Plastic modifications induced by object recognition memory processing

Results

We examined whether training in the OR paradigm is capable of inducing changes in synaptic strength in the hippocampal CA3-CA1 synapse. C57BL/6 mice had stimulation and recording electrodes stereotaxically implanted in the CA3 and CA1 regions, respectively (Fig. 1). When recovered from surgery, mice were trained (Fig. 2A, Tr) in the OR paradigm by using two different objects, denoted by A and B. During the training session, evoked fEPSPs were recorded at the moment that animals explored each object. No changes in the amplitude of recorded fEPSPs were observed during exploration of the two objects used for training (Fig. 2B; 107.3 ± 4.7% and 99.33 ± 7.4% for A and B, respectively, n = 12). Animals were removed from their home cages and taken to the recording room at different times after training (1.5, 3, or 6 h). At each of these events, mice were submitted to a 5-min recording session, where fEPSPs evoked at the CA3-CA1 synapse were again recorded. These recording sessions were meant to detect any late training-induced change in synaptic strength. LTP-like enhancements in synaptic efficacy were detected 6 h posttraining (110.4 ± 4.4%, P < 0.05, as compared with 100, n = 13; Fig. 2C) lasting <24 h (Fig. S1B Left). To find out whether these observed changes were really a consequence of OR training, and not to human handling or to exposure to the OR box alone, a group of electrode-implanted mice were submitted to the same protocol, with one difference: no objects were presented at the training session—i.e., one extra habituation session was performed instead. Stimuli were presented at 0, 1.5, 3, 6, and 24 h for 5 min (every 20 ± 5 s). Quantitative analysis showed that the percentage of variation of fEPSP amplitudes across this time, as compared with the mean value (100%) computed during the baseline (fourth day of habituation) period, was ≤11.2%, with no statistically significant tendency toward a decrease or increase (P = 0.67; Fig. S2).

In addition, we investigated whether a pharmacological treatment capable of hindering OR LTM formation was also capable of preventing training-related synaptic plasticity events in the hippocampus. Pretraining systemic treatment with the NMDA-receptor antagonist MK-801 hindered OR LTM formation and concomitantly blocked the late training-induced changes in CA3-CA1 fEPSP enhancements recorded in vehicle-treated animals (P > 0.05; Fig. S3).

OR memory retention was assessed in a test session performed 24 h after training (Fig. 2A, Test) in which a familiar and a novel object were used (denoted by A and C, respectively). Successful learning was obtained when animals explored the novel object for a significantly longer time than the familiar one, and only data from these animals were analyzed. Approximately 10% of all animals assigned to the experiment were excluded for not having learned the task effectively (no significant changes in fEPSP amplitudes were observed in those animals; P > 0.05). In contrast to the training session, exploratory behavior toward both familiar and novel objects employed in the test session was accompanied by significant increases in CA3-CA1 synaptic efficacy (139.4 ± 17.6% and 133.6 ± 12.8%, respectively; P < 0.05 compared with baseline values collected during the last habituation session, n = 12; Fig. 2D). Interestingly, there were no changes in the hippocampal fEPSPs in animals submitted to retraining instead of a test session (Fig. S4). In that retraining session, there was no novel feature relating to training, because the same environment and objects were used in both sessions.

For those animals submitted to regular test sessions (a novel and a familiar object), 5-min recording sessions were carried out at different times posttest to evaluate test-induced late changes in fEPSP (Fig. 2A, right). We found that a short period of depotentiation occurs at ~1.5 h posttest, followed by an even later potentiation phenomenon (6 h; Fig. 2C). No such modifications were observed when the test session was omitted (Fig. S1B Right). Importantly, none of the recorded changes in fEPSP was accompanied by a significant alteration in the hippocampal theta or gamma rhythms (Fig. S5), indicating that all data were collected from similar alertness states.

In a second set of experiments, we investigated whether saturation of the CA3-CA1 synapse by experimental high-frequency stimulation (HFS) hindered OR-evoked changes in synaptic strength. On their last day of habituation to the open field, one group of mice received a HFS capable of causing a 5-day persistent LTP (197.3 ± 19.1% at 20 min after HFS, P < 0.05 compared with baseline; Fig. 3B, filled circles; see ref. 17 for details). Twenty-four hours after HFS presentation, mice were trained for OR. For as long as they expressed HFS-evoked potentiation of hippocampal CA3-CA1 synapses (Fig. 3D, filled circles), mice spent approximately the same amount of time exploring both novel and familiar objects in a test session conducted 24 h after training (Fig. 3E, right, LTP), suggesting that saturation of that pathway by external HFS makes animals incapable of acquiring OR memory. However, when the same group of mice was trained and tested 2 weeks after HFS—a situation in which no significant increase in fEPSP could be distinguished (Fig. 3F, filled circles)—they were able to acquire OR LTM normally (Fig. 3G, right, LTP). The control group exhibited fEPSPs similar to baseline in all recording sessions (Fig. 3 C, D, and F, open circles) and learned normally (Fig. 3 E and G, right, Contr).

