Knockdown of Nurr1 in the rat hippocampus: implications to spatial discrimination learning and memory

Abstract

Nurr1 expression is up-regulated in the brain following associative learning experiences, but its relevance to cognitive processes remains unclear. In these studies, rats initially received bilateral hippocampal infusions of control or antisense oligodeoxynucleotides (ODNs) 1 h prior to training in a holeboard spatial discrimination task. Such pre-training infusions of nurr1 antisense ODNs caused a moderate effect in learning the task and also impaired LTM tested 7 d later. In a second experiment, ODN infusions were given immediately after the animals had received two sessions of training, during which all animals showed normal learning. Although antisense treated rats were significantly impaired during the post-infusion stages of acquisition of the task, no group differences were observed during the LTM test given 7 d later. These animals were subjected 3 d later to reversal training in the same maze in the absence of any additional treatments. Remarkably, rats previously treated with antisense ODNs displayed perseveration: The animals were fixated with the previously learned pattern of baited holes, causing them to be significantly impaired in the extinction of acquired spatial preferences and future learning. We postulate that Nurr1 function in the hippocampus is important for normal cognitive processes.

Figure 1.

Experimental design for pre- and mid-training hippocampal Nurr1 knockdown. (A) Experiment 1. Effects of pre-training nurr1 antisense hippocampal infusions on acquisition and retention of spatial learning. Animals were able to freely explore the baited maze 1 d before spatial training. On the training day, bilateral infusions of saline, sense, or antisense ODNs were directed to the CA3 region 1 h before training. Animals received four training sessions each separated by a 1-h rest period. One group of rats was sacrificed immediately after the fourth session, and brains were used for Nurr1 immunohistochemical analysis, while a second group of rats was subjected to a retention test 7 d post-training to assess LTM. (B) Experiment 2. Effects of mid-training nurr1 antisense hippocampal infusions on acquisition and retention of spatial learning. Animals were habituated as in Experiment 1. However, on the training day, animals received two training sessions, followed immediately by bilateral infusions of sense or antisense ODNs to the CA3 region. Training was continued as in Experiment 1. A group of animals was sacrificed immediately after the fourth session, and brains were used for immunoblotting analysis. A separate group of animals was subjected to a retention test 7 d post-training to assess LTM. (C) Experiment 3. Long-term impact of mid-training nurr1 antisense hippocampal infusions on spatial learning and memory. Rats in Experiment 2 were allowed to rest for 3 d after their retention test. These rats were then subjected to a reversal task requiring them to learn a new pattern of food location within the same maze, but without additional ODN treatment. Animals received four training sessions, each separated by a 1-h rest period. A retention test was given 7 d after reversal training.

Figure 2.

In vivo brain diffusion studies and in vitro studies of antisense ODN molecular efficiency. (A) Verification of cannulae placements. Rats were decapitated immediately after the end of all behavioral treatments. Brains were dissected, serial coronal sections were obtained, and these were subjected to Nissl staining with thionin. Drawings represent the area of −3.3 mm from bregma. The dots represent the estimated sites of cannula placements. The analysis presented includes representative data from our studies with CA3 sense ODN-treated rats (N = 7) and CA3 antisense ODN-treated rats (N = 7). (B) The composite photomicrograph depicts the infused fluorescein isothiocyanate-labeled ODN (FITC-ODN) detected at 3 h post-injection within CA3 pyramidal cells of the dorsal hippocampus (white arrows). Occasional diffusion extended to the DG, a brain region shown to display low Nurr1 protein expression. Red and yellow arrows display little or no diffusion within the CA1 or CA2 regions, respectively. (C) Alignment of the designed nurr1 antisense ODN sequence shows a perfect match with rat nurr1 mRNA (shaded area), but not for the other two members of the Nur family, ngfi-b and nor-1. The start codon on each sequence is in bold. (D) In vitro nurr1 antisense ODN treatment in cotransfected AD293 kidney cells. Nurr1 antisense ODN-treated cells had lower levels of luciferase activity than sense ODN-treated cells (*P < 0.05). Luciferase activity was normalized to Renilla luciferase activity. Data are presented as the means ±SEM (bars) of independently transfected cultures (N = 6 transfections).

