The hippocampus is extremely sensitive to cardiac arrest, in general, slightly more than even the cerebral cortex, so that changes are often limited to the hippocampus in global ischemia in humans (DeJong et al. 1969; Zola-Morgan et al. 1986) or animals (Auer et al. 1989). It is important to remember, however, that sometimes the hippocampus can be selectively spared (Adams et al. 1966) in global ischemia. In animals, physiologically controlled experiments demonstrate that ischemic periods as short as 2 min are capable of causing hippocampal CA1 pyramidal cell neuronal necrosis (Smith et al. 1984). Like the cerebral cortex, the hippocampus in global ischemic damage can show asymmetric damage, probably due to asymmetries in the human vasculature. These include the size of the carotid arteries and, more importantly, asymmetries in the circle of Willis (Riggs and Rupp 1963; Stehbens 1963).
Of all the brain regions, the hippocampus demonstrates within it a most heuristic selective vulnerability, and an order of vulnerability which is quite specific. Whereas some diseases, such as neurolipi-doses, affect the CA3 pyramidal cells first, most diseases such as Alzheimer's disease and ischemia first affect the smaller CA1 pyramidal cells within the hippocampus, which are less rich in RNA and protein synthetic capability.
Ischemic CA1 necrosis (Figs. 13.2, 14.9, 14.14) can occur after only 2 min of global ischemia. After 2-4 min of global cerebral ischemia, the CA1 pyramidal cells regularly undergo neuronal necrosis (Smith et al. 1984). However, this does not occur at the time of ischemia or shortly thereafter as was once believed. The seminal experiments of Kirino in the 1980s have conclusively shown that, in gerbils (Kirino 1982) and rats (Kirino et al. 1984), neuronal necrosis after global cerebral ischemia takes place between the second and fourth days. Delayed neuronal death may occur earlier with greater degrees of insult, and even later with insults that are mitigated by hypothermia, or insults that are shorter. But the basic principle in global ischemia is that there is delayed neuronal death, especially in the hippocampus.
The phenomenon of delayed neuronal death has also been demonstrated in humans, with roughly the same time course of 2-4 days in the hippocampus (Petito et al. 1987; Horn and Schlote 1992). The principles outlined are thus especially important for medicolegal timing of an ischemic insult in relation to the death. In other brain regions such as neocor-tex and especially striatum, neuronal death occurs more quickly than in hippocampus (Pulsinelli et al. 1982).
The hilus cells of the dentate, also known as CA4 (Fig. 13.6a), is phylogenetically the oldest, reticular nervous tissue of the hippocampus. In CA4, cells die after only small global ischemic insults, as they do in CA1, and calcium accumulates there early (Ben-veniste and Diemer 1988). This is of importance because of the large number of cell types in CA4 (Amaral 1978), many of which are inhibitory (Ribak and Anderson 1980). Loss of such inhibitory cells in the dentate hilus might worsen ischemic damage by removing inhibition through the trisynaptic chain from the dentate granule cells to the CA3 cells, in turn to the CA1 pyramidal cells. The pathologist should examine the dentate hilus if there is suspicion of short periods of global cerebral ischemia, especially in children where microglial activation may be the only change seen (Del Bigio and Becker 1994).
After CA1 cells are affected, progressively longer or more severe ischemic insults will cause recruitment of CA3 cells into the necrotic process. Lastly, the dentate gyrus, very resistant to cerebral ischemia, is recruited in the most severe cases of insults. This contrasts with the often seen vulnerability of the dentate gyrus to hypoglycemia, but the latter is based on the massive release of aspartate into the extracellular fluid of the brain (see Sect. 13.5). The above pattern of damage in the hippocampus can be seen unilaterally or bilaterally. This asymmetry has been alluded to above, and its basis may be either vascular or in an asymmetric timing of onset of spreading depression and other prerequisite events which precede and lead to neuronal necrosis in the ischemic process later.
In addition to sampling both hemispheres, the hippocampus should ideally be sampled at several points along its septo-temporal axis. The classic section of the hippocampus is seen when a coronal section is taken from the temporal lobe at the level of the lateral geniculate body. This produces a microscope section including the dentate gyrus in its classic C-shape, the hilus or CA4, the CA3 pyramidal cell band, and CA1 cells. CA2, the zone generally ignored, is theoretically a very narrow zone where the zinc-containing mossy fibers do not terminate from dentate granule cells on CA3 cells, yet they have the morphology of CA3 cells.
The order of vulnerability within the hippocampus deserves further discussion. CA1 cells are exqui-
sitely sensitive to ischemic insults (as well as hypo-glycemic and epileptic insults). This may be related to a combination of cellular features leading to selective vulnerability: small amounts of Nissl substance relative to neuronal size, and high concentration of NMDA receptors (Greenamyre et al. 1985) leading to long-term changes in neuronal excitability that can run down ion gradients of the neuron, without the protein synthetic machinery necessary to ensure a robust cellular response (Bodsch et al. 1986).
Since memory formation in the human brain depends on the hippocampus, and the hippocampus can be selectively involved in global brain ischemia or bilateral posterior cerebral artery occlusion (Victor et al. 1961), it is not surprising that selective
amnesia occurs after global ischemia (Woods et al. 1982; Zola-Morgan et al. 1986). However, amnesia of events prior to 15 years before the insult are preserved, as storage and retrieval of distant memories slowly migrate out of the hippocampus into the temporal lobe (Squire et al. 1989).
Was this article helpful?