The subjects were 83 Long-Evans rats from our breeding colony, of which 52 were female and 31 males, that conformed mixed groups in some experiments (Figures 1, 2, 3B, C). The subjects weighed 200–300 g at the start of testing. The rats were housed on a reversed 12-h light/12-h dark cycle (lights on 19:00–07:00), in groups of two or four. All behavioral testing was conducted during the dark phase of the cycle. Rats were food deprived to 85–90% of their free feeding weight to increase spontaneous exploration, except during recovery from surgery, where food was available ad libitum. The water remained available ad libitum throughout the study. All experimentation was conducted in accordance with the Institutional Animal Care and Use Committee of the Favaloro University.

All rats that were used for pharmacological infusions were implanted bilaterally in Prh with 22-gauge indwelling guide cannulas. Subjects were anesthetized with ketamine (Holliday, 74 mg/kg, i.p.) and xylazine (Konig, 7.4 mg/kg, i.p.) and placed in a stereotaxic frame (David Kopf Instruments, Tujunga, CA, USA) with the incisor bar set at −3.2 mm. Guide cannulas were implanted according to the following coordinates, measured relative to the skull at bregma (Paxinos and Watson, 1998) anteroposterior −5.5 mm, lateral ± 6.6 mm, dorsoventral −7.1 mm. The cannulas were secured to the skull using dental acrylic. Obturators, cut to sit flush with the tip of the guide cannulas and with an outer diameter of 0.36 mm, were inserted into the guides and remained there until the first infusions. At the completion of each surgery, an antibiotic was applied for 3 days (Enrofloxacin; 0.27 mg kg-1, Vetanco, Arg). Animals were given approximately 7 days to recover before behavioral testing and drug infusions.

A Tat-conjugated peptide, designed to block the binding of dynamin to amphiphysin and thus prevent endocytosis (Gout et al., 1993; Lissin et al., 1998; Kittler et al., 2000; Lin et al., 2011), was infused in order to block endocytosis. Depending on the experiment, rats received bilateral infusions of Tat P4 peptide and Tat Scrambled control peptide (S) (60μg/μl/0.5 μl side; Cambridge, UK), human recombinant BDNF or saline (0.5 μg/μl/0.5 μl side), ANA-12 or saline at different times during the behavioral task. The injection volume was always 0.5 μl/side. Sequences are as follows: amino acid sequence for the dynamin inhibitory peptide (P4) is QVPSRPNRAP, and for the Scrambled control peptide (S) is QPPASNPRVR.

Bilateral infusions were conducted simultaneously using two 5-μl Hamilton syringes that were connected to the infusion cannulas by propylene tubing. Syringes were driven by a Harvard Apparatus precision syringe pump, which delivered 0.5 μl to each hemisphere over 1 min. The infusion cannulas were left in place for an additional minute to allow for diffusion. At least 3 days were allowed for washout between repeated infusions.

For the behavioral procedures, an open triangular acrylic field was used as an arena, each wall 60 cm long by 60 cm high. The walls of the triangular open field were higher to minimize the visual access to the distal cues in the room. The arena was located in the middle of a room with dim lighting, and the floor was always covered with wood shavings. A video camera was positioned on the arena in order to record both the sample session and the evaluation session for later analysis. The objects to be used were created by attaching together two small objects, depending on the condition to be studied, similar or dissimilar. Different objects were used for our within-subject design, all of them made of different materials, such as metal, glass or plastic. All of the object were approximately between 8 and 15 centimeters tall, and 4 to 7 centimeters width. All objects were adhered to the open field floor with reusable adhesive putty and cleaned with 50% ethanol solution between sessions, both sample and choice phases. For the task, the three composite objects were aligned closely to one of the arena walls and the position of each object was counterbalanced.

For the Spontaneous Object Recognition (SOR) task (Figure 1A), each rat was handled for 3 days and then habituated to the arena for 5 min per day for 3 days before exposure to the objects. After habituation, the rats were exposed during a 5 min duration sample phase to three objects made of two features depending on the condition. For the similar condition, two of the objects shared one feature (AB and BC) and the third object was made of two other different features (EF). For the dissimilar condition, all three objects were made of different features (AB, CD, and EF). Twenty-four hours after sample phase, we conduct a choice phase, of 3 min duration, in which the animals were exposed to two objects, a novel one and a familiar one, and depending on the condition evaluated, the objects varied in composition. For the similar condition, the novel object was made of the two non-shared features of the objects presented in the sample phase (AC), and the familiar object was a copy of the third object (EF). For the dissimilar condition, the novel object was made of two novel features (KI) and the familiar object was a copy of the object presented during the sample phase (AB, CD, and EF).

