17β Estradiol increases resilience and improves hippocampal synaptic function in helpless ovariectomized rats

Abstract

Memory impairment is the most commonly reported cognitive symptom associated with major depressive disorder. Decreased hippocampal volume and neurogenesis in depression link hippocampal dysfunction with deficits in memory. Stress decreases hippocampal dendritic spine density and long-term potentiation (LTP) at glutamate synapses, a cellular correlate of learning and memory. However, elevated plasma levels of 17β estradiol (E2) during proestrus increase hippocampal structure and function, directly opposing the negative consequences of stress. In women, significant fluctuations in ovarian hormones likely increase vulnerability of hippocampal circuits to stress, potentially contributing to the greater incidence of depression compared to men. Using the learned helplessness model of depression and ovariectomized female rats, we investigated whether acquisition of helplessness and hippocampal synaptic dysfunction is differentially impacted by the presence or absence of plasma E2. We find that inescapable shock induces a greater incidence of helplessness in vehicle- versus E2-treated OVX rats. In the vehicle-treated group, LTP was absent at CA3-CA1 synapses in slices only from helpless rats, and CA1 spine density was decreased compared to resilient rats. In contrast, significant LTP was observed in slices from E2-treated helpless rats; importantly, spine density was not different between E2-treated helpless and resilient rats, dissociating spine density from the LTP magnitude. We also find that E2 replacement can reverse previously established helpless behavior. Thus, our results show that E2 replacement in OVX rats increases resilience and improves hippocampal plasticity, suggesting that E2 therapy may increase resilience to stress and preserve hippocampal function in women experiencing large fluctuations in plasma estrogen levels.

Keywords: Dendritic spines; Estrogen; Hippocampus; LTP; Learned helplessness; Stress.

Figure 1

Learned helplessness protocol and experimental timeline. (A and B) On Days 1 and 2, inescapable shock (IES, 0.65 mA) was conducted in a dark, solitary cage and on Day 3, rats experienced escape testing to determine helplessness or resilience (see “Materials and Methods”). (C) Graph shows average escape latency for each bin of 5 trials during escape testing in untreated OVX rats that were not exposed to inescapable shock. (D) Experimental timeline for all treatment groups: IES, inescapable shock; Test, escape testing; OVX + VH, vehicle injection (INJ); OVX + E2, estradiol injection; LTP/Golgi, sacrifice for LTP recordings or Golgi staining. Reversal: untreated OVX rats that met criteria for helplessness were divided into two separate groups for VH or E2 injections before 2nd escape testing. (E and F) Left bar charts show percentage of helpless rats in male (■ helpless, n = 16; 55.2%; □ resilient n = 13, 44.8%) and untreated OVX control groups (■ helpless, n = 58 56.3%; □ resilient n = 45, 43.7%). There was a significant difference in mean escape latency during the 30 trials of escape testing between helpless and resilient rats in both groups (middle bar charts, *p < 0.0001). Right graphs show average escape latency in helpless and resilient rats for each bin of 5 consecutive trials during escape testing in male (panel E) and untreated OVX rats (panel F).

Figure 2

LTP is absent and spine density is decreased only in vehicle-treated OVX rats reaching criteria for helplessness. (A) Left bar chart shows percentage of helpless (■, 50.8%, n = 44) and resilient (□, 49.2%, n = 41) rats, with no difference in numbers between phenotypes. Middle bar chart shows mean escape latency during the 30 trials of escape testing. Right graph shows average escape latency in helpless and resilient rats where each bin is an average of 5 consecutive trials across the 30 trials. (B) Summary plot shows a lack of significant LTP in slices from vehicle-treated helpless rats (●, n = 7 slices/5 rats) in contrast to significant LTP in rats meeting criteria for resilience (○, n = 14 slices/9 rats). Inset shows representative fEPSP waveforms at baseline (a) and 40 min post tetanus (b) in helpless and resilient animals. Right bar chart shows average LTP magnitude from each recording included in the data sets. (C) Left, representative images of a tertiary CA1 dendrite from a helpless (■) and resilient (□) rat. Right, bar chart shows CA1 pyramidal cell dendritic spine density is decreased in vehicle-treated helpless versus resilient rats. Spines were counted from 10 mm sections above and below the plane of focus. N = 8 sections/4 rats per group; images show deconvolved, stacked images of all focal planes collapsed into maximum 3D projection. Scale bar represents 2 μm. LTP and spine analysis was performed 24 h following escape testing. Asterisk indicates significance at p < 0.001 for 2A, 2B, and p < 0.05 for 2C. Error bars represent SEM.

