Human Molecular Genetics Advance Access originally published online on July 6, 2005
Human Molecular Genetics 2005 14(16):2335-2347; doi:10.1093/hmg/ddi236
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Somatodendritic accumulation of misfolded SOD1-L126Z in motor neurons mediates degeneration: 
B-crystallin modulates aggregation
1Department of Pathology and 2Department of Neuroscience, The Johns Hopkins University School of Medicine, 720 Rutland Avenue, Room 558, Baltimore, MD 21205, USA and 3Mouse Cancer Genetics Program, NCI-Frederick Cancer Research and Development Center, Frederick, MD 21702, USA
* To whom correspondence should be addressed at: Department of Neuroscience, College of Medicine, McKnight Brain Institute of the University of Florida, University of Florida, 100 Newell Drive, Room L1-100H, PO Box 100244, Gainesville, FL 32610-0244, USA. Tel: +1 3522940105; Fax: +1 3523928347; Email: borchelt{at}mbi.ufl.edu
Received April 6, 2005; Revised June 16, 2005; Accepted June 28, 2005
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONCONCLUSIONSMATERIALS AND METHODSSUPPLEMENTARY MATERIALSREFERENCES |
|---|
Mice expressing variants of superoxide dismutase-1 (SOD1) encoding C-terminal truncation mutations linked to familial amyotrophic lateral sclerosis (FALS) have begun to define the role of misfolding and aggregation in the pathogenesis of disease. Here, we examine transgenic mice expressing SOD1-L126Z (Z=stop-truncation of last 28 amino acids), finding that detergent-insoluble mutant protein specifically accumulates in somatodendritic compartments. Soluble forms of the SOD1-L126Z were virtually undetectable in spinal cord at any age and the levels of accumulated protein directly correlated with disease symptoms. Neither soluble nor insoluble forms of SOD1-L126Z were transported to distal axons. In vitro, small heat shock protein (Hsp)
B-crystallin suppressed the in vitro aggregation of SOD1-L126Z. In vivo, B-crystallin immunoreactivity was most abundant in oligodendrocytes and up-regulated in astrocytes of symptomatic mice; neither of these cell-types accumulated mutant SOD1 immunoreactivity. These results suggest that damage to motor neuron cell bodies and dendrites within the spinal cord can be sufficient to induce motor neuron disease and that the activities of chaperones may modulate the cellular specificity of mutant SOD1 accumulation. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONCONCLUSIONSMATERIALS AND METHODSSUPPLEMENTARY MATERIALSREFERENCES |
|---|
A large number of dominantly inherited mutations in Cu/Zn-superoxide dismutase-1 (SOD1) cause familial amyotrophic lateral sclerosis (FALS). Over 100 different missense substitutions distributed at more than 60 residues throughout the protein (153 amino acids) have been described (www.alsod.org). A small number of mutations lead to early translation termination in the last of the five exons. SOD1 is one of the principle anti-oxidant enzymes of the cytoplasm, but it does not appear that partial loss of this enzymatic activity plays a significant role in disease pathogenesis (1
–3). Instead, data from transgenic and knockout mice argue that the mutant protein gains a property that is particularly toxic to motor neurons (1,4–6). Indeed, some FALS mutations induce little change in the structure or activity of the protein; these mutants (termed wild-type-like mutants) retain superoxide dismutase activity (G37R and G93A) (1,4) and most properties of wild-type enzyme (7,8).
Several theories, which are not entirely mutually exclusive, have been put forward to explain the disease. One theory, termed ‘the Cu hypothesis’, states that damage to neurons is mediated by the Cu cofactor of mutant enzyme, which can react with cellular ions to produce radical species and/or nitrate proteins (9–15). Wild-type-like mutants have been reported to engage in Cu-mediated toxic reactions that produce hydroxyl-like radicals (9,12,16) or to catalyze the covalent nitration of tyrosine residues (14,15). However, we have created experimental mutants that lack critical Cu-binding histidine residues and dismutase activity [H46R/H48Q and H46R/H48Q/H63G/H120G (or Quad)], which when expressed in transgenic mice induce typical motor neuron disease (17,18). Moreover, eliminating the Cu chaperone for SOD1, by deletion mutation, lowers the loading of Cu into wild-type-like mutant enzymes (G37R and G93A) without proportionately diminishing toxicity (19). Together, these studies provide in vivo evidence that motor neuron toxicity does not require Cu to be correctly bound in the active site of mutant SOD1 (reviewed in 8).
Other recent theories involve an activity of mutant SOD1 upon some aspect of mitochondrial function. Liu et al. (20) recently reported that mutant SOD1 was selectively associated with mitochondria in spinal cords of FALS mice. Importantly, over-expressed wild-type human SOD1 was not associated with mitochondria and the association of mutant SOD1 with mitochondria was not found in tissues not directly affected by the disease (muscle and liver) (20). However, a recent paper by Vijayvergiya (21) reported that mitochondria from brain and liver contain both wild-type (over-expressed) and mutant human SOD1 in both the matrix and outer membrane space of mitochondria. The basis for this discrepancy is not currently known but these latter data suggest that an association of mutant SOD1 with mitochondria may not fully explain why SOD1 is specifically toxic to motor neurons.
