Frequency of C9orf72 and SOD1 mutations in 302 sporadic ALS patients from three German ALS centers

Abstract

Background: ALS patients with a negative family history (sporadic ALS, SALS) represent more than 90% of all ALS cases. In light of the gene-specific therapies that are currently in development for ALS, knowledge about the genetic landscape of SALS in Germany is urgently needed. Objectives: We aimed to determine the frequency of C9orf72 hexanucleotide repeat expansion (HRE) and SOD1 mutations among patients in Germany with a diagnosis of sporadic or idiopathic ALS. Methods: We genotyped SALS patients from three German ALS centers. Sanger sequencing, fragment length analysis, and repeat-primed PCR technologies were used to detect mutations in SOD1 and C9orf72 HRE. Pathological C9orf72 HRE results were confirmed in an independent laboratory. Results: In 302 patients with SALS, 27 (8.9%) patients with a C9orf72 HRE mutation were detected. Moreover, we identified two patients with a pathogenic SOD1 mutation, one patient with a heterozygous p.D91A mutation in SOD1, and three additional patients with rare SOD1 variants not predicted to change the amino acid sequence. Conclusions: According to our data, the proportion of SALS patients with SOD1 mutations is in the expected range, whereas that with C9orf72 HRE is higher, suggesting a reduced penetrance. A considerable number of SALS patients can be amenable to gene-specific therapies.

Keywords:

• Amyotrophic lateral sclerosis
• neurodegeneration
• sporadic ALS
• motor neuron disease
• neurogenetics

Introduction

About 5% of German patients with the motor neuron disease amyotrophic lateral sclerosis (ALS) have a positive family history of ALS or the genetically related frontotemporal dementia (FTD) (familial ALS, FALS) (1). More than 90% do not report a familial incidence and have historically been denoted sporadic ALS (SALS).

Currently, there are about 30 genes known to be involved in ALS causation. The most frequently mutated ALS disease gene in patients of European descent is C9orf72. A heterozygous large expansion of a hexanucleotide repeat sequence GGGGCC in the intron 1A of is pleiotropic and has been linked to ALS, FTD, and ALS/FTD (2,3). Up to approximately 30 repeats have been observed in healthy individuals whereas C9orf72 HRE-mediated ALS patients typically have between 800 and 2000 repeats when assayed using Southern blot using DNA purified from peripheral blood leukocytes. Furthermore, there may be extensive mosaicism across different tissues, with a greater expansion in neural tissues (4). A pathogenic C9orf72 HRE can be found in about 35% of FALS cases in the European population, whereas it is rare in Asians, only making up to 2.3% (5). On the other hand, C9orf72 HRE is estimated to be found in about 5% of European and 0.3% of Asian SALS patients according to previous reports (5,6). In this context, the discovery of a higher percentage of C9orf72 HRE in the Mongolian population was surprising (7).

The second most commonly mutated ALS gene in Europe is superoxide dismutase 1 (SOD1). Mutations resulting in a change of the coding sequence have been identified in 14.8% of FALS and 1.2% of SALS patients in the European population (5). Although the exact mechanism of how SOD1 mutations cause ALS is unclear, a toxic gain of function involving the formation of SOD1 prion species was proposed (8,9).

There has been a major increase in trials of therapeutic approaches for ALS in recent years (10). In particular, gene-specific therapies (antisense oligonucleotides; ASOs) for Mendelian cases of ALS have had a rapid development, specifically for SOD1- and C9orf72-linked ALS. ASOs are short, single-stranded deoxyribonucleotides hybridizing to a complementary target mRNA. This binding results in the degradation of the RNA or prevents its translation, leading to reduced protein expression (11). The ASO tofersen has been under study to treat ALS patients harboring SOD1 mutations. In phase I/II studies, the highest dose of tofersen led to reduced SOD1 and neurofilament levels in cerebrospinal fluid (12). Similarly, several clinical studies using ASOs to target different ALS genes are underway, with the most advanced being ASOs against the HRE in C9orf72. Thus, considering the gene-specific therapies, we here set out to determine the epidemiology of mutations in these two most frequently mutated ALS genes among German SALS patients. A robust genetic data collection already exists with regard to the mutation frequencies in FALS patients in Europe and Germany (5,13). However, knowledge about the genetic landscape of SALS in Germany is largely lacking. In this study, we investigated the frequency of mutations of SOD1 and C9orf72 in 302 SALS patients from three specialized ALS outpatient clinics in Germany.

