The hippocampus is composed of anatomically heterogeneous subregions, including CA1, CA3, dentate gyrus and subiculum. A major feature of the hippocampal area CA3 is its recurrent collateral circuitry, by which the CA3 pyramidal cells make excitatory synaptic contacts with each other, while CA1 pyramidal cells are not extensively interconnected. These anatomical differences inspired David Marr (1971), conceptualizing the hippocampus, and in particular, area CA3, as an autoassociator network that performs pattern storage and retrieval. In 1996, McClelland et al. proposed that this autoassociator network could mediate a dynamic competition between pattern completion and pattern separation, two complementary processes of associative memory. Pattern separation and completion are considered to be the two main complementary associative memory processes. Pattern completion is the process by which the network can recall an entire memory using only degraded input, while pattern separation is the ability of the network to distinguish similar inputs. The use of pattern completion and separation is important for hippocampus dependent contextual learning to decrease the possibility of error in memory recall. While Nakazawa et al. (2002) demonstrated that area CA3 is crucial for spatial pattern completion using CA3 NMDA receptor 1 (NR1) KO mice, the evidence for the involvement of CA3 in pattern separation is scarce. Further, there are few relevant tasks that can test pattern separation. First, Catherine Cravens developed a new contextual pattern separation behavioral paradigm using C57BL/6N (B6) mice. A mouse was placed in fear conditioning chamber A for 3 min. At the end of exposure in chamber A, which served as the conditional stimulus (CS), the mouse received a single foot shock as the unconditioned stimulus (US) to learn the association between context A and the aversive foot shock. Three hours or 24 hours later, freezing levels of mice were observed either in chamber A where mice were conditioned, or in chamber B, in which the contextual cues were completely changed. Under this condition, both B6 wild type mice and the floxed-NR1 control mice showed robust freezing to context A, but not to context B 3 hr later. However, while CA3-NR1 KO mice showed a comparable level of freezing to A as the control mice, she found that the mutants froze in context B at the same level, suggesting that CA3-NR1 KO mice are impaired in contextual pattern separation and thus unable to distinguish the two contexts. Furthermore, she found that this mutant phenotype disappeared when post-shock period was elongated to 3 min with 1 min-preshock period during the one-trial CS-US conditioning, suggesting mutants are specifically impaired in automatic or incidental encoding of context acquisition. In other words, it is suggested that an attentional or arousal increase during the conditioning restores the context acquisition in the CA3-NR1 KO animals (Cravens et al., 2006). Second, Kimberly Christian and Angela Miracle have initiated a new study on CA3-NR1 KO mice to examine the role of CA3 NMDA receptors in chronic stress response. Chronic stress induces behavioral, neuroendocrine and structural changes in rodents that mirror those observed in patients with clinical depression. To date, the bulk of morphological research following chronic stress has focused on dendritic changes in the hippocampus, particularly subfields CA3 and CA1, and the prefrontal cortex. In particular, it has been shown that the dendritic atrophy observed in CA3 following chronic stress can be prevented by the infusion of NMDA antagonists. Therefore, we hypothesized that NMDA receptors of area CA3, a region which is particularly vulnerable to the effects of stress, may modulate the stress response of efferent and/or interconnected regions, such as CA1 and prefrontal cortex. First, Angela Miracle learned a method of Glaser and Van der Loos? modified Golgi-Cox staining under Dr. Cara L. Wellman at Indiana University to analyze dendritic morphology. They then subjected male and female CA3 NR1-KO mutant mice and their homozygously floxed-NR1 littermates to immobilization stress in plastic rodent restraint bags for 2 hours per day for 10 days. After brains were processed using Glaser and Van der Loos? modified Golgi stain, neuronal reconstruction and analysis was performed with Neurolucida. Preliminary data showed a significant reduction in the length of CA1 apical dendrites in the control mice following chronic stress compared to stressed mutants and non-stressed animals. This result is the first demonstration that CA3 may exert control over stress-related changes in efferent targets and is an important initial step in understanding how the integrated response to chronic stress develops and can be manipulated. If this receptor is permissive in allowing for the induction of stress-related structural changes or conversely, its absence initiates compensatory processes, we can hypothesize that functional ablation of this receptor may preserve some cognitive abilities compromised by stress. Indeed, in a pilot behavioral experiment, Kimberly Christian observed a modest protection in the CA3 NR1-KO mice against the effects of stress in a novel spatial working memory task. Briefly, mice are placed in a transparent circular arena with a shockable grid floor (~45cm diameter) in the presence of extra-maze cues. Animals receive mild foot-shocks every 2 seconds except in an unmarked 6 cm diameter zone defined by experimenter. Mice must explore the arena and use spatial cues to find and remember the location of this ?hidden? safety zone. During a 12 minute acquisition trial, stressed CA3 NR1-KO mice show delayed acquisition compared to non-stressed controls but significantly improved performance over stressed control mice in the last 3 minutes of the trial. Further analysis of depression-related behavior will be conducted in collaboration with Dr. Husseini Manji at NIMH. The third research focus of this CA3 project is to clarify the neuronal activity alteration in area CA3 as a consequence of CA3 NMDA receptor ablation. Previously, it was shown that CA1 place cell activity of CA3-NR1 KO mice was transiently impaired upon novelty exposure while this phenotype disappeared in a familiar environment (Nakazawa et al., 2003), suggesting CA3 NR networks support intact spatial representation in CA1 in response to novelty. However, it has never been investigated whether spatial representations in CA3 are altered as a consequence of CA3 NR ablation. We have predicted that the CA3 recurrent network is unable to form efficient attractors without its NRs, thereby causing impaired spatial tuning with multiple place field peaks. We have set up an in vivo multi-electrode recording system to isolate neural activity in awake behaving mice, using the Neuralynx system (Tucson, AZ). This system allows us to simultaneously record unit activity, local EEG, and animal head location and direction. While the experiments conducted by Juan Belforte are still in progress, preliminary results showed that the spatial tuning of dorsal CA3 place cell activity is impaired in familiar linear tracks. This result suggests that NRs in dorsal CA3 participate in the formation of spatial representations, a novel finding in that previous studies have shown nearly normal spatial tuning of CA1 place cells despite impaired synchronized firing following genetic ablation of NRs in CA1 (McHugh et al., 1996).
| Kishimoto, Yasushi; Nakazawa, Kazu; Tonegawa, Susumu et al. (2006) Hippocampal CA3 NMDA receptors are crucial for adaptive timing of trace eyeblink conditioned response. J Neurosci 26:1562-70 |
| Nakazawa, Kazu; McHugh, Thomas J; Wilson, Matthew A et al. (2004) NMDA receptors, place cells and hippocampal spatial memory. Nat Rev Neurosci 5:361-72 |
