вход по аккаунту


Epidemiology of mutations in superoxide dismutase in amyotrophic lateal sclerosis.

код для вставкиСкачать
Epidemiology of Mutations in Superoxide
Dismutase in Amyotrophc Lateral Sclerosis
M. E. Cudkowicz, MD,*T D. McKenna-Yasek, RN,* P. E. Sapp, BS,*$ W. Chin, BS,*$ B. Geller, BS,*$
D. L. Hayden, MS,$ D. A. Schoenfeld, PhD,S B. A. Hosler, PhD,* H. R. Horvitz, PhD,$
and R. H. Brown, MD, PhD't
We registered 366 families in a study of dominantly inherited amyotrophic lateral sclerosis. Two hundred ninety families
were screened for mutations in the gene encoding copper-zinc cytosolic superoxide dismutase (SODl). Mutations were
detected in 68 families. The most common SODl mutation is an alanine for valine substitution in codon 4 (50%). We
present clinical and genetic data concerning 112 families with 395 affected individuals. The clinical characteristics of
patients with familial amyotrophic lateral sclerosis arising from SODl mutations are similar to those lacking SODl
defects. Mean age at onset was earlier (Wilcoxon test, p = 0.004) in the SODl group (46.9 years [standard deviation,
12.51 vs 50.5 years [11.5] in the non-SOD1 group). Bulbar onset was associated with a later onset age. The presence
of either of two mutations, G37R and L38V, predicted an earlier age at onset. Kaplan-Meier plots demonstrated shorter
survival in the SODl group compared with the non-SOD1 group at early survival times (Wilcoxon test, p = 0.0007).
The presence of one mutation, A4V, correlated with shorter survival. Q7R, G41D, and G93C mutations predicted
longer survival. This information suggests it will be productive to investigate other genetic determinants in amyotrophic
lateral sclerosis and to use epidemiological characteristics of the disease to help discern molecular mechanisms of motor
neuron cell death.
Cudkowicz ME, McKenna-Yasek D, Sapp PE, Chin W, Geller B, Hayden DL, Schoenfeld DA,
Hosler BA, Horvitz HR, Brown RH. Epidemiology of mutations in superoxide dismutase
in amyotrophic lateral sclerosis. Ann Neurol 1997;41:210-221
Amyotrophic lateral sclerosis (ALS) is a progressive,
uniformly lethal neurodegenerative disorder of the voluntary motor system [l]. The prevalence of ALS is 4
to 6 cases per 100,000, with an annual incidence of
0.4 to 1.8 per 100,000 [2-61. In the majority of patients ALS is sporadic. However, in 10 to 15% of patients it is inherited as an autosomal dominant trait
(familial ALS or FALS) [7].
A subset of these are associated with a defect in the gene encoding cytosolic
copper-zinc superoxide dismutase (SOD1) [8]. Nearly
45 different mutations in SODl have been found exclusively in FALS. That sporadic and familial forms of
ALS are clinically and pathologically similar [2, 7, 91
suggests that an understanding of the familial disease
will illuminate possible epidemiological and pathophysiological mechanisms in sporadic ALS.
One approach to understanding the pathobiology of
a disease is to use epidemiological analyses to delineate
parameters that distinguish subpopulations of patients.
Ultimately, such parameters may identify critical determinants that govern the evolution of a disease. Toward
this end, several studies described the epidemiological
features of both sporadic and familial ALS. Age and
gender are the only risk factors repeatedly documented
in epidemiological studies of sporadic ALS [ 101. Sporadic ALS and FALS rarely occur before the age of 30
years; in sporadic ALS, the median age at onset is 55
and the median survival time is approximately 3 years
[7,1I]. There is no racial or geographic predisposition
for either familial or sporadic ALS, with the exception
of an increased incidence in ALS-related disorders in
certain isolated regions in the Western Pacific rim including the Marianas Islands of Guam after World
War I1 [3]. Some epidemiological studies suggested
that trauma and exercise may be risk factors for the
development of sporadic ALS, but these results were
not widely confirmed [ l o , 121.
The natural history of ALS varies widely from patient to patient, resulting in a 20-fold difference between the slowest and the most rapid courses [ 131. A
few natural history studies examined factors that influence age at disease onset and survival in sporadic ALS;
the majority of these studies were retrospective, and the
conclusions differed. One study found no correlation
From the *Day Neuromuscular Research Laboratory, Massachusetts
General Hospital East, Charlestown; ?Department of Neurology
and $General Clinical Research Center, 'Massachusetts Gener-a1
Hospital, Harvard Medical School, Boston; and $Howard Hughes
Medical Institute, Department of Biology, Massachusetts Institute
of Technology, Cambridge, MA.
Accepted for publication May 1, 1996, and in revised form Ju1 24.
Accepted for publication 1ul 25, 1996.
Add,& corre~pondence Dr Cudkowicz, Department of
ogy, Massachusetts General Hospital, 32 Blossum Street, Room
VBK 81 Boston, MA 0211 4 ,
210 Copyright 0 1997 by the American Neurological Association
between the rate of deterioration or the site of symptom onset and age at onset [14]. Another study of patients with sporadic ALS documented that age at onset,
delay from first symptom to entering an ALS clinic,
and rates of change in motor function and respiratory
function are significant covariates of survival [151.
In sporadic ALS, shorter survival time is found in patients with older age at onset and bulbar onset [16201.
The natural history of FALS is not well studied. Although numerous mutations in the gene for SODl
have now been documented in FALS, the overall frequency and distribution of SODl mutations in FALS
are unknown. In addition, it is unclear whether FALS
patients with and those without mutations of SODl
differ clinically. A literature review of FALS in 1991
demonstrated that older age at disease onset was associated with shorter survival time [21]. Also in FALS,
there is a greater variability in onset age between than
within families, suggesting a role of the specific genetic
defect in determination of age at disease onset [2, 7,
15, 161. Variability in disease duration between families with FALS occurs with a range from less than 1
year to more than 20 years [22]. These studies were
completed prior to the knowledge of mutations in
SODl as a cause of some cases of FALS.
In the present study of genetic linkage in FMS, we
acquired information on a large number of ALS pedigrees. We performed analyses of the gene for SOD1,
screening for FALS-associated SOD 1 mutations in at
least 1 affected member of most of these pedigrees. The
combined epidemiological and genetic database affords
an excellent opportunity both to compile a comprehensive description of the clinical phenotype of individuals
bearing SODl mutations and more generally to conduct a retrospective cohort study of patients with
FALS. Our analysis examined the hypothesis that some
clinical features may be stratified according to specific
SODl mutations in inherited ALS.