The present study focuses on the notion that abnormal aggregation of SOD1 is a key mechanism of toxicity (6,17,18,22,23). In previous work, we have shown that the two main classes of FALS-SOD1 variants (metal-binding region and wild-type-like) are prone to aggregate (17,18,23). Here, we focus on mice that express the FALS variant, SOD1-L126Z, which is a simple C-terminal deletion mutant (24). Recent studies have described two other C-terminal deletion mutants that arise by insertion or deletion mutations that shift reading frames and add non-native sequence prior to premature termination (25,26). Mice expressing these truncation mutants, including the L126Z mutant we describe here, develop motor neuron disease. Similar to what has been described in mice expressing other truncation mutants, the steady-state levels of SOD1-L126Z protein in young asymptomatic mice were much lower than endogenous SOD1 protein. However, spinal cords of paralyzed mice accumulated relatively high levels of detergent-insoluble forms of the L126Z variant. Indeed, spinal cords of terminally ill mice expressing SOD1-L126Z protein and SOD1-G37R contained similar levels of insoluble mutant SOD1. The SOD1-L126Z protein appeared to be so prone to misfold that detergent-soluble species were virtually undetectable in spinal cord at any age. Histologically, SOD1-L126Z immunoreactivity accumulated, as animals aged, in somatodendritic compartments of spinal motor neurons. Importantly, for consideration of mechanisms of toxicity, we failed to find evidence that SOD1-L126Z was transported to distal nerve fibers, indicating that direct damage to axons by mutant protein is not a required mechanism of toxicity. In previous studies, we have noted the spinal cords of symptomatic FALS mice contain increased levels of insoluble B-crystallin (17). Here, we show that B-crystallin, normally absent in motor neurons and expressed in oligodendrocytes, was up-regulated in astrocytes in symptomatic SOD1-L126Z mice. In vitro, B-crystallin was found to prevent the conversion of detergent-soluble SOD1-L126Z protein in brain extract into detergent insoluble species. Collectively, these data demonstrate that the natural FALS mutation L126Z creates a protein that accumulates only as an insoluble, sedimentable, structure in spinal cord tissue and which specifically accumulates in somatodendritic compartments of neurons. The absence of mutant SOD1 accumulation in other cells of the spinal cord, such as oligodendrocytes and astrocytes, correlates with the expression pattern of B-crystallin. We find no evidence for direct damage of axonal compartments by the mutant protein. The accumulation of detergent insoluble protein in somatodendritic compartments of spinal motor neurons appears to be sufficient to induce degeneration.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONCONCLUSIONSMATERIALS AND METHODSSUPPLEMENTARY MATERIALSREFERENCES |
|---|
The truncation mutant SOD1-L126Z induces motor neuron disease
Subsets of FALS mutations, including rare frameshift mutations, introduce early stop codons in the last coding exon to generate truncated variants. The FALS mutation L126Z is a simple stop mutation introduced at the codon for leucine 126 [(24
)also see www.alsod.org]. Three lines of the transgenic mice expressing SOD1-L126Z were established, which developed stereotypical ALS-like disease as early as seven months of age (Fig. 1). In young, asymptomatic animals, the levels of SOD1-L126Z mRNA were comparable to those of other previously established SOD1 models of FALS, but the steady-state levels of SOD1-L126Z protein were much lower, suggesting rapid degradation of the mutant protein (Fig. 1A). Similar findings were reported in mice that express the truncation mutant G127insTGGG (truncated to 133 with the mutations G127W, K128Q, G129R, G130W, N131K, E132Z) (25). Moreover, CNS tissues from humans with either the L126Z or L126insTGGG have been reported to contain low levels of mutant protein when compared with protein derived from the normal allele (24,25). Despite the relatively low level of protein in pre-symptomatic states, survival was not extraordinarily different from that of other FALS models expressing similar levels of transgene mRNA. For example, the age to paralysis in lines of mice expressing the L126Z variant was similar to that of mice (line 29 with similar levels of mRNA) expressing the far more stable G37R variant of SOD1 [age of onset in line 29 mice is 9–11 months (4)] (Fig. 1B).
|
Histological and immunochemical studies of symptomatic SOD1-L126Z mice demonstrated pathology similar to that described for other mutant SOD1 models. Abnormalities included occasional circular eosinophilic inclusions (Fig. 2A, inset), massive loss of motor neurons with loss of fiber tracts (compare Fig. 2C with D), microglia activation (Fig. 2E) and astrogliosis (Fig. 2F). Spinal cords of symptomatic SOD1-L126Z mice also accumulated ubiquitin immunoreactive inclusions (Fig. 2G), similar to what we have previously described in other FALS mice (18
). However, unlike other FALS mice (18), we did not find thioflavin-S positive inclusions in brain stem or spinal cords of symptomatic SOD1-L126Z mice (Fig. 2I compare to J—SOD1-H46R/H48Q mouse).
|
Accumulation of non-ubiquitinated detergent-insoluble SOD1-L126Z in spinal cords
In previous work, we have developed a protocol for detergent extraction that involves sequential solubilization and centrifugation in buffers containing non-ionic detergent (first pellet=P2 fraction), followed by buffers containing SDS (pellet=P3 fraction) (17
). In spinal cord of either pre- or post-symptomatic mice, very little detergent soluble SOD1-L126Z protein was detected (in the first supernatant from non-ionic detergent) (Fig. 3A). Even in young animals, SOD1-L126Z protein from spinal cord was found in the detergent-insoluble P3 fraction. To illustrate the tight association between insoluble forms of SOD1-L126Z and disease, we examined the levels of mutant protein in the detergent insoluble fraction (non-ionic P2 fraction) of spinal cords from two of our lines of mice at different ages (Fig. 3B). In the P2 fraction, which is the first detergent-insoluble fraction in sequential extraction, only samples from symptomatic animals contain mutant protein. As shown in A, mutant protein was undetectable in soluble fractions at any age. As previously described in other mouse models of FALS (17,18,23), forebrain from symptomatic animals lacked detectable insoluble (or soluble—not shown) mutant protein. Note the considerable accumulation of SDS-insoluble SOD1-L126Z protein in older symptomatic mice; the levels approximate the levels of insoluble SOD1-G37R found in spinal cords of symptomatic G37R mice (Fig. 3A). Similar data have been reported for mice that express the truncation mutant G127insTGGG (25).
|
Interestingly, in comparison with SOD1-G37R mice, the insoluble fractions of spinal cords from SOD1-L126Z mice accumulate much less SDS/ßME-resistant species of high-molecular-weight mutant protein (Fig. 3A, two right lanes). We have previously shown that some of the high-molecular-weight species of SOD1 that accumulate in the detergent-insoluble fraction of diseased spinal cords from G37R, G85R, G93A, H46R/H48Q and H46R/H48Q/H63G/H120G mice are ubiquitinated. For example, two major ubiquitin immunoreactive bands at 28 and 36 kDa in the G37R mice contain both SOD1 and ubiquitin (Fig. 3C) (17
). In contrast, the SOD1-L126Z that accumulates does not appear to be significantly ubiquitinated (Fig. 3C). Thus, in the SOD1-L126Z mice, we find biochemical evidence for the accumulation of detergent insoluble forms of mutant SOD1, which increase in abundance as symptoms become severe. The accumulated mutant SOD1 does not appear to be heavily ubiquitinated or organized into thioflavin-S positive structures. The origin of the ubiquitin immunoreactivity visible under light microscopy is unclear (Fig. 2G), but it does not appear to be ubiquitinated mutant SOD1. Importantly, we find that very little of SOD1-L126Z protein in spinal cord is in a detergent soluble state.
