Better survival in female SOD1-mutant patients with ALS: a study of SOD1-related natural history

Background SOD1 mutations are the most common cause of amyotrophic lateral sclerosis (ALS) in non-Caucasian patients. Detailed natural history profiles of SOD1-mutant patients will be beneficial for the strategy and interpretation of future SOD1-targeted clinical practice. Methods Mutational distribution, age at onset (AAO), site of onset, diagnostic delay, disease progression (rate of ALSFRS-R decrease, ΔFS) and survival were analysed. Further comparisons between heredity of disease, gender, and mutations were performed. Results Sixty-six cases with 43 SOD1 mutations were included and analysed, with p.His47Arg as the leading mutation and seven novel variants identified. The mean (SD) AAO was 43.92 years (9.24) for all subjects, with a significant difference between patients carrying mutations in exon 2 (n = 24,46.83, 8.31) and exon 4 (n = 18, 37.75, 7.67) (p = 0.002). The median (IQR) diagnostic delay from symptom onset was 14.50 (6.00–36.50) months for all SOD1-mutant patients, 9.50 (4.75–24.25) months for males and 24.00 (9.50–47.50) months for females, revealing a gender difference (p = 0.009). Similar advantages in median (IQR) ΔFS [male: female, 0.55 (0.24–0.94) vs 0.19 (0.06–0.90), p = 0.041] and mean (95% CI) survival [57.4 (38.90–75.90) months vs 125.6 (99.80–151.50) months, p = 0.006] were also observed in females, both of which existed in sporadic ALS only when stratified by familiar or sporadic ALS. Conclusions The results highlight a distinct mutational distribution and natural history spectrum in ALS patients carrying SOD1 mutations in China. A prominent mild disease progression was observed in female patients, which had rarely been reported in the previous literature. This finding, together with the detailed analysis of natural history among each mutation, can have important implications for future genetic counselling and SOD1-targeted clinical trials.

In the present study, we described the genetic and natural history profiles of ALS patients with SOD1 mutations obtainted from a national referral hospital site in China. The effort to clinically stratify mutations in SOD1 according to the patients' natural histories should be beneficial for future genetic counselling and the selection of genetically and clinically homogeneous patients for SOD1-targeted clinical trials for ALS.

Subjects
Patients were recruited from the national referral Amyotrophic Lateral Sclerosis Clinic at the Department of Neurology, Peking University Third Hospital (PUTH), Beijing from 2007 to 2013. The patients were examined and diagnosed by board-certified neurologists as having definite, probable, or possible ALS according to Airlie House diagnostic criteria [17]. All patients included provided written informed consent to participate in the clinical and genetic studies, which were approved by the institutional ethics committee of PUTH, during their first visit to the hospital. Only patients with genetically confirmed SOD1 mutations/variants were included in further clinical analysis.

Clinical data collection and analysis
Demographic and clinical patient data, including gender, age at onset (AAO), site at onset, diagnostic delay, and ALS Functional Rating Scale -Revised (ALSFRS-R) score at diagnosis, were collected during the first visit to the hospital,. Outcome/endpoint event data were updated during telephone follow-ups every three months. Diagnostic delay from onset was identified as the interval from symptom onset to diagnosis. Disease progression was defined as the rate of decrease in the ALSFRS-R score at enrolment (ΔFS), which was calculated as follows: ΔFS = (48-ALSFRS-R score at diagnosis) / diagnostic delay (months) [18]. Cut-off ΔFS values of 0.5 and 1.0 were applied to divide the patients into three subgroups: slow progression (< 0.5), intermediate progression (0.5-1.0), and fast progression (> 1.0) [19,20]. Survival time was defined as the interval from symptom onset to endpoint events from any cause or the last follow-up, where death and tracheotomy were defined as endpoint events. The censoring date for survival data was May 31, 2016. Patients lost to follow-up were censored at the last known living data point. Additional clinical features, such as cognitive status and presence of sensory symptoms and signs, were not analysed in this study.

