OUP user menu

Motor Neuron Disease in Mice Expressing the Wild Type-Like D90A Mutant Superoxide Dismutase-1

P. Andreas Jonsson MD, PhD, Karin S. Graffmo MD, Thomas Brännström MD, PhD, Peter Nilsson MS, Peter M. Andersen MD, PhD, Stefan L. Marklund MD, PhD
DOI: http://dx.doi.org/10.1097/01.jnen.0000248545.36046.3c 1126-1136 First published online: 1 December 2006


Mutant human CuZn-superoxide dismutases (hSOD1s) cause amyotrophic lateral sclerosis (ALS). The most common mutation is the wild type-like D90A and to explore its properties, transgenic mice were generated and compared with mice expressing wild-type hSOD1. D90A hSOD1 was both in vivo in mice and in vitro under denaturing conditions nearly as stable as the wild-type human enzyme. It appeared less toxic than other tested mutants, but mice homozygous for the transgene insertion developed a fatal motor neuron disease. In these mice, the disease progression was slow and there were bladder disturbances similar to what is found in human ALS cases homozygous for the D90A mutation. The homozygous D90A mice accumulated detergent-resistant hSOD1 aggregates in spinal cords, and abundant hSOD1 inclusions and vacuoles were seen in the ventral horns. Mice expressing wild-type hSOD1 at a comparable rate showed similar pathologic changes but less and later. Hemizygous D90A mice showed even milder alterations. At 600 days, the wild-type hSOD1 transgenic mice had lost more ventral horn neurons than hemizygous D90A mice (38% vs 31% p < 0.01). Thus, wild-type hSOD1 shows a significant neurotoxicity in the spinal cord, that is less than equal but more than half as large as that of D90A mutant enzyme.

Key Words
  • Amyotrophic lateral sclerosis (ALS)
  • D90A
  • Inclusions
  • Mice
  • SOD1
  • Stereology
  • Transgenic


Amyotrophic lateral sclerosis (ALS) is characterized by the death of motor neurons in the motor cortex, brainstem, and spinal cord, causing progressive paralysis and death. Approximately 10% of ALS cases are familial (1) and in some of these, the disease has been linked to mutations in the CuZn-superoxide dismutase (hSOD1) gene (2, 3). Overall, approximately 5% of all ALS cases show mutations in hSOD1 (4). Currently, over 100 mutations in the hSOD1 gene have been found to be associated with ALS (4). The mutations confer the enzyme a cytotoxic gain of function of unknown nature (5-7).

The Asp90Ala (D90A) hSOD1 mutation, which likely causes more ALS cases than any other hSOD1 mutation (4, 8), is unusual in several aspects. Unlike most other positions with ALS-associated mutations, D90 is not conserved (4, 9). Whereas other hSOD1 mutations cause dominantly inherited disease, ALS caused by the D90A mutation shows a unique pattern of inheritance (8, 10). In Scandinavia, the D90A mutation is associated with recessive inheritance (6). In these pedigrees, ALS has not developed in heterozygous individuals (11). Remarkably, pedigrees with a dominant inheritance have also been found as have D90A heterozygous ALS cases in populations in which the mutation is rare (4, 12, 13). In the recessive families, ALS shows a slowly progressing characteristic phenotype, whereas in heterozygous/dominant cases, the disease is, unexpectedly, more aggressive (4, 11-13). In the few ALS cases with other hSOD1 mutations so far examined, the content of mutant hSOD1 is very low in the central nervous system (14-17). In a large survey of erythrocytes, the content of mutant hSOD1 in patients with ALS varied from nondetectable up to half of that of the wild-type hSOD1 (18). In contrast, in patients homozygous for the D90A mutation, the hSOD1 activity in the central nervous system (Graffmo et al, unpublished data) as well as in erythrocytes is approximately 90% of controls (6). D90A mutant hSOD1 protein isolated from erythrocytes of patients with ALS is fully metallated and shows full specific enzymic activity. Furthermore, catalysis of hydroxyl radical formation from hydrogen peroxide is equal to that of the wild-type enzyme, and there is only a minor reduction in stability under strongly denaturing conditions (19). Thus, the D90A mutant hSOD1 appears to represent a link between other ALS-associated hSOD1s and wild-type hSOD1.

The peculiar inheritance and the similarities to wild-type hSOD1 prompted us to examine the in vivo properties and effects of the D90A mutant protein more closely. To this end, we generated transgenic mice expressing D90A mutant hSOD1 and compared these strains with mice overexpressing wild-type hSOD1.