Aggregates of SOD1-L126Z are localized to motor neuron somatodendritic compartments
Our biochemical studies described earlier demonstrate that in asymptomatic mice, the steady-state levels of SOD1-L126Z are very low, much lower than endogenous SOD1, and that as the animals develop symptoms, the protein accumulates to significant levels. The low abundance of this variant in young animals presents a unique opportunity to visualize the cellular location of the protein as it accumulates over the disease course. In younger pre-symptomatic animals, immunostaining with a human SOD1-specific peptide anti-serum demonstrated that SOD1-L126Z immunoreactivity accumulates initially in the cell bodies of neurons (Fig. 4B and C). Faint, but discernible staining of motor neurons was evident throughout the ventral horn of the spinal cord (lumbar). In older symptomatic mice, staining of neuronal soma intensified in remaining motor neurons, with staining in dendritic profiles also increasing (Fig. 4F). Cross-sections of dendrites create small circular profiles of immunoreactivity (Fig. 4E). In symptomatic mice, structures that appear to be motor neuron corpses, filled with SOD1 immunoreactivity, are visible (Fig. 4F). To better define whether mutant SOD1 was localized in astrocytes, free-floating frozen sections of spinal cord from symptomatic mice were immunostained with GFAP and hSOD1 antibodies, using green (GFAP) or red (SOD1) fluorescently labeled secondary antibodies (Fig. 5). The patterns of GFAP and hSOD1 immunostaining show little similarity (compare Fig. 5A and B) and although GFAP positive astrocytes can be seen in proximity to SOD1 stained structures (Fig. 5C), we do not find evidence that SOD1 is located within astrocytes. In brain, there was no evidence of accumulation of SOD1-L126Z immunoreactivity in neurons of the forebrain (Supplemental Fig. 1). These data indicate that SOD1-L126Z accumulates primarily in motor neuron cell bodies and dendrites of spinal cord. Notably, absent are well-defined inclusion-like structures, such as Lewy-body-like structures reported in the G85R-FALS model (27) or in mice expressing another truncation mutant L126delTT (26). Instead, the reactivity appears to fill neuronal compartments. A similar pattern of immunoreactivity was reported in mice expressing the truncation mutant L126insTGGG (25).
|
|
To determine whether the SOD1-L126Z protein is also present in axons, we analyzed segments of sciatic nerve. In previous studies, we have demonstrated that SOD1-G37R is anterogradely transported down the sciatic nerve in slow component B (28
). Given the slowness of the anterograde transport process and the short half-life of SOD1-L126Z, we considered the possibility that SOD1-L126Z may not be present at appreciable levels in the axonal compartment. Sciatic nerve segments were dissected and fractionated into detergent-soluble and -insoluble materials. SOD1-L126Z was not found in either the soluble or insoluble fraction of sciatic nerve (Fig. 6). In the detergent soluble fraction, SOD1 levels were consistent with levels in total spinal cord extracts (WT>H46R/H48Q>G37R>L126Z). In the insoluble fraction, there was a marked accumulation of insoluble SOD1-H46R/H48Q, but the G37R variant was not more abundant than WT SOD1, and the L126Z variant was undetectable. This result clearly shows that SOD1-L126Z is not transported to distal axons and hence the mechanism by which this mutant damages motor neurons does not involve a direct disruption of axonal processes. Similarly, the accumulation of detergent-insoluble SOD1-G37R in nerve is also not a pronounced feature of disease for this variant.
|
Small Hsp’s are up-regulated and able to suppress SOD1 aggregation
We have previously demonstrated that Hsp25 and
B-crystallin are specifically up-regulated in spinal cord tissues from symptomatic G37R, G85R, G93A, H46R/H48Q and Quad mice (17). We investigated whether similar up-regulation occurs in the SOD1-L126Z model, which as described earlier does not form thioflavin-S staining inclusions. Histologically, antibodies to Hsp25 robustly stained the cell bodies and dendrites of motor neurons of the ventral lumbar spinal cord in young or old non-transgenic animals (Fig. 7A and 7B) and in young SOD1-L126Z transgenic animals (Fig. 7C). However, in older, symptomatic SOD1-L126Z mice, the neuronal pattern of staining was replaced by a more diffuse neuropil pattern, and remaining motor neurons were much less intensely stained (relative to neuropil) (Fig. 7D, inset).
|
Immunostaining of spinal cord sections with antibodies to
B-crystallin specifically stained cells resembling oligodendrocytes in both young or old non-transgenic animals (Fig. 7E and 7F) and in younger SOD1-L126Z transgenic animals (Fig. 7G). In older, symptomatic SOD1-L126Z mice, the selective staining of oligodendrocytes was less apparent with a subset of astrocytes demonstrating intense immunoreactivity (Fig. 7H, inset). General neuropil staining for B-crystallin was also increased (Fig. 7H), which we interpret as probably astrocytic in origin because of the appearance of some immunoreactive astrocyte profiles and the lack of staining in motor neuron cell bodies of either young or older SOD1-L126Z mice (Fig. 7H). Biochemical analyses of spinal cords from symptomatic mice demonstrated increased levels of Hsp25 in both the detergent-soluble and insoluble fractions, but only moderate increases in B-crystallin were noted (Fig. 7I).