Statistical analysis
The mean (SD) AAO, median (IQR) diagnostic delay, mean (IQR) ΔFS, and mean (SD) disease duration were straightforwardly calculated [29] for each mutation. Descriptive statistics were provided for the bulbar site of onset, AAO, diagnostic delay, ΔFS and survival time by total patients, heredity of disease (familial or sporadic ALS), and gender. Moreover, Student's t-test or the Mann-Whitney test was applied to compare continuous data, while a standard chi-square test or Fisher's exact test was used to analyse dichotomous variables, such as gender and site of onset. Survival time was determined using Kaplan-Meier analysis, and differences were determined by log-rank testing. A two-tailed p < 0.05 was considered statistically significant. All analyses were performed using GraphPad Prism 5.0 (GraphPad Software, CA, USA). A total of 39 missense, 1 nonsense, and 3 deletion/insertion mutations/variants were found spanning all five exons, only one of which was in exon 3. The most frequent SOD1 mutation was p.His47Arg, traditionally named H46R, which was found in 6 fALS patients and 2 sALS patients (8/66, 12%), followed by p.Gly42Asp, p.Gly42Ser, p.His44Arg, p.Leu107Phe, and p.Gly142Ala in 3 patients each (3/66, 4.5%). Seven novel variants, including p.Phe21Val, p.Trp33Gly, p.Arg80Ser, p.Gly86Cys, p.Ala90Phe, p.Val95Gly, and p.Glu133Ter, were identified as variants of unknown significance (VUS) or likely pathogenic variants in the present study ( Table 1). Some of the SOD1 mutations identified in fALS patients were previously reported [30].

Overview of demographic and genetic features
We found that the mutation p.His47Arg exhibited a relatively consistent and mild phenotype, expect for in a male patient who was 69 years old at onset and died within 14 months after onset. The other seven patients presented a mean (SD) AAO of 48.14 (7.47) years, a median (IQR) diagnostic delay of 62 (45-84) months, and a disease course between 65 and 155 months, with no endpoint events.
The AAO varied among different mutations, with p.Asn140Lys and p.Ser106Leu being associated with the oldest and youngest AAOs (patients in their sixties and twenties, respectively), albeit with only one case for each mutation ( Table 2). Ten mutations/variants presented a relatively older AAO (more than the third quartile of 50 years, highlighted in green in Table 2), and eleven showed relatively younger AAO (less than the first quartile of 38.25 years, highlighted in yellow in Table 2).

Disease progression
Because ALSFRS-R scores at enrolment were available for 57 patients, the rate of disease progression was analysed in these patients.   [21] and [22], values less than 0.50 were defined as slow progression and highlighted in green, whereas values greater than 1.00 were defined as fast progression and highlighted in yellow # For disease duration, there were 61 available cases, and 60% were censored (lost to follow-up or still surviving). Therefore, 25% of the patients survived less than 31 months, and 50% of the patients survived less than 89 months. Values greater than the second quartile (89 months) are highlighted in light green, whereas values less than the first quartile (31 months) are highlighted in yellow 32%) carrying 12 mutations were identified with intermediate progression, and 9 patients (9/57, 16%) carrying 9 mutations deteriorated more rapidly (highlighted in yellow in Table 2).

Disease duration and survival
Course of disease data were available for 61 subjects, including 12 subjects censored for lost to follow-up and 26 subjects still surviving at the censoring date. Based on statistical analyses, the median survival time was 89.0 months, the mean survival time was 97.  (Fig. 2). Among patients with different mutations, those with p.Ala5Val and p.Ile150Val had the shortest and longest disease durations, respectively; notably, the latter group was still alive at the censoring date. Because 60% of the cases were censored, we were able to only calculate only the first and second quartiles of disease duration as 31.0 months (mutations highlighted in yellow, Table 2) and 89.0 (mutations highlighted in light green, Table 2), respectively.