Materials and Methods
Patient Selection
From 1985 to the present, we registered 366 families with
inherited ALS in a study of genetic linkage in this disease.
Genomic DNA from 308 individuals in 290 of these families
was screened for SODl mutations [23], as were samples
from 23 individuals with familial and 90 with sporadic parkinsonism and nearly 200 individuals without neurological
diseases. From 112 of our 366 ALS families, genetic, genealogical, and clinical data were collected in a database of 395
FALS patients. Of these 395 FALS patients, 194 were in 49
families with SODl mutations, while 146 were in 43 families without S O D l mutations. Fifty-five patients were in 20
families from which the D N A of an affected patient was not
available for screening. The patients were obtained through
the weekly Neuromuscular Clinic at the Massachusetts General Hospital, and referrals from outside physicians and both
the Muscular Dystrophy Association and the ALS Association. Families were from throughout the United States and
Europe. The single area from which we recruited the largest
number of families (10% of total) was Massachusetts. Families included in the dataset contained multiple individuals
affected with ALS representing one, two, or three generations. Clinical and genealogical information was obtained
through interviews with family members and health care personnel, medical records, and autopsy reports. The criteria for
diagnosis were the presence of progressive, selective motor
dysfunction, signs of both lower and upper motor neuron
involvement, and at least 2 affected members in a family.
Clinical Features
Data on the following clinical features were collected for all
patients: gender, age at symptom onset, site of onset, predominance of lower or upper motor neuron involvement,
inheritance pattern, presence of dementia or parkinsonism,
date of death, date of ventilatory support, or date last known
alive. The month and year of onset of symptoms and death
or ventilatory assistance were recorded. Age at onset was defined as the age when first symptoms were noted by the
patient. Symptoms included motor weakness but not cramps.
Disease duration was determined as the time from onset of
symptoms to death, or an equivalent censoring variable (date
placed on ventilation, date last known alive). Age at onset
and disease duration were recorded in years.
Data Analysis
A Unix-based Statistical Analysis System (SAS, Cary, NC)
program was used for descriptive statistics, significance testing, and survival analysis modeling. Baseline characteristics
including median and mean ages at onset and disease durations, male-female ratio, onset site, and inheritance were determined for the FALS S O D l group and FALS non-SOD1
group. The male-female ratio was determined for each mutation group. Mean duration was determined using maximum
likelihood estimates including censored data (SAS, proc lifereg). Data were censored at the last date of data entry (December 1995) if the patient was alive, or at the last date the
subject was known to be alive if no additional data were
available as of December 1995. Median duration was derived
from Kaplan-Meier estimates of the survival curves (SAS,
proc lifetest). This facilitated comparison with published reports of mean duration with uncensored data. The frequency
of specific SOD 1 mutations was determined and the median
and mean ages at onset and disease durations were calculated
for each SODl mutation type.
Survival analysis was used to analyze time to onset of disease symptoms and disease duration. Univariate survival
modeling was performed using Kaplan-Meier analysis [24].
To determine whether there was a difference in time to disease onset or survival between groups, two nonparametric
statistical tests were performed-the log-rank test and Wilcoxon test [25]. The Wilcoxon test places more emphasis on
shorter survival times [25, 261. A significance of 0.025 was
used to correct for the use of two tests. Patients with and
those without SOD1 mutations were compared with respect
to both age at onset and disease duration. All variables con-
Cudkowicz et al: Epidemiology of S O D l in ALS
rained in the database were examined individually for relationship to age at symptom onset and disease duration. The
Cox proportional hazards regression mode1 was used for multivariate survival modeling to examine the relationship between potential predictors and the outcome of interest (either
time to disease onset or survival) [27]. Data from the 55
patients whose mutation status was unknown were not included in the Cox proportional hazards regression models.
Gender, inheritance, and onset site were treated as dichotomous variables (male vs female, maternal vs paternal inheritance, bulbar vs limb onset site). Age at onset was treated
as a continuous variable. Mutation status was treated as a
categorized variable. Missing data were included as separate
categories. Groups of FALS S O D l mutations that had less
than 5 patients were combined into one group. A forward
step-selection algorithm was used [25].
There were 49 families with SODl mutations in this
study. T o check whether the effect of the mutation arose
entirely from nongenetic differences in families, we calculated the mean disease duration for each family using maximum likelihood estimation that accounts for censoring [28]
and tested for an effect of mutation using a van der Waerdin
test [29]. In this analysis, affected family rather than affected
individual is the statistical unit. For onset age a mixed model
with a random family effect was used since there are no
censored data. The results of these analyses are only menrioned in the text when they differ substantially from the
per-individual analyses.
Our data collection methods make it impossible to estimate the disease onset and duration distributions without
bias. Onset distribution estimates d o not take into account
family members who do not yet have the disease. Unbiased
estimates would require knowing the ages of each family
member at risk who had not yet developed the disease. The
duration distribution would also be biased if it depended o n
the age at onset. Our primary interest is the effect of SODl
mutations on disease onset and duration. The significance
of SODl mutations determined by the statistical tests we
used is not affected by the biases described above because
the biases would equally affect each group of patients in the
Descriptive Features
Of 290 ALS families whose DNA was analyzed, 68
(24.3%) had mutations in SOD]. Of 395 individuals
with FALS included in this clinical study, 194 (49.1%)
had mutations in SODI; 146 (37.0%) did not and 55
patients (13.9%) were from 20 families for which there
was no DNA from affected members for screening (Table 1). This group of 395 FALS patients is from 49
S O D l (43.8%) families, 43 (38.4%) non-SOD1 families, and the 20 families of unknown S O D l mutation
status (17.9%).
Regarding the FALS patients, the male-female ratio
was 1 : 1 in the SODl group, and 1.4: 1 in the nonS O D l group (see Table 1). In all the mutation groups
except G41 D (Gly-Asp), the gender ratios did not differ significantly from 1: 1. All 7 subjects with the
Annals of Neurology
Vol 41
No 2 February 1997
Tdle 1. Clinical Characteristics by SOD1 Mutation Status
in Familial Amyotrophic Lateral Sclerosis
No. of subjects
Male-female ratio
Mean age (yr) at
onset (SD)
Mean duration (yr)
Non-SOD1 p Value"
194 (49.1)
46.9 (12.5)
146 (37.0)
50.5 (11.5)
3.9 (5.5)
3.3 (2.8)
0.1 1
Median duration
% Bulbar onset
Yo Limb onset
Yo Onset site
Yo Paternal
Yo Maternal
Yn Inheritance
'Sratistical comparisons were made using
"Fifty-five subjects (13.9%) were from 20 families from which the
DNA of an affected patient was not available.
standard deviation.