Materials and methods

Study cohort

Patients were recruited in three German ALS research centers in the cities of Essen (n = 241; Alfried-Krupp-Krankenhaus Essen) and Mannheim (n = 44; University Clinic Mannheim, Division for Neurodegeneration, and n = 17; Diakonissen Hospital Mannheim, Department of Neurology) in outpatient and inpatient clinic settings from July 2020 to October 2022. The study was a scientific investigation and patient identity was anonymized. The study was approved by the relevant local ethics committees and performed in accordance with the Declaration of Helsinki (WMA, 1964). Written informed consent was obtained from all participants prior to enrollment. The examination and diagnosis were performed by neurologists according to the revised El Escorial criteria (14). Diagnostic management was in accordance with the EFNS guidelines (15). We defined ALS as familial when at least two individuals within a 3rd-degree relationship were affected with ALS or FTD. When the medical records were not available, we relied on patients’ or other family members’ reports of symptoms for the diagnosis of FALS. Thus, FALS patients were not included in this study. DNA was extracted from EDTA-venous blood by standard methods.

C9orf72 analysis

Combined fragment length analysis (FLA) and repeat-primed PCR (RP-PCR) were performed to genotype for a hexanucleotide repeat expansion in intron 1A of C9orf72. The primers for FLA were described earlier (Forward primer: FAM-CAAGGAGGGAAACAACCGCAGCC and reverse primer: GCAGGCACCGCAACCGCAG) (2). For RP-PCR, the primers were forward primer: FAM-AGTCGCTAGAGGCGAAAGC, reverse primer TACGCATCCCAGTTTGAGACGGGGGCCGGGGCCGGGGCCGGGG, and anchor primer: TACGCATCCCAGTTTGAGACG (3).

Concentrations of the ingredients as well as protocols for FLA and RP-PCR were followed as performed by laboratory A in the international C9orf72 HRE validation study of Akimoto et al. (16). Briefly, the FLA reaction was set to a final volume of 15 µL using OneTaq Hot Start 2X Master Mix with GC buffer (New England Biolabs), 0.4 µM final concentration of each primer, and 35 ng final DNA amount. A touchdown PCR program was used for amplification (65 °C–55 °C). For RP-PCR, the following final amounts were used: 1X FastStart PCR master mix, 7% DMSO, 1 M betaine, 0.18 mM 7-deaza-2-dGTP, 0.89 mM MgCl₂, and 100 ng of DNA in a total volume of 14 µL. The final concentrations of forward and anchor primers were 1.4 µM and for the reverse primer 0.7 µM. After the amplification reaction, the samples were mixed with GeneScan 600 LIZ dye size standard (Thermo Fisher Scientific, Damstadt, Germany) and denaturated at 95 °C for 5 min before they were processed on the genetic analyzer. Data were analyzed using the Peak Scanner software v2.0. If a hexanucleotide repeat expansion was detected, the genotype was confirmed by analysis in an independent laboratory (Department of Clinical Sciences, Neurosciences, Umeå University, Umeå Sweden).

SOD1 sequencing

To screen for SOD1 mutations, we used M13-tailed oligonucleotides flanking the five exons and at least 80 base pairs from the exon-intron junctions. The primer sequences are listed in the supplementary data. The PCR products were enzymatically cleaned up using ExoSAP according to the manufacturer’s protocol. The BigDye v1.1 Cycle Sequencing Kit (Thermo Fisher Scientific) was used for the sequencing reaction. Sanger Sequencing was performed on SeqStudio™ Genetic Analyzer System with SmartStart (Applied Biosystems). Sequences were analyzed using CLC Main Workbench 7 (Qiagen). Variant nomenclature is according to the SOD1 transcript NM_000454.5.

Results

One hundred seventy-nine male and 123 female patients with sporadic ALS (SALS) (male:female ratio = 1.5) were recruited from the ALS centers in Essen (Alfried-Krupp-Hospital; PI T.G.) or Mannheim (Diakonissen-Hospital and University Clinic Mannheim; PIs J.W. and J.H.W., respectively). The mean age at onset of the disease was 60.4 ± 12.4 years, and the disease duration (data available from 51 deceased patients) was on average 29.1 ± 40.2 months and the median was 20 months. Screening of the 302 SALS patients for a C9orf72 HRE or a non-synonymous SOD1 mutation revealed likely causative genetic variants in a total of 30 patients (Table 1). Twenty-seven patients were heterozygous for C9orf72 HRE with full concordance between the two involved laboratories. Patients with C9orf72 HRE had a relatively faster disease progression, more frequent bulbar onset of paresis, and FTD co-morbidity (Table 1).

Table 1 Overview of the patient information.