Materials and Methods


A 0.6-kb polymerase chain reaction fragment containing HindIII-NsiI exon 4, amplified from genomic DNA of a D90A patient with ALS (primers 5′-CAC TAG CAA AAT CAA TCA TCA-3′ and 5′-TCT TAG AAT TCG CGA CTA ACA ATC-3′), was ligated into a HindIII-NsiI cleaved PvuII-PstI subclone of the hSOD1 gene (20). This subclone includes exons 2-4. Exon 5 with flanking sequences was added by ligating a PstI-BamHI fragment to it. Finally, a complete hSOD1 genomic fragment with a D90A mutation in exon 4 was completed by ligating the PvuII-BamHI to an exon 1 containing EcoRI-PvuII subclone. The 11.6-kb EcoRI-BamHI fragment was then excised from an agarose gel, electroeluted, and used for microinjection into ova from C57Bl6/CBA mice. Transgenic mice were identified by Southern blots. The mice were then backcrossed with C57Bl/6JBom mice.

Mice transgenic for the wild-type and G93A (the high-level Gur strain) mutant hSOD1 were acquired from Jackson Laboratories (Bar Harbor, ME) (7) and backcrossed into C57Bl/6JBom background. Mouse genotyping was performed on erythrocytes using an enzyme-linked immunosorbent assay (ELISA) for hSOD1 (21) or by measuring the SOD1 activity (16). Throughout the study, nontransgenic littermates were used as control mice. The mice were considered terminally ill when they no longer could reach the food in their cages. The animal care and experiments were carried out in accordance with the European Communities Council Directive (86/609/EEC).

Homogenization of Tissues

For biochemical analysis, mice at approximately 50, 100, 200, 400, 600, and 700 days (if applicable) of age and at terminal disease as defined previously were used. The mice were killed by an intraperitoneal injection of sodium pentobarbital. After dissection, pieces of tissues were frozen in liquid N2 and kept at -80°C. The tissues were homogenized in 25 volumes of 50 mM potassium phosphate, pH 7.4, 3 mM DTPA, 0.3 M KBr, and Complete with EDTA (Roche Diagnostics, Mannheim, Germany) using an Ultraturrax (IKA, Staufen, Germany) for 2 minutes followed by sonication using a Sonifier Cell Disruptor (Branson, Danbury, CT) for 1 minute.

Analysis of Detergent-Resistant Aggregates

For analysis of aggregates, tissues were homogenized in 25 volumes of phosphate-buffered saline (10 mM potassium phosphate, pH 7.0, in 0.15 M NaCl) with Complete without EDTA and 0.1% Nonidet P40 (NP40) (Roche Diagnostics). The homogenates were centrifuged at 20,000 g for 30 minutes. The supernatants were removed and the pellets were resuspended and sonicated in double the original volume of homogenizing solution. This washing of the pellets was repeated 5 times and the final pellets were analyzed by Western blotting as previously described (16).


Polyclonal rabbit antibodies were raised against keyhole limpet hemocyanin-coupled peptides corresponding to amino acids 4-20, 24-39 (human specific), 43-57, 57-72, 80-96, 100-115, and 131-153 in the hSOD1 sequence. Antibodies specific for the murine SOD1 (mSOD1) were raised against a peptide corresponding to amino acids 24-36 in the murine sequence. The antisera were then affinity purified in 2 steps as previously described (16).

Blotting Procedures

Southern, Northern, and Western blots were carried out as described previously (16). For Northern blots, at least 3 different mice were analyzed 3 times or more. For quantification of hSOD1 protein by Western blotting, wild-type hSOD1 with the concentration determined by quantitative amino acid analysis was used as an original standard (19). The hSOD1-specific antibody, 24-39, was used for the quantifications and blots were performed in duplicate at least.

SOD Analysis

The SOD activity was determined with the KO2 assay (22). One unit corresponds to 4.25 ng wild-type and D90A mutant hSOD1 (19, 23).