G41D mutations were female. In 11.9% of SODl and
18.5% of non-SOD1 individuals, FALS began with
bulbar onset. Limb symptoms heralded the start of the
disease in 51.0% of S O Dl and in 47.3% of nonS O D l patients. Onset site was unknown in the remaining individuals. In 13.3% of female and 15.3%
of male patients, FALS began with bulbar onset.
Paternal transmission in the FALS group was evident
in 36.1% of SODl and 36.3% of non-SOD1 patients.
Maternal transmission occurred in 46.7% of SOD 1
and 30.8% of non-SOD1 patients. In 17.2% of S O D l
and 32.9% of non-SOD1 patients it was not known
whether the disease was transmitted maternally or paternally.
Only 6 patients with FALS demonstrated signs of
dementia; all 6 were in the non-SOD1 group.
The most common mutation in the FALS SODl
group was the A4V (Ala-Val) mutation, which occurred in 27 (55.1%) of the 49 SODl families included in this study. The frequencies and clinical characteristics of the different mutations among these 49
families are shown in Table 2. The A4V mutation was
also the most common mutation among the entire 68
S O D l families in our database (50%). A typical pedigree with the A4V mutation is shown in Figure 1.
Age at Onset
Age at onset data for the FALS group were available
for 190 (97.4%) SODl and 146 (100%) non-SOD1
Table 2. Disease Characteristics by SOD I Mutation l j p e in Familial Amyotrophic Lateral Sclerosis
No. of
Families (Yo)
No. of
27 ( 5 5 . 1 )
4 (8.0)
2 (4.0)
2 (4.0)
1 (2.0)
1 (2.0)
1 (2.0)
1 (2.0)
1 (2.0)
1 (2.0)
1 (2.0)
1 (2.0)
1 (2.0)
1 (2.0)
1 (2.0)
1 (2.0)
1 (2.0)
1 (2.0)
49 ( 1 00%)
Mean Age
(yr) 2 SD (n)
at Onset
Mean Durationb
(yr) 2 SD (4
47.8 5 13.3 (84)
58.9 t 12.6 (7)
46.9 ? 12.0 (26)
41.5 t 8.1 (14)
(29-5 8 )
47.9 -+ 17.7 ( 9 )
40.0 ? 9.9 ( 8 )
46.0 t 7.3 (7)
47.4 t 12.4 (7)
50.8 t 19.9 (4)
49.8 t 3.9 (4)
35.5 2 7.2 (4)
43.5 ? 8.5 (4)
48.3 t 16.2 (3)
44.0 t 7.1 (2)
42.5 t 10.6 (2)
48.0 2 2.8
54.0 (2)
44.0 ( 1 )
1.4 2 0.9 (84)
3.5 t 2.8 (7)
4.0 t 2.3 (22)
2.8 t 1.9 (12)
2.2 2 1.5 ( 9 )
18.7 2 11.4 ( 8 )
17.0 t 6.3 (7)
10.1 t 6.2 (7)
0.9 t 0.5 (4)
2.8 t 1.5 (4)
2.3 5 1.3 (4)
2.3 % 2.2 (4)
10.5 t 5.5 (3)
0.9 t 0.1 (2)
12.3 2 3.7 ( 2 )
1.6 t 0.5 ( 2 )
2.8 2 0.3 (2)
2.4 ( 1 )
I .o
1 .o
'The mutations are ordered by decreasing number of affected families.
bCensored data are included in determination of mean or median duration.
'Range for age at onset and disease duration are listed for the four most common mutations.
SD = standard deviation.
patients. The distribution of onset age across all FALS
individuals is shown in Figure 2A. The mean age at
onset was 46.9 years (SD, 12.5) for SODl and 50.5
years (SD, 11.5) for non-SOD1 patients. Average age
at onset varied with different mutations: thus, the
Ll06V (Leu-Val) and the I113T (Ile-Thr) mutations
were associated, respectively, with the youngest and
oldest mean ages at onset (see Table 2, Fig 2B).
A Kaplan-Meier analysis of time to disease onset was
performed for the S O D l and non-SOD1 FALS
groups. This allows one to plot the cumulative probability of onset at any age (1 - the probability of surviving free of disease: Fig 3). To determine whether there
was a difference in time to disease onset between these
two groups, the log-rank test and Wilcoxon test were
n s
Fig I , Representativepedigree. Affected individuals in this
pedigree have the common ahnine-to-valine substitution at
codon 4 of the SODl gene. They are descendants of a family
first described by William Osler in 1880 with inherited,
adult-onset, rapidly progressive muscular atrophy [33].
Cudkowicz et al: Epidemiology of SODl in ALS 213
SODl (190)
100 -
Nan-SOD1 (146)
80 -
60 40
Age of Onset (years)
Fig 3. Plot of the probability o f disease onset by age.fir all
familial amyotrophic lateral sclerosis (FALS) patients divided
into non-SOD1 and SODl groups. The median age at disease onset was 47 years in the S O D l group and 52 years in
the non-SOD1 group. There was a signijcant diference in
time to disease onset between SODl and non-SOD1 (logrank test, p = 0.05; Wilcoxon test, p = 0.004). Each point
may rqresent multiple subject events.
Age of Onset (years, by decade)
" .
u c = J
; ;~ z "~ z~ " -
o - a r a z
SODl Mutation Status
by decade, f o r all patients with familial amyotrophic lateral
sclerosis (FALS). Thisfollows approximately a normal distribution with a mean age at onset of 49.3 years (standard deviation, 12.4 years). (B) Histogram illustrating the mean age
at onset f i r all patients with FALS without S O D l mutations
(non-SODI) and with SODl mutations, and individual
SODl mutations. The mean age at onset was 50.5 years
(standard deviation, 11.5 years) for non-SOD and 469 years
(standard deviation, 12.5 years) f i r SODI patients. Average
age at onset varied with different mutations, with the Ll06V
and the 1113 T mutations associated, respectively, with the
youngest and oldest mean ages at onset.
