Estrous Cycle-Dependent Phasic Changes in the Stoichiometry of Hippocampal Synaptic AMPA Receptors in Rats

Abstract

Cognitive function can be affected by the estrous cycle. However, the effect of the estrous cycle on synaptic functions is poorly understood. Here we show that in female rats, inhibitory-avoidance (IA) task (hippocampus-dependent contextual fear-learning task) drives GluA2-lacking Ca2+-permeable AMPA receptors (CP-AMPARs) into the hippocampal CA3-CA1 synapses during all periods of the estrous cycle except the proestrous period, when estrogen levels are high. In addition, IA task failed to drive CP-AMPARs into the CA3-CA1 synapses of ovariectomized rats only when estrogen was present. Thus, changes in the stoichiometry of AMPA receptors during learning depend on estrogen levels. Furthermore, the induction of long-term potentiation (LTP) after IA task was prevented during the proestrous period, while intact LTP is still expressed after IA task during other period of the estrous cycle. Consistent with this finding, rats conditioned by IA training failed to acquire hippocampus-dependent Y-maze task during the proestrous period. On the other hand, during other estrous period, rats were able to learn Y-maze task after IA conditioning. These results suggest that high estrogen levels prevent the IA learning-induced delivery of CP-AMPARs into hippocampal CA3-CA1 synapses and limit synaptic plasticity after IA task, thus preventing the acquisition of additional learning.

Fig 1. Acquisition of IA learning during the estrous cycle.

Latency before entering the dark side of the box was consistently longer in IA-conditioned rats. (A) Rats successfully acquired IA memory throughout the estrous cycle. ProE (proestrous), n = 4; E (estrous), n = 6; D1 (diestrous day 1), n = 5; D2 (diestrous day 2), n = 6. (B) OVX rats successfully acquired IA memory with (n = 5) or without (n = 5) estrogen (E2) priming. *P < 0.05, unpaired Student’s t test. Error bars represent SEM.

Fig 2. Changes in learning-driven synaptic delivery of CP-AMPARs in CA1 pyramidal neurons during the estrous cycle.

(A) There was no difference in rectification in between untrained (n = 13) and IA-trained (n = 32) ProE rats. However, rectification increased notably after IA conditioning during other periods of the estrous cycle (untrained/ IA-trained: E, n = 11/ 32; D1, n = 7/ 16; D2, n = 9/ 28). (D) Rectification increased after IA conditioning in OVX rats that were not primed with E2 (untrained, n = 19; IA-trained, n = 16), but rectification did not change after IA conditioning in the presence of E2 (untrained, n = 13; IA-trained, n = 11). Scale bars: vertical, 20 pA; horizontal, 20 ms. (B, E) GluA1 and (C, F) GluA2 content in the synaptoneurosome fraction obtained from the rat dorsal hippocampus in ProE (untrained, n = 6; IA-trained, n = 6), D2 (untrained, n = 11; IA-trained, n = 16), and OVX rats with (untrained, n = 7; IA-trained, n = 7) or without E2 priming (untrained, n = 7; IA-trained, n = 7). *P < 0.05; Mann-Whitney U test (A, D) or unpaired Student’s t test (B-F). Error bars represent SEM.

Fig 3. LTP was attenuated after IA task in ProE rats.

(A, B) LTP was induced at hippocampal CA3-CA1 synapses from untrained ProE (n = 8) and D2 (n = 6) rats, and from (C, D) IA-trained D2 rats (n = 6), but not in synapses from IA-trained ProE rats (n = 6). The mean amplitude 30–40 min after LTP induction was normalized to the baseline amplitude. *P < 0.05, unpaired Student’s t test. Error bars represent SEM.

Fig 4. E2 priming prevented LTP after IA task in OVX rats.

(A, B) LTP was induced at hippocampal CA3-CA1 synapses from untrained OVX rats with (n = 8) or without (n = 6) E2 priming. (C, D) LTP was also induced in IA-trained OVX rats without E2 (n = 8), but not in IA-trained OVX rats primed with E2 (n = 7). The mean amplitude 30–40 min after LTP induction was normalized to the baseline amplitude. *P < 0.05, unpaired Student’s t test. Error bars represent SEM.

Fig 5. Reduced IA-induced mushroom-spine formation in CA1 pyramidal neurons in ProE rats and E2-primed OVX rats.

(A, B) The number of each type of dendritic spine in untrained and IA-trained rats during different periods of the estrous cycle. IA-learning tasks significantly increased the number of mushroom-shaped spines on CA1 pyramidal neurons of D2 rat dendrites (untrained, n = 27; IA-trained, n = 25) but not in ProE rat dendrites (untrained, n = 17; IA-trained, n = 15). (C, D) The number of each type of dendritic spine in untrained and IA-trained rats with or without E2 priming. IA-learning tasks significantly increased the number of mushroom-shaped spines on the dendrites of unprimed OVX rats (untrained, n = 31; IA-trained, n = 44), but not on those of E2-primed OVX rats (untrained, n = 18; IA-trained, n = 20). *P < 0.05, unpaired Student’s t test. Error bars represent SEM. (Bottom) Representative photomicrographs (×400) of tertiary apical dendrites of CA1 pyramidal neurons. Black arrowheads indicate mushroom-shaped spines; white arrowheads indicate filopodial spines. Scale bar: 3 μm.

Fig 6. Prior IA-task experience impaired hippocampus-dependent learning in ProE rats.

(A) Percentage of spontaneous alternation in the Y-maze task before and after IA training. The alternation rate did not change after IA learning in the D2 period (untrained, n = 8; IA-trained, n = 8), but the alternation rate was reduced after IA learning during the ProE period (untrained, n = 6; IA-trained, n = 10). (B) Latency in the second IA learning session after the initial IA training. The latency to re-enter the dark box in the second IA task was shorter during the ProE period (n = 5) than in the D1 period (n = 4). *P < 0.05, unpaired Student’s t test. Error bars represent SEM.