Discussion
We identified 66 patients harbouring 43 confirmed SOD1 mutations among 923 sALS and 159 fALS patients. Because of the number of patients available, the present study represents the largest effort to study SOD1 mutations to date in a non-Caucasian ALS population, including 47 fALS patients and 19 sALS patients with SOD1 mutations. The extensive survey and analysis of ALS-related SOD1 mutations and the patients' natural histories provide data that could serve in the design and interpretation of future clinical trials targeted at patients with SOD1 mutations.
In Chinese SOD1 mutant patients with ALS, p.His47Argwas most frequently identified (8/66, 12%); this mutation is found mainly in Asian ALS patients (more than 15 pedigrees have been reported, only 3 of which are of Caucasian descents) [31]. P.His47Arg was also the leading mutation in Japanese SOD1-mutant patients [32]. In our study, almost all patients with p.His47Arg presented with the characteristic phenotype of exclusive spinal muscle initiations but mild disease course, i.e., an older AAO, extended diagnostic delay, slower progression, and longer survival time (Table 2), with the exception of one patient with a survival time of only 14 months. Conversely, the most globally predominant SOD1 gene mutation, D90A [32], was absent in our study, and only one case with A4V (p.Ala5Val, 1/66) and one case with I133T (p.Ile114Thr, 1/66) were found; these two mutations account for~50% and 65% of SOD1 mutations in the United States [29] and Canada [33], respectively. This result is not uncommon since the frequency of specific mutations can vary among different countries or even among different regions of the same country. For example, the A4V mutation has also been rarely observed in Europe. Another case in point is the mutation cluster of L48F-SOD1, which has been described in central Italy but is quite rare in other Italian regions [34].
Furthermore, significant beneficial effects of female gender on diagnostic delay, disease progression and survival time were observed, and we postulate two explanations for this result. 1) The mutations with poorer prognosis presented more often in male patients. For example, 17 patients had a diagnostic delay of less than 6 months (the first quartile), 13 of which were males; similarly, 13 patients had an ΔFS of more than 0.90 (the third quartile), 8 of which were males. Finally, 23 patients had a survival time of less than 31 months (the first quartile), 16 of which were males. 2) For the identical mutation, the male patients presented a poorer prognosis than females. In total, 8 patients carried the p.H47R mutation. Among these patients, the 3 male patients presented a mean diagnostic delay of 54.7 months, an ΔFS of 0.12, and a survival time of 70 months, while these values in the other 5 female patients were 67.4 months, 0.08, and 120.5 months, respectively. The difference in the diagnostic delay between genders existed in both fALS (p = 0.042) and sALS (p = 0.027), while gender differences in disease progression (p = 0.031) and survival time (p = 0.009) were present in only sALS. The finding of a longer survival time in female patients is consistent with what we previously found in the overall sALS population (female: male, 87 months vs 63 months, p = 0.008) [38]. To our knowledge, such a gender difference in SOD1-related ALS patients' natural histories has not been previously described. The rarity of this finding may be due to having an insufficient number of ALS cases with SOD1 mutations or a lack of detailed follow-up data. It remains unclear why female patients present a more remarkable survival advantage in the sporadic population. A possible explanation for this discrepancy is that fALS is inherited as a mendelian disease (monogenic), while sALS is likely an oligogenic disease [39]. We speculate that in fALS cases, the impact of a causative gene is predominant; while in sALS cases, the effect of each variant in certain gene is minor, and the modifying effect of gender could therefore be manifested. Since extended diagnostic delay has been associated with longer survival in ALS patients in several studies [38,40], it is reasonable that females with a longer diagnostic delay had better survival in the present study. The effect of gender on survival time has been mentioned multiple times with contrasting findings in ALS populations, with some studies suggesting worse survival for women [41,42] and others suggesting better survival [43,44]. One possible explanation for gender differences could be the role of gonadal hormones since these hormones, particularly oestrogen, have been proven to