For the extra-similar condition, the process was the same as the similar condition, differing only in the objects used. During sample phase, animals were exposed during 5 min to three different objects, and two of those shared one feature (ABB and BBC), while the third object was made of different features (EFG). During the choice phase, 24 h later, the animals were exposed to a novel object was made of a novel combination of familiar features (ABC), and the familiar object was a copy of the third object presented in the sample phase (EFG) (Figure 3D). Exploration was recorded and later scored manually for both the sample and choice phases. For all experiments, exploration of a particular object was defined as the rat having its nose directed at the object at 2 cm or less or touching the object with its nose. Rearing with the head oriented upward did not count as exploration. Climbing over or sitting on the object was not included.

In every trial, objects were pseudorandomly assigned to a different location in the arena to avoid a bias for locations within the arena.

For all the experiments, the results were expressed as a discrimination ratio calculated as the time exploring the novel object minus the time exploring the familiar object divided by the total exploration time [(tnovel-tfamiliar)/(ttotal)].

For the sample phases, the percentage of time spent exploring each object was compared using a repeated-measures two-way ANOVA or one-way ANOVA. For the choice phases, we performed one-sample t-test for every discrimination ratio to analyze whether control animals learned the task, verifying that the ratio was different from zero. Discrimination ratios were compared within subject using a paired t-test, one way ANOVA or two-way ANOVA.

Some experiment has female and male animals, and the statistical analyses was made with pooled data, except for the first experiment (Figure 1). The experiments in which there are animals of both sexes, the number of animals of each sex is detailed in the results.

It has already been shown that object exploration and preference is driven by novelty in the modified version of the SOR task in male rats (Miranda et al., 2017). This task includes a similar and a dissimilar condition in which the load on pattern separation is different (Figure 1A; see M and M). The first goal of this work was to test the performance of female rats in the same SOR task version and compare the result with male rats (Figure 1).

Six male Long Evans and 6 female Long Evans were used for this experiment, all animals underwent both the similar (s-SOR) and dissimilar (d-SOR) conditions. We did not find a difference in the percentage of time the animals spend exploring the objects during the sample phase {repeated measures one way ANOVA: females s-SOR [F (1.588, 7.942) = 1.829, p = 0.2211]}; males s-SOR [F (1.147, 5.734) = 3.364, p = 0.1171]; females d-SOR [F (1.457, 7.287) = 3.202, p = 0.1081]; males d-SOR [F (1.438, 7.190) = 2.240, p = 0.1779)]. There was no significant interaction between the objects by condition and sex (repeated measures two-way ANOVA: psex = 0.6109, pcondition + object = 0.0745, pinteraction = 0.2793).

Also, we did not find any differences in total exploration times (female s-SOR = 39.78 ± 5.09, male s-SOR = 39.63 ± 5.21, female d-SOR = 30.08 ± 3.94, male d-SOR = 36.88 ± 1.92) comparing sexes in the same condition (paired t-test: female versus male similar, p = 0.9840; female versus male dissimilar, p = 0.1518).

The choice phase was conducted 24 h after the sample phase for both conditions and memory was evaluated by comparing the amount of time spent exploring a novel object and familiar object. This comparison was expressed in the discrimination ratio. There were no significant differences between the discrimination ratio for both sexes in the similar and the dissimilar condition (pconditions = 0.0819, psex = 0.6047, pinteraction = 0.4822) (Figure 1B, Supplementary Table 1). Also, we did not find any differences in total exploration times comparing sexes in the same condition (paired t-test: female versus male similar, p = 0.9840; female versus male dissimilar, p = 0.1518) (Table 1). One-sample t-tests against a value of zero indicated that both sexes were able to learn the task in both conditions [One-sample t-test (s-SOR females, t = 6.146), p = 0.0017; one- sample t-test (s-SOR males, t = 6.740), p = 0.0011; d-SOR females, t = 3.277), p = 0.0220; one- sample t-test (d-SOR males, t = 5.128, p = 0.0037)].

These results indicate that intact female rats were able to spontaneously disambiguate the representations of two similar object seen 24 h before, and that there is no difference between sexes for performance in this specific task.