Figure 3.

Results of in vivo studies addressing the effectiveness of nurr1 antisense ODN. (A) Bar graphs demonstrate Nurr1 immunohistochemical analysis of pre-training ODN infused rats. Antisense ODN treatment significantly decreased the number of immunopositive nuclei in the CA3 region compared to the sense ODN-infused rats (**P < 0.01). (B) Bar graph (mean ±SEM) depicting the results of Nurr1 immunoblotting analysis of mid-training ODN-infused rats normalized to β-Actin. The graph displays reduced Nurr1 levels within protein extracts prepared from tissue adjacent to the injection site of rats infused with antisense ODN, compared to sense ODN-treated rats (*P < 0.05). (C) Bar graph of β-Actin normalized c-Fos immunoblotting analysis revealed no significant differences in protein expression between the nurr1 antisense and sense ODN-treated rats. (D) Representative Western blot showing (top) Nurr1 (66 kDa), (middle) c-Fos (60 kDa), and (bottom) β-Actin (42 kDa) levels in dorsal hippocampus extracts of sense and antisense treated rats. Both β-Actin and c-Fos showed no significant differences in protein expression between the ODN antisense and sense treated rats. Results show that the antisense treatment reduced Nurr1 levels without affecting c-Fos or actin expression.

Figure 4.

Pre- and mid-training microinfusions of antisense ODNs into hippocampal CA3 impair both the acquisition and LTM of spatial discrimination. Only the data for animals subjected to both acquisition and retention are shown. (A) Plot depicting the decreases in searching times during acquisition and retention for rats subjected to pre-training microinfusions of either saline, sense, or antisense ODNs. No overall difference was detected in latency to complete the task between groups throughout acquisition or retention. (B) Plot depicting the number of errors committed by rats in the three groups throughout acquisition and retention. Overall, antisense ODN-treated rats committed significantly more errors compared to sense ODN- or saline-treated animals during acquisition and retention (Two-Way ANOVA with repeated measures, *P < 0.05), although no specific group differences were identified in particular sessions when using multiple comparison testing. (C) Behavioral analysis of rats receiving mid-training microinfusions confirms that sense and antisense ODN-treated animals displayed comparable spatial learning during sessions 1 and 2 (prior to ODN treatment). Antisense ODN-treated rats committed significantly more errors than sense ODN-treated rats at the latter stages of acquisition (Two-Way ANOVA with repeated measures, *P < 0.05). Post hoc analysis identified specific differences between the groups during session 4 of acquisition (*P < 0.05). No significant differences between the groups were seen during the retention test. (D) Bar graph depicting the results of an LTM test for intact rats subjected to one or two training sessions during acquisition. The rats that received two sessions of training committed significantly fewer errors during their retention test than the rats that received only one training session (Student t-test, **P < 0.005). Data are presented as the ±SEM of rats trained with one (gray bar) or two (black bar) sessions.

Figure 5.

Effects of mid-training CA3 knockdown of Nurr1 on reversal learning. (A) Plot depicting the difference in reversal learning and memory for rats treated with sense or antisense ODNs during their first spatial learning experience. The results show that animals previously receiving antisense ODN treatment were significantly less accurate than those receiving sense ODN treatment during reversal learning and retention test (Two-Way ANOVA with repeated measures, *P < 0.05). (B) Bar graphs display the perseveration index (relative preference to use old spatial pattern of baited holes vs. the new pattern) for animals of each group during the retention test of the reversal-learning task. Rats that previously received the antisense ODN treatment had a significantly higher preference for the old pattern of baited holes compared to those that previously received sense ODN treatment (*P < 0.05). Data are presented as the ± SEM of sense (gray bar) or antisense (black bar) treated rats. (C) Representative trajectories of hole visits during the retention test of the reversal task for sense ODN and antisense ODN treated rats, respectively. The holes baited in the reversal protocol are shown in gray, whereas those baited during the first training experience contain an inner black circle.