The genetic causes of FALS are partly known. In this study, we found that 22 out of 393 FALS index cases without FTD or PD, 2 out of 76 patients with FALS/FTD, and 1 out of 7 FALS/PD/DE index cases had
FUS gene mutations. The
FUS gene mutation frequency of non-
SOD1 and non-
TARDBP cases of FALS was 5.6%, and of all cases of FALS was about 4.79%. It was reported that 4.5% of 198 index cases without
SOD1, VAPB, ANG, DYNACTIN, CHMP2B, or
TARDBP mutations
22 and 5.3% of 94 cases of FALS without
SOD1,
TARDBP, and
ANG mutations had
FUS mutations.
20 Both reports estimated the
FUS mutational frequency of the general FALS population to be ~4%. FUS mutation frequency was found to be 5.8% in 293 FALS, but it was unclear which gene mutations were excluded, and 209 cases of the 293 cases of FALS were screened only for exon 15.
21 FUS and
TARDBP are both DNA/RNA binding proteins.
33 The prevalence of mutations of
TARDBP in ALS varied from 0.65% to 4.85% in studies with 80 to 154 patients with FALS examined and 0% to 5% in studies with 86 to 541 patients with SALS screened
4–14; the variation might be related to clinical heterogeneity of patients. To date, 21 mutations in the
TARDBP genes were found in 643 cases of FALS with a mutational frequency of 3.27%.
FUS mutations are more frequent than
TARDBP mutations, and appear to be the second most frequent cause of disease after
SOD1 mutations in FALS. For SALS, we found no
FUS gene mutations in 41 cases; however, the sample size was very small. One report found no
FUS mutation in 293 patients with SALS,
21 but another found 3 out of 405 SALS cases had
FUS gene mutation.
18We found that 1 FALS/PD/DE index case out of 7 and 2 ALS/FTD index cases out of 76 had
FUS gene mutations. The R521C mutation in the index case of F8828 with FALS/PD/DE was the most frequent mutation documented in FALS.
21,22 The G206S mutation in F7543 with ALS/FTD has not been reported. G174-G175 mutation in exon 5 was first reported in a FALS pedigree (F213) with a screening of 176 controls.
21 We found it in 1 ALS/FTD and 2 FALS index cases, and it was not present in the sequencing of 726 controls. No patients with FALS with cognitive deficiency were reported in the first 2
FUS mutation reports,
21,22 but
FUS immunoreactive neuronal inclusions have been reported in FTD.
34 However, a G156E mutation was found in an Italian FALS index case who developed FTD in his fourth decade.
20 The proband with S96del had mental retardation, the proband with G174-G175 del in F7390 had the diagnosis of schizophrenia, and the proband with R524S had cerebellar atrophy on MRI, but no additional clinical records were available for further characterization. All the affected individuals in F8726 with G497AfsX527 mutation had learning disabilities; 1 died at age 14; the 2 adults were drug users. It was suggested that
FUS mutations might result in cognitive dysfunction, which merits further study.
The deletion of 4 glycine residues (G223–G226del) in exon 6 occurred in a 10-glycine stretch from amino acid residue 222 to 231. Six out of the 10 glycine residues were coded by 6 GGC repeats. A loss of 3 glycine residues was found in a SALS case and in 1 out of 190 controls. A glycine insertion in the same region was found in a patient with schizophrenia and autism.
18 The 4-glycine residue deletion in F1090 segregated with the patients and was not present in 700 controls (appendix e-1). In this study, the controls were mainly European Americans, and we had no autopsy tissue for pathologic evaluation; the glycine stretch changes could be polymorphism, and pathogenicity of some of the novel mutations awaits further verification.
Identification of 4 novel frameshift mutations and 1 nonsense mutation in exon 14 suggests that the c-terminal amino acid residues may be of key biologic relevance if the pathogenesis of
FUS mutation was due to loss of function. It may also imply that a shorter N-terminal peptide of 494 amino acids is sufficient to cause neuronal toxicity via a gain of function mechanism (figure e-1). Such a phenomenon was previously observed in truncated
SOD1-mediated ALS in humans and transgenic mice.
35 A nonsense change at Y374X in the
TARDBP gene resulting in a truncated C-terminal has also been reported in SALS.
5 The nonsense R495X and frameshift mutations in
FUS may suggest that haploinsufficiency rather than a gain of function of
FUS could cause ALS. Truncated transcript and protein would be degraded or functionless, nevertheless,
FUS knockout mice had perinatal mortality, sterility, and radiation sensitivity, but had no obvious neurologic manifestation.
36,37The distribution of FUS mutations reported to date (figure e-1) may delineate 2 major mutation clusters in the FUS gene: 1 in exon 4 to 6 and the other in exon 14 to 15. The apparent grouping of mutations in these 2 clusters may imply the functional importance of these regions in triggering motor neuron degeneration. It may also indicate regions of interest for genetic screening of FUS mutations in patients with ALS.
In previous studies,
FUS mutations were found in ALS cases of European American and Cape Verdean Island origins.
21,22 We identified
FUS mutations in additional ethnic groups, including European American, African American, Asian (Chinese, Korean, and Cambodian), and Latino. Most
FUS mutations identified in European Americans were also identified in patients from other ethnicity except G206S, which was found in a family of South Korean origin. Although this study is not a population study, it suggests that
FUS mutations may be a globally distributed genetic cause of FALS in patients of different genetic backgrounds.
When comparing patients with ALS with
SOD1 and
TARDBP mutations, patients with
FUS mutations had earlier symptom onset, a higher rate of bulbar onset, and a shorter duration of symptoms in general. No report has compared the phenotype of
FUS, SOD1, or
TARDBP mutations, but the average age at symptom onset and duration of symptoms of patients with
FUS mutations from previous studies were close to those of this study.
21,22 The reported average age at symptom onset of patients with TARDBP mutations was 55.6 years (n = 8),
6 close to the 54.7 years observed in the 34 cases in this study. The variability in survival noted in
FUS families may pose difficulties in assessing response to treatment. We found a patient with the p.S96del mutation who survived for 18 years, suggesting that patients with certain specific
FUS mutations may have a better prognosis than others. This variation in survival has also been observed in some mutant
SOD1-mediated ALS cases. Patients with the H46R mutation have a much longer survival (>17 years) than those with the A4V mutation (1 year) in
SOD1.
38,39 We also noted that patients in the same pedigree may have significant differences in age and site of symptom onset. Some individuals with
FUS mutations had no symptoms even as their children were getting affected. This phenomenon suggests that other factors, including genetic background and environmental exposure, may modulate the clinical course. Differences in age at onset between
FUS and
TARDBP mutations are interesting as these 2 molecules share similar domains, and both are RNA/DNA binding proteins.
33 Our study has enlarged the spectrum of neurodegenerative phenotype associated with mutations in
FUS.