We then proceeded to study the role of the endocytosis in the formation of differentiated representations in the Prh in male and female rats. If mechanisms of receptor internalization are specifically required in the Prh for memory differentiation, then blocking internalization should alter the similar version of the SOR task, without affecting the dissimilar version. To test this hypothesis, we blocked the putative receptor internalization immediately after the sample phase. We used a Tat-conjugated peptide (Tat-P4) and a scrambled control peptide (Tat-S) design to prevent endocytosis by blocking the binding of dynamin (Gout et al., 1993; Lissin et al., 1998; Lin et al., 2011). Tat-P4 or Tat-S was injected in Prh immediately after the sample phase, and the memory of animals was test 24 h later in both conditions, similar (n = 12, 6 male and 6 female) and dissimilar (n = 11, 5 male and 6 female) (Figures 2A, D). Animals from both sexes were pooled and analyzed altogether, and all animals underwent the experimental (Tat-P4) and the control (Tat-S) conditions. A discrimination index above zero indicates a significant discrimination and a reasonable memory retention. One-sample t-tests against a value of zero indicated that Tat-S injected animals were able to learn both the s-SOR and the d-SOR [One-sample t-test (d-SOR Tat-S, t = 7.735), p < 0.001; one- sample t-test (s-SOR Tat-S, t = 3.071), p = 0.0106] (Figures 2C, F), whereas Tat-P4 injected animals only learned the d-SOR version [One-sample t-test (d-SOR Tat-P4, t = 4.655), p = 0.0009; one- sample t-test (s-SOR Tat-P4, t = 1.142), p = 0.2776)] (Figures 2C, F). We found a significant difference between Tat-S and Tat-P4 injected animals in the choice phase in the s-SOR [Tat-S vs. Tat-P4 paired t-test (t = 2.899), p = 0.0145, n = 12] (Figure 2C). There were no differences in total exploration times between groups (see Table 1). These results indicate that endocytosis is important to spontaneously disambiguate the memory representations of two similar objects.

Memory consolidation is a process that occurs during a restricted time window (McGaugh, 2000; Winters and Bussey, 2005b). To test whether Tat-P4 interfered with memory during a restricted delay after the sample phase, Tat-P4 or Tat-S was injected into the Prh 5 h after the sample phase, and rats were tested 24 h later, all animals underwent both drug conditions (n = 7, 2 males, 5 female) (Figure 3A). Since in the previous experiment we found an effect only in the similar condition, we decided to test this time window in this specific condition. Injection of the TatP4 did not change total exploration times compared with Tat-S (see Table 1). We did not observe any memory impairments in the s-SOR when Tat-P4 was injected 5 h after sampling the objects [One-sample t-test (Tat-S, t = 13.88), p < 0.0001; one- sample t-test (Tat-P4, t = 7.377), p = 0.0003], indicating that Tat-P4 injected animals were able to learn the similar condition as successfully as Tat-S-injected animals [Tat-S vs. Tat-P4 paired t-test (t = 0.2985), p = 0.7754, n = 7] (Figure 3C). In sum, this result indicates that extending the time interval between sample phase and Tat-P4 infusion reduces the disrupting effect of the peptide on the choice phase.

Brain derived neurotrophic factor enhances memory consolidation in several tasks if injected exogenously (Alonso et al., 2002; Peters et al., 2010; Bekinschtein et al., 2013). We wondered if the enhancing effect found in previous studies could be prevented if we blocked the internalization of receptors. We have previously shown that BDNF in Prh is critical for this task. In particular, injection of exogenous BDNF into Prh enhanced discrimination of similar object memories (Miranda et al., 2017). To be able to see any memory improvement induced by BDNF in pattern separation, we used a slightly different version of the SOR task, the extra-similar SOR (xs-SOR) (Miranda et al., 2017) in which we make discrimination more difficult by bringing the performance of the control animals down. The key modification to the task is making the objects more similar during the sample phase. In this particular experiment, we used three groups of rats, a group injected in Prh with Tat-S and saline 15 min prior to sample phase, a second group injected with Tat-S 15 min prior and human recombinant BDNF (hrBDNF) immediately after the sample phase, and a third group injected with Tat-P4 15 min prior and hrBDNF immediately after sample phase. All animals were exposed to the three treatments (n = 12, all males) (Figure 3D).