Stabilities of Human and Murine SOD1s

Pools of packed EDTA-anticoagulated erythrocytes from 3 homozygous D90A humans, 3 control humans, and 3 control C57Bl6 mice were mixed with 1.6 volumes of 37.5/62.5 vol/vol chloroform/ethanol at -20°C to precipitate and remove the hemoglobin. After vortexing and centrifugation (2,500 g, 10 minutes), 200 μL of the SOD1-containing upper phase was added to 400 μL 3.75 M guanidinium chloride in 0.1 M Na HEPES, pH 7.4, with 3 mM DTPA and incubated at 37°C. The denaturation was stopped after different times by addition of 50-μL aliquots of the mixture to 400 μL 50 mM Na HEPES pH 7.4 with 0.25% BSA followed by analysis of the SOD activity.

For analysis of the thermal stabilities, 100 μL of the chloroform/ethanol upper phase was added to 200 μL 0.1 M Na HEPES, pH 7.4, with 3 mM DTPA and the tube was placed in a 70°C water bath. After different times, 50-μL aliquots were added to the HEPES buffer at room temperature and analyzed as described previously.


For histopathologic investigation G93A, hemi- and homozygous D90A as well as wild-type hSOD1 transgenic mice and nontransgenic control mice were anesthetized by an intraperitoneal injection of midazolam, fentanyl, and fluanisone and killed by perfusion fixation through the heart with 4% paraformaldehyde in phosphate buffer (pH 7.6).

Animals were studied at approximately 50, 100, 200, 400, and 600 days (if applicable) of age and at terminal disease as defined previously. Five mice of each age and genotype were studied. After fixation, the central nervous system was dissected in situ and the cervicothoracic and thoracolumbar borders as well as the junction between the brainstem and spinal cord were identified by guidance of the ventral roots. The cervical, thoracic, and lumbosacral spinal cord was then dissected free, and these parts were then equally divided into 6 blocks, which were separately embedded in paraffin wax.

The uppermost block of the cervical, thoracic, and lumbosacral spinal cord as well as the second lowermost block of the lumbosacral spinal cord were used for the histopathologic investigations. Sections from each piece were stained with hematoxylin and eosin as well as with antibodies toward GFAP (Dako, Glostrup, Denmark) and the 4-20 and 131-153 anti-SOD1 peptide antibodies described previously. The 4-20 anti-SOD1 peptide antibody was used at 2 different dilutions to check for background murine SOD1. In the "Results" section, findings from the higher-dilution stainings are presented if not otherwise stated. Furthermore, the results presented in the results section correspond to findings in all 3 levels of the spinal cord as described previously if not stated otherwise. Tests were performed with the other anti-hSOD1 peptide antibodies available on one terminal mouse of each strain, which gave similar results as the 4-20 and 131-153 anti-hSOD1 peptide antibodies.

The immunohistochemistry was performed using the Ventana immunohistochemistry system using the standard protocol and was preceded by microwave irradiation of the sections in citric acid buffer for 3 × 5 minutes.


From the upper end of each of the 6 paraffin blocks containing the thoracic spinal cord of terminally ill mice or of mice approximately 600 days old, 50-μm-thick sections were cut on a sliding microtome, mounted on glass slides, and stained with cresyl violet (24). Using a Cast-Grid system (Olympus, Solna, Sweden) connected to a BX61 motorized microscope, the number of neurons in the thoracic ventral spinal cord was calculated. For the calculations, the ventral horn was defined as the area enclosed by the gray-white matter border of the spinal cord ventral to a plane at right angles to the long axis of the spinal cord and passing through the midpoint of the central canal. For each block, this area was calculated using a 10× objective. Using a 60× oil objective, the numeric density of neurons in the ventral horn was calculated using the optical dissector technique (25). Neurons were defined as cells possessing a nucleus with an identifiable nucleolus (24). The number of neurons in the thoracic ventral horn in each block was then calculated as follows: the density of neurons in that block × the area of the ventral spinal cord in that block × one sixth of the length of the entire thoracic spinal cord. The number of neurons in the thoracic ventral horn was found by adding the numbers in all 6 blocks.

Statistical Calculations

For statistical analysis, SPSS 13.0.1 software was used (SPSS, Kista, Sweden). Groups were compared with analysis of variance and, if significant, Tukey post hoc test was used to delineate significant differences.


