Emotional comorbidities in epilepsy result from seizure-induced corticosterone activity

Abstract

People with epilepsy often have psychiatric comorbidities that can significantly impair their quality of life. We previously reported that repeated seizure activity persistently alters endocannabinoid (eCB) signaling in the amygdala which accounts for comorbid emotional dysregulation in rats, however, the mechanism by which these alterations in eCB signaling within the epileptic brain occur is unclear. Endocannabinoid signaling is influenced by corticosterone (CORT) to modulate cognitive and emotional processes and a hyperactive hypothalamic-pituitary-adrenal (HPA) axis occurs in both people with epilepsy and nonhuman animal models of epilepsy. We employed selective pharmacological tools and a variety of approaches including whole-cell patch-clamp electrophysiology, behavioural paradigms and biochemical assays in amygdala kindled adult male Long-Evans rats. We aimed to determine whether seizures induce hypersecretion of CORT and the role this plays in eCB system dysregulation, impaired fear memory, and anxiety-like behaviours associated with seizure activity. Plasma CORT levels were significantly and consistently elevated following seizures over the course of kindling. Pre-seizure administration with the CORT synthesis inhibitor metyrapone prevented this seizure-induced CORT increase, prevented amygdala anandamide downregulation, and synaptic alteration induced by seizure activity. Moreover, treatment with metyrapone or combined glucocorticoid receptor (GR)/mineralocorticoid receptor (MR) antagonists prior to each elicited seizure were equally effective in preventing chronically altered anxiety-like behaviour and fear memory responses. Inhibiting seizure-induced corticosterone synthesis, or directly blocking the effects of CORT at GR/MR prevents deleterious changes in emotional processing and could be a treatment option for emotional comorbidities in epilepsy.

Keywords: Corticosterone; Emotional behaviour; Endocannabinoids; Epilepsy; Seizure; Stress.

Fig. 1

Pre-seizure administration of metyrapone alone, CORT113176 alone, spironolactone alone, and CORT113176 with spironolactone did not alter seizure parameters. Repeated electrical kindling of the BLA results in longer, more severe seizures over multiple sessions. A) There was no statistical difference in seizure duration at each session for rats administered vehicle alone, metyrapone alone, CORT113176 alone, spironolactone alone, or both CORT113176 and spironolactone (two-way RM ANOVA, day effect: F_(20,1776) = 131.60, p < 0.0001; drug effect: F_(4,1776) = 5.52, p = 0.0002; interaction: F_(80,1776) = 0.65, p = 0.99. B) There was no statistical difference in seizure stage at each session for rats administered vehicle alone, metyrapone alone, CORT113176 alone, spironolactone alone, or both CORT113176 and spironolactone (two-way RM ANOVA, day effect: F_(20,1776) = 377.20, p < 0.0001; drug effect: F_(4,1776) = 6.64, p < 0.0001; interaction: F_(80,1776) = 1.25, p = 0.007. “T” on each graph denotes the initial thresholding session. Data are expressed as means ± SEM. n = 15–22 per group.

Fig. 2

Kindled seizures produce a large increase in CORT. Kindled seizures generate a large CORT response that does not habituate over repeated sessions. These changes could be prevented using the 11-betahydroxylase inhibitor, metyrapone (50 mg/kg; data not shown). A) Experimental schematic illustrating the timeline of CORT extraction/analysis experiments. White filled arrows represent a 7-day break from experimentation. B) Kindled BLA seizures trigger a significant elevation in the stress hormone, CORT (two-way RM ANOVA, seizure effect: F_(1,22) = 39.31, p < 0.0001; time effect: F_(1,22) = 48.84, p < 0.0001; Interaction: F_(1,22) = 48.56, p < 0.0001). Rats do not habituate to the seizure-induced increases in CORT as there are no differences measured over kindling days (two-way RM ANOVA, day effect: F_(2,33) = 0.61, p = 0.55; Interaction: F_(2,33) = 0.70, p = 0.51). The increase in CORT following each kindled seizure is completely abolished using metyrapone (data not shown). Overall ANOVA statistics are reported by comparing day 10 of each experimental group. Significance in figures is represented using a Tukey's multiple comparisons follow up test. Data are expressed as means ± SEM. ∗∗∗∗p < 0.0001; n = 6–7 per group. Graphics created with Biorender.com.

Fig. 3

Metyrapone prevents seizure-induced emotional dysfunction. Kindled rats display both an anxious-like phenotype and have impaired fear memory retention measured 24- and 48-h post conditioning; these behaviours are reverted when metyrapone is administered prior to each kindled seizure. A) Experimental schematic illustrating the timeline of elevated plus maze experiments. White filled arrows represent a 7-day break from experimentation. (B-D) Kindled BLA seizures promote anxious-like behaviours on the EPM which are prevented with pre-seizure metyrapone administration B) Effects of metyrapone administration and repeated kindled seizures on total distance travelled, (two-way ANOVA, seizure effect; F_(1,36) = 0.10, p = 0.75; drug effect: F_(1,36) = 1.21, p = 0.28; interaction: F_(1,36) = 0.67, p = 0.42), C) percentage of time spent in the open arms (two-way ANOVA, seizure effect: F_(1,36) = 4.71, p = 0.037; drug effect: F_(1,36) = 5.28, p = 0.028; interaction: F_(1,36) = 3.79, p = 0.060), and D) percentage of open arm entries (two-way ANOVA, seizure effect: F_(1,36) = 17.85, p = 0.0002; drug effect: F_(1,36) = 9.31, p = 0.0043; interaction: F_(1,36) = 2.53, p = 0.12). E) Experimental schematic illustrating the timeline of auditory fear conditioning experiments. White filled arrows represent a 7-day break from experimentation. (F-H) Kindled seizures impair fear memory behaviours at 24- and 48-h post-conditioning; these effects were fully prevented using metyrapone. F) Effects of metyrapone administration and repeated kindled seizures on freezing rates expressed as a percent during cued fear conditioning (two-way RM ANOVA, seizure effect: F_(3,224) = 0.62, p = 0.62; CS effect: F_(3,224) = 133.3, p < 0.0001; interaction: F_(3,224) = 0.2943, p = 0.98), G