Annals of Neurology
Vol 41
Fig 2. (A) Histogram showing the distribution of age at onset
performed. The Wilcoxon test finding was significant
at the 0.025 level (Wilcoxon test
= 8.051, df = 1,
p = 0.004; log-rank test = 3.81, df = 1 , p = 0.05).
There was a significant difference in time to disease
onset between FALS patients with and those without
mutations in SODl. The non-SOD1 group had a later
disease onset than did the SODl group (a lower onset
probability at each onset age < 70). The age at 50%
onset probability was 47 years in the SODl group and
52 years
in- the non-SOD1 group.
No 2 February 1997
Covariates uf Age a t Symptom Onset
Three clinically significant variables were covariates of
age at symptom onset. These include bulbar onset, and
the following two mutations: G37R (Gly-Asp) and
L38V (Leu-Val). The G37R and L38V mutations were
associated with earlier disease onset. Bulbar onset, independent of SODl status, was associated with longer
intervals before disease onset. The subjects whose inheritance data were unknown had later disease onset
than did those for whom these data were known. This
may reflect a recall bias in our data collection methods.
Figure 4 plots the probability of onset by age for each
subgroup defined by these risk variables. The following
variables were not significant covariates of age at symptom onset: maternal inheritance, paternal inheritance,
gender, absence of mutation in SOD1, or any of the
remaining SOD1 mutations. These findings were not
present when family was taken as the unit of observation. This occurred because there was substantial variation in age at onset between families that had the same
Bulbar onset ( 5 7 )
Limb onset (190)
Age of Onset (years)
G37R ( 8 )
SODl Mutation Status
All Other SODI mutations (168)
Non-SOD1 (145)
Age of Onset (years)
Fig 5. Histogram illustrating the mean disease duration for
all patients with familial amyotrophic lateral sclerosis (FALS),
non-SOD1 patients, S O D l patients, and individual SODI
mutations. The mean disease duration was 3.5 years (standard deviation, 4.2 years) for all FALS, 3.9 (standard deviation, 5.5)for SOD, and 3.3 years (standard deviation, 2.8
years) for non-SOD1 patients. Average disease duration varied with different mutations, with the G41S and the G37R
mutations associated, respective&, with the shortest and longest
disease duration.
Fig 4. Plots of the probability of disease onset in all familial
amyotrophic lateral sclerosis (FALS) patients for each subgroup defined by the signijicant covariates of age at onset.
The risk ratiosfor L38K G37R, and bulbar onset, and the
corresponding 95% confidence limits, reqective&, were 2.2
(1.28, 3.79), 2.22 (1.09, 4.53), and 0.67 (0.50, 0.91). The
risk ratio is the relative risk of onset for an individual with a
given variable (mutation G37R) relative to an individual
without that variable, holding the other two variables (Le.,
mutation L38V and bulbar onset) constant. (A) Onset site:
All FALS patients were divided into bulbar and limb onset
groups. Bulbar onset is associated with longer intervals before
disease onset. (B) SODl mutation: All FALS patients were
grouped by mutation status (G37R, L38K and all remaining
FALS patients). The G37R and L38V mutations were associated with earlier disease onset.
mutation. Thus, the per-family analysis did not reveal
any predictors of early onset. This may be due to the
data collection scheme which would affect the age at
onset distribution for each family.
(SD, 5.5) for S O D l and 3.3 years (SD, 2.8) for nonS O D l patients. The mean disease duration varied depending on mutation type, with a range of 0.9 to 18.7
years (see Table 2). Median survival time was 2.0 years
in the S O D l group and 2.5 years in the non-SOD1
group. The distribution of disease duration across all
FALS individuals is shown in Figure 5.
Figure 6 shows a Kaplan-Meier survival plot comparing the S O D l and non-SOD1 groups with FALS.
The Wilcoxon test result was significant at the 0.025
level (x' = 11.5, df= 1, p = 0.0007); however, the
log-rank test result was not significant (x'= 11.5, df
= 1, p = 0.30). The two groups differed primarily at
early survival times, with a lower probability of survival
in the S O D l group. The plots cross, and at long survival times a lower probability of survival occurs in the
non-SOD1 group.
Disease Duration
Disease duration was known for 169 (87.1%) S O D l
and 130 (89.0%) non-SOD1 patients with FALS. An
additional 15 of the S O D l and 13 of non-SOD1 data
points were censored. Mean duration was 3.9 years
Covariates of Survival
The following four mutations-A4V, G41 D, G37R,
and G93C (Gly-Cys)-were the only significant covariates of survival. The A4V mutation was associated
with shorter survival, while this was significantly longer
Cudkowicz et al: Epidemiology of SODl in ALS
SODl (184)
A4Y (84)
Non-SOD1 (143)
G37R (8)
G41D (7)
0 9 % (7)
All other SODl mutations (79)
Non-SOD1 (143)
Duration (years)
Duration (years)
Fig 6. Kaplan-Meier plot o f survival probabiliq from time of
disease onset for all familial amyotrophic lateral sclerosis
(FALS) patients divided into non-SOD1 and SODl groups.
The median disease duration inas 2.0 years in the SODl
group and 2.5 years in the non-SOD1 group. There was a
sign$cant difference in survival between non-SOD1 and
SODl patients only at early survival times, with a lower
probability of survival in the SODl group (Wilcoxon test, p
= 0.0007). Each point may represent multiple subject events.
The graph was truncated at 10 years f o r better visual clarity
at shorter time points.
in patients with one of the following three mutations-G37R,
G41D, and G93C. Figure 7 plots the
probability of survival for each significant covariate of
disease duration. The following variables were not significant covariates of survival: age at onset, maternal
inheritance, paternal inheritance, site of disease onset,
gender, absence of mutation in SOD1, or any of the
remaining SODl mutations. Comparison of the survival probability between the non-SOD1 group and
the SODl group after removal of patients with the
mutations associated with shortest survival (A4V), and
longest survival (G37R, G41D, and G93C) revealed
that in general, FALS patients with SODl mutations
had similar survival probability to FALS patients without mutations in SODl (see Fig 7).
Over the last decade we have accumulated epidemiological data on an extensive group of FALS pedigrees.
