This study was carried out following the protocol approved by the Hospital Universitari Vall d’Hebron Institutional Review Board with written informed consent from all subjects in accordance with the Declaration of Helsinki.

Our pedigree originated in north-western Spain. The simplified pedigree of the family is shown in Figure 1. We examined two ALS patients and two healthy individuals in the pedigree after obtaining informed consent. There were no skipped generations. The affected individuals were clinically characterized according to gender, age at onset, initial topography, signs of dementia, and survival time (Supplementary Table S1).

The proband, a 53-year-old male (II:1), noticed weakness in his left leg beginning 7 months previously, which he attributed to a traumatism he had experienced when working in construction. An initial electromyographic study showed denervation in all four extremities. Transcranial magnetic stimulation and brain and spinal MRI were normal. Cerebrospinal fluid, hemogram, biochemical screening, and serological tests for neurotropic infectious diseases all presented normal values, except for creatine kinase, which was 735 U/l (normal values < 135). The patient subsequently experienced pain, with frequent falls in the months after his first examination, when muscle strength in his upper limbs and right lower limb was completely normal. His Amyotrophic Lateral Sclerosis Functional Rating (ALSFRS-R) score was 44. In the next visit, 3 months later, we observed weakness and amyotrophy in the left upper limb, particularly in the proximal muscles. We observed generalized fasciculations, which were more marked in the muscles on the left side of the body, where it was difficult to elicit deep tendon reflexes. Eight months after the initial visit, the patient was admitted to the emergency ward complaining of dyspnea and orthopnea, and non-invasive mechanical ventilation was fitted. The patient was offered a gastrostomy, but declined. The patient died 19 months after symptom onset.

Two years later, the proband’s 47-year-old brother (II:5) came to our clinic due to presenting clumsiness and difficulties climbing stairs, running and walking. He had experienced recurring episodes of cramp in both calves in the 2 months before his visit. We observed weakness and amyotrophy in the proximal muscles of the right lower limb and a minimal loss of muscle mass in the first interosseous of the right hand. Deep tendon reflexes were normal, except for the patellar and Achilles reflexes. Fasciculation was observed in the musculature of the four extremities. His ALSFRS-R rating at this point was 39. Like his brother, neurophysiological studies showed signs of denervation in three regions and an absence of upper motor neuron signs. The patient declined any follow-up, life support, ventilation or nutritional measures. He died 17 months after clinical onset.

In their family, their two sisters requested genetic screening for familial forms of ALS, as they remembered that their father had died at 57 years of age due to respiratory insufficiency after 2 years of weakness in the upper limbs. To the best of our knowledge, this pedigree contains no other cases of neurodegenerative disorders, including frontotemporal dementia and Parkinson’s disease.

The clinical phenotype in the affected members in our family—a mean age of onset of 51.7 years, spinal onset, predominant lower motor neuron signs, and a survival time of 20 months—is consistent with the phenotype in the American, Italian, and Swedish families. No cognitive impairment in either brother was detected in the neuropsychological test (Supplementary Table S1).

Mutation analysis of the SOD1 gene by direct PCR sequencing revealed a C-to-T transition at nucleotide position 14 (c.14C > T, Supplementary Figure S1) leading to a p.A5V (rs121912442, A4V in the old nomenclature) sequence change at protein level in the two affected ALS patients: the proband (individual II:1) and his brother (individual II:5) (Table 1). This exon 1 mutation was not found in the two healthy members of the family (individuals II:3 and II:4) who requested details of their genetic situation (Supplementary Figure S1).

No mutations were detected in the FUS, TARDBP, and PFN1 genes, and no C9orf72 expansions were identified in the two ALS patients.

We analyzed the 15 genetic variants spanning the human chromosome 21 from the SOD1 gene to the SCAF4 gene (Supplementary Table S2). The genetic markers around the SOD1 gene were genotyped to infer the SOD1 haplotypes of the ALS patients and the family members. Based on the genotypes of the four siblings, it was possible to identify four different inferred haplotypes present in the family and propose the haplotypes of the affected father and the mother (Figure 1).

Fourteen of the 16 SNPs analyzed in the SOD1 gene were detected in subjects of the IBS population from the 1K Genomes Project. Eight phased haplotypes of the SOD1 region were inferred in the IBS population with different frequencies (Figure 2). The haplotypes were named (1 to 8) according to their frequency in the population, ranging from higher to lower frequency (Figure 2). The mutation p.A5V-SOD1 detected in the ALS patients is only compatible with haplotype 1 (haplo1V), which is the most common in the IBS population (haplo1A) (Figure 2 and Table 1). According to the nomenclature proposed, the family has the IBS haplotypes 1A, 1V, 4, and 1A* (Figure 1).

The linkage disequilibrium of these 14 SNPs was calculated for the IBS population (Supplementary Figure S2). A conserved haplotype block, in high linkage disequilibrium, was detected between the start of intron 1 and the end of SOD1, and including the SCAF4 gene. The SOD1 exon 1, including the c.14C > T (p.A5V) mutation, is located outside this conserved haplotype in the IBS population (Supplementary Figure S2). Using data for all the populations deposited in the 1K Genomes Project, the analysis of linkage disequilibrium showed the same results for the conservation of this SOD1 haplotype block (data not shown).

The founder p.A5V-SOD1 haplotype in our patients is the same as the one found in the Swedish population, and differs from the founder haplotype observed in patients from North America (Table 1).

We determined all the inferred SOD1 haplotypes for these 14 SNPs in the IBS, European, Asian, African, and Mixed American populations, using all the available subjects from the 1K Genomes Project as a control group (Figure 3). A total of 21 phased inferred haplotypes were detected in all the populations. The European haplotype (p.A5V-EUR) was the most common in all the populations. The American haplotype (p.A5V-USA) was the second most common worldwide. However, both haplotypes appear with similar frequencies in the Asian population (0.386 for p.A5V-EUR, 0.385 for p.A5V-USA) (Figure 3).