Generation and Phenotype of D90A Transgenic Mice

Mice carrying an 11.6-kb genomic hSOD1 sequence containing the D90A mutation were generated. Twelve pups were identified as being positive for the D90A construct. The mice with the highest copy numbers (mice 134 and 154) were chosen as founders. None of the hemizygous offspring from the 2 founders developed any phenotype, even at very old age (800 days). We therefore generated mice homozygous for the transgene insertions as well as mice hemizygously expressing both the 134 and 154 insertion sites. At approximately 350 days of age, homozygous 134 mice began to show signs of weakness with abnormal extension of one of the hind limbs when lifted by the tail. Within 1 to 2 weeks, the symptoms progressed to the other hind limb and the mice became progressively paralyzed and muscle wasting was evident. During this period, the mice developed a rough coat probably because of poor grooming. After a disease duration of approximately 50 days, the mice could no longer reach the food in the cages and were therefore considered terminally ill and killed. The mean survival time was 407 ± 46 days. Homozygous 154 mice and the 134/154 crossbred mice displayed a similar picture with a somewhat later onset (Fig. 1A). All the analyses in this article were carried out on strain 134 D90A mice.


(A) Kaplan-Meier diagram showing the cumulative survival of homozygous D90A hSOD1 transgenic mice. Two strains, denoted 134 and 154, developed motor neuron disease when homozygous for the D90A transgene insertion site. Mice hemizygous for both of the insertion sites (134/154) also developed motor neuron disease. (B) Photograph of the urinary bladder of a terminal D90A homozygous mouse.

A feature not previously reported in other SOD1 transgenic mice nor seen by us (in G93AGur, G93AGurdl, G85R, G127insTGGG, and wild-type hSOD1 transgenic mice), was the common occurrence of a urinary bladder distended by urine in terminally ill homozygous D90A transgenic mice (Fig. 1B). When pressure was applied over the bladder of live mice, leakage of urine was seen. This resembles the situation in homozygous D90A patients, who frequently experience urgency to micturition and/or difficulty initiating micturition (11) and could thus be specific to D90A mutant hSOD1.

SOD1 in Transgenic Mice

The SOD1 activity levels in the central nervous system (brain, cerebellum, and spinal cord) of homozygous D90A mice were 6 to 8 times higher (between 60-90 kU/g ww) than those seen in nontransgenic mice (9-14 kU/g ww) and somewhat lower than the SOD1 activity seen in the G93A (60-140 kU/g ww) and wild-type hSOD1 (70-160 kU/g ww) transgenic mice (Table 1). The SOD1 activity in the central nervous system did not change with age in the homozygous D90A transgenic mice, whereas the wild-type hSOD1 transgenic mice showed a slight increase in SOD1 activity in older mice. The SOD1 activity in liver, kidney, and heart tissue were in all transgenic strains higher than in the central nervous system, whereas skeletal muscle had similar activity levels. The amount of hSOD1 protein was also analyzed using the hSOD1-specific 24-39 antibody (Table 1). By calculating the amounts of hSOD1 protein from the SOD1 activities and the specific activity of human hSOD1 (19, 23), it is found that only 20% to 40% of the hSOD1 protein in the central nervous system of D90A, G93A, and wild-type hSOD1 transgenic mice is enzymatically active. The proportion of active hSOD1 is much higher in liver, kidney, heart, and skeletal muscle (50-90%).

View this table:

In Western blots from the homozygous D90A mice, one band at the expected molecular weight of hSOD1 was seen. After prolonged exposure, however, a second band with higher molecular weight was also visible (Fig. 2A). This pattern was seen with all the anti-hSOD1 antibodies and is consistent with previous observations in other murine models (16, 26). The high-molecular-weight hSOD1 species was seen almost exclusively in spinal cord homogenates and the amount increased with the age of the mice. Similarly, a high-molecular-weight band was seen in spinal cord homogenates from wild-type hSOD1 and G93A transgenic mice (Fig. 2B, C).


Western blots of homogenates from different organs of transgenic and nontransgenic mice using the 24-39 hSOD1-specific antibody. The arrows indicate a high-molecular-weight species of human SOD1 occurring predominantly in spinal cords. The liver shows a crossreacting band at approximately 25 kDa. The figures to the right of the Western blots indicate the positions of molecular weight markers in kDa. (A) A terminal homozygous D90A transgenic mouse (369 days) and a nontransgenic mouse (407 days). (B) A 404-day-old wild-type hSOD1 transgenic mouse and a nontransgenic mouse (407 days). (C) A terminal G93A transgenic mouse (143 days) and a nontransgenic mouse (105 days).

In Figure 3A, the enzyme-linked immunosorbent assay analyses for hSOD1 protein in erythrocytes used for typing the transgenic mice are plotted against the age at bleeding. Whereas newborn mice have a large pool of freshly produced erythrocytes, the mean age of the cells will become higher the older the mice are. The erythrocyte survival time in mice is approximately 60 days (27). Proteins are synthesized only during erythropoiesis, but the ability to degrade misfolded proteins remains high in mature erythrocytes. The constant level of wild-type hSOD1 thus indicates a high in vivo stability. D90A mutant hSOD1 appears to be nearly as stable, whereas G93A mutant hSOD1 appears much less stable (Fig. 3A).


(A) hSOD1 content of erythrocytes measured by enzyme-linked immunosorbent assay at the time the mice were bled for genotyping. The amount of hSOD1 in erythrocytes of transgenic mice at different ages reflects the stability of the protein because newborn mice have a higher proportion of freshly synthesized SOD1 than mice at higher ages. The amount of hSOD1 has been corrected for the amount of hemoglobin. Note the logarithmic scale. (B) Stabilities of different SOD1s treated with 2.5 M guanidinium chloride (□, ○, ▵) or high temperature (70°C) (▪, •, ▴) in the presence of the chelator DTPA (2 mM). The loss of SOD enzymic activity versus time was monitored. Wild-type hSOD1 (▪, □), D90A-mutant hSOD1 (•, ○) and mSOD1 (▴, ▵). The points represent the means of 3 separate experiments, each carried out with different pooled SOD1 preparations from different individuals.

Murine SOD1 Is Structurally More Stable Than D90A and Wild-Type hSOD1

All ALS-associated mutant hSOD1s tested have been found to be less stable than the wild-type enzyme (28, 29), and the noxious effects may be linked to the structural instability of the mutants. The stabilities of D90A, wild-type hSOD1, and mSOD1 were compared by incubation in guanidinium chloride or at high temperature in the presence of the chelator DTPA (Fig. 3B). As previously found for the purified human enzymes (19), D90A mutant hSOD1 was nearly as stable as wild-type hSOD1 and both were considerably less stable than mSOD1.

The Spinal Cords of Both D90A and Wild-Type hSOD1 Transgenic Mice Accumulate Detergent-Resistant hSOD1-Containing Aggregates

In terminal mice of other mutant SOD1 transgenic models, large amounts of detergent-resistant aggregates have been demonstrated (16, 30-32). Figure 4A, B shows that there is a marked terminal accumulation of hSOD1 containing detergent-resistant aggregates of monomeric as well as high-molecular-weight species of mutant hSOD1 also in the spinal cords from homozygous D90A mice. Although the aggregates may appear abundant, they account for less than 3% of the D90A mutant hSOD1 present in the spinal cord (Fig. 4B; Table 1). In brains from the terminal mice, no such hSOD1 accumulation was seen. In hemizygous D90A mice, which do not develop an ALS phenotype during their lifespan, a minor detergent-resistant hSOD1 accumulation in spinal cords was seen in aging 700-day-old mice (Fig. 4C, D). Remarkably, in terms of accumulation of aggregates, the wild-type hSOD1 transgenic mice fell in between the homozygous and hemizygous D90A transgenic mice (Fig. 4B, D, F).


Time course of the occurrence of SOD1 in detergent-resistant aggregates in the spinal cords of D90A and wild-type hSOD1 transgenic mice. (A, C, E) Western blot patterns of SOD1 in aggregates isolated from equal amounts of tissue (wet weight). The upper blot shows hSOD1 (human-specific 24-39 antibody) and the arrow indicates the high-molecular-weight species also indicated in Figure 2. For the lower blot, the murine-specific 24-36 antibody was used, and the arrowhead indicates the mSOD1 band. The figures to the right of the Western blots indicate the positions of molecular weight markers in kDa. (B, D, F) Quantifications of hSOD1 in aggregates from spinal cords expressed as amount per tissue gram wet weight. For comparison, brains from the oldest mice in each group were included in the Western blots and the quantifications.

The murine SOD1 is structurally more stable (Fig. 3B) and it was therefore of interest to compare the amounts of mSOD1 with those of hSOD1s in the detergent-resistant aggregates. Separate Western blots of the detergent-resistant aggregates were stained with the mSOD1-specific antibody. Notably, there was much less age-related accumulation of detergent-resistant mSOD1 aggregates in the spinal cords (Fig. 4A, C, E). At 600 days, like at earlier ages, the detergent-resistant part accounted for approximately 0.06% of the mSOD1 in the spinal cord homogenate, whereas at 600 days, the corresponding figure for wild-type hSOD1 was 0.3% (Fig. 4E).

Histopathology of Spinal Cords from D90A and Wild-Type hSOD1 Transgenic Mice

Ventral horn neurons were counted in the thoracic spinal cord, because that is the part that most precisely can be delineated. All neurons were counted to circumvent bias resulting from cell size or phenotype changes. All transgenic strains showed a considerable cell loss at around 600 days of age or at the terminal stage (Table 2, p < 0.001 for all transgenic strains compared with nontransgenic mice). Compared with 700-day-old nontransgenic control mice, the terminal homozygous D90A mice had lost 40% of their ventral horn neurons and terminal G93A mice had lost 45%. At 600 days of age, wild-type hSOD1 transgenic mice had lost more neurons than hemizygous D90A mice (38% vs 31%, p < 0.01) The neuron loss in terminal homozygous D90A mice was larger than in 600-day-old hemizygous D90A mice (p < 0.01), but the difference versus 600-day-old wild-type hSOD1 and terminal G93A transgenic mice was not significant. Terminal G93A mice had lost more neurons than both hemizygous D90A mice (p < 0.001) and wild-type hSOD1 transgenic mice (p < 0.01). Motor neurons with pyknotic nuclei were seen in all strains but in highest numbers in the homozygous D90A mice. These were seen already in 100-day-old mice and were not overtly related to cell loss.

View this table:

Diffuse SOD1 staining of motor neuron somata was seen in all 3 strains (homozygous D90A, hemizygous D90A, and wild-type hSOD1 transgenic mice) at all ages studied and increased with time (Fig. 5A-D, F-J, K-O). This staining was also seen in the cells of Clarke's nucleus. No corresponding staining was seen in other cells of the spinal cord even in the sections stained with the lower dilution of the 4-20 anti-SOD1 peptide antibody and no SOD1 staining was seen in the nontransgenic mice (Fig. 5E).


Photomicrographs of immunohistochemistry sections from spinal cord ventral horns from the various transgenic strains at ages as indicated. The sections were stained with the 4-20 anti-SOD1 peptide antibody. A section from a 600-day-old nontransgenic mouse stained with the antibody was included for comparison (E). Scale bars = 60 μm.

In the homozygous D90A mice, a progressive astrogliosis in the ventral neuropil was seen (Fig. 6D), which could be discerned already at 50 days. Large SOD1-staining inclusions were seen in the ventral neuropil as early as 100 days of age (Fig. 5B), and these increased in number and size with age (Fig. 5C, D). Throughout the ventral neuropil, vacuoles were seen from 200 days of age, increasing in size and numbers at later time points (Fig. 5C, D). The motor neuron SOD1 staining was localized into large inclusions at 200 days of age (Fig. 5C) and these were pronounced in terminal homozygous D90A mice (Fig. 5D). At 100 days and later, vacuolization and rim SOD1 immunopositivity were seen in the ventral funiculi and roots (not shown). This suggests that at least some of the vacuoles seen in the ventral neuropil were derived from swollen motor axons. In contrast to the findings in D90A hemizygous mice (Fig. 6B), SOD1 inclusions were also seen in dorsal roots, indicating that such alterations also occur in peripheral neurons (Fig. 6A).


Photomicrographs of sections from spinal cord dorsal roots and ventral horns from the various transgenic strains at ages in days as indicated. The sections were stained with the 4-20 anti-SOD1 peptide antibody or the anti-GFAP antibody. Scale bars = 60 μm.

In the hemizygous D90A mice, a progressive astrogliosis was seen in the ventral neuropil from 400 days on (Fig. 6E). Some defined and inspissated somal SOD1 staining in motor neurons was seen at 100 days and increased with age (Fig. 5F-J). Vacuoles in the ventral neuropil were seen at 400 days (Fig. 5I) and these were increased in number and size at 600 days (Fig. 5J) but always less pronounced than those seen in the homozygous D90A mice. The vacuoles had a rim of SOD1 immunopositivity (Fig. 5I, J). The intensity of the SOD1 staining was higher in the neuropil than in the motor neuron somata at 400 and 600 days (Fig. 5I, J). From 200 days on, progressive vacuolization and rim SOD1 immunoreactivity were also seen in the ventral roots and ventral funiculi but not in the dorsal roots (Fig. 6B).