More recently we examined the status of the SODl
gene in affected members from these families. Because
there has been no comprehensive description of the
clinical phenotype of individuals bearing SOD 1 mutations, we analyzed our family database to study both
the influence of SODl gene mutations on FALS and
FALS epidemiology more generally. Within the context
216 Annals of Neurology Vol 41
No 2
February 1397
Fig 7 . Kaplan-Meier plots of survival probability for each signi$cant covariate of survival. All familial amyotrophic lateral
sclerosis (FALS) patients were divided into six groups: A4K
G37R, G41D, G%C, all remaining SOD1 patients, and all
non-SOD1 patients. The presence of mutations A4V was associated with shorter survival; the G37R, G41D, and G93C
SODl mutations were associated with longer survival. The
risk ratiosfor A4K G93C, G37R, and G41D and the corresponding 95% conjidence limits, respectively, were 3.02
(2.30, 3.98), 0.28 (0.11, 0.69), 0.16 (0.06 0.41), and
0.14 (0.05, 0.38).
of these two broad issues, several points have been discerned.
First, mutations in SOD1 are highly specific to patients with FALS. These mutations were not detected
in 200 normal individuals, in nearly 100 patients with
sporadic or familial parkinsonism, or in patients with
other neurological diseases.
Second, the mutations are found only in 23.4% of
ALS families. Thus, there must be at least one additional gene defect responsible for FALS. At our center
and elsewhere, collaborative studies are under way to
identify these other genes. We and colleagues detected
genetic linkage of a juvenile-onset form of recessively
inherited ALS to a locus on chromosome 2q33. However, no specific gene defect has yet been identified
in that region [30] nor has this locus been linked to
any pedigrees with adult-onset, dominantly inherited
FALS. It was recently demonstrated that even in the
absence of a specific, disease-causing gene defect, it may
be instructive to identify genes whose polymorphisms
may confer heightened risk for specific diseases. Thus,
the E4 allele of the protein apolipoprotein E is associated with earlier onset of Alzheimer’s disease [311. We
reported previously that this allele is not overrepresented in early-onset or more rapidly progressive forms
of ALS [32];moreover, in our FALS patients, it was
not associated with any particular pattern of disease
onset. However, the allele reportedly is overrepresented
in sporadic ALS with bulbar onset [33].
Third, as first reported by Deng and associates [34],
the most common mutation is the substitution of alanine for valine at codon 4 (A4V), occurring in 50%
of our 68 families with SODl mutations. The second
most common among our ALS families is the substitution of isoleucine for threonine at codon 113 (1113T)
(11.7% of our 68 families with S O D l mutations). We
note in historical perspective that the first detailed description of inherited, adult-onset, rapidly progressive
muscular atrophy was published by William Osler in
1880 1351. We have now had the privilege of following
affected members of this family (see Fig 1); these individuals have the common A4V mutation. The family
members we followed demonstrated onset and survival
characteristics typical of A4V patients in our present
survey (see Table 2).
Fourth, we demonstrated that in FALS there is
marked variation in disease severity as measured by age
at symptom onset and disease duration; this variability
is evident when one compares SODl to non-SOD1
FALS pedigrees. It is also evident when one compares
the clinical characteristics of patients with the different
mutations. For example, age at onset is younger in the
FALS S O D l group, which also shows a higher probability of dying in the first months of the illness than
is seen in the non-SOD1 group (see Table I , Fig 6).
Moreover, our data confirm earlier reports [7,
211 that
FALS patients (in the aggregate) have earlier onset and
shorter courses than do those with sporadic ALS [ 151.
In our series, another factor that appears to modulate
onset age is bulbar onset; in general, bulbar-onset patients are older when their illness begins (see Fig 4A).
No other tested clinical variable was a significant predictor of age at symptom onset or survival.
The variations in disease onset and course are also
evident within the population of patients with S O D l
mutations: These parameters stratify by mutation type
and location. The variation in disease duration with
different mutations is more striking than the variation
in disease onset. Indeed, there was substantial variation
in age at onset between families that had the same
mutation. This raises the possibility that the variation
in onset might be due to other genetic or environmental factors rather than the SODl mutations themselves.
A summary of age at onset and disease duration from
published studies of families with SOD mutations is
found in Table 3. The data we present here on age at
onset and disease duration by mutation status (see Table 2) are similar to those in Table 3 [34, 36-40],
except that our data are based on larger sample sizes.
It is intriguing that mutations that confer a risk for
earlier onset (G37R and L38V) differ from those associated with shorter duration of illness (A4V). Indeed,
G37R is associated with a longer duration. These features suggest that the physicochemical properties of a
mutant enzyme that modify the timing of disease onset
are different from those modulating the rate of disease
progression. In this context, it recently was shown in
Table 3. Disease Characteristics by SODl Mutation q p e (Published by Other Investigators)
Mean (yr) t SD (n)
at Onset
47.0 (4)
40.0 (1)
36-62b (2)
44.9 (7)
46.8 t 13.5 ( 8 )
42.8 (2)
48.6 (6)
53.8 2 15.3 (5)
35.8 t 4.3 (4)
46-49" ( 2 )
40 ( 1 )
46-48b (3)
42 ( 1 )
Mean Duration
(yr) 2 SD (n)
N /A
0.75 (1)
4-8b ( 2 )
11.6 t 1.7 ( 8 )
N/ A
17.0 t 11.0 ( 5 )
1.8 t 5
5.7 2 4.5 (3)
N /A
2-20h (3)
1.5 (2)
'The mutations are ordered by codon position.
"Only range of data available.
standard deviation; N / A
not available.
Cudkowicz et al: Epidemiology of SODl in ALS 217
an animal model of ALS that agents which slow the
progression (riluzole, gabapentin) do not modify the
timing of onset [47, 481. That some mutations have
significantly different effects on a given disease parameter may assist in the analysis of biophysical defects in
the mutant molecule affecting that parameter. If a putatively critical defect determines the timing of disease
onset, for example, one anticipates that there will be a
correlation between the extent of the abnormal chemical property, at the molecular level, and the effect on
disease onset, at the epidemiological level.
In this regard, we note that it remains unclear why
the mutant S ODl molecule is injurious to motor neurons. Numerous studies indicate that the mutations
confer an adverse, toxic property on the mutant enzymes [49]. The mutations also variably reduce the dismutase activity of the molecule and generally reduce its
overall stability. In three tissues (brain, lymphoblastoid
cells, and red blood cells), S O D l activity is selectively
reduced in patients with SODl mutations [36, 50, 511.