The small Hsp B-crystallin is believed to bind to misfolded proteins and inhibit aggregation (reviewed in 29,30). To study whether B-crystallin could modulate SOD1 aggregation in ALS, we performed cell-free aggregation assays (see Materials and Methods) in the presence or absence of purified B-crystallin (Fig. 7J and K). As a source of soluble SOD1-L126Z, we utilized high-speed supernatant fractions of homogenized forebrain from mutant mice. Although the steady-state levels of SOD1-L126Z are low in brain (Supplemental Fig. 2), all detectable protein is initially soluble and remains soluble throughout the course of disease (See Fig. 3B). Incubation of these clarified brain lysates at 37°C resulted in the conversion of a fraction of the soluble protein to insoluble protein (Fig. 7J). The addition of B-crystallin to cell-free lysates significantly reduced the aggregation of SOD1-L126Z (Fig. 5J and K). This suppression was completely neutralized by an anti-B-crystallin monoclonal antibody (Fig. 7J and K). These data suggest that cells that express B-crystallin may be able to attenuate aggregation of SOD1-L126Z.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONCONCLUSIONSMATERIALS AND METHODSSUPPLEMENTARY MATERIALSREFERENCES |
|---|
In the present study, we establish linkage between motor neuron disease and aggregation of mutant SOD1. We found that a very unstable FALS variant of SOD1 (SOD1-L126Z) induced rapidly progressing motor neuron disease in transgenic mice. In spinal cord of both young and old mice, soluble forms of SOD1-L126Z were virtually undetectable. In contrast, detergent-insoluble forms of SOD1-L126Z accumulated to high levels as disease symptoms worsened. In this regard, the SOD1-L126Z mice are very similar to recently described mice that express SOD1-G127insTGGG (25
). Here, we demonstrate for the first time that SOD1-L126Z protein is not transported to distal nerve fibers in either soluble or insoluble states and that accumulating mutant protein appeared to be primarily localized in neuronal cell bodies and dendrites. The specific somatodendritic accumulation of mutant SOD1-L126Z suggests that direct damage to these neuronal compartments may be sufficient to induce disease.
We also demonstrate for the first time that the availability of Hsp’s may influence the cellular targets of mutant SOD1 aggregation. In vitro, the Hsp B-crystallin, which was expressed in oligodendrocytes and up-regulated in some astrocytes, was able to slow the rate of SOD1-L126Z aggregation. However, motor neurons do not express this protein. Motor neuron cell bodies of younger transgenic and non-transgenic animals contain abundant Hsp25 immunoreactivity, which in symptomatic mice was redistributed to a more neuropil pattern. Despite the apparent high level of Hsp25 expression in motor neurons, the presence of this Hsp was not sufficient to prevent the accumulation of misfolded SOD1-L126Z protein. Together, these studies provide the strongest linkage yet between aggregation of mutant SOD1 and the pathogenesis of FALS. A framework for explaining why specific populations of neurons selectively accumulate mutant SOD1 is also revealed, in which we suggest that HSP-like chaperone activities may, in part, govern the susceptibility of certain cells types to mutant SOD1 aggregation (see what follows).
Is protein aggregation the cause of SOD1-induced ALS?
The reader may note that we have used the terms aggregate and detergent-insoluble species interchangeably in describing the sedimentable forms of mutant SOD1. We assume that the process by which mutant SOD1 becomes detergent-insoluble involves specific self-association of the protein. Elam et al. (31) have demonstrated that certain types of self-association can occur between dimeric forms of soluble wild-type-like mutants. However, such structures are unlikely for the SOD1-L126Z mutant because the deletion eliminates part of the dimer interface. Additional study will be required to discern the structure(s) of detergent-insoluble SOD1-L126Z protein that accumulates in these animals.
Notably, we do not equate biochemically detectable aggregates with histological inclusions. The sedimentable species of mutant SOD1 found in tissue extracts are in essence forms of the protein, which are organized into structures that are not readily dispersed by detergent. The size of these structures is likely to be heterogeneous with the only commonality being a size or shape that pellets in a small volume (<200 µl) after centrifugation for 10 min at >100 000g. The relationship between these structures and histological inclusions is unclear. The immunostaining patterns of SOD1-L126Z in symptomatic mice do not resemble well-defined inclusion-like structures seen in other FALS mouse models (27,32). Instead, it appears that the entire somatodendritic compartment is filled with diffuse immunoreactivity.
The biochemical correlate to the histological findings is that all SOD1-L126Z appears to be organized into detergent-insoluble, sedimentable structures. Over time, and as symptoms become apparent, the amount of this material increases. In previous studies of mice that express SOD1-G93A, it has been noted that the accumulation of detergent-insoluble protein is not due to altered transcription of the mutant gene, instead it appears to be due to the accumulation of forms of the protein, which are more slowly turned over (33). We interpret the accumulation of detergent-insoluble truncated SOD1 in our SOD1-L126Z and the SOD1-L126insTGGG mice (25) as consequence of changes in the degradation and clearance of the aggregated mutant proteins. Importantly, the L126Z mutant does not ever appear to exist in a soluble state in spinal cord and therefore the misfolded detergent-insoluble molecules must be a primary mediator of toxicity.
Our data in the present work and in previous studies of FALS models (17,23) demonstrate that the accumulation of biochemically distinct species of SOD1 (sedimentable and detergent-insoluble) are a hallmark of disease without exception. Similarly, Johnston et al. (22) have noted an accumulation of high-molecular-weight species of SOD1-G85R and SOD1-G93A in symptomatic mice of these FALS models. With present-day technology, it is not possible to discern which species of conformationally distinct SOD1, on the continuum between misfolded monomer and assembled oligomer, is responsible for toxicity. Although we do not yet fully understand how these accumulating species of mutant SOD1 may be toxic, one possibility is that aggregates could entrap critical proteins. Interestingly, recent studies suggest that mutant SOD1 binds the anti-apoptotic protein bcl-2 (34), which could be entrapped in these structures. It has also been suggested that the accumulation of misfolded mutant SOD1 overwhelms the chaperone and/or proteasome systems of these neurons (22,35).
