Amyotrophic lateral sclerosis is a disease not only of lower motor neurons but also of the upper motor neurons (UMNs), but whether UMN degeneration in mouse models closely mirrors human pathology has been unclear. In today's Journal of Neuroscience, researchers outline a detailed characterization of the neural degeneration that occurs in the brain of one type of ALS mouse model. Hande Ozdinler of Northwestern University in Evanston, Illinois, who led the study in collaboration with her former advisor Jeffrey Macklis of the Massachusetts General Hospital in Boston, reports that accompanying lower motor neuron deficits, there is early and specific loss of upper motor neurons in these animals. The findings may give researchers a handle on degeneration of those same neurons in humans.

Superoxide dismutase-1 (SOD1), when mutated, can cause ALS in people, and mice carrying human SOD1-G93A are a common model for the disease. Scientists knew that the mice, in addition to their well-known spinal cord motor neuron loss and associated paralysis, had upper motor neuron problems (Zang and Cheema, 2002), but as Ozdinler wrote in an e-mail to ARF: “We lacked a detailed study that showed cell-type-specific vulnerability and degradation of upper motor neurons.” The identification of new markers for neural cell types allowed the researchers to analyze corticospinal motor neurons and their neighbors more carefully than in the past.

The caveat is that the mouse’s neurobiology makes it a poor model for the upstairs portion of ALS, said Andrew Eisen of the University of British Columbia in Vancouver. “Mice and rats do not have a cortical motor neuron system like humans do,” he told ARF.

In contrast, Ozdinler believes that understanding the how upper motor neurons decay in the mice will be informative and may suggest therapeutics for people with upper motor neuron symptoms, such as stiff muscles and overactive reflexes. She used retrograde labeling to tag upper motor neurons in the SOD1-G93A mice. Even at 30 days of age—before symptoms begin—the animals evinced fewer and smaller corticospinal motor neurons than did wild-type mice or those carrying non-mutant human SOD1. The defects progressed over time as the motor neurons succumbed to apoptosis, or cell death.

Examining a variety of cell-type markers, Ozdinler found that the neurodegeneration was limited to corticospinal motor neurons and other developmentally related non-motor, subcerebral neurons. Other types of neurons and interneurons were unaffected. This is a pattern similar to that seen in human ALS. Although lower and upper motor neurons are not near neighbors, or develop along the same tracks, the authors suggest that they share something in common that makes them particularly vulnerable to SOD1 mutations.—Amber Dance


  1. This new study by the lab of Jeffrey Macklis assesses if, beside the well-established degeneration of (lower) spinal cord motor neurons in the widely used mutant SOD1/G93A mouse line, there is also a loss of (upper) corticospinal motor neurons (CSMNs), as is the case in actual human ALS. This is an important question and essential in order to judge the accuracy of this mouse ALS model.

    Already in 2002, a study published in Neuroscience Letters by the lab of Surindar Cheema in Australia attempted to answer this question (Zang and Cheema, 2002). Both studies injected retrograde fluorogold labeling (at cervical levels) at different disease stages into high-expressing G93A mice (that reach endstage at 120 days) in order to mark the CSMNs and to assess their loss. The former study assessed loss at 60, 90, and 110 days of age (with negative littermates as controls), while the Macklis study did a more extensive approach starting at 30 days of age (then 60, 90, and 120 days) and comparing to wild-type overexpressing SOD1 mice. Interestingly, the Cheema study could identify significant degeneration of CSMNs at 90 days of age (30 percent loss) with first signs starting at 60 days. The current Macklis study clearly can identify already a significant 30 percent loss at 60 days of age, when G93A mice are still without apparent signs of hindlimb paralysis. Even more so, there is already clearly reduced soma diameter (25 percent) of CSMNs as early as 30 days of age (a classic pre-symptomatic stage) as this study finds. Importantly, the degeneration of CSMNs seems to be rather selective, as other cortical projecting neurons or even cortical interneurons are not affected.

    As is often the case, an interesting finding raises an even larger number of interesting questions: In this same ALS line, if a similar precise study would be done with spinal cord MNs (maybe even with retrograde labeling), how would the two degenerative phases temporarily compare? Which neurons degrade first? Likewise, if mutant SOD1 toxicity could be reduced specifically in one of the two locations, how would it influence the other region? With modern cell type-specific gene expression profiling approaches (e.g., fluorescence-activated cell sorting or laser-microdissection), how do the two neuronal degenerative programs compare?

    Despite the significant loss of CSMNs, the reactive gliosis is rather small—a strange finding if one compares to the early and fulminant gliosis in the spinal cord. Likewise, the early Cheema study actually showed degeneration of rubrospinal neurons in the midbrain as well (up to 40 percent at endstage), which is not detected in the current Macklis study.

    As we often are biased by the obvious phenotype of ALS mice (hindlimb paralysis), we probably overlook many other less or even equally touched neuronal systems in these model mice. Despite the interest of analyzing the degeneration of CSMNs in ALS, it is mostly the loss of the spinal cord motor neurons that makes the disease so deadly; slowing motor neuron degeneration in the spinal cord remains, therefore, a very important goal. However, insights from CSMNs could well reveal additional candidate pathways for intervention in both neuronal populations.


    . Degeneration of corticospinal and bulbospinal systems in the superoxide dismutase 1(G93A G1H) transgenic mouse model of familial amyotrophic lateral sclerosis. Neurosci Lett. 2002 Oct 31;332(2):99-102. PubMed.

    View all comments by Christian Lobsiger

Make a Comment

To make a comment you must login or register.


Paper Citations

  1. . Degeneration of corticospinal and bulbospinal systems in the superoxide dismutase 1(G93A G1H) transgenic mouse model of familial amyotrophic lateral sclerosis. Neurosci Lett. 2002 Oct 31;332(2):99-102. PubMed.

Further Reading


  1. . Upper motor neuron and extra-motor neuron involvement in amyotrophic lateral sclerosis: a clinical and brain imaging review. Neuromuscul Disord. 2009 Jan;19(1):53-8. PubMed.
  2. . The syndromes of frontotemporal dysfunction in amyotrophic lateral sclerosis. Amyotroph Lateral Scler. 2008 Dec;9(6):323-38. PubMed.
  3. . Cortical hyperexcitability may precede the onset of familial amyotrophic lateral sclerosis. Brain. 2008 Jun;131(Pt 6):1540-50. Epub 2008 May 9 PubMed.

Primary Papers

  1. . Corticospinal motor neurons and related subcerebral projection neurons undergo early and specific neurodegeneration in hSOD1G⁹³A transgenic ALS mice. J Neurosci. 2011 Mar 16;31(11):4166-77. PubMed.