	Sex	Age at onset (y)	Disease duration (mo)^a	Site of onset	FTD co-morbidity	Bulbar symptoms	UMN/LMN predominancy
C9orf72 HRE+	17 m: 10 f	59.8 ± 7.8	19 ± 8.1	11 spinal, 9 bulbar, 1 FTD, 1 hypoventilation	2 yes, 20 no	9 no, 6 dysarthria, 2 dysarthria and dysphagia	11 LMN, 3 UMN, 2 both
SOD1 mutation+	2 m: 1 f	35 ± 2.8	–	2 spinal, 1 bulbar	1 no	–	1 LMN
No SOD1/C9orf 72 mutation	160 m: 112 f	60.6 ± 12.5	30.9 ± 43.5	159 spinal, 62 bulbar, 6 hypoventilation, 2 FTD	5 yes, 218 no	11 yes, 134 no, 32 dysarthria, 6 dysphagia, 41 dysarthria and dysphagia	147 LMN, 56 UMN, 20 both
Total^b	179 m: 123 f	60.4 ± 12.4	29.1 ± 40.2	170 spinal, 72 bulbar, 3 FTD, 7 hypoventilation	7 yes, 239 no	11 yes, 143 no, 38 dysarthria, 6 dysphagia, 43 dysarthria and dysphagia	159 LMN, 59 UMN, 22 both

HRE: hexanucleotide repeat expansion; y: year; mo: months; m: male; f: female.

^aBased on 51 deceased patients (8 with C9orf72 HRE and 43 without SOD1 or C9orf72 mutation). ^bBased on the patient clinical data when available.

Three patients carried a previously described pathogenic missense mutation in SOD1 (Tables 1 and 2). We identified one female patient with a heterozygous and a male patient with a homozygous c.272A > C/p.D91A variant, which is the most common SOD1 mutation globally (17). Although SOD1 mutations are usually autosomal dominantly inherited, p.D91A is peculiar in that it is also inherited in the autosomal recessive form to cause ALS (18). In accordance with the previous literature about homozygous p.D91A mutations, it was associated with an exceptionally slow disease duration of more than 207 months in the patient described here (Table 2). However, distinct from other patients with a homozygous p.D91A mutation (19), the patient exhibited lower motor neuron predomination and bulbar symptoms.

Table 2 Patients with SOD1 mutations identified in this study.

Variant	Sex	Age at onset (y)	Disease duration (mo)^a	Site of onset	FTD co-morbidity	Bulbar symptoms	UMN/LMN predominancy
c.272A > C, p.D91A (hom)	m	53	207	spinal	No	Yes	LMN
c.272A > C, p.D91A (het)	f	33	57	spinal	No	Yes	LMN
c.434T > C, p.L145S (het)	m	37	79	spinal	No	No	LMN

y: year; mo: months; m: male; f: female.

^aAll three patients are still alive at the time of writing.

The penetrance of heterozygous p.D91A mutations in SOD1 can be population-dependent and patients exhibit variable clinical manifestations. The patient found to carry the heterozygous p.D91A mutation had an overall milder ALS with a spinal site of onset, and weakness in the right arm and shoulder, which subsequently spread to the left side (Table 2).

A third patient carried a c.434T > C/p.L145S variant in SOD1, which turned out to be absent from the gnomAD reference database but was reported in earlier ALS studies in various ethnic backgrounds including individuals of African origin (20,21). This male patient presented with the first signs of ALS with spinal onset early at the age of 37 years and was still alive after 79 months.

Of note, we identified three very rare additional variants in the SOD1 predicted not to alter the amino acid sequence, specifically two variants preceding the first exon (c.-55C > T, c.-8A > C) and a synonymous variant c.253T > C/p.Leu85Leu. The pathogenicity of these variants is unknown at the moment, thus they were not included in the percentage analysis. The variants c.-55C > T and c.-8A > C were in a heterozygous state. Three and two alleles in the heterozygous state and none in the homozygous state are present out of 31,388 and 250,362 alleles, respectively, in the gnomAD database. The variant c.253T > C/p.Leu85Leu was present in a homozygous state in a patient in this study. It was found only once in a heterozygous and not in a homozygous state out of 251,428 alleles in gnomAD.

Discussion

We performed a multicenter study to measure the frequency of SOD1 and C9orf72 mutations in our cohort of 302 sporadic ALS patients from three German ALS clinics in two large cities. Despite the negative family history that defines sporadic ALS, we discovered that almost 10% of SALS patients carry a mutation in C9orf72 or SOD1.

A genetic inventory of 301 German familial ALS patients has been taken before, reporting 25% and 11% of FALS cases to have a C9orf72 or SOD1 mutation, respectively (13). Corresponding numbers have not been determined yet for a larger cohort of sporadic ALS patients in Germany. However, this would be relevant for the general understanding of the genetic architecture of ALS in Germany and Europe. Moreover, data describing the proportion of SALS patients with a monogenic cause of the disease is important for planning of upcoming gene-specific therapy studies and healthcare structures providing such therapies (e.g. antisense-oligonucleotides against SOD1 (tofersen) or C9orf72).

