There are a number of important limitations to the present review. First, our discussion on the morphological features of cell death after global ischemia was restricted to neuronal injury. Death of glia (astrocytes, oligodendroglia and microglia) must also be considered in the quest to limit ischemic injury with cytoprotectants. Indeed, it has been argued that glial but not neuronal apoptosis occurs in post-ischemic brain (Martin et al. 1998; Petito et al. 1998). Furthermore, the death of invading leukocytes (e.g., macrophages) must be considered as their apoptotic remnants might be mislabeled as apoptotic neurons. Second, we have limited our discussion to adult animals. Programmed cell death, specifically apoptosis, is more likely in developing organisms. Third, neurogenesis has been reported after cerebral ischemia, including in the CA1 zone (Bendel et al. 2005; Nakatomi et al. 2002; Schmidt and Reymann 2002), although many reports do not find this. An interesting question is whether the death of these newly-generated neurons (Bueters et al. 2008) occurs via programmed cell death mechanisms and whether this is apoptotic or necrotic in appearance.
While insight into the mechanisms of ischemic injury is gained from morphology studies, such findings must be considered in tandem with biochemical data. Notably, the consistent observation of neuronal necrosis after global ischemia does not exclude a contribution of programmed cell death mechanisms. There are many deleterious mechanisms set in motion by ischemia, and this undoubtedly includes programmed cell death mediators (e.g., apoptosis-inducing factor release from mitochondria (Boujrad et al. 2007). The question is whether such pathways are causal to cell death or are merely activated concomitantly with many other pathways. We expect, certainly in cases of more severe ischemia mimicking cardiac arrest, that the latter situation is true. If so, the activation of programmed cell death pathways is akin to taking a suicide pill just prior to being executed by firing squad. The reliance upon mild global ischemic insults in arterial occlusion models is conceptually similar to the use of very brief focal ischemic insults. In both cases, fewer cells tend to die, they take longer to succumb, and perhaps cells express more markers for programmed cell death. However, is this situation really going to predict clinical outcome? Thus, we encourage investigators to consider using more clinically relevant insults (e.g., cardiac arrest models, permanent focal ischemic insults) in their assessment of treatments targeting programmed cell death. Such studies must also be careful to avoid problems encountered with other neuroprotectants (e.g., NMDA-receptor antagonists). For instance, physiological variables should be controlled (e.g., to prevent unintentional hypothermia) and long-term functional and histological outcome must be assessed.
In summary, the majority of studies examining the ultrastructural morphology of cell death following global ischemia report features more consistent with neuronal necrosis; although some differences exist among studies and brain regions. Only a few studies claim to have observed apoptotic neuronal morphology, but they either do not present convincing evidence or they misinterpret the morphology. Further work is needed to link the morphology of ischemic cell death with mechanisms of action. Such studies should take into account the pattern and extent of injury commonly found after cardiac arrest and other causes of global ischemia in humans.
Was this article helpful?