Figure 3

E2 replacement increases behavioral resilience. (A) Left bar chart shows percentage of helpless (■, 22.8%, n = 16) and resilient rats (□, 77.2%, n = 54) with significantly more resilient than helpless animals. Middle bar chart shows mean escape latency during the 30 trials of escape testing. Right graph shows average escape latency in helpless and resilient rats where each bin is an average of 5 consecutive trials across the 30 trials. (B) Summary plot shows significant LTP in E2-treated helpless OVX rats (●, n = 5 slices/4 rats), however, the magnitude is significantly less than in E2-treated resilient OVX rats (○, n = 12 slices/10 rats). Inset shows representative fEPSP waveforms at baseline (a) and 40 min post tetanus (b) in helpless and resilient animals. Right bar chart shows average LTP magnitude from each recording included in the data sets. (C) Left, representative images of a tertiary CA1 dendrite from a helpless (■) and resilient (□) rat. Right, bar chart shows no significant difference in dendritic spine density between E2-treated helpless and resilient rats. Spines were counted from 10 μm sections above and below the plane of focus. N = 8 sections/4 rats per group; images show deconvolved, stacked images of all focal planes collapsed into maximum 3D projection. Scale bar represents 2 μm. LTP and spine analysis was performed 24 h following escape testing. Asterisk indicates significance at p < 0.001 except potentiation from baseline in helpless rats where p < 0.05. Error bars represent SEM.

Figure 4

No change in the paired-pulse facilitation ratio or steady-state depolarization in helpless versus resilient rats. (A1) No significant difference in the paired pulse facilitation ratio between vehicle-treated resilient (□, n = 14 slices/9 rats) and helpless (■, n = 7 slices/5 rats) OVX rats. (A2) Left bar chart shows no significant difference in average steady state depolarization (arbitrary units) between helpless and resilient rats during the 4th round of tetanus in vehicle-treated rats. Right, representative waveforms from a helpless and a resilient rat. (B1) No significant difference in the paired-pulse facilitation ratio between E2-treated resilient (□, n = 12 slices/10 rats) and helpless (■, n = 5 slices/4 rats) OVX rats. (B2) Left bar chart shows no significant difference in steady-state depolarization during the 4th round of tetanus in E2-treated rats. Right, representative waveforms from a helpless and a resilient rat. Bar charts show averaged data from all recordings used in LTP data sets in Fig. 4. Error bars represent SEM.

Figure 5

E2 replacement can reverse previously established helpless behavior. (A) Experimental paradigm. (B and C) Left bar chart shows the percentage of helpless rats that remain helpless following a 2nd round of escape testing on Day 6. Middle bar chart shows the average escape latency from all 30 trials during escape testing on Day 6. Right graph shows escape latency where each bin is an average of 5 consecutive trials across the 30 trials. All vehicle-treated rats (n = 17 of 17) remained helpless while only 62.5% of E2-treated rats were helpless and 37.5% were resilient (6 of 16 rats). In the E2-treated group, mean escape latency significantly decreased during the 2nd round of escape testing on Day 6 due to some rats becoming resilient and learning to escape shock. Asterisk indicates p < 0.001. Error bars represent SEM.