Discussion

It is currently believed that experience-dependent changes in synaptic strength are the underlying physiological mechanism of mnemonic processes. In the hippocampus, the two known phenomena of synaptic plasticity modification—LTP and LTD—have been correlated with different types of memory formation (18 –20). LTP, in particular, was shown to accompany associative aversive learning in the hippocampus (2 –4).

When considering OR memory, hippocampal functionality is necessary for the acquisition of a temporal sequence of events (21), as well as for the distinction of spatial information about objects (22, 23). Although it may not play a direct role in distinguishing the different features of each object, it is fundamental as a novelty detector because of its role in comparing previously stored information with new incoming aspects of one particular situation. The hippocampus receives inputs from the perirhinal cortex, which is itself the site of entrance for visual, olfactory, and somatosensory information, all of which are relevant for object recognition. As happens when other behavioral paradigms are employed, OR also originates strong long-lasting mnemonic traces that can be accessed for >24 h after the acquisition phase (24). It would be expected, therefore, that this type of memory would also induce biochemical and electrophysiological changes in specific structures, including the hippocampus. Many biochemical aspects of OR memory have been described (24 –27), and we now provide evidence that OR LTM formation induces synaptic efficacy changes comparable to those that underlie the lasting storage of other memory types.

Here, we show the induction of object-dependent LTP-like synaptic changes, given that simple exposure to the environment does not induce similar modifications. This endogenous phenomenon reaches its peak at 6 h after the acquisition phase (training). These results are at variance with those of previous studies in which acquisition of information about novel objects (28, 29), or object configurations (18), facilitates LTD expression, while information about novel environments favors LTP. Interestingly, our results show that training-induced hippocampal LTP lasts <24 h, although animals do still express OR memory for 24 h or more (24). This observation suggests that although memory consolidation is hippocampus-dependent, this may not be true for memory persistence, which has been studied in other paradigms (30, 31) but not in the OR task. One possible explanation is that information about spatial and contextual characteristics of objects could end up being relocated to other parts of the brain once memory is consolidated (32 –34).

When a given memory is retrieved in the presence of novelty, it is set into a labile phase and requires stabilization to persist (35). This phase of memory processing is called reconsolidation. It is considered an active phase that takes place to allow reorganization of the already formed memories, allowing incorporation of new information (36, 37).

From this point of view, and considering the OR paradigm (see Methods), the test phase could itself act as a trigger for a reconsolidation-like labile phase of memory, because it necessarily involves a novel object being presented simultaneously to a familiar one (26). Present results constitute evidence that reactivation of a consolidated memory induces changes in synaptic efficacy. Here, we show that reactivation of OR memory induces novelty-dependent synaptic modifications in the hippocampus, suggesting that similarly to memory consolidation, the reconsolidation phenomenon is also capable of modifying hippocampal plasticity. In some way, reactivation-induced changes are similar to those observed after a training session (6 h posttraining) but occur only after a brief period of depotentiation observed at 1.5 h after reactivation. A possible explanation for this rapid LTD phase is that memory updating after reactivation requires a transient protein degradation period. In fact, this theory has already been discussed in a recently published paper showing that the process of adding new information to a preexisting memory requires a protein degradation phase that takes place ≈2 h after retrieval (38). In that study, basal protein levels were reestablished at 6 h after retrieval, a fact that could account for the depotentiation/potentiation pattern observed in our study.

If we assume that OR memory formation is totally reliant on CA3-CA1 synaptic functionality, then any experimental procedure capable of disturbing hippocampal patterns of synaptic strengths would be enough to prevent memory consolidation. The assumption is that the huge wave of plasticity generated by experimentally evoked LTP produces retrograde amnesia by interfering with the activation of hippocampal memory networks (39). This was proved correct for other types of memory (2, 19, 40) and is considered a relevant argument to support the theory that memory acquisition depends on hippocampal LTP. In our experiments, the application of an external HFS in a certain time window before OR training reversibly hindered OR LTM formation—i.e., animals were incapable of learning while CA3-CA1 synapses were potentiated by the HFS-evoked LTP. Once fEPSP values had returned to basal level and animals were retrained in the OR task, they showed normal acquisition.

Taken together, our results indicate that the OR memory processing, as well as the acquisition of conditioned fear responses, is accompanied by an enhancement in synaptic strength (2, 3, 41) and that NMDA receptors are involved in this adaptive process (2, 42). The fact that saturating LTP is able to occlude the learning-induced synaptic modifications further reinforces the hypothesis that LTP-like processes are involved in actual learning (42, 43). Moreover, our results suggest that OR LTM formation depends on hippocampal integrity and is capable of inducing, in the CA3-CA1 synapses, long-lasting plastic changes that play a role in memory codification for the first few hours. Reconsolidation of OR memory also leads to important synaptic modifications, although with a different pattern.

Methods

Supplementary Material