Limbic progesterone receptors regulate spatial memory

Abstract

Progesterone and its receptors (PRs) participate in mating and reproduction, but their role in spatial declarative memory is not understood. Male mice expressed PRs, predominately in excitatory neurons, in brain regions that support spatial memory, such as the hippocampus and entorhinal cortex (EC). Furthermore, segesterone, a specific PR agonist, activates neurons in both the EC and hippocampus. We assessed the contribution of PRs in promoting spatial and non-spatial cognitive learning in male mice by examining the performance of mice lacking this receptor (PRKO), in novel object recognition, object placement, Y-maze alternation, and Morris-Water Maze (MWM) tasks. In the recognition test, the PRKO mice preferred the familiar object over the novel object. A similar preference for the familiar object was also seen following the EC-specific deletion of PRs. PRKO mice were also unable to recognize the change in object position. We confirmed deficits in spatial memory of PRKO mice by testing them on the Y-maze forced alternation and MWM tasks; PR deletion affected animal's performance in both these tasks. In contrast to spatial tasks, PR removal did not alter the response to fear conditioning. These studies provide novel insights into the role of PRs in facilitating spatial, declarative memory in males, which may help with finding reproductive partners.

Figure 1

PR expression in the EC and hippocampus of adult male mice. (A–E) Images from representative Pgr-Cre animals injected with AAV9 expressing a flexed-tdTomato. The AAVs were injected into the hippocampus and EC. The sections were counterstained with a neuronal marker protein NeuN (Green) to label neurons. The tdTomato labeled terminals of the perforant path (white arrows), temporoammonic pathway (white arrowheads), and temporoammonic Alvear path fibers (yellow arrowheads) are also marked in 1C. (a1, a2, b1–e1) Magnified boxed regions from A-E showing neuronal tdTomato expression. (F) PR mRNA expression in the EC and hippocampus (Hp) of male mice, n = 7 for EC and 5 for Hp. The PR mRNA expression (1.12 ± 0.29, n = 7) in the uterine tissue isolated from females in the estrus stage of the cycle was used for normalization (dotted black line). The expression of PR mRNA in the EC and Hp tissue isolated from female mice in the estrus stage is shown for comparison (n = 5 EC and 7 Hp).

Figure 2

PR expression in the principal neurons vs interneurons. (A–C) Representative images of the dentate hilus of Pgr-Cre mice injected with AAV expressing flex tdTomato. The images show colocalization of tdTomato expression with that of the immunoreactivity of the GluA2 subunits of AMPA receptors, parvalbumin (PV), and somatostatin (Som) neuropeptides respectively. The expression of the GluA2 subunit marked the hilar mossy cells, whereas, parvalbumin and somatostatin labeled two types of interneurons. The scale bars in these and all other images in this figure correspond to 50 μm. The hilar neuron marked by an arrow is magnified in the inset in (A) to show punctate GluA2 subunit immunoreactivity in a tdTomato-expressing neuron. The arrow heads in (B, C) mark the parvalbumin and somatostatin-expressing neurons respectively, none of which colocalized with tdTomato fluorescence. (D–F) Representative images showing colocalization of tdTomato with that of CamKII, PV, and Som respectively in the DGCs. The inset corresponds to the boxed region and shows perisomatic CamKII immunoreactivity in the tdTomato-positive DGCs. The arrows in (E, F) point to the interneurons. (G–I) Colocalization of tdTomato and CamKII, PV, and Som in the CA1. The boxed region in (G) is magnified in the inset, arrows in (H, I) point the interneurons. (J–L) Colocalization of tdTomato with CamKII, PV, and Som expression respectively. The boxed region in (J) is magnified as an inset, arrows in (K, L) point to the interneurons. (M–O) Colocalization of tdTomato in the medial EC with CamKII, PV, and Som respectively. Please note that som labeling in the superficial layers of EC was sparse, but som-positive neurons were present in the deeper layers (marked by arrows). Also please note that the number of som-expressing neurons seems to be substantially lower than those expressing PV.

Figure 3

The effect of PR activation on the number of active neurons in the EC and hippocampus of adult male mice. (A, B) Representative images from segesterone- (10 mg/kg, subcutaneous) and vehicle- (20% β-hydroxycyclodextrin) treated male mice respectively. 4OHT was administered 2 days after the treatment in home cages to TRAP the active neurons (red). The green fluorescence corresponds to the expression of neuronal marker protein NeuN. (A′, B′) Magnified boxed areas from A and B respectively. The CA1, subiculum (sub), central (cEC), medial (mEC) and dorsolateral (dlEC) EC regions are marked. (C, D) Representative images from segesterone- and vehicle-treated mice show active DGCs, the NeuN immunoreactivity is shown in green. (C′, D′) Images showing the TRAPed DGCs in the segesterone- and vehicle-treated animals. (E, G) and (F, H) Images showing the TRAPed neurons in the septum, paraventricular thalamic nucleus, and colliculus of segesterone- and vehicle-treated animals respectively. (I–K) The number of TRAPed neurons in the entorhinal cortex, combined CA1 and subiculum, and DGCs in segesterone- and vehicle-treated animals, n = 5 each, **p = 0.0098 for EC, p = 0.008 for CA1 and subiculum, and p = 0.0094 for DGCs, student’s t-test.

