470 1357-4310/99/$ - see front matter © 1999 Elsevier Science Ltd. All rights reserved. PII: S1357-4310(99)01592-0 Opinion MOLECULAR MEDICINE TODAY, NOVEMBER 1999 (VOL. 5) Neurodegenerative diseases of the central ner- vous system (CNS) and acute traumatic injuries to the CNS, despite their differences in etiology, share common features. In both cases, neurons often continue to die even after the primary risk factor, whatever its nature, is removed. This ongoing degeneration is attributable to the col- lective effect of a number of different processes and can occur independently of the primary risk factors. Scientists have attempted to identify the mediators of such degeneration, with the aim of either neutralizing them and/or their effects, or rendering the neurons more resistant to them – approaches known as neuroprotection 1,2 . Neuroprotection after acute nerve insult is help- ful only for the protection of neurons which, after the initial injury, are still viable but will undergo de- layed degeneration unless adequately treated. We now have a clearer understanding of how nerve damage spreads, and we know that differ- ent types of primary insult can have common mediators of toxicity, some of which have been identified. It might therefore be feasible to develop neuroprotective therapies for many chronic dis- eases with different causes but similar patterns of progression. Recognition of the neurodegenerative nature of certain diseases that were not formerly re- garded as such (Box 1) raises the question of whether these syndromes resemble the classic neurodegenerative diseases. In this article, we distinguish between neurodegenerative diseases on the basis of their site of origin: ‘axogenic’ dis- eases, which originate in the axons (and were not necessarily classified as neurodegenerative in the past), and ‘somagenic’ diseases, which orig- inate in the neuronal cell bodies (soma). We will also examine differences, other than site of ori- gin, between the two groups and attempt to de- termine whether such differences might affect the choice of the neuroprotective approach. Somagenic versus axogenic degeneration From the clinician’s perspective, once a disease has been diagnosed as neurodegenerative, it matters little whether the neurodegeneration is initiated at the cell body or the axon. In both cases, the outcome is functional loss and even- tual cell death. Despite these similarities, there are notable differences between the time course and biochemical nature of the cell death that are caused by axonal damage and by direct injury to the cell body. For example, following ischemia, which affects the cell bodies, death of neurons is usually inevitable and rapid; retinal ischemia last- ing only 90 min results in the loss of virtually all neurons in the inner retina 3 . By contrast, tran- section of the optic nerve in a rat results in the loss of only half of the retinal ganglion cells after one week 4,5 . Thus, the nature of the insult – whether it is axogenic or somagenic – dictates the kinetics of cell death. The delay in cell death following axonal injury is due, at least in part, to the relatively long dis- tance between the site of injury and the cell body. This slower process of degeneration leaves a larger window of opportunity for treatment, so that it might be possible to rescue the cell bodies of ‘Axogenic’ and ‘somagenic’ neurodegenerative diseases: definitions and therapeutic implications Michal Schwartz, Eti Yoles and Leonard A. Levin Neurodegenerative diseases are characterized by a relentless loss of specific groups of neuronal subtypes. Many of these diseases share similar molecular mechanisms and extracellular mediators of neuronal loss. We now suggest that neurodegeneration originating in the neuronal cell bodies (e.g. in Alzheimer’s disease, Parkinson’s disease and amyotrophic lateral sclerosis) should be distinguished from that originating in the axons (e.g. in glaucoma, certain peripheral neuropathies and spinal stenosis). We propose that the former group of diseases be defined as ‘somagenic’ and the latter as ‘axogenic’. Although axogenic disorders may share common symptoms and mediators of toxicity with somagenic disorders, they have distinct temporal, subcellular and signal-transduction features. We further suggest that, by adopting this classification of disorders based on pathophysiological processes, we will come to recognize additional diseases (in particular, those defined as axogenic) as being neurodegenerative and therefore possibly amenable to neuroprotective therapy.