Over 186 SOD1 mutations have been reported in amyotrophic lateral sclerosis (ALSoD). Previous epidemiological studies suggest that p.D91V is the most common worldwide, followed by p.I113T and p.A5V. The latter is the most commonly reported in the United States, accounting for 41% of the mutations identified in that population. It is one of the most aggressive mutations, as the mean survival time is under 2 years. The mutation has rarely been described among European populations. This is the first time that the p.A5V-SOD1 has been reported in the Spanish population.

With the rare exception of a 73-year-old male presenting facial diplegia and unilateral vocal cord paralysis as the initial symptoms of ALS (Salameh et al., 2009), the phenotype associated with p.A5V in the American population is characterized by rapid progression and short survival. Patients present predominantly lower motor neuron signs, with limb weakness in 51% of patients, and bulbar symptoms in less than 11% of cases (Cudkowicz et al., 1997; Juneja et al., 1997; Andersen, 2003b; Andersen, 2006; Broom et al., 2008; Saeed et al., 2009). In a recent review of 57 p.A5V-SOD1 cases, the mean survival time was 1.4 years, with clinical onset at a mean age of 50.0 years, and a male-female ratio of 1.3:1. Other characteristics associated with this SOD1 mutation include fast progression, predominantly spinal onset, high penetrance, involvement of primarily lower motor neuron signs, and an absence of dementia (Bali et al., 2017).

However, the cases reported in European families differ slightly from those in America. For example, Andersen reported the clinical characteristics of a Swedish family (Andersen et al., 1997; Andersen, 2007). One case had leg onset, another had hand onset, another had shoulder onset, and a further case had bulbar onset. These individuals were aged between 56 and 68 years old, and survival ranged from 8 months to 2 years. In the members of the two Italian families reported in 2001, information was available for two probands. One had shoulder onset at 55 years old, and died 8 months later. The other proband presented hand onset at 57 years old. No information about this individual’s survival time is available, although the course was rapidly progressive (Gellera et al., 2001). Our family’s phenotype—a mean age of onset of 51.7 years, spinal onset, predominant lower motor neuron signs, no cognitive alteration, and a survival time of 20 months—is consistent with the phenotype in these European families. Interestingly, none of the women inherited the mutation. Whether this was due to chance or a protective factor is unknown.

The reasons why p.A5V-SOD1 predicts fast progression and short survival are unknown. In vitro studies have pointed to this mutation leading to a decline in dimerization capacity, a loss of metalation (a 30-fold decrease in zinc-binding affinity), and aberrant oligomerization, leading to misfolding and aggregation in the form of insoluble toxic inclusions within motor system cells, a key pathological hallmark of ALS. There is some variability in the propensity of SOD1 mutants to aggregate, which could be related to the duration of the disease. High-aggregation propensities have been described for p.A5V, which could be the factor responsible for the disease’s short duration (Prudencio et al., 2009; Zhao et al., 2014; Farrawell et al., 2018; Maurel et al., 2018; Srinivasan and Rajasekaran, 2018).

Another alternative and complementary hypothesis to explain the aggressive nature of p.A5V is its position in the tertiary structure of the protein. Interestingly, of the 30 mutations reported in exon 1, most of those in residue A5 (A4 in the previous nomenclature) predispose to a rapid progression and survival of under 24 months (Syriani et al., 2009). According to the tertiary structure of SOD1, residue A5 is located in the first β-strand that forms part of the dimer interface. A5 mutations affect SOD1 dimerization and/or destabilize the Greek key β-barrel because A5 packs into the SOD1 monomer’s hydrophobic core (Getzoff et al., 1990; Deng et al., 1993; Cardoso et al., 2002; Galaleldeen et al., 2009; Schmidlin et al., 2009; Schmidlin et al., 2013; Kumar et al., 2018).

Two earlier studies investigated a possible founder effect in the cases of North American and European origin. Following these haplotype studies, there is evidence to suggest at least two different origins: one European (probably a Scandinavian founder effect) and one Amerindian (Native Americans). The latter presents in 82% of North American families with the p.A5V-SOD1 mutation. In European populations, p.A5V-SOD1 has been reported in Scandinavia (three cases) and Italy (six cases). An isolated case has recently been reported in China, which shares most of the European haplotype. These findings could be interpreted in terms of the existence of various founders for p.A5V patients worldwide, rather than one founder, as previously assumed (Broom et al., 2008; Saeed et al., 2009). Some authors suggest that the European founder effect is older than the Amerindian, hypothesizing that the American families were in fact descendants of the European founder. Other authors argue that the European origin is different from the American one (Rosen, 2004; Armon, 2009).

After analyzing the SOD1 haplotypes in different populations (European, Asian, African, and all the populations analyzed in the 1K Genomes Project), we found that the SOD1 haplotype in which the p.A5V mutation is located in European fALS patients (p.A5V-EU) is the most common in the population worldwide, and the haplotype for American p.A5V-SOD1 patients (p.A5V-USA) is the second most common. These variants appear to be linked to the most common haplotypes, and not linked to rare haplotypes. The haplotype p.A5V-USA is the most common (0.385) in the East Asian population and the p.A5V-EU haplotype is the second most common (0.306), with little difference between their allelic frequencies. However, in the European population, the haplotype p.A5V-EU is the most common haplotype (0.298), far below the frequency of the haplotype p.A5V-USA (0.062). The haplotypes p.A5V-EU (0.037) and p.A5V-USA (0.045) are rare in the African populations (as shown in Figure 3).

Based on these data, we propose two possible hypotheses for the origin of this mutation in two different haplotypes: