Wild-Type Nonneuronal Cells Extend Survival of SOD1 Mutant Motor Neurons in ALS Mice

See allHide authors and affiliations

Science  03 Oct 2003:
Vol. 302, Issue 5642, pp. 113-117
DOI: 10.1126/science.1086071

This article has a correction. Please see:


The most common form of amyotrophic lateral sclerosis (ALS), a neurodegenerative disease affecting adult motor neurons, is caused by dominant mutations in the ubiquitously expressed Cu-Zn superoxide dismutase (SOD1). In chimeric mice that are mixtures of normal and SOD1 mutant–expressing cells, toxicity to motor neurons is shown to require damage from mutant SOD1 acting within nonneuronal cells. Normal motor neurons in SOD1 mutant chimeras develop aspects of ALS pathology. Most important, nonneuronal cells that do not express mutant SOD1 delay degeneration and significantly extend survival of mutant-expressing motor neurons.

Amyotrophic lateral sclerosis (ALS) is a progressive neurodegenerative disorder in which motor neurons die beginning in mid–adult life. About 10% of cases are dominantly inherited; about 20% of these arise from mutations in the gene for Cu-Zn superoxide dismutase (SOD1) (1). Transgenic mice (24) and rats (5, 6) that express mutant SOD1 develop a progressive motor neuron disease that shares many features with human ALS; the complete absence of SOD1 in mice does not cause such disease (7). Because toxicity is neither accelerated nor ameliorated by reducing wild-type SOD1 activity (8) and is either unaffected (8) or enhanced (9) by increasing wild-type SOD1 activity, mutant SOD1 must cause disease through acquisition of toxic properties. These may include aberrant oxidative chemistry catalyzed by SOD1-bound copper (1014) or poisoning of a cellular process (or processes) by abundant SOD1 protein aggregates (1517). This triggers a cell death pathway in motor neurons that includes activation of caspase 3 (18, 19).

Damage to nonneuronal cells may be involved in toxicity. Before onset of disease in SOD1 mutant mice, there is an inflammatory response, including activation of microglia (2022) and astrocytosis (3, 20); the anti-inflammatory compound minocycline extends survival in mouse models of ALS (2325), although whether this reflects action on microglia, astrocytes, or more directly on motor neurons is not established (24). Although accumulation of mutant SOD1 damages motor neurons in culture (26), SOD1 mutant expression only in neurons (27, 28) or glia (29) has not provoked disease in mice. Thus, fundamental unanswered questions are whether motor neuron death is caused by toxicity of mutant SOD1 acting solely within motor neurons, whether cells expressing mutant SOD1 damage neighboring wild-type motor neurons, and whether wild-type nonneuronal cells can protect motor neurons expressing ALS-causing SOD1 mutations.

To resolve these issues, we generated chimeric animals composed of mixtures of normal cells and cells that express a human mutant SOD1 polypeptide at levels sufficient to cause fatal motor neuron disease when expressed systemically in mice. Forty-two chimeras were produced by injection of wild-type embryonic stem (ES) cells that constitutively express yellow fluorescent protein (YFP) (30) into SOD1G85R (line 148) (3) or SOD1G37R (line 42) (4) mutant blastocysts (Fig. 1, A and B). Percent chimerism (ranging from ∼5 to ∼90% wild-type cells) was determined by multiple measures (table S1), including assessment of coat color, immunoblotting of tail extracts for accumulation of mutant SOD1 and YFP (Fig. 1B), and by the proportion of cross-sectional area in spinal cords with detectable YFP (Fig. 1, C and D). Wild-type and SOD1 mutant–expressing cells contributed to multiple cell types (fig. S1).

Fig. 1.

Wild-type cells increase survival of mice expressing ALS-linked SOD1 mutations. (A) Chimeras generated by injection of E-YFP–expressing ES cells into blastocysts of SOD1G85R or SOD1G37R transgenic mice. (B) Degree of chimerism determined by immunoblotting of tail extracts from mice in (A) with antibodies having equal affinity for human and mouse SOD1 (3, 4) or to GFP. Spinal cord sections from mice with (C) low and (D) high contribution of wild-type cells identified with a GFP antibody (green). DH, Dorsal horn; VH, ventral horn. (E and F) Delayed onset of disease and extended survival of SOD1 mutant chimeras. (E) Disease onset and (F) survival times for SOD1G37R/YFP chimeras. (G) All wild-type cells or all neurons, respectively, identified in lumbar spinal cord sections of SOD1G37R/YFP chimeras identified with GFP (green) or neurofilament (red) antibodies. Ages shown are at end-stage disease. (H to K) Chimeras generated by aggregation of morulas from normal mice and mice heterozygous for SOD1G93A and lacZ transgenes. (H to J) Spinal cords from a (I) chimera, (J) wild-type mouse, and (K) germline SOD1G93A, lacZ mouse after hematoxylin-and-eosin staining and assay for lacZ (brown). Black and white arrows point to mutant and wild-type neurons, respectively. (L) Survival versus age for SOD1G93A/lacZ chimeras and germline SOD1G93A animals. (M) Survival of SOD1G93A/lacZ chimeras versus percent wild-type cells. (N and O) SOD1G93A mutant motor neurons identified in ventral horns of a SOD1G93A/lacZ chimera (N) versus a germline SOD1G93A mouse (O). Mutant cells were identified by antibodies specific for human SOD1 (green) (3, 4); neurons were identified with a neurofilament antibody (SMI32) (red). Wild-type motor neurons (red arrows); SOD1G93A mutant–expressing motor neurons (white arrows). A non-neuronal, SOD1G93A-expressing cell is marked by an asterisk.

An additional 23 chimeras (Fig. 1H) were produced by using aggregation (31) of morulae from wild-type embryos with morulae carrying transgenes for another mutant SOD1 (SOD1G93A) (2) and for ubiquitously expressed β-galactosidase (lacZ) (32). Comparable estimates were obtained for the amount of mutant SOD1 by using either coat color or immunohistochemistry of spinal cord sections to identify cells expressing lacZ. For example, animals determined to be ∼50% chimeric by coat color had a corresponding ∼50% of spinal cord areas expressing lacZ, including large spinal motor neurons (Fig. 1I). For comparison, spinal cord sections from wild-type and germline SOD1G93A mice are unstained or completely stained for lacZ (Fig. 1, J and K).

The presence of wild-type cells in the SOD1G37R/YFP and SOD1G85R/YFP chimeras delayed disease onset with average extensions of 1.6 months (P = 0.0033; n = 17; one tailed Mann-Whitney test) for SOD1G85R and 1.2 months (P = 0.0007, n = 17) for SOD1G37R chimeras (Fig. 1E). Compared with germline SOD1 mutant mice, there was an average extension of life-span of 1.8 months (P = 0.04, n = 13) for SOD1G85R and 1.1 months (P = 0.0001, n = 17) for SOD1G37R chimeras with a maximum delay of 7.8 and 3.3 months, respectively; Fig. 1F). Both measures correlated well with the proportion of wild-type (YFP-expressing) cells within each spinal cord (Fig. 1G).

A robust extension in life-span was also seen in the SOD1G93A/lacZ chimeras. Eleven of the 23, including 5 with <30% contribution from wild-type cells, survived disease-free until they were killed at >10 months of age (Fig. 1, L and M), an age at least twice that of the longest lived germline SOD1G93A littermates (Fig. 1L). Examination of spinal cord and motor roots of two of these revealed that 67% (chimera 45; Fig. 1N) and 77% (chimera 67) of motor neurons contained mutant SOD1, but there was no degeneration or axonal loss in thoracic roots of either chimera and only the earliest signs of degeneration in some lumbar roots of chimera 67. This contrasts with germline SOD1G93A animals in which 100% of the motor neurons express mutant SOD1 (Fig. 1O), and half of these are lost by 5 months (2).

Extended survival of mutant-expressing motor neurons was also seen in chimeras generated by aggregation of morulae from a SOD1G37R transgenic line (line 29, which has a later disease onset) (4) with morulae whose wild-type neurons were marked by expression of very low levels of the smallest human neurofilament subunit (hNF-L) (33) (Fig. 2A). Immunoblots for the SOD1G37R mutant and hNF-L in spinal cord extracts revealed ∼30 and ∼90% mutant cells in two SOD1G37R/hNFL chimeras (Fig. 2B). Chimera 7, with the higher wild-type content, did not develop disease even 5 months beyond the age of the longest-lived germline SOD1G37R mice. As seen with antibodies specific for human SOD1 (fig. S2), hNF-L, or all neurofilaments (to identify mutant and wild-type axons), 30% of ventral root axons (Fig. 2, C to E) were mutant. However, there was no sign of axonal degeneration or loss in the L5 ventral root (Fig. 2F), in which 978 axons remained, a number consistent with the 927 ± 99 (n = 26) seen in age-matched wild-type mice. Furthermore, in contrast to parental SOD1G37R mice (Fig. 2, H, K, and N), even the earliest pathologic signs of disease, including astrocytosis and microgliosis, were absent in this chimera (Fig. 2, I, L, and O), just as they were in normal mice (Fig. 2, G, J, and M).

Fig. 2.

Absence of motor neuron pathology or degeneration despite 30% SOD1G37R mutant–expressing motor neurons. (A) Chimeras generated by aggregation of morulas derived from hNF-L and SOD1G37R (line 29) mice. (B) Chimerism was determined by immmunoblotting spinal cord extracts with antibodies to hNF-L, actin, and human/mouse SOD1. The amount of hNF-L is a measure for the contribution of wild-type neurons. (C to F) Sections of an L5 ventral root of SOD1G37R/hNF-L chimera 7 simultaneously labeled for (C) neurofilaments (antibody SMI-32) and (D) human SOD1 (3, 4). (E) The genotype of all axons in the same root were identified as in (C and D) with antibodies specific for human SOD1 (green), hNF-L (red) (to identify wild-type axons), and myelin basic protein (blue). (F) No sign of neurodegeneration was visible in a toulidine blue–stained, semithin section of the same root. (G to O) Spinal cord sections of (G, J, and M) wild-type, (H, K, and N) SOD1G37R (line 29), and (I, L, and O) chimera 7 immunostained with antibodies against MAC-2 (G to I), glial fibrillary acidic protein (GFAP) (J to L), and cyclin-dependant kinases (Cdk1, 2, 3) (M to O), which are markers of activated microglia, astrocytes, and proliferating cells, respectively.

Extended survivals of SOD1 mutant–expressing motor neurons in the chimeras could arise, at least in part, from a protective effect of wild-type motor neurons. To test this, two SOD1G37R/YFP chimeras were identified that developed without wild-type motor neurons; all motor neurons of multiple lumbar levels (Fig. 3, B and C, arrows) and motor roots (Fig. 3E) accumulated mutant SOD1. Both of these also displayed a striking left-right asymmetry in the proportion of wild-type (YFP-positive) nonneuronal cells in the two halves of their spinal cords. Germline SOD1G37R mice at end-stage disease uniformly exhibit symmetric loss of two-thirds of their large motor neurons in both halves of the lumbar spinal cord (Fig. 3A). However, although all motor neurons were mutant in these two SOD1G37R/YFP chimeras, there was an asymmetric loss of motor neurons (Fig. 3A; P < 0.001; paired Student's t test with n > 4 sections per animal) and axons, with more than twice as many large-caliber (>3.5 μm diameter) surviving axons (Fig. 3D; 187 on the left versus 89 on the right) in the less-affected side. In both chimeras, the side with higher neuronal survival had a higher proportion (25 versus 2% in chimera 646; 30 versus 10% in chimera 213) of wild-type (YFP-expressing) nonneuronal cells throughout the lumbar cords. Thus, even when all motor neurons are mutant, an environment having a higher proportion of wild-type, nonneuronal cells reduces motor neuron mortality.

Fig. 3.

Wild-type cells that are not motor neurons extend survival of SOD1 mutant motor neurons. (A) Number of ventral horn motor neurons on both sides of the lumbar spinal cord was determined for germline SOD1G37R mice, age-matched wild-type mice, and two chimeras with asymmetric distribution of wild-type cells (n ≥ 4 sections per animal). (***P ≤ 0.001 with Student's t test for paired samples; error bars are SEM). (B) Lumbar spinal section demonstrating an asymmetric distribution of wild-type (green, GFP-immunopositive) cells. (C) Higher power view of (B). Motor neurons are mutant (blue, stained with a human-specific SOD1 antibody), not wild type (GFP-specific antibody, green). (D) Asymmetric loss of large-caliber axons seen in toluidine blue–stained semithin sections of left and right L5 ventral roots of chimera 646. (E) Triple-immunofluorescence staining for (green) wild-type YFP-expressing cells, (red) mutant human SOD1, and (blue) myelin in an L5 ventral root of chimera 646. Green staining of a wild-type Schwann cell body (left) serves as a positive control for detection of YFP-containing wild-type cells. Examination of the entire root revealed that all axons were mutant bearing (red) (fig. S2 for roots from other chimeras stained contemporaneously).

To assess whether SOD1 mutant nonneuronal cells can influence neighboring wild-type neurons, spinal cord sections of chimeric animals were analyzed at end-stage disease for pathologic signs of neurodegeneration. A hallmark for damage to neurons in human patients is the appearance of ubiquitin-positive protein aggregates (34, 35). These are also seen as an early sign for damaged neurons in SOD1G85R (3, 8) and SOD1G37R (4) mice, but do not appear in motor neurons of wild-type mice (Fig. 4, A to C). Ubiquitin aggregates appear in neuronal processes and, less prominently, in cell bodies (Fig. 4B; fig. S3). Similar ubiquitinated epitopes were never seen in age-matched wild-type littermates (Fig. 4A; fig. S3). In contrast, in both SOD1G37R and SOD1G85R chimeras (n = 4), some wild-type neurons (YFP-containing; arrows in Fig. 4, D and E, and G and H) in end-stage chimeras accumulated ubiquitinated epitopes in neuronal processes (Fig. 4F, arrows) and cell bodies (Fig. 4I, arrow), which indicates that a deficit in ubiquitin-dependent protein degradation is acquired by these wild-type neurons. The intensity of such ubiquitin staining in wild-type axons (Fig. 4F, arrows) and motor neuron cell bodies (Fig. 4I, arrow) frequently exceeded that of neurons expressing mutant SOD1 (Fig. 4F, boxed areas; Fig. 4I).

Fig. 4.

Acquisition of abnormal ubiquitination in wild-type neurons adjacent to SOD1 mutant–expressing cells. Confocal micrographs of spinal cord cross sections from the lumbar region of (A) normal (C57/B6), (B) SOD1G85R (line 148), and (C) SOD1G37R (line 42) mice after staining with (red) neurofilament (SMI32) and (blue) ubiquitin antibodies. Ubiquitin-containing aggregates in neuronal processes (arrows) are present in SOD1G85R and SOD1G37R germline animals, as well as in the cell bodies (block arrow). (D to I) Triple labeling of spinal cord sections from (D to F) SOD1G37R and (G to I) SOD1G85R chimeras stained with (D and G) SMI32 to identify neurons (red), (E and H) GFP to detect wild-type cells (green), and (F and I) ubiquitin (blue). Arrows in (D to F) point to a wild-type axon with elevated levels of ubiquitin compared with a mutant axon highlighted in the boxed region. Arrow in (G to I) points to a wild-type neuronal cell body with high ubiquitin accumulation compared with an adjacent SOD1G85R motor neuron. Additional examples are in fig. S3.

We found that expression of mutant SOD1 in motor neurons at levels that cause disease in parental mice is not sufficient to trigger their degeneration or the development of pathologic abnormalities. Rather, wild-type nonneuronal cells, in some cases representing a small minority of total cells, can ameliorate degeneration and death of SOD1 mutant–expressing motor neurons compared with those in parental SOD1 mutant mice. That SOD1 mutant neurons survive longer when surrounded by a wild-type environment supports the view that damage to adjacent nonneuronal cells by mutant SOD1 is a major contributor to disease caused by SOD1 mutations. Damaged glial cells and neurons, therefore, could act in concert to provoke disease, consistent with failure of mutant expression in single cell types to induce motor neuron degeneration (2729). It is also consistent with the failure of increased levels of mutant SOD1 within neurons to accelerate disease caused by ubiquitous expression of SOD1G93A (28). Indeed, we know of no compelling in vivo evidence that the genotype of the motor neurons themselves has any bearing on the probability of their death in ALS; motor neuron death could in principle be provoked solely by damage to multiple types of adjacent cells such as interneurons, astrocytes, and microglia. Further work is critical to evaluate this possibility.

Supporting Online Material

Materials and Methods

Figs. S1 to S4

Table S1

References and Notes

References and Notes

View Abstract

Navigate This Article