Apoptosis and Stroke Pathogenesis
To the Editor:
The articles by Kitagawa et al1 and Pulera et al2 in the same issue of Stroke permit one to make some generalizations about publications implicating apoptosis in the pathogenesis of “stroke” or the “response” to cerebral ischemia. Numerous publications fail to make a clear distinction between infarction, which is total necrosis of the affected area, and delayed neuronal necrosis in areas of selective vulnerability. In the latter there is frequently no pannecrosis of tissue, only loss of neurons with, perhaps, reactive astrocytosis. These two different consequences of ischemic hypoxia, and also of hypoxic hypoxia, may overlap3 ; however, this fact should not obscure the fact that different as well as identical pathogenetic (ie, biochemical) factors may be involved in each outcome, so that therapies reported to be of value in models of delayed neuronal death in restricted brain area might not, in fact, be useful in treating or preventing infarction. I would suggest that the potential limitations of an experimental report could best be “flagged” by insisting that all reports concerned with delayed necrosis in selectively vulnerable areas include the words “selective necrosis” or “selectively vulnerable neurons” in their title. This could then be paralleled with the word “infarction” in the title of articles dealing with that subject. I believe this would be of particular importance for neurologists and other nonneuropathologists who may not be attuned to the distinction.
The article by Kitagawa et al1 clearly deals with selective death of neurons in a limited area of temporal lobe even though global ischemia was used. The brains were examined 7 days later, and the findings correspond to delayed neuronal cell death.3 In the study of Pulera et al,2 there was apparently both infarction in the frontotemporal brain and also selective, delayed neuronal death in areas of the hippocampus. After describing both findings, the results concentrate only on the latter and report morphological evidence for a nonnecrotic, probably apoptotic, mode of delayed death in the neurons of the dentate gyrus. It is unclear from the text whether evidence for apoptosis was also found in the CA1 region where “ischemic” neurons were also found. Nor can I find a clear statement concerning the presence or absence of evidence for apoptosis in the truly infarcted portion of the brain.
A lack of clarity in terminology, at least in my mind, is also illustrated by the legend to Figure 3.1 At the bottom of the figure we are told that we are looking at the difference in time courses of “ischemic” cell death in two brain regions. Do the authors mean death “caused” by ischemia, whether or not the death is delayed, or do they mean death as indicated by the appearance of “ischemic” neurons in the two areas? This distinction is important because eosinophilia of neuronal cytoplasm was used as an important criterion for ischemic cell death in this article. This is the classic hallmark of what has been called “acute ischemic cell change.” The word “acute” implied that the ischemic/hypoxic event occurred less than 24 hours before death. Only recently has a review3 appeared which clearly demonstrates that the “red,” “eosinophilic,” or “acidophilic” neuron can also appear days after the ischemic event. Thus, these neurons may signify not only recent ischemia but also a delayed ischemia, which might be a second episode of ischemia or “something else.” This is not merely important for clinical neuropathologists who are trying to date lesions and their time of onset, but also for experimentalists, because the “something else” means that the neuron can respond in at least a superficially identical manner to very different challenges—eg, ischemia itself versus excitotoxic damage. Moreover, the red neuron has conventionally been thought to die by necrosis, not by apoptosis or some related mechanism with orderly rather than random breakdown of DNA. If at least some red neurons are destined for the latter “apoptotic”-like pathway and other red neurons are not, this is important information. The question appears to have been first explicitly raised in the review cited above.3 The data in the study of Pulera et al2 clearly suggest that at least some red neurons are destined not for necrosis but for a death that at least shares characteristics with apoptosis, and that their appearance is “delayed.”
A recent study by Ni et al4 clearly shows this. Like Kitagawa et al,1 Ni et al used caspase activity as one marker of the apoptotic path. Ni et al found that in an area where virtually all neurons appear to be marked for apoptotic death, as indicated by caspase activity, the same neurons go through the red stage. This point was not made by Ni and his coworkers.
The literature cited by both Kitagawa et al1 and Pulera et al2 notes that apoptosis can occur in or around areas of frank infarction. Are these neurons red but undergoing death with a caspase-driven orderly degradation of DNA? What about the actual mode of cell death for the red neurons in the heart of an infarct?
- Copyright © 1999 by American Heart Association
Kitagawa K, Matsumoto M, Tsujimoto Y, Ohtsuki T, Kuwabara K, Matsushita K, Yang G, Tanabe H, Martinou J-C, Hori M, Yanagihara T. Amelioration of hippocampal neuronal damage after global ischemia by neuronal overexpression of BCL-2 in transgenic mice. Stroke.. 1998;29:2616–2621.
Pulera MR, Adams LM, Liu H, Santos DG, Nishimura RN, Yang F, Cole GM, Wasterlain CG. Apoptosis in a neonatal rat model of cerebral hypoxia-ischemia. Stroke.. 1998;29:2622–2630.
Dr Rosenblum has proposed that two different types of ischemic lesions in experimental animals, selective neuronal death and infarction, should be separately investigated when the involvement of apoptosis or necrosis in ischemic brain damage is examined. We agree with his notion and have focused on the effect of BCL-2 overexpression in selective neuronal death.R1
Certainly, there exist several differences underlying between infarction and selective neuronal death.
One of them is the astroglial reaction. In infarction, astrocytes as well as neurons are destroyed, but in selective neuronal death they are viable and reactive after ischemia. If apoptosis is the critical determinant for infarction, the question of whether both neurons and astrocytes are killed by the same mechanism, apoptosis, will arise. To the best of our knowledge, there is no clear explanation. However, the mechanism underlying neuronal apoptosis may cause secondary damage, including energy depletion and acidosis in astrocytes in infarcted tissue.