One concern about the biochemical detection of mutant SOD1 aggregates is that the organization of the mutant protein into sedimentable structures could occur ex vivo. Many years ago, McKinley et al. (36) established that the organization of conformationally altered prion protein into amyloid fibrils occurs ex vivo after extraction of tissues with detergent and limited proteolysis. However, as occurs with prion disorders, we have demonstrated that the levels of those species of mutant SOD1 that partition into the detergent-insoluble fraction increase as animals age and become symptomatic (17,23) (Fig. 3A). Thus, like the prion disorders, the appearance and severity of disease in FALS mice correlate with the accumulation of conformationally distinct species of mutant SOD1 (17,18,23).
Cellular control of protein aggregation
The ubiquitin-proteasome system and molecular chaperones are clearly key systems a cell uses to control the abundance of misfolded proteins and their aggregation. Wild-type SOD1 is a long-lived protein, with a half-life >48 h (7). All FALS mutations compromise protein half-life to some extent, with the truncated mutants being the most short-lived (7). Previously, we noted that mono- and di-ubiquitinated forms of mutant SOD1 accumulate in all FALS models that express G37R, G85R, G93A, H46R/H48Q and Quad variants (17). However, this feature has been largely uncoupled from the disease in the L126Z model. It is possible that the L126Z variant is so unstable that mono- and di-ubiquitinated intermediates have no time to accumulate and participate in aggregation. Alternatively, ubiquitination may occur only after mutant proteins form disease-specific structures and thus does not directly influence aggregation rates.
As occurs in most major neurodegenerative diseases, ubiquitin-positive deposits are a feature of ALS pathology in humans and mice (18,32). Despite the absence of evidence that the accumulated SOD1-L126Z protein in symptomatic mice is ubiquitinated, ubiquitin-positive deposits are also found in diseased spinal cords of this model. We had previously thought that the ubiquitinated inclusions evident in pathology may be composed primarily of mutant SOD1 (17). It now appears that other proteins, to which ubiquitin may attach, are either additional components of such structures or are forming distinct inclusions.
Up-regulation of small Hsp’s has long been associated with stress conditions in CNS (29). In a previous study, we found that up-regulation of Hsp25 and B-crystallin is a consistent feature of all SOD1 mutant mouse models (17). Others have demonstrated in a similar way upregulation of Hsp25 and B-crystallin in the G93A, G85R and G37R mouse models (35,37) and in the SOD1-L126insTGGG model (25). Other Hsp’s, including Hsp70 and Hsp40, are not up-regulated in these mice (17,35,37). Although Tummala et al. (35) reported induction of Hsp25, but not B-crystallin, in the G85R and G93A mouse models, we have previously demonstrated that the total amount of B-crystallin is not as dramatically increased as Hsp25, rather that the most dramatic alteration is an increase in the levels of sedimentable B-crystallin (17). Oligomerization of B-crystallin is associated with its activation as a chaperone (38) and thus we interpret our findings as evidence that the activity of this particular chaperone is increased in symptomatic FALS mice.
Hsp25 and B-crystallin are thought to bind to and stabilize misfolded proteins (29,30). Recent work by several laboratories have demonstrated that B-crystallin primarily acts by binding to exposed hydrophobic surfaces on denatured or misfolded proteins (39–42). Although there have been reports that B-crystallin can assist in the refolding of denatured proteins (43,44), whether B-crystallin is capable of disassembling aggregated forms of protein is unclear. In intact cells over-expressing human glial fibrillary acid protein (GFAP), transfection of astrocytes with vectors expressing B-crystallin was reported to induce the dissociation of inclusion structures (43). However, in these intact cells, it is not possible to distinguish the activity of B-crystallin alone as other Hsp’s and chaperones would also be present. Interestingly, Tummala et al. (35) reported that spinal cords of symptomatic mice from the G85R and G93A models contain lower levels of chaperone activity as measured by the ability of cell lysates to prevent the heat denaturation of catalase. The relative importance of the various Hsp activities, including Hsp 40 and 70, in this assay is not entirely clear to us. It seems important to directly assay B-crystallin activity, such as binding to denatured T4 lysozyme (40), in order to assess whether all chaperone activities are diminished in these animals. Our demonstration here in the SOD1-L126Z mice, and previously in other FALS mouse models (17), that oligomeric sedimentable B-crystallin is up-regulated in symptomatic mice argues that the activity of this Hsp is not among those diminished in symptomatic mice. Here, we demonstrate for the first time that, in vitro, the aggregation of mutant SOD1 is suppressed by the activities of B-crystallin. We believe the most likely explanation for this observation is that B-crystallin is binding to exposed surfaces on misfolded SOD1-L126Z to prevent self-assembly of the mutant protein.
We also demonstrated that B-crystallin is normally expressed in oligodendrocytes and up-regulated in some astrocytes in symptomatic mice (Fig. 7). Jonsson et al. (25) noted a similar pattern of B-crystallin immunostaining in symptomatic mice expressing SOD1-L126insTGGG. The cellular distribution of B-crystallin along with the observation that SOD1-L126Z and SOD1-L126insTGGG appear to specifically accumulate in neuronal cell bodies leads us to speculate that the activities of B-crystallin may protect astrocytes and oligodendrocytes by inhibiting the aggregation of these truncation mutants in these cells. We note, however, that in mice expressing the SOD1-G85R variant, mutant SOD1 immunoreactive inclusions have been reported to occur in both astrocytes and neurons (27). Whether B-crystallin can similarly inhibit the aggregation of the G85R variant is unknown. Also, whether a lack of induction of this Hsp in some astrocytes leaves them vulnerable is unknown. The L126Z and L126insTGGG variants may be unique in that their poor stability help preclude accumulation to levels that promote self-association in many cell types, including astrocytes. Alternatively, the L126Z variant may uniquely be a good substrate for B-crystallin, allowing us to discern the potential contribution of neurons, astrocytes and oligodendrocytes to motor neuron degeneration.