No correlation between disease severity and red blood
cell S O D l activity has been defined [52]. This may
reflect the fact that in this anuclear cell type with a
long cellular lifespan (120 days) there is little mutant
protein present. Given the uncertainties inherent in
gauging SOD 1 activity in red blood cells, it is probably
more meaningful to evaluate enzyme activity either in
nucleated cells such as lymphocytes or lymphoblastoid
cells, or in in vitro systems. One study quantitated the
specific activity and molecular half-lives of SOD 1 protein expressed in COS cells [53]. The normal or wildtype S O D l molecule has a half-life of about 30 hours
or longer, while five of the six mutant SOD 1 molecules
studied show a reduced half-life. While one cannot
generalize from this initial study, we note that the halflife of the A4V molecule is markedly reduced (only 7
hours); families with this mutation have extremely
short survival. Reciprocally, a mutation associated with
longer survival, G37R, produces a mutant molecule
with a relatively longer half-life (13 hours). By the
same token, the specific molecular activity of A4V is
reduced in this system, while that of G37R is not.
Given the small number of mutants analyzed in this
system, it is hazardous to draw firm conclusions from
these data, although the study suggested the hypothesis
that the mutant proteins that are least stable (shorter
half-life, less enzyme activity) may generally have a
more severe clinical phenotype. Clearly, it will be imperative both to examine the molecular stabilities of a
wide range of mutant SODl proteins and then to correlate this with the clinical phenotype of large numbers
of families with the mutations.
Our extensive database providing clinical information on more than 113 ALS families allows us to study
the epidemiology of this illness. In particular, we have
an excellent opportunity to compare the epidemiologi-
218 Annals of Neurology
Vol 41 No 2
February 1997
cal features of FALS arising with and without SODl
mutations. In our view, the clinical phenomenology of
inherited ALS is quite similar in these two populations
of ALS patients. We found that there was a statistically
significant, but slight clinical difference in the mean
age at onset in these groups, with onset earlier in the
FALS S O Dl group (46.9 years [SD, 12.51 vs 50.5
years [SD, 11.51 in the FALS non-SOD1 group). In
a combined analysis of all of our FALS pedigrees, bulbar onset predicted a significantly later age at onset (see
Fig 4A). In any epidemiological study there is potential
for ascertainment bias. We minimized the potential for
geographic ascertainment bias by recruiting patients
from throughout the United States and a few European
countries. The rate of clinical data collection was
greater in the SODl families than the non-SOD1 families and this may pose a bias in our data collection.
FALS patients with S O D l defects survived approximately as long as did FALS patients without SODl
mutations. Thus, the mean survival time from symptom onset in the S O Dl subset was 3.9 years (SD, 5.5)
compared with 3.3 (SD, 2.8) for the non-SOD1
group. The median survival time from symptom onset
was 2.0 years in the SODl subset and 2.5 in the nonSODl group. Kaplan-Meier survival plots demonstrated shorter survival in the S O Dl compared with
the non-SOD1 group at early survival times.
As summarized already, our observations of the clinical characteristics of FALS are, in the main, consistent
with those previously published in careful studies by
several investigators. We specifically addressed three issues: the general descriptive features of our population
of FALS patients, factors that modify the timing of
onset of the disease, and modifiers of the disease
course. The mean age at onset in our total FALS group
was 49.3 years (12.4, SD); this is close to values of
45.7 (11.3, SD), 50.3 (12.4, SD), and 48.3 (11.7)
years previously reported [7, 21, 541. The median survival time in our total FALS group was 2.0 years; this
corresponded to values of 2.0, 2.0, and 2.4 years in
earlier studies [7, 2 I , 541. Like previous investigators,
we also found that in FALS, the male-female ratio was
approximately 1 : 1. In our series, the FALS gene defect
appeared to be highly penetrant. The major risk factor
for its phenotypic expression was age. In our series,
92% of FALS individuals were affected by the age of
70. This was determined by calculating the proportion
of at-risk FALS individuals in each family with members who developed symptoms by the age of 70. Due
to the small sample sizes, we were able to determine
mutation-specific penetrance only for the four most
common SOD1 mutations (see Table 2). The estimated penetrance in the group of subjects with the
most common mutation, A4V, was 91%. The estimated penetrance for I1 13T, ElOOG, and L38V mutations was 100%. The clinical features of our total FALS
group were slightly different from those of sporadic
ALS. The mean age at onset in our total FALS group
(49.3 years [SD, 12.31) was younger than that for sporadic ALS patients (55.7 years [SD, 12.91); the median
survival was shorter in our FALS group (2.0 years) than
in the sporadic ALS group (2.5 years) [14-161.
In our FALS material, only one variable other than
SODl mutation status was a covariate of age at onset:
In the subset of FALS patients with bulbar onset (with
or without SODl mutations) the disease began later
than in the group with limb onset. The association
of bulbar onset with later age at onset was described
previously in sporadic ALS but not FALS [ 15, 16, 181.
In their analysis of FALS, Strong and colleagues [34]
detected no relationship between site of onset or gender and age at onset. We do not have a clear explanation for this discrepancy.
As in earlier analyses, we examined clinical parameters that modify patient survival. Only SODl mutation
status was a covariate of survival. By contrast with prior
reports based predominantly on sporadic ALS [l5, 16,
19, 201, neither our study of FALS nor that of Mulder
and associates [7] confirmed the observation that patients with younger onset survive longer [15, 16, 19,
201 or that bulbar onset shortens survival [16, 181. O n
the other hand, in their analysis of FALS, Strong and
colleagues [21] did report that younger age at onset
predicts longer survival. Our study of FALS and that
of Strong and colleagues [34] did not find that site of
onset predicted survival. However, the Mulder study
[7] did find an effect of the site of disease onset on
survival. The basis for these discrepancies remains to
be defined.
A few studies examined the role of the gender of
the parent transmitting the disease on disease course
in FALS. Orell and colleagues [55] examined the pedigrees of 17 families with autosomal dominant ALS
with 106 affected subjects and found no significant
change in the proportion of patients inheriting the disease from female parents or grandparents. The median
age at onset was lowest in the grandfather to father
line (42 years) and highest in the grandfather to
mother line (53 years). Leone [56] examined the parental gender effect in 145 patients with FALS in a literature review. In Leone’s study, the age at onset in offspring was correlated with age at onset in the affected
parent only when the affected parent was the mother.
In addition, the percentage of maternally transmitted
cases increased significantly with age [56].
In conclusion, we demonstrated that in a population
of FALS patients the clinical parameters of affected individuals with SODl mutations are broadly similar to
those of individuals without the mutations. The only
distinguishing parameter is that the onset of symptoms
was slightly earlier in the FALS group with SODl mutations. The clinical features of the SODl group varied
to some degree with the specific SODl mutation. Two
specific SODl mutations, and an initial bulbar distribution of disease, are significant predictors of age at
onset in our combined group of all FALS patients.