Another difference is recruitment of inflammatory cells into tissue. Granulocytes are dominant in infarction, but they are scanty in the area with selective neuronal death. Several studies, including our recent article,R2 have shown the importance of microcirculatory disturbance and the pathogenic role of granulocytes in the expansion of infarction after focal cerebral ischemia. In contrast, microcirculatory disturbance is unlikely to be involved in the course of delayed neuronal death.
In spite of the differences mentioned above, however, several important findings have been first discovered with the model of delayed neuronal deathR3 and then confirmed in focal cerebral ischemia producing infarction.R4 R5 R6 Therefore, there seems to be no doubt that the model of delayed selective neuronal death is valuable when investigating the molecular mechanism underlying ischemic neuronal damage. Although the mechanism of selective neuronal death was examined mostly in the hippocampus, we would like to emphasize that delayed neuronal death was observed not only in the hippocampus but also in the whole brain after brief periods of ischemia.R7
Kitagawa K, Matsumoto M, Tsujimoto Y, Ohstuki T, Kuwabara K, Matsushita K, Yang G, Tanabe H, Martinou JC, Hori M, Yanagihara T. Amelioration of hippocampal neuronal damage after global ischemia by neuronal overexpression of BCL-2 in transgenic mice. Stroke.. 1998;29:2616–2621.
Sheardown MJ, Nielsen EO, Hansen AJ, Jacobsen P, Honore T. 2,3-dihydroxy-6-nitro-7-sulfamoyl-benzo(F)quinoxaline: a neuroprotectant for cerebral ischemia. Science.. 1990;247:571–574.
Hatakeyama T, Matsumoto M, Brengman JM, Yanagihara T. Immunohistochemical investigation of ischemic and postischemic damage after bilateral carotid occlusion in gerbils. Stroke.. 1988;19:1526–1534.
We thank Dr Rosenblum for his interest in our report and thank the editor for the opportunity to provide the following clarifications.
First, it is apparent that there is a lack of consensus regarding the terminology of apoptosis. Fundamental concepts such as the definition of apoptosis are not agreed upon. Consequently, the literature is full of inconsistencies, and experimental results are difficult to interpret. In our recent article,1 we attempted to minimize these inconsistencies.
The stated purpose of our study was to determine whether evidence of apoptosis was present in the cerebral cortex and dentate gyrus in our model of cerebral hypoxia-ischemia. In both the dentate gyrus and frontotemporal cortex, we described evidence of apoptosis, such as a DNA laddering pattern on agarose gel electrophoresis, the detection of apoptotic bodies by electron microscopy and ethidium bromide staining, and tdt-mediated dUTP-biotin nick-end labeling. The stated purpose of the study did not include examining the CA1 region for apoptosis, and our data indicate a relative paucity of cell death in the CA1 in our model. Consequently, we did not investigate the CA1 in this article. In addition, we had no data regarding cerebral blood flow or a physiological definition of the “penumbra” or “core” of the infarction area in our model, and the precise definition of these concepts is under debate. Therefore, we could not comment on the presence of apoptosis in the ischemic penumbra or core. Developing experimental techniques to explore these concepts is of paramount importance.
Furthermore, we believe we have used the term “ischemic cell change” appropriately. In our article, we described ischemic cell change as “eosinophilic cytoplasm with pyknotic nuclei.” We derived the definition of ischemic cell change from the classic description of Brown and Brierley.2 They described ischemic cell change by stating “the nucleus is shrunken, dark-staining and often triangular” and “the cytoplasm is acidophilic, staining pink with eosin.” Brierley et al3 go on to say that ischemic cell change is “not, as the term suggests, a neuronal alteration due to anoxia-ischemia alone” and “is the neuropathologic common denominator in all types of hypoxia.” These authors emphasize that ischemic cell change implies a process that transforms a normal cell into an essentially naked, shrunken nucleus and then results in cell loss. Regions of excessive cell loss result in infarction area. This process occurs after hypoxia-ischemia in regions of selective vulnerability. Therefore, it seems appropriate to describe cells in our study with pyknotic nuclei and eosinophilic cytoplasm generated from a hypoxic-ischemic insult as having ischemic cell change. Brierley et al3 indicate that the process of ischemic cell change occurring in regions of selective vulnerability “is most readily seen when survival after the hypoxic stress is 48 hours or more.” This concept is further supported in the literature cited in the review of Rosenblum.4
Our article provides evidence that apoptosis occurs in regions of ischemic damage defined by the criteria described above. Since techniques such as double-labeling of neurons undergoing ischemic cell change were not performed, we cannot make definitive statements regarding the fate of these neurons. However, we can speculate that perhaps some of these neurons die by apoptosis as well as necrosis. Clearly, additional investigation into these areas could provide useful information, especially regarding the development of new therapies for stroke.
Pulera MR, Adams LM, Liu H, Santos DG, Nishimura RN, Yang F, Cole GM, Wasterlain CG. Apoptosis in A neonatal rat model of cerebral hypoxia-ischemia. Stroke.. 1998;29:2622–2630.
Brown AW, Brierley JB. The nature, distribution and earliest stages of anoxic-ischaemic nerve cell damage in the rat brain as defined by the optical microscope. B J Exp Pathol.. 1968;49:87–106.
Rosenblum WI. Histopathologic clues to the pathways of neuronal death following ischemia/hypoxia. J Neurotrauma.. 1997;14:313–326.