The Hsp25 protein is thought to be a homolog of B-crystallin, sharing a conserved domain. We show here that in non-transgenic animals, Hsp25 immunostaining is most abundant in neuronal cell bodies. In contrast, in symptomatic FALS mice, Hsp25 staining in neuronal cell bodies diminishes dramatically. Despite the apparent high levels of Hsp25 expression in motor neuron cell bodies, mutant SOD1-L126Z abundantly accumulates in the same somatodendritic compartment. Hence, in vivo, Hsp25 does not appear to possess activities to prevent the misfolding and accumulation of SOD1-L126Z.
Primary site of SOD1-L126Z toxicity
The strongest correlation between the increased levels of biochemically detectable SOD1-L126Z protein and immunohistological staining is the increase in neuronal immunoreactivity. It is possible that some very small amount of mutant SOD1-L126Z is present in astrocytes or microglia, but the material that accumulates and correlates with disease appears to be concentrated in motor neurons. Similarly, Jonsson et al. (25) described an age-dependent accumulation of SOD1-L126insTGGG in neuronal cell bodies. In our opinion, it would seem that the most likely site of cell toxicity for the L126Z and 126insTGGG mutants is the somatodendritic compartment of the motor neuron.
Clearly, however, there are FALS models where mutant SOD1 accumulates in glial cells, and processes in these cells could be involved in pathogenesis (27). Our data, however, suggest that such processes may not be central to pathogenesis and that damage to neuronal cell bodies can be sufficient. Damage to non-neuronal cells may contribute to the pathogenesis of disease in other FALS mouse models, but we would suggest that neuronal damage may be crucial.
There are studies that suggest that the foregoing conclusion cannot be true. Two previous reports have described mice where mutant SOD1 has been expressed via promoters that specifically target neurons (45,46). In neither case did these animals develop motor neuron disease. In a recent paper by Liu et al. (20), the levels of expression of SOD1-G85R in mice described by Lino (46) were examined (see Fig. 2D of Liu et al.). Relative to other stains of mice that express the same mutant, the levels of mutant protein in the neuronal promoter mice were very low. Similarly, in the mice that express SOD1-G37R via the neurofilament light chain promoter, the steady-state levels of SOD1-G37R are less than endogenous SOD1 levels (see Fig. 1 of 45), a level too low to induce disease in mice made with the human SOD1 genomic construct (4). However, in the mice expressing mutant SOD1 by neurofilament promoters, there was relatively strong staining of motor neurons, providing evidence that expression was most robust in motor neurons, which comprise a relatively small proportion of the total mass of the spinal cord. Nevertheless, with these negative outcomes, it is difficult to rule out the possibility that the lack of disease in these mice is because expression levels are below the threshold for toxicity.
It is possible that in mice expressing the G37R and G85R mutants, the contribution of non-neuronal cells to pathogenesis is greater. Notably, the G85R mice display robust glial pathology (27).
Another approach to examine the cellular site of toxicity has been to create chimeric mice that are derived from fusions of blastocysts from mice that express mutant SOD1 (G37R or G85R) and non-transgenic mice (47). Analyses of these animals suggested that neurons derived from mutant mice can survive considerably longer when surrounded by non-neuronal cells derived from non-transgenic animals. These data could be interpreted as evidence that the disease has a feed-forward component and that in the chimeric mice that do not develop disease, the secondary components of the process, such as inflammatory insults, cytokines, and so on, are absent. We demonstrate that astrocytic reaction and microglial activation are components of the disease in the SOD1-L126Z mice, but it appears that SOD1-L126Z does not directly mediate these processes in these cells. Instead, it appears that damage to neuronal cell bodies initiates events that result in both astrocytic and microglia responses. If these responses were attenuated, even in these L126Z mice, it remains possible that the course of disease would be altered.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONCONCLUSIONSMATERIALS AND METHODSSUPPLEMENTARY MATERIALSREFERENCES |
|---|
In this analysis of mice that express a very unstable FALS variant of SOD1, we find very compelling evidence for the notion that the accumulation of conformationally distinct, insoluble, aggregation-prone species of mutant SOD1 are directly toxic to motor neurons. It appears to us that damage to neuronal cell bodies and dendrites can be sufficient to initiate the disease process. As SOD1-L126Z is not transported to axonal compartments, it appears that one can rule out direct damage to axons by mutant protein as a required mechanism of toxicity in FALS. We show that
B-crystallin, normally expressed in oligodendrocytes and up-regulated in astrocytes, possesses activities that can inhibit the aggregation of SOD1-L126Z. It is possible that one mechanism by which FALS mutant SOD1 proteins selectively injure motor neurons could involve the relative abundance, activity or subcellular location of chaperone activities in this vulnerable population of neurons. We speculate that Hsp activity, perhaps in combination with variations in the efficiency of mutant SOD1 degradation, may ultimately provide an explanation for the selectivity vulnerability of motor neurons to FALS variants of SOD1. |
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONCONCLUSIONSMATERIALS AND METHODSSUPPLEMENTARY MATERIALSREFERENCES |
|---|
Transgenic mice
To create genomic SOD1 constructs for the L126Z mutant, we re-engineered the 12 kb human SOD1 genomic fragment we have previously used to make SOD1 transgenic mice (4
,17,18) such that the introns between exons 3, 4 and 5 were removed, creating a fragment with just two introns; one between exon 1 and 2 and one between exon 2 and 3 (exons 3, 4 and 5 are fused to make a new exon). The 5' and 3' untranslated portions of the genomic fragment were unaltered. We took this approach, in part, because of a desire to produce even shorter C-terminal truncation mutants in future and the need to eliminate introns that could trigger nonsense-mediated decay (48–50). The L126Z mutation was engineered by replacing the L126 codon with a stop codon. Each coding exon and flanking intronic sequences were verified by sequence analysis before injection into mouse embryos (C3H/HeJxC57BL/6J F2). Transgenic founder mice were identified by PCR amplification of DNA extracted from tail biopsies, using the primers Mo-S: GTTACATATAGGGGTTTACTTCATAATCTG; Hu-S: CAGACAGTAGTTAGAACTTGGT; Mo/Hu AS: CAGCAGTCACATTGCCCARGTCTCCAACATG (R=A/G).