In the wild-type hSOD1 transgenic mice, the alterations seen were intermediate between those in the hemizygous and homozygous D90A mice. Astrogliosis was seen in the ventral neuropil and ventral roots where it was more pronounced than in the hemizygous D90A mice (Fig. 6F). SOD1-specific neuropil immunostaining was very scarce at 100 days of age and steadily increased in intensity and size thereafter (Fig. 5K-O). At 600 days, the neuropil staining exceeded that seen for the hemizygous D90A mice. Furthermore, the vacuolization seen in the ventral neuropil was established at 100 days of age and exceeded that seen in the hemizygous D90A mice at 400 days on (Fig. 5L-O). The background intensity of SOD1-specific motor neuron somal staining was similar to that seen for the hemizygous D90A mice, but more localized staining could be seen. The changes in ventral funiculi and roots were somewhat more pronounced than in the hemizygous D90A mice.

In the sacral spinal cord in terminal homozygous D90A mice, extensive vacuolization and cell loss was seen in the ventral neuropil exceeding the corresponding changes seen at cervical, thoracic, and lumbar levels. Because these findings might explain the finding of distended bladders in the terminal homozygous D90A mice (Fig. 1B), terminal G93AGur mice (which do not show distended bladders) were similarly examined. The same extensive vacuolization and cell loss in the sacral spinal cord was seen in these mice. Thus, we were unable to find a specific histopathologic correlate to the bladder disturbances in the homozygous D90A mice. The sacral spinal cord in 600-day-old hemizygous D90A and wild-type hSOD1 transgenic mice showed similar changes with SOD1-positive inclusions extensively distributed throughout the ventral neuropil and some vacuolization in the ventral roots and ventral neuropil.


D90A-Induced Motor Neuron Disease in Mice and Humans Show Similarities

The noxious activity of mutant SOD1s shows a distinct gene dosage effect. In 2 lines of G127insTGGG transgenic mice, the survival times are nearly twice as long in hemizygous mice as in mice homozygous for the insertions (16). The G93AGurdl mice, which we find to have mRNA levels 50% of that of G93AGur mice, survive in our laboratory 253 days as compared with 124 days (16). When comparing various transgenic mouse models, it is thus important to relate phenotypes with gene expression (mRNA) levels. Compared with G93AGur mice, the hSOD1 mRNA levels in G127insTGGG, G85R, and homozygous D90A mice are 63%, 43%, and 51% and their lifespans are 250, 345, and 406 days, respectively (33). Therefore, among the murine ALS models (G93A, G85R, G127insTGGG, and D90A) that have been analyzed under identical conditions, the D90A mutant hSOD1 appears least neurotoxic. This is in accord with the fact that ALS caused by the D90A mutation as a rule shows recessive inheritance. The homozygous D90A transgenic mice also showed a slow disease progression and bladder disturbances, which further mirrors the phenotype of human ALS associated with homozygosity for the D90A hSOD1 mutation (11). ALS cases heterozygous for the D90A mutation lack bladder disturbances and display a more aggressive disease phenotype similar to those seen with most other hSOD1 mutations. The present data in the mouse suggest that the conundrum of different inheritance patterns of D90A-associated ALS might be explained by additional susceptibility factors in dominant cases rather than protective factors inherited along with the D90A locus in recessive pedigrees (34).

The wild-type hSOD1, which in the transgenic mice have an mRNA level of 60% of G93AGur, is obviously less toxic than D90A and the other mutant SOD1s. However, given the distinct gene-dosage relationships, comparison with homo- and hemizygous D90A mice could yield information on the degree of cytotoxicity.

Only Minor Differences in Turnover and Biochemical Properties Between D90A and Wild-Type hSOD1

Assuming that mSOD1 shows a specific enzymic activity similar to that of hSOD1, the D90A hSOD1 protein levels at 100 days in the spinal cord were 20-fold higher than the background mSOD1 levels (Table 1). The wild-type hSOD1, which according to the mRNA levels, is expressed at a rate similar to that of D90A mutant hSOD1 in homozygous D90A mice, was approximately 25-fold higher than the mSOD1. In both cases, the major parts of the hSOD1s were enzymatically inactive. The small differences in these steady-state central nervous system levels and also the analysis of erythrocytes of young mice (Fig. 3A) indicate that D90A mutant hSOD1 is in vivo nearly as stable as the wild-type hSOD1. There was also only a small difference under denaturing conditions in vitro (Fig. 3B). Notably, compared with the mSOD1, the hSOD1 variants were similar and much less structurally stable.