be neuroprotective. A very recent report [45] demonstrated a negative association between ALS and hormonal contraception use in women, reporting a dose-response effect. Another potential reason for the gender difference is that smoking is a risk factor for ALS; patients who smoke may have a shorter survival time [46,47], and there is a higher percentage of male smokers in the Chinese ALS  [38]. Another supporting and interesting interpretation is from a meta review [48] of a SOD1 G93A mouse model of ALS. That report was suggestive of gender-and genetic background-related effects on disease course, i.e., female-related neuroprotective effects on lifespan and disease duration were observed for B6SJL mice but not for mice on a C57BL/6 background, implying that the inherent genetic differences observed between backgrounds had some interactive effect on the presentation of the female hormone-related protective effect.  [29,50]. As described in a previous study, SOD1-mutant patients in a specific population were also younger than the overall ALS patients in the same population (total in Canada: 59.5) [49]. Patients carrying SOD1 mutations with bulbar onset, who usually present a later AAO, were less common than overall ALS patients in China according to our previous report [38] (8.1% vs 14%), which may partially explain why the mean AAO of SOD1-carrying patients was younger than that of the overall ALS patients in the same population.
Another striking finding of this study is the distinctive profile of disease progression and survival in the overall SOD1-mutant patients. The proportion of subpopulations was approximately 5:3:2 (slow: intermediate: fast). The overall median survival time of 89 months, much longer than that of 17.5 months reported in a recent SOD-related clinical trial [15], could be due to the rarity of the A4V mutation in the Chinese ALS population, which was characterized by a survival time of 1-2 years. The median survival of non-A4V SOD1 patients in the United States was 6.8 years, comparable to that of our patients. Nevertheless, the discrepancy in the SOD1-related natural histories between Chinese and Caucasians should be recognized in the design and strategy of SOD1-targeting clinical trials.
The present study distinguished seven novel variants in the included ALS patients. These variants were evaluated as VUS or likely pathogenic according to the ACMG Standards and Guidelines [28] (Table  1). It is noteworthy that the nonsense variant of p.Glu133Ter was not classified as evidence of very strong pathogenicity because loss of function (LOF) is not believed to be the primary pathomechanism of SOD1-related ALS.
Since this cohort was based on a national referral site for ALS, caution should be applied when using these data to deduce the frequency of SOD1 mutations in Chinese ALS patients. Additionally, the relatively younger mean AAO of SOD1-mutant patients and the longer median survival time compared with those of overall ALS patients in China (mean AAO: 43.92 years vs 49.8 years, median survival time: 89 months vs 71 months) made it more likely that these patients reported to the referral hospital. The calculated frequencies of 30% for fALS and 2% for sALS were likely an overestimation. Nevertheless, these numbers are the result of the greatest effort to date to explore the prevalence of SOD1 mutations in the Chinese ALS population.

Conclusions
In conclusion, the data shown herein demonstrate that p.H47R is the leading mutation in SOD1-mutant patients in China, and that female patients with SOD1 mutations have notable advantages in disease progression and survival over male patients. Detailed analyses of patients' natural histories revealed a discrepancy not only between Chinese and Caucasian ALS SOD1-mutant patients but also among different mutations. These findings may serve as a reference database and will be helpful for future SOD1-targeted clinical trials. intellectual content. Liu X.L., study design, genetic analysis, and critical revision of the manuscript for intellectual content. Chen L., patient follow-up and data collection and critical revision of the manuscript for intellectual content. Fan D.S., study concept and design, acquisition of data, study supervision, and critical revision of the manuscript for intellectual content. All authors read and approved the manuscript.
Ethics approval and consent to participate The study was approved by the ethics committee of Peking University Third Hospital (PUTH) and written informed consent was obtained from each subject included in the study after the procedure was fully explained.

Consent for publication
Not applicable.