There were no differences in exploration of the three objects during the sample phase (repeated-measures two-way ANOVA, xs-SOR: F = 1.000, pdrug = 0.1748; F = 0.1089 pobjects = 0.8668, F = 1.191, pinteraction = 0.3282) (Figure 3E). A one-way between-subjects ANOVA was conducted to compare the effect of the treatments in the choice phase. There was a significant difference between treatments at the p < 0.05 level [F (Rolls and Kesner, 2006; Miranda et al., 2017) = 3.838, p = 0.043]. Post hoc comparison using Tukey’s multiple comparison test indicated that the mean score for the Tat-S/Veh condition (M = −0.02, SEM = 0.037) was significantly different than the Tat-S/BDNF condition (M = 0.11, SEM = 0.036) and different than the Tat-P4/BDNF condition (M = 0.08, SEM = 0.038). However, the Tat-S/BDNF condition did not significantly differ from the Tat-P4/BDNF condition (Figure 3F). This experiment was inconclusive, as it appears that BDNF was not able to improve discrimination in Tat-P4-injected animals, but the discrimination index did not differ from that of the Tat-S/BDNF group.

We wanted to use a different strategy to evaluate the possible interaction between BDNF and endocytosis by blocking both the BDNF receptor TrkB and endocytosis in a molecular disconnection experiment (Miranda et al., 2017). We first tested the effect of Prh injection of ANA-12, a selective non-competitive antagonist of TrkB, BDNF receptor (Cazorla et al., 2011). We injected all animals with ANA-12 or saline in Prh 15 min prior to the sample phase and evaluated memory 24 h after training in both similar and dissimilar version (Figures 4A, D). There were no differences in total exploration times in neither the similar (n = 11, all female) nor the dissimilar (n = 10, all female) version of the task (paired t-test, similar, p = 0.6742; dissimilar, p = 0.8804) (Table 2). We did not find any interactions between drugs or objects (repeated-measures two-way ANOVA, d-SOR: pdrug = 0.1934, pobjects = 0.1557, pinteraction = 0.4504; s-SOR: pdrug = 0.3409, pobjects = 0.7420, pinteraction = 0.2839) (Figures 4B, E). We observed a memory impairment in the s-SOR version of the task (paired t-test, p = 0.0005, t = 4.995) (Figure 4C), but not in the d-SOR version (paired t-test, p = 0.7462, t = 0.3337) (Figure 4F), Thus, blocking TrkB only generated a deficit in the “similar” condition, disabling animal’s capacity of discrimination of overlapping memories.

We next evaluated whether BDNF pathway and endocytosis interacted during consolidation of similar object memories in the SOR task. We used a protocol of molecular disconnection that we have carried out in previous studies (Miranda et al., 2017). The logic underlying this is the same that in any brain disconnection experiment that tries to determinate if, during a specific behavioral manipulation, two brain structures are connected (Gaffan et al., 1989; Ito et al., 2008). If we assume that the principal connection between two structures is in the same hemisphere (ipsilateral), the deactivation or lesion of the two regions in the same side will keep the behavior intact, but the contralateral deactivation will affect the performance. If we consider two molecules or gene expression pathways in specific given structure instead of two regions, a similar method of reasoning can be used. If two molecular pathways interact to produce a specific behavior, blocking both pathways in that area of only one hemisphere will have no effect; but if one pathway is blocked in one hemisphere and the second pathway in the other, we would see a deficit. Thus, we evaluated if BDNF and endocytosis interacted in Prh during consolidation of similar memories by blocking both pathways in the same hemisphere or blocking BDNF in one hemisphere and endocytosis in the other. We injected ANA-12/Veh or Tat-P4/Tat-S in Prh 15 min before the sample phase and evaluated memory 24 h after it (n = 8, all female) (Figure 5A). All animals underwent both treatment conditions. There were no differences in total exploration time between the two groups (paired t-test, p = 0.5426, t = 0.6399, n = 8), nor between objects (repeated-measures two-way ANOVA, s-SOR: pcondition = 0.7318, pobject = 0.1943, pinteraction = 0.4355) (Figure 5B).

We found no effect in the similar SOR task when ANA-12 and TatP4 were injected in the same hemisphere (and their corresponding vehicles were injected in the other hemisphere) (Figure 5C), however, when ANA-12/TatS and Veh/TatP4 where injected in different hemispheres in Prh, we found a significant impairment in the similar SOR task (Figure 5C) (paired t-test unilateral vs. contralateral, p = 0.0090, t = 3.574). There were no differences in total exploration times between the two groups (see Table 1). One sample t-test showed that the discrimination ratio from the “unilateral” group was different from zero, whereas the discrimination ratio from the “contralateral” group was not (p_unilateral = 0.0210, t = 2.96; p_{contralateral} = 0.0686, t = 2.150). This result suggests that BDNF and endocytosis interacts during consolidation of similar memories in Prh.