Four specific SOD 1 mutations determined survival;
age at onset and site of onset were not predictors of
disease duration. Further study of other genetic mutations or gene modifiers as well as biochemical studies
of the specific mutated proteins are under way and
should yield further insights of the pathogenesis of
FALS. While the mechanism whereby the mutant
SOD 1 molecules impair neuronal function remains
unknown, we anticipate that there will ultimately be a
correlation between one or more measures of clinical
severity and biochemical parameters of the offending
chemical processes. Thus, this type of epidemiological
study may lay the groundwork for differentiating between different molecular mechanisms of neurotoxicity
in this disease and ultimately for initiating trials of
therapy in both familial and the more common forms
of sporadic ALS.
D t Brown is supported by grants from the Muscular Dystrophy
Association, the Amyotrophic Lateral Sclerosis Association, the
Pierre L. De Bourgknecht ALS Research Foundation, the Myrtle
May MacLellan ALS Research Foundation, the C.B. Day Investment Company, the National Institute on Aging (grant
lPOIAg12992-01), and the National Institutes of Health (NIH)
(grant IPOlNS31248-01). Dr Cudkowicz is the recipient of the
Clinical Invesiigator Training Program Award: HarvardiMIT Division of Health Sciences and Technology-Beth Israel Hospital, in
collaboration with Pfizer and the Pharmaceutical Research and
Manufacturers of America Foundation Award. Drs Hayden and
Schoenfeld receive NIH grant support from the Division of Research Resources (grant RR01066-17). Dr Hosler is supported by
NIH grant lPOlNS31248-01. Dr Horvitz is an investigator of the
Howard Hughes Medical Institute.
We thank the patients and their families for their assistance with
this project.
1. Adams RD, Victor M. Degenerative diseases of the nervous
system. In: Adam RD, Victor M, eds. Principles of neurology.
New York: McGrall-Hill, 1989:953-956
Horton WA, Eldridge R, Brody JA. Familial motor neuron
disease: evidence for at least three different types. Neurology
Emard JF, Thouez JP, Gauvreau D. Neurodegenerative disease
and risk factors: a literature review. SOCSci Med 1995;40:847858
Kurtzke JF, Kurland LT. The epidemiology of neurologic disease. In: Joynt R, ed. Clinical neurology. Philadelphia: JB Lippincott, 1983:l-43
Annegers JF, Appel S, Lee JF, Perkins P. Incidence and prevalence of amyotrophic lateral sclerosis in Harris County, Texas,
1985-1988. Arch Neurol 1988;48:589-593
Forsgren L, Almay GL, Holmgren G, Wall S. Epidemiology
of motor neuron disease in northern Sweden. Acta Neurol
Scand 1983;68:20-29
Mulder DW, Kurland LT, Offord KP, Beard CM. Familial
Cudkowicz et al: Epidemiology of SODl in ALS
adult motor neuron discase: amyotrophic lateral sclerosis. Neuroloby 1386;36:511-517
8. Kosen DR, Siddique T, Patterson U. et a]. Mutations in Cu/
Zn superoxide dismutase gene are associated with fdmilial
amyotrophic lateral sclerosis. Nature 1993;362:59-62
9, Swerts I., VanUenBerg 1’. Sclerose laterale atnyotrophique
familiale. J Genet 1976;24: 147-1 55
10. Kurtzke IF. Risk facton in amyotrophic lateral sclerosis. Adv
Neurol 1991;56:245-270
1 1. Munsat TL. Adult motor neuron disease. In: Kowland LP, ed.
Merritt’s textbook of neurology. I’hil‘1delphia: Lea Sr Febiger,
12. Sienko DG, Davis JP, ’Taylor JA, Brooks B. Amyotrophic lateral sclerosis: a case-control study following derection of a cluster in a small Wisconsin community. Arch Neurol 1990;47:
13. Appel V, Stewart SS, Smith G, Appel SH. A raring scale for
amyotrophic lateral sclerosis: description and preliminary experience. Ann Neurol 1987;22:328-3.33
14. Munsat ‘TL, Andrcs I’L, Finison I., et al. ‘The natural history
of mororneuron loss in ALS. Neurology 1988;38:452-458
15. Haverkainp LJ, Appe! V, Appel SH. Natural histoiy of amyotrophic lateral sclerosis i i i a database population. Validation of
a scoring system and ‘1 model for survival prediction. Brain
16. Norris F, Shepherd R, Ilenys E, et al. Onset, natural hisroiy
and outcome in idiopathic adult motor neuron disease. J Neurol Sci 1993;118:48-55
17. Jablecki CK, Berry C , Leach J. Survival prediction in amyotrophic lateral sclerosis. Muscle Nerve 1989;12:833-841
18. Gubhay SS, Kahana E, Zilber N, et al. Amyotrophic lateral
sclerosis: a study of its presentation and prognosis. J Neurol
13. Eiseti A, Schtilzer M, MacNeil M, et al. Ihration of amyotrophic lateral sclerosis is age dependent. Muscle Nerve 1993;16:
20. 1.ouverse ES, Bossuyt PMM, Weverling GJ, Jong JV. Mortality
risk stratification in amyotrophic lateral sclerosis. J Neurol Sci
21. Strong MJ, Hudson AJ, Alvord W G . Familial amyotrophic lateral sclerosis, 1850-1989: a statistical analysis of the world
literature. Can J Neurol Sci 1991;18:45-58
22. Kurland Ll‘, Mulder DW. Epidemiologic investigations of
amyotrophic lateral sclerosis. Neurology 1 355;5:249-268
23. Rosen DR, Brown KH Jr. Dinucleotide repeat polymorphism
in the HOX4E locus. Hum Mol Genet 1993;2:617
24. Kaplan EL, Meier 1’. Nonparamerric esrimution from incomplete observations. Am Stat Assoc J 1958;53:457-481
25. Lee ET. Statistical methods for survival data analysis. Belmont,
MA: Wadsworth, 1980
26. Peto R, Pike MC, Armitage I’, et al. Design and analysis of randomized clinical trials requiring prolonged observation of each