Mouse lines were maintained by crossing transgenic males to non-transgenic (C57BL/6J x C3/HeJ F1) females (Jackson Laboratories, Bar Harbor, ME, USA). All other lines of SOD1 transgenic animals have been previously characterized: the G93A variant [B6SJL-TgN (SOD1-G93A)1Gur; Jackson Laboratory, Bar Harbor, ME, USA], the G85R variant [line 164 (27)], the G37R variant [line 29 (4)], the H46R/H48Q variant [line 139 (18)], the H46R/H48Q/H63G/H120G [line 87 or 125 (17)] and wild-type SOD1 [line 76 (4)]. All studies involving mice were approved by the JHU Institutional Animal Care and Use Committee.
Immunoblotting
Nervous system tissues were homogenized with a probe sonicator as previously described (17). For analysis of protein in nerve fibers, 1 cm of sciatic nerve, located about 1 cm from spinal cord, was taken. Standard SDS–PAGE was performed with 18% Tris–HCl criterion gels (Bio-Rad, Hercules, CA, USA); samples were boiled for 10 min in the Laemmli sample buffer with 2.5% ß-mercaptoethanol (ßME). Immunoblots were probed with the hSOD1 antibody (hAb), a rabbit peptide anti-serum binding to amino acids 24–36 (not conserved between mouse and human SOD1). The following antibodies were also used: an antibody against ubiquitin (5–25, Signet, Dedham, MA, USA) and antibodies from Stressgen (Victoria, BC, Canada) against Hsp25 (SPA-801) and B-crystallin (SPA-222).
Histopathology and Immunocytochemistry
Mice anesthetized with ethyl ether were sacrificed by transcardial perfusion with phosphate buffered saline (PBS, pH 7.4) (Amresco Inc., Solon, OH, USA), followed by 4% paraformaldehyde in 1X PBS. Brains and spinal cords were removed, post-fixed in the same fixative, embedded in paraffin and sectioned for histological and immunological staining. Sagittal sections of the brain and brain stem, and coronal sections of spinal cord (10 µm) were stained with hematoxylin and eosin, silver impregnation and thioflavin-S. Deparaffinized sections were processed for immunocytochemistry, using antibodies against ubiquitin (DAKO, Carpenteria, CA, USA), human SOD1, glial fibrillary acidic protein (GFAP, from DAKO) and Iba-1 (Wako, Osaka, Japan). The immune reaction was visualized with diaminobenzidine, and sections were counterstained with hematoxylin.
For immunofluorescence staining, tissue sections from spinal cord were cut from frozen tissue on a vibratome. Free-floating spinal sections were blocked by 2.5% normal goat serum and treated with 0.1% Triton for 1.5 h, before incubating the anti-human SOD1 serum (1:2500) and a monoclonal antibody against GFAP (1:500, Chemicon, Temecula, CA, USA) at 4°C for overnight. Human SOD1 and GFAP were visualized by subsequent incubation with Cy3 (red)-conjugated goat anti-rabbit IgG (1:300, Jackson Immunoresearch, West Grove, PA, USA) and Cy2 (green)-conjugated goat-anti-mouse IgG (1:200, Jackson Immunoresearch) for 2 h, respectively.
Cell-free protein aggregation assay
Brains were homogenized using a Dounce homogenizer in phosphate-buffered saline (PBS, pH 7.4) with proteinase inhibitors. The homogenate was centrifuged at 150 000g at 4°C for 50 min, and the supernatant was adjusted to 2 mg/ml of protein with PBS. To induce aggregation, 50 µl of sample was added to a 200 µl tube and shaken on a titer plate shaker (750 rpm) at 37°C for 12 h. The resulting samples were centrifuged in an AirFuge (Beckman Coulter) (>100 000g) for 10 min, and the pellets were washed twice with PBS. The resulting pellets were considered aggregated proteins and resuspended in PBS with 2% SDS for further analyses. In vitro assays included the addition of -crystallin (predominantly B-crystallin, SPP-225, Stressgen) with or without a monoclonal antibody against B-crystallin (SPA-222, Stressgen).
|
SUPPLEMENTARY MATERIALS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONCONCLUSIONSMATERIALS AND METHODSSUPPLEMENTARY MATERIALSREFERENCES |
|---|
Supplementary material is available at HMG Online.
|
ACKNOWLEDGEMENTS |
|---|
We are very grateful to Ms Debbie Swing for her help in transgene injections and to Ms Hilda Slunt for help in the construction of the genomic construct for SOD1-L126Z. This study was supported by grants from the National Institutes of Neurologic Disease and Stroke (R01 NS 37225 and R01 NS 047225), the ALS Association, the Muscular Dystrophy Association and by the Robert Packard Center for ALS Research at The Johns Hopkins University.
Conflicts of Interest statements: None declared.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONCONCLUSIONSMATERIALS AND METHODSSUPPLEMENTARY MATERIALSREFERENCES |
|---|
-
Gurney, M.E., Pu, H., Chiu, A.Y., Dal Canto, M.C., Polchow, C.Y., Alexander, D.D., Caliendo, J., Hentati, A., Kwon, Y.W., Deng, H.-X. et al. (1994) Motor neuron degeneration in mice that express a human Cu, Zn superoxide dismutase mutation. Science, 264, 1772–1775.
[Abstract/Free Full Text] -
Borchelt, D.R., Lee, M.K., Slunt, H.H., Guarnieri, M., Xu, Z.-S., Wong, P.C., Brown, R.H., Jr., Price, D.L., Sisodia, S.S. and Cleveland, D.W. (1994) Superoxide dismutase 1 with mutations linked to familial amyotrophic lateral sclerosis possesses significant activity. Proc. Natl. Acad. Sci. USA, 91, 8292–8296.