Pathologic Alterations in D90A Mutant and Wild-Type hSOD1 Transgenic Mice

Although the nature of the noxious property of (mutant) hSOD1 has not been elucidated, the preponderance of evidence suggests that instability, misfolding, and possibly aggregation of the hSOD1 protein is involved (16, 26, 28, 30, 32, 35, 36). Abundant aggregates of mutant SOD1 were also seen in the spinal cord ventral horns of a patient with ALS carrying the G127insTGGG mutation (16). In homozygous D90A mice, we found a distinct terminal accumulation of detergent-resistant hSOD1 aggregates in spinal cords (Fig. 4A, B). Notably, the mice expressing wild-type hSOD1 also showed a late accumulation of detergent-resistant aggregates, but these did not reach the same levels within the short lifespan of the mice (Fig. 4E, F). Previous studies have failed to show detergent-resistant aggregates in wild-type hSOD1 transgenic mice apparently because only younger mice were examined (30-32). To explore the significance of this disease hallmark in the wild-type hSOD1 transgenic mice, hemizygous D90A mice were examined. These mice also showed accumulation of hSOD1 aggregates (Fig. 4C, D) but later and in lower amounts.

The histopathologic examination showed significant loss of ventral horn neurons in all the transgenic strains earliest and most pronounced in the homozygous D90A mice followed by the wild-type hSOD1 and the hemizygous D90A mice (Table 2). There were also inclusions of SOD1 in motor neurons, neuropil, and axons. Astrogliosis, pyknosis, and vacuolization were also seen. These changes were again clearly most pronounced in the homozygous D90A mice followed by the wild-type hSOD1 transgenic mice and the hemizygous D90A mice. The histopathologic picture of the D90A mice is reminiscent of those previously reported for G93A (37-39) and G37R mice (38, 40), and the changes here found in wild-type hSOD1 transgenic mice are similar to those previously reported (41).

Could Wild-Type hSOD1 Cause Amyotrophic Lateral Sclerosis?

Overall, approximately 5% of all ALS cases show mutations in hSOD1 (4). A vital question is whether wild-type hSOD1 plays a pathogenic role in some of the remainder. In several other neurodegenerative conditions such as Alzheimer, Parkinson, and Creutzfeldt-Jakob diseases, the proteins that are mutated in some of the familial cases are also involved in the disease process of sporadic cases lacking such mutations. Mice expressing wild-type hSOD1 have previously been found to develop impaired motor performance as tested on rotating rods combined with a loss of lumbar motor neurons at high age (41). Moreover, coexpression of wild-type hSOD1 has been found to exacerbate disease in transgenic mice expressing mutant hSOD1s (41, 42). We here show that the structural stabilities of wild-type hSOD1 and the D90A mutant are similar both in vitro and in vivo. Compared with SOD1 from another species, the mouse, both appear relatively unstable in vitro and both show much higher propensities to form aggregates in the compromised spinal cord (Fig. 4A, C, E). We also show that the wild-type hSOD1 has a definite neurotoxic effect in murine spinal cords, that is less than equal but more than half of that of the D90A mutant enzyme (Table 2). Notably, the existence of several ALS-linked C-terminal truncation mutations shows that the sequence elements required for ALS provocation are present in the wild-type hSOD1 (4). All these data provide circumstantial evidence that the wild-type hSOD1 also has the potential to participate in the pathogenesis of ALS.

D90A families with both recessive and dominant inheritance exist (4, 6, 11-13). Furthermore, the disease phenotype can be very variable within families with SOD1 mutations and several mutations show a partial penetrance (4). This indicates that there are genetically determined as well as environmental and lifestyle factors that influence the susceptibility to the noxious effects of mutant hSOD1s. Such putative factors may conceivably also induce the ubiquitously and highly expressed wild-type hSOD1 to occasionally cause ALS.


The authors thank Eva Bern, Ingalis Fransson, Ulla-Stina Spetz, Karin Hjertkvist, Ann-Charlott Nilsson, Karin Wallgren, Jörgen Andersson, and Agneta Öberg for expert technical assistance.


  • This study was supported by the Swedish Science Council, the Swedish Brain Fund/Hållsten Foundation, the Swedish Medical Society including Björklunds Fund for ALS Research, the Swedish Association of Persons with Neurological Disabilities, the Kempe Foundations, King Gustaf and Queen Victoria Foundation, the Umeå University Foundation for Medical Research, and the Västerbotten County Council, Sweden.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
View Abstract