patient. 11. Analysis and examples. Br J Cancer 1977;35: 1-39
27. Cox DR. Regression models and life-tables. J R Stat Soc B
I972;34: 187-202
28. Lawless JE. Statistical niodels and merhods for lifetime data.
New York: Wiley, 1982
29. Hollander M, Wolfe DA. Nonparametric sratistical methods.
New York: Wiley, 1973:73
30. Henrati A, Bejaoui K, I’ericak-Vance MA, et al. Linkage of
recessive familial amyotrophic lateral sclerosis to chromosome
2q33-35. Nat Genet 1994;7:425-428
31. Stritrmatrer WJ, Saunders AM, Schrnechel 11, et al. Apolipoprotein E: high avidity binding ro fi amyloid and increased
freqwnce of type 4 Aele in late-onset familial Alzheimer disease. I’roc Natl Acad Sci USA 1093;90:1977-1981
Annals of Neurology
Vol 41
No 2
February 1397
32. Mui S, Rebeck GW, McKenna-Yasek D, et al. Apolipoprotein
E ~4 allele is not associated with earlier age at nnset in amyotrophic lateral sclerosis. Ann Nrurol 1995;38:460-463
33. Ammar A, Enayar ZE, Bakker MC, et al. Early reports: association of apolipoprorein E epsilon 4 allele with bulbar-onset motor neuron disease. Lancet 1996;347:159-160
34. Deng H-X, Hentati A, Tainer JA, et al. Amyotrophic lateral
sclerosis and structural defects in Cu,Zn superoxide dismutase.
Science 19%;261: 1047-1 051
35. Osler W . Heredity in progressive muscular atrophy as illustrated in Farr family of Vermont. Arch Med 1880;4:316-320
36. Robberecht W, Sapp P, Viaene MK, et al. C u / Z n superoxide
dismutase activity in familial and sporadic amyotrophic lateral
sclerosis. J Neurochem 1994;62:384-387
37. Takahashi H, Makifuchi T, Nakano R, et al. Familial amyotrophic lareral sclerosis with a mutation in the Cu/Zn superoxide
dismutase gene. Acra Neuropathol (Berl) 1994;88:185-188
38. Hirano M, Fujii J, Nagai Y, et al. A new variant Cu/Zn superoxide dismutase (Val 7 ro Glu) deduced fi-om lymphocyte
mRNA seqcences from Japanese patients with hmilial amyorrophic lateral sclerosis. Biochem Biophys Res Commun 1394;
39. Raiiiero I, Pinessi I,, Tauda ’r, et al. S O D l missense mutation
in an Italian family with ALS. Neurology 1934;44:347-349
40. Aoki M, Ogasawara M, Matsubara Y, et al. Familial amyorrophic lateral sclerosis (ALS) in Japan associated with H46R mutation in Cu/Zn superoxide dismutase gene: a possible new
subtype of familial ALS. J Neurol Sci 1994;126:77-83
41. Aoki M, Abe K, Houi K, er al. Variance of age at onset in a
Japanese family with amyotrophic lateral sclerosis associated
with a novel CulZn superoxide dismurase mutation. Ann Neurol 1995;37:676-679
42. Calder VL, Uoniigan N M , George PM, et al. Superoxide dismutase (glu100gly) in a family with inherited motor neuron
disease: detection of mutant superoxide dismutase activity and
the presence of heterodimers. Neurosci I.ett 1995;189:143I46
43. Kawamata J. Hasegawa H, Shimorhama S, et al. I.eul06 to
Val (CTC+GTC) mutation of superoxide dismutase-1 gene
in patients with familial amyotrophic lateral sclerosis in Japan.
Lancet 1994;343:1501
44. Orrcll IIW, King AW, Hilton DA. Familial amyotrophic lateral sclerosis with a point mutation of SOD1: intrafamilial heterogeneity of disease duration associated with neurofibrillary
tangles. J Neurol Neurosurg Psychiatry 1395;59:266-270
45. Suthers G , Laing N, Wilton S, er al. “Sporadic” motorneuron
disease due to familial SODl mutation with low penetrance.
Lancet 1994;344:1773
46. Ikeda M, Abe K, Ogasawara M, et al. A novel point mutation
in the CuiZii superoxide disniutase gene in a patient with
familial amyotrophic lateral sclero
H u m Mol Genet 1995;
4:49 1-492
47. Gurney ME, Cutting FB, Zhai l’, et al. Benefit of vitamin E,
riluzole and gahapentin in a transgenic model of familial ALS.
Ann Neurol 1996;39:147-157
48. Bensinion G, Laconiblez L, Meininger V, et al. A controlled
trial of riluzole in amyotrophic lateral tclerosis. N Etigl J Med
49, Brown RH. Superoxide dismutase in familial amyotrophic lateral sclerosis: models for gain of function. Cu1-r Opin Neurob i d 1995;5:841-846
50. Bowling AC, Schulz JB, Brown RH Jr, Beal MF. Superoxide
dismutase activiry, oxidative damage and mitochondria1 energy
metabolism in familial and sporadic amyotrophic lateral sclerosis. J Neurochem 1993;61:2322-2325
5 1. Siddique TS, Figlewicz DA, Pericak-Vance MA, et al. Linkagc
of a gene causing familial amyotrophic lateral sclerosis to chromosome 21 and evidence of genetic-locus heterogeneity. N
Engl J Med 1991;324:1381-1384
52. Cleveland DW, Laing N , Hurse PV, Brown RH. Toxic mutants in Charcor’s sclerosis. Nature 1995;378:342-343
53. Borchelt DR, Lee MK, Slunt HS, et al. Superoxide dismutase
1 with mutations linked to familial amyotrophic lateral sclerosis
possesses significant activity. Proc Natl Acad Sci USA 1994;
54. Chio A, Brignolio F, Meinere D, Schiffer D. Phenotypic and
genotypic heterogeneity of dominantly inherited amyotrophic lateral sclerosis. Acta Neurol Scand 1986;75:277282
55. Orrell RW, King AW, deBelleroche JS. Parental influence on
inheritance of familial amyotrophic lareral sclerosis. Lancet
56. Leone M. Parental sex effect in familial amyotrophic lateral
sclerosis. Neurology 1991;41:1292-1294
57. Orrell RW, de Belleroche JS, Marklund S, et al. A novel S O D
mutant and ALS. Nature 1995;374:504-505
Cudkowicz et ai: Epidemiology of SOD1 in ALS
Без категории
Размер файла
1 118 Кб
epidemiology, mutation, superoxide, dismutase, lateal, sclerosis, amyotrophic
Пожаловаться на содержимое документа