patho)physiology by animal models. Furthermore, currently published findings in humans made use of low doses of silencing agents not resulting in complete suppression of SOD1 activity (Miller et al., 2013). Given the absence of an observable neurological phenotype in heterozygous carriers of the variant described in our study, these results do not exclude adverse events by SOD1 silencing.

In conclusion, our data characterize the effects of a truncating mutation of SOD1, leading to total absence of SOD1 activity in the affected patient. The resulting phenotype is severe with tetraspasticity mainly affecting the upper limbs and hyperreflexia reflecting an affection of the first motor neuron. Antioxidant supplementation may represent a therapeutic approach, although further research is needed to characterize the effects of the variant on a deeper level. The results of this study call for a cautious approach to SOD1 silencing as a therapeutic concept for ALS.

Supplementary Material

Acknowledgements

Glossary

References