That 8.9% of patients have C9orf72 HRE is surprisingly high when compared to the 5% of European SALS patients detected by Zou et al. (5), while three out of 302 patients with a SOD1 variant was close to the expected range of an estimated 1.2–1.5% global frequency of SALS (22). The comparably high number of C9orf72 mutant patients in SALS cohorts can at least partially be attributed to the incomplete penetrance of C9orf72 that is age-dependent (23). Other relevant factors are most likely partially missing family history, missed diagnosis, pleiotropy, or demise of family members due to other reasons before the onset of ALS. Moreover, C9orf72 HRE can manifest as both ALS, frontotemporal dementia or movement disorder, thereby obscuring the familial aspect. In this study, we excluded patients who had a history of neurodegenerative disorders in the family such as dementia. It is also important to emphasize that both SOD1 mutations that we detected in our cohort belong to the rare instances of SOD1 variants that can cause also recessively inherited ALS, and may therefore be enriched in ALS cases categorized as sporadic.

We took advantage of targeted gene sequencing that allows efficient in-house screening (Division for Neurodegeneration, Mannheim) for the two clinically most relevant ALS disease genes in our clinical routine setting. To identify any additional variants in other ALS-linked genes, broader analysis such as whole genome/exome sequencing or panel testing would be required.

The clinical characteristics of C9orf72 HRE mutant patients were in agreement with the previously reported faster disease course and the more frequent bulbar onset of paresis. The phenotype of the patients with SOD1 p.D91A mutation reported here are in line with the previously described slow disease progression for homozygous SOD1 mutation, while heterozygous p.D91A mutations are usually associated with a faster disease course and a much shorter survival time than observed in patients with the homozygous p.D91A mutation. The discordance between p.D91A SOD1 gene dosage and phenotype and survival time is enigmatic but has been observed in other European populations. Speculatively, since the concentration of mutant SOD1 monomers is a critical factor in the formation of SOD1 prion strains, different SOD1 prion strains may be involved in p.D91A SOD1-mediated ALS. Alternatively, since the pathogenicity of heterozygous p.D91A is under debate (24–26), the mutation reported here may end up being non-pathological in the future.

Also the third SOD1 mutation c.434T > C/p.L145S has previously been reported in a single homozygous case, causing a severe disease course with an onset already in the second decade and fast progression (27), which suggests a gene dosage-dependency of the phenotypic manifestation. In heterozygous form, this mutation gives rise to Mendelian dominant ALS with the lower-limb onset of paresis, mid-early disease onset, and a very slow disease progression often over decades (28,29).

We are currently unable to judge the pathogenicity of the three observed variants not predicted to alter the SOD1 amino acid sequence. Consequently, they were not taken into account for calculating the percentage of SOD1 mutant patients. It is nevertheless remarkable that they are very rare in reference databases. Moreover, two of the variants are located 8 and 55 nucleotides upstream of the first exon, thus opening the possibility that they may influence the expression of SOD1. Moreover, the exonic but synonymous variant c.253T > C/p.L85L is found only once in a heterozygous state out of 251,428 alleles in gnomAD. While also for this variant the pathogenicity remains eventually unclear, it is surprising to note that we found it in a homozygous state in the patient reported here, while consanguinity was denied. The altered nucleotide is located 15 nucleotides downstream of the splice acceptor site and may influence splicing. Moreover, a pathogenic missense mutation affecting the same codon (p.Leu85Phe) was reported by others in both FALS and SALS patients (30).

The fact that all three ALS centers contributing to this study are located in the South or West of Germany is a limiting aspect for the interpretation of the data. Thus, although all three centers frequently provide second opinions also for patients from other regions of Germany, further studies and more expanded gene testing are needed for a complete genetic characterization of the German SALS population.

In conclusion, we found a substantially higher than expected proportion of sporadic ALS patients with an HRE in C9orf72 in Germany, which could be a result of incomplete penetrance of this mutation. Taken together, almost 10% of the SALS patients carried a mutation in either C9orf72 or SOD1, while the relevance of the non-coding variants remains to be shown. Yet, with genetic screening for SALS patients, a considerable proportion of all ALS patients will turn out to be amenable to gene-specific therapies.

Acknowledgements

We are indebted to the patients and their families for their participation in this project. We are thankful to Helena Alstermark for her technical assistance and our study nurse Antje Knehr. We thank Dr. Rosanna Parlato for the discussions and critical reading of the manuscript.

Declaration of interest

There are no relevant financial or non-financial competing interests to report.

Additional information

Funding

RY is supported German Research Foundation (Deutsche Forschungsgemeinschaft, DFG) Walter Benjamin Programme (YI 209/1-1, AOBJ 680080).