29 July 2011. Although it seems that many deficiencies can lead to amyotrophic lateral sclerosis, the enzyme superoxide dismutase 1 (SOD1) has long received the lion’s share of attention. Mutations in SOD1 cause disease in a subset of people with familial amyotrophic lateral sclerosis (ALS), and the majority of animal models for the disease carry one or another SOD1 mutation. The mutant protein aggregates, suggesting that it might avoid normal degradation pathways. But this is not the case, according to the authors of a Protein Science paper published online July 20. Researchers in the laboratory of Arthur Horwich at Yale University School of Medicine in New Haven, Connecticut, found that mutant SOD1 (mSOD1) protein rapidly turns over in mice, even while aggregates are accumulating.
An alternative explanation for mSOD1’s toxicity might lie in its propensity to hook up with mitochondria. In the July 11 PLoS One online, first and senior author Christine Vande Velde and members of her lab at the University of Montréal, Canada, reported a thorough analysis of mSOD1’s interaction with the organelles. The paper is the first, Vande Velde said, to show that misfolded SOD1 associates with mitochondria specifically in motor neurons, and in vivo. While the study is strictly observational, it suggests that mitochondrial function is somehow altered in mSOD1 motor neurons.
Speedily Cycling SOD1
There are more than 100 different SOD1 mutations linked to ALS. While the wild-type protein is quite stable, several of the mutant proteins exhibit rapid turnover, Horwich told ARF in an e-mail. Pathological aggregates of these mutant SOD1 proteins form in people with ALS and in mouse models that express mSOD1 (see ARF related news story on Bruijn et al., 1998). The G85R mutant that Horwich studied, for example, is only present at low levels in soluble form (Jonsson et al., 2006), but steadily forms oligomers and insoluble aggregates as mice carrying the mutation age (Wang et al., 2009; Matsumoto et al., 2005).
Horwich and first author George Farr suspected that, with aging, reduced turnover of SOD1-G85R turnover might lead to an excess of the protein. The abnormally high levels of mSOD1 might then be cause for aggregation, they hypothesized. To examine turnover in mice carrying yellow fluorescent protein (YFP)-labeled human SOD1, they performed a pulse-chase experiment. Starting at 1.5 months, they provided the mice with drinking water containing 8 percent deuterium to label proteins with the heavy hydrogen. When the researchers washed out the deuterium at four months of age, they determined that YFP-tagged wild-type SOD1 protein had a half-life of 22 days. In contrast, half of the G85R mutant version was gone in fewer than three days. “That implies that it is a small fraction of G85R-YFP that somehow goes on to be insoluble,” Horwich wrote. When they started the experiment later, commencing deuterium treatment at 4.5 months and washing it out at seven, the mutant protein was similarly unstable. Thus, aging did not affect mSOD1 turnover.
Inhibitors of the proteasome and autophagy blocked the mSOD1 turnover, showing that it is mostly degraded by these usual pathways. The researchers do not yet know what determines whether mSOD1 is degraded or aggregates; it could be that aggregation happens to a random fraction of the mutant protein or that a specific misfold is most likely to aggregate, he added.
mSOD1 Mangles Mitochondria
Vande Velde started her project in the University of California, San Diego, laboratory of Donald Cleveland. Several studies, many from the Cleveland lab, have hinted that mSOD1 spells trouble for mitochondria, which are defective in the spinal cords of ALS mice. Mitochondria swell up in pre-symptomatic animal models, and proceed to disintegrate into vacuoles once symptoms arise (Martin et al., 2009; Kong and Xu, 1998). Mutant SOD1 binds to mitochondria and the vacuoles derived from them in the spinal cord (Vande Velde et al., 2008; Higgins et al., 2003). However, Vande Velde said, most studies have examined mitochondria from mashed-up spinal cords, not specifically from motor neurons, or they have focused on cultured cells.
To take a look in vivo, the researchers developed a new reporter mouse line that produces mitochondria-targeted green fluorescent protein in cells that express the motor neuron marker, Hb9. They crossed these mice with mSOD1 lines expressing either an inactive (G85R) or active (G37R) enzyme, and then examined tissue from mice that were sacrificed a few weeks before obvious symptoms arise. Using an antibody specific for misfolded SOD1 (Gros-Louis et al., 2010), Vande Velde showed that both mutant proteins colocalize with mitochondria in motor neuron axons (see ARF related news story on Liu et al., 2004).