[Abstract/Free Full Text] -
Borchelt, D.R., Guarnieri, M., Wong, P.C., Lee, M.K., Slunt, H.S., Xu, Z.-S., Sisodia, S.S., Price, D.L. and Cleveland, D.W. (1995) Superoxide dismutase 1 subunits with mutations linked to familial amyotrophic lateral sclerosis do not affect wild-type subunit function. J. Biol. Chem., 270, 3234–3238.
[Abstract/Free Full Text] - Wong, P.C., Pardo, C.A., Borchelt, D.R., Lee, M.K., Copeland, N.G., Jenkins, N.A., Sisodia, S.S., Cleveland, D.W., Price, D.L. (1995) An adverse property of a familial ALS-linked SOD1 mutation causes motor neuron disease characterized by vacuolar degeneration of mitochondria. Neuron, 14, 1105–1116.[CrossRef][ISI][Medline]
- Reaume, A.G., Elliott, J.L., Hoffman, E.K., Kowall, N.W., Ferrante, R.J., Siwek, D.F., Wilcox, H.M., Flood, D.G., Beal, M.F., Brown, R.H., Jr. et al. (1996) Motor neurons in Cu/Zn superoxide dismutase-deficient mice develop normally but exhibit enhanced cell death after axonal injury. Nat. Genet., 13, 43–47.[CrossRef][ISI][Medline]
-
Bruijn, L.I., Houseweart, M.K., Kato, S., Anderson, K.L., Anderson, S.D., Ohama, E., Reaume, A.G., Scott, R.W. and Cleveland, D.W. (1998) Aggregation and motor neuron toxicity of an ALS-linked SOD1 mutant independent from wild-type SOD1. Science, 281, 1851–1854.
[Abstract/Free Full Text] -
Ratovitski, T., Corson, L.B., Strain, J., Wong, P., Cleveland, D.W., Culotta, V.C. and Borchelt, D.R. (1999) Variation in the biochemical/biophysical properties of mutant superoxide dismutase 1 enzymes and the rate of disease progression in familial amyotrophic lateral sclerosis kindreds. Hum. Mol. Genet., 8, 1451–1460.
[Abstract/Free Full Text] -
Valentine, J.S. and Hart, P.J. (2003) Misfolded CuZnSOD and amyotrophic lateral sclerosis. Proc. Natl. Acad. Sci. USA, 100, 3617–3622.
[Abstract/Free Full Text] -
Yim, H.S., Kang, J.H., Chock, P.B., Stadtman, E.R. and Yim, M.B. (1997) A familial amyotrophic lateral sclerosis-associated A4V Cu, Zn-superoxide dismutase mutant has a lower Km for hydrogen peroxide. Correlation between clinical severity and the Km value. J. Biol. Chem., 272, 8861–8863.
[Abstract/Free Full Text] -
Yim, M.B., Chock, P.B. and Stadtman, E.R. (1990) Copper, zinc superoxide dismutase catalyzes hydroxyl radical production from hydrogen peroxide. Proc. Natl. Acad. Sci. USA, 87, 5006–5010.
[Abstract/Free Full Text] -
Yim, M.B., Chock, P.B. and Stadtman, E.R. (1993) Enzyme function of copper, zinc superoxide dismutase as a free radical generator. J. Biol. Chem., 268, 4099–4105.
[Abstract/Free Full Text] -
Yim, M.B., Kang, J.H., Yim, H.S., Kwak, H.S., Chock, P.B. and Stadtman, E.R. (1996) A gain-of-function of an amyotrophic lateral sclerosis-associated Cu, Zn-superoxide dismutase mutant: an enhancement of free radical formation due to a decrease in Km for hydrogen peroxide. Proc. Natl. Acad. Sci. USA, 93, 5709–5714.
[Abstract/Free Full Text] - Beckman, J.S., Carson, M., Smith, C.D. and Koppenol, W.H. (1993) ALS, SOD and peroxynitrite. Nature, 364, 584.[CrossRef][Medline]
- Crow, J.P., Ye, Y.Z., Strong, M., Kirk, M., Barnes, S. and Beckman, J.S. (1997) Superoxide dismutase catalyzes nitration of tyrosines by peroxynitrite in the rod and head domains of neurofilament-L. J. Neurochem., 69, 1945–1953.[ISI][Medline]
-
Estévez, A.G., Crow, J.P., Sampson, J.B., Reiter, C., Zhuang, Y., Richardson, G.J., Tarpey, M.M., Barbeito, L. and Beckman, J.S. (1999) Induction of nitric oxide-dependent apoptosis in motor neurons by zinc-deficient superoxide dismutase. Science, 286, 2498–2500.
[Abstract/Free Full Text] - Wiedau-Pazos, M., Goto, J.J., Rabizadeh, S., Gralla, E.B., Roe, J.A., Lee, M.K., Valentine, J.S. and Bredesen, D.E. (1996) Altered reactivity of superoxide dismutase in familial amyotrophic lateral sclerosis. Science, 271, 515–518.[Abstract]
-
Wang, J., Slunt, H., Gonzales, V., Fromholt, D., Coonfield, M., Copeland, N.G., Jenkins, N.A. and Borchelt, D.R. (2003) Copper-binding-site-null SOD1 causes ALS in transgenic mice: aggregates of non-native SOD1 delineate a common feature. Hum. Mol. Genet., 12, 2753–2764.
[Abstract/Free Full Text] - Wang, J., Xu, G., Gonzales, V., Coonfield, M., Fromholt, D., Copeland, N.G., Jenkins, N.A. and Borchelt, D.R. (2002) Fibrillar inclusions and motor neuron degeneration in transgenic mice expressing superoxide dismutase 1 with a disrupted copper-binding site. Neurobiol. Dis., 10, 128–138.[CrossRef][ISI][Medline]