Figure 4

The effect of PR deletion on novel object recognition in adult male mice. (A) A schematic showing habituation, familiarization, and testing of animals in a novel object recognition test. The assay was performed as described in the methods section. (B) A heat map of the time spent by representative PR knockout (PRKO) and littermate wildtype (WT) mice during the testing phase performed 8 h after the familiarization. The yellow arrows denote the position of novel objects. (C) Discrimination index in the WT and PRKO mice, n = 7 WT and 6 PRKO, *p = 0.023, student’s t-test. (D) The average and SD of the percent novel object exploration time evaluated after a delay of 8 h, n = 7 WT and 6 PRKO, *p = 0.0167, one-sample t-test. The black line denotes 50%, corresponding to chance performance. (E) Discrimination index in the WT and PRKO mice, ***p = 0.0006, student’s t-test. (F) The percent novel object exploration time evaluated after one-hour delay, n = 7 WT and 10 PRKO, *p = 0.0106, one sample t-test for WT and p = 0.0311, one sample t-test for PRKO. The black line denotes 50%, which is chance performance. (G) Movement of representative WT and PRKO mice in the open-field arena. (H) Mean and standard deviation of the total exploration time in the WT and PRKO mice during familiarization and (1-h delay) testing phases. (I) Mean and standard deviation of the total distance traveled by 7 WT and 10 PRKO mice. (J) Discrimination index in the vehicle- and RU-486-treated mice, *p = 0.016, student’s t-test. The animals were treated daily with vehicle (20% β-hydroxycyclodextrin daily subcutaneous, sc) and RU-486 (10 mg/kg, sc) for a week. The experiment was performed a day after the last injection. There was a one-hour delay between the familiarization and testing phases. (K) The percentage novel object exploration of the vehicle- and RU-486-treated animals, n = 9 vehicle-treated and 7 RU-486-treated, *p = 0.026, one sample t-test for WT mice. The black line denotes 50%, corresponding to performance by chance. (L) Average and standard deviation of the total exploration time in the WT and PRKO mice during familiarization and testing phases.

Figure 5

The effect of PR deletion on object place recognition and EC-specific deletion on novel object recognition. (A) The schematic showing object place recognition testing. (B) Heat map from representative WT and PRKO mice during the testing phase. The arrows mark the displaced object; purple squares mark their original position. The animals were tested after an hour. (C) The percent time spent in exploring the displaced object, n = 8 each, *p = 0.016 student’s t-test. (D) Images from representative animals injected with AAVs to express CamKII-driven Cre-GFP or GFP alone. Sections from the ventral and dorsal EC cortex are shown, the green fluorescence corresponds to Cre-GFP or GFP expression, and DAPI (blue) was used as a counterstain. (E) Discrimination indices in the Cre-GFP and GFP mice, *p = 0.0212, student’s t-test. (F) Percent novel object exploration in the animals injected with AAVs to express CamKII driven Cre-GFP or GFP alone, n = 8 for GFP alone and 12 for GFP-Cre, *p = 0.017 one sample t-test for Cre-GFP mice. The black line marks a chance performance 50%.

Figure 6

Effect of PR deletion on Y maze forced alternation and Morris Water Maze tasks. (A) A schematic of the Y maze forced-alternation test. (B) The number of entries in the novel arm by PRKO and WT males in a Y maze forced-alternation task, n = 6 each, *p = 0.0074, ANOVA with Sidák's multiple comparison WT familiar vs WT novel. (C) The time spent exploring the novel and familiar arms. (D) Latency to reach the platform during training, n = 8 each. (E) Track plots from representative WT and PRKO mice during the probe trial on day 6. The arrow marks the quadrant that held the escape platform during training sessions. Please note that the WT animal kept swimming in the region around the location of the platform (marked by the circle) during the probe trial whereas, the PRKO mouse swam in all the quadrants. (F) The time spent in the target quadrant during the probe trial, *p = 0.0277, student’s t-test.

Figure 7

The cued-contextual fear conditioning in PRKO male mice. (A) A schematic showing fear conditioning and recall. Following an initial 3 min of chamber exploration, the animals received five 2-s footshocks that were contiguous with a 20-s tone. The animals were removed from the chamber two min after the last footshock. (B) Percent freezing in the WT and PRKO male mice following the five footshocks, n = 6 WT and 7 PRKO. (C) Percent freezing during the context recall was performed 24 h later. The number of replicates is the same as in E.

Figure 8

Validation of the qRT-PCR assay. (A) A PCR was performed using primers amplifying a fragment of PR and GAPDH using cDNA as a template. The PCR reaction generated a single band of the expected size in the WT animals; an amplification product was not seen in the PRKO animals. (B) Melting curves of PR and GAPDH qPCR from representative PRKO and WT mice. (C) Efficacy of amplification of PR and GAPDH primers. (D) Western blots of brain proteins using two anti-PR antibodies. The expression of β-actin was used as a loading control. Molecular weight markers are shown in lane M. Please note that the antibodies reacted with multiple proteins suggestive of lower specificity. The β-actin blot is cropped to show only the signal, full blot is included as a supplementary file.