In cells with normal SOD1, mitochondria formed tubular networks in the soma, while long, skinny, individual mitochondria were evenly distributed along the axons. In the mSOD1 mice, somal mitochondria were fatter and rounder, with fewer interconnections. In the axons, some mitochondria appeared as small, round organelles linked together like “pearls on a string,” the authors write, suggesting that there might be defects in mitochondrial fusion.
There were also differences between the G85R and G37R phenotypes. In the dismutase-inactive G85R mice, those axonal mitochondria that did not line up like pearls were often longer than the organelles in control animals. In the G37R mice, axonal mitochondria appeared shorter and fatter. Further, in the dismutase-active G37R mice, mitochondria tended to accumulate near Schmidt-Lanterman incisures. These spots represent gaps in the myelin sheath where axons are thinned, and may be involved in axon-glia communications (Price et al., 1990; Campana, 2007). Similarly, mitochondria in mice with the dismutase active SOD1-G93A mutation cluster at regular axonal intervals (Sotelo-Silveira et al., 2009).
“The effects of G85R on axonal mitochondria seem to be much less pronounced than in G37R,” wrote Kurt De Vos, of King’s College London, U.K., in an e-mail to ARF. That could be because the enzyme activity of G37R makes pathology worse, he suggested, or simply because disease develops more quickly in that animal model. The aberrant clustering could be related to defects in axonal transport, added Gerardo Morfini of the University of Illinois at Chicago (see ARF related news story). Vande Velde examined G85R mice several months before symptoms arose, and discovered misshapen mitochondria in those axons as well. This suggests that mitochondria falter quite early in disease, noted Jordi Magrane of Weill Cornell Medical College in New York.
Just what mSOD1 is doing with mitochondria, and exactly how that relates to disease, remains unknown. Vande Velde was cautious about speculating on mechanisms from this purely descriptive study. However, she said, when mitochondria are misshapen, it usually indicates a disruption in their function.
There are many ways in which mSOD1 and mitochondria could, theoretically, lead to motor neuron degeneration. Alterations in mitochondrial channel conductance or calcium regulation could muck up cellular activities, Vande Velde suggested. Mutant SOD1, by binding the mitochondrial protein Bcl-2, promotes apoptosis (see ARF related news story on Pasinelli et al., 2004). The mutant protein also blocks mitochondrial pores (Li et al., 2010), including voltage-dependent anion channel 1 (see ARF related news story on Israelson et al., 2010). Drosophila lacking that channel exhibit elongated mitochondria, presumably due to lack of fission or increased fusion amid the organelles (Park et al., 2010).
The current work describes mitochondrial pathology in two mSOD1-based animal models; the majority of people with ALS have idiopathic, not SOD1-based disease. “How generalizable is this mitochondrial pathobiology?” asked Lee Martin of the Johns Hopkins School of Medicine in Baltimore, Maryland, in an e-mail to ARF. Some research suggests that SOD1 might misfold in idiopathic ALS, too (see ARF related news story on Bosco et al., 2010). “We can only hope that mitochondria do have an important role in human ALS,” wrote Martin, noting that a mitochondrial-targeted drug, olexisome, is already in clinical trials in Europe (Martin, 2010; Kirk, 2010). Such a treatment might help, agreed Mark Cookson of the National Institute on Aging in Bethesda, Maryland. However, he noted, the organelle has proven to be a difficult drug target for researchers seeking to cure mitochondrial diseases. In addition, the efficacy of a mitochondrial therapy would depend on whether the mitochondrial breakdown is actually a key cause of ALS.
Vande Velde’s future work should help discover how crucial the mitochondrial alterations are in ALS. “We definitely can get at that mechanism,” she said. The researchers are now analyzing their mice for defects in mitochondrial functions such as ATP production, transmembrane potential, calcium handling, and production of reactive oxygen species.—Amber Dance.
Vande Velde C, McDonald KK, Boukhedimi Y, McAlonis-Downes M, Lobsiger CS, Bel Hadj S, Zandona A, Julien JP, Shah SB, Cleveland DW. Misfolded SOD1 associated with motor neuron mitochondria alters mitochondrial shape and distribution prior to clinical onset. PLoS One. 2011;6(7):e22031. Abstract
Farr GW, Ying Z, Fenton WA, Horwich AL. Hydrogen-deuterium exchange in vivo to measure turnover of an ALS-associated mutant SOD1 protein in spinal cord of mice. Protein Sci. 2011 Jul 20. Doi: 10.1002/pro.700. Abstract