All About Multiple Sclerosis

More MS news articles for May 2004

Spinal Cord Alive and Kicking

Blocking activation of a death receptor enhances axonal regeneration and functional recovery in an animal model of spinal cord injury

May 13, 2004
Catherine Barthélémy; Christopher E Henderson
Developmental Biology Institute of Marseille, Marseille, France
Nat Med 10(4):339-340, 2004

Spinal cord injury will permanently handicap about 1 in 1,000 individuals over the course of their lifetime.[1] Much effort has been devoted to understanding the complex cellular changes that develop after injury, and to inventing ways to overcome the poor capacity of the adult spinal cord for spontaneous regeneration.[2] Programmed cell death within the damaged tissue is one of these changes, but whether it hinders recovery has been controversial.

In this issue, Demjen et al.[3] show that the death receptor CD95/Fas is a major trigger of cell death in the injured spinal cord. They find that mice treated with neutralizing antibodies to CD95/Fas recover from a spinal cord lesion that leaves untreated animals heavily handicapped. This shows that keeping cells alive at the lesion site promotes subsequent recovery, and provides a potential tool for doing so in human patients.

Programmed cell death is a tightly regulated process that can be initiated by diverse stimuli, including extracellular signals acting at specific receptors such as CD95/Fas and tumor necrosis factor (TNF) receptor.[4] Both of these receptors have a predominant role in apoptosis (hence the commonly used term 'death receptor'), but they can activate a variety of other pathways and cellular phenomena.[5] In the nervous system, CD95/Fas activation can lead to cell death of neurons[6] and glial cells,[7] but also to enhanced axonal growth.[8] CD95/Fas may have an important role in pathological degeneration in vivo: CD95/Fas and its ligand, CD95L/FasL, have been implicated in cell death after nerve transection,[9] ischemia,[10] amyotrophic lateral sclerosis,[11] multiple sclerosis, Parkinson disease[12] and Alzheimer disease.

After spinal cord injury in human patients and in animal models, some cells at the lesion site die by post-traumatic necrosis, whereas others, including oligodendrocytes, die by apoptosis.[13] In line with this, spinal cord trauma leads to increased expression of death receptors and their ligands, as well as to activation of death-promoting caspases.[14] Initial experiments showed that caspase inhibition has a protective effect in models of spinal cord injury, providing the first indication that reducing cell death might be beneficial.[15] Others, however, have observed no protection.[16] Further uncertainty about the role of cell death arose when it was shown that inactivation of TNF or its receptor did not improve prognosis.[17]

These ambiguous results probably reflect the known capacity of TNF to be both pro- and antiapoptotic. Demjen et al. therefore asked instead whether a less ambivalent death receptor, CD95/Fas, might be involved. They first confirmed that in their model, spinal cord injury led to rapid upregulation of the components of the CD95/Fas system and to increased apoptosis (measured as the number of neurons and oligodendrocytes with fragmented nuclear DNA). Intraperitoneal injection of blocking antibodies to CD95L/FasL at the time of injury led to a 60% reduction in the number of dying cells three days later. When analyzed up to one month after injury and treatment, animals showed significant functional recovery in behavioral tests of reflexive and voluntary motor function. This recovery was correlated with increased sprouting and regrowth of descending corticospinal fibers toward the lesion, and with decreased loss of staining for myelin basic protein, a marker for myelinating oligodendrocytes. On the basis of these results, Demjen et al. propose that the CD95/Fas system be considered a new therapeutic target.

A true understanding of how antibodies to CD95/Fas reduce cell death and enhance recovery will require more detailed knowledge. The cellular source and target of the ligand in damaged spinal cord need to be identified, and the links between axonal sprouting and recovery of motor function need to be defined.

Nonetheless, the following model (Fig. 1) may apply: After injury, CD95L/FasL is upregulated at the lesion site on neurons, astrocytes and invading lymphocytes and microglia. CD95L/FasL triggers death of oligodendrocytes and local spinal interneurons, resulting in an environment that is not favorable to regrowth of descending corticospinal axons. Preventing cell death provides an environment more conducive to spontaneous regeneration, and formation of new circuits leads to functional recovery in voluntary motor tests.

Figure 1. Suicide prevention. (a) As reported by Demjen et al., spinal cord injury induces increased expression of the death receptor CD95/Fas by neurons, lymphocytes, microglia and oligodendrocytes at the lesion site. In parallel, the natural ligand CD95L/FasL is upregulated on neurons, astrocytes, lymphocytes and microglia. CD95/Fas activation by paracrine or autocrine mechanisms, or both, results in cell death of neurons and oligodendrocytes, considerably modifying the environment in which reparative axon regeneration needs to occur. (b) Blocking CD95L/FasL by intraperitoneal injection of a specific antibody protects spinal neurons and oligodendrocytes from CD95/Fas-induced death and permits axonal regrowth and functional recovery.

This model may in some respects be reductionist. For example, it ignores the possible role of autocrine CD95/Fas activation, and possible effects of CD95L/FasL or the blocking antibodies on more distant targets such as circulating lymphocytes or the corticospinal neurons themselves.

What remains to be done before these findings can lead to tests in human patients, in parallel with the wide range of other strategies being developed? Several differences with the clinical setting are apparent. First, the lesion used here (surgical transection of the spinal cord) is quite different from human cases of spinal cord injury, many of which result from local compression. Second, for the behavioral experiments mice were treated with antibodies immediately before injury. Clearly, protocols will need to be developed to deliver antibodies to the lesion site within the human CNS at later stages. Third, the oncogenic risks of blocking cell death cannot be ignored, although the authors point out that these risks should be minor over short periods of treatment.

As with other experimental strategies, many questions remain to be addressed. The experimental paradigm of Demjen et al. is particularly stringent, meaning that functional recovery may be greater in other models. The new work thus firmly places a new actor on the list of players in spinal cord repair.


  1. Sekhon, L.H. & Fehlings, M.G. Spine 26, S2-S12 (2001).
  2. Schwab, M.E. Science 295, 1029-1031 (2002).
  3. Demjen, D. et al. Nat. Med. 10, 389-395 (2004).
  4. Wallach, D. et al. Annu. Rev. Immunol. 17, 331-367 (1999).
  5. Pettmann, B. & Henderson, C.E. Nat. Cell Biol. 5, 91-92 (2003).
  6. Raoul, C., Pettmann, B. & Henderson, C.E. Curr. Opin. Neurobiol. 10, 111-117 (2000).
  7. D'Souza, S.D. et al. J. Exp. Med. 184, 2361-2370 (1996).
  8. Desbarats, J. et al. Nat. Cell Biol. 5, 118-125 (2003).
  9. Ugolini, G. et al. J. Neurosci. 23, 8526-8531 (2003).
  10. Martin-Villalba, A. et al. J. Neurosci. 19, 3809-3817 (1999).
  11. Raoul, C. et al. Neuron 35, 1067-1083 (2002).
  12. Hayley, S. et al. J. Neurosci. 24, 2045-2053 (2004).
  13. Emery, E. et al. J. Neurosurg. 89, 911-920 (1998).
  14. Casha, S., Yu, W.R. & Fehlings, M.G. Neuroscience 103, 203-218 (2001).
  15. Li, M. et al. Neuroscience 99, 333-342 (2000).
  16. Ozawa, H., Keane, R.W., Marcillo, A.E., Diaz, P.H. & Dietrich, W.D. Exp. Neurol. 177, 306-313 (2002).
  17. Farooque, M., Isaksson, J. & Olsson, Y. J. Neurotrauma 18, 105-114 (2001).

Copyright © 2004, Nature Publishing Group