Pii: S0197-4580(99)00074-3

The data reported in Nicolle et al. [30] suggest a preservation of three hippocampal synaptic proteins (synaptophysin, synaptotagmin, and synaptosomal associated protein-25) in old, memory-deficient rats. These are interesting and important data, which, at first consideration, seem to raise a question about whether alterations of synaptic connectivity can contribute in a substantial way to the cognitive changes that occur during the normal process of aging. One myth about the effects of normal aging on brain anatomy that has clearly been shattered recently is the idea that there is widespread cell loss. When modern stereological methods are used for cell counting, neither rats [29,35,36] nor humans [41] show substantial principal cell loss over the lifespan. If the principal cells in the hippocampus do not die, then what anatomical substrate may be responsible for the cognitive changes that are observed in hippocampal-dependent behavioral tasks during aging? The logical place to look, of course, is at changes in synapse numbers. Connectivity changes, rather than cell loss, could be the anatomical alteration responsible for age-related memory changes. Several studies have conducted synapse counts in the rat hippocampus in various synaptic subfields. Some studies have reported decreases in synapse numbers [2,3,12,18–20,21, 23,28,39], whereas others have reported no change [1,13, 14,27,37]. There is, however, only one study that has counted synapses with unbiased stereological methods in behaviorally-characterized rats. Geinisman et al. [21] found a 24% decrease of axospinous synapses in the middle and inner molecular layers of the fascia dentata (FD) of old, memory-impaired rats. Because the amount of data on synapse counts in the multiple synaptic input regions of the hippocampal formation is limited at present, it remains an open question about how many synaptic populations might be changed during aging. Clearly, Geinisman et al. [21] have shown that synaptic contacts in the inner and middle molecular layer are reduced, but definitive answers (using the most convincing methods) about other regions remains for future experiments to resolve. What about functional synaptic connectivity that can be measured in electrophysiological experiments? There have been several experiments that address this issue across the lifespan of rats, using in vivo anesthetized, in vivo awake, and in vitro preparations. In order to better understand how to place the results of the Nicolle et al. [30] experiment into the context of existing data on synaptic function, a short review of synaptic physiology in the aging hippocampus is in order. There are two synaptic connections that have been studied extensively with electrophysiological methods in the aged rat hippocampus. This includes the medial perforant path input from the entorhinal cortex to the granule cells of FD, for which it is known that there are fewer morphological synaptic contacts in old rats [21]. The CA3 Schaffer collateral input to stratum radiatum of CA1 pyramidal cells has also been studied electrophysiologically; however, the data on whether, in memory-impaired rats, there is an anatomically detectable reduction in synapse numbers [2] or no change [37,39], have yet to be established. For FD, electrophysiological measurements of the presynaptic fiber potential (the medial perforant path input fibers) suggest that there is a pruning of functional axon collaterals [5,16] from the entorhinal cortex in old rats (i.e., a smaller fiber potential response at stimulus levels above threshold). Furthermore, extracellular [4,5,16] and intracellular [5] excitatory postsynaptic potentials (EPSPs) elicited to medial perforant path stimulation are smaller in old rats over the full range of stimulus intensities from threshold to maximal response. These electrophysiological data are consistent with the anatomical observation of fewer synaptic contacts made by medial perforant path inputs to FD. The ratio of field or intracellular EPSP magnitude to the magnitude of the pre< C.A.Barnes’ salary is supported by MH01227. * Corresponding author. Tel.: 11-520-626-2312; fax: 11-520-6262618. E-mail address: [email protected] (C.A. Barnes) Neurobiology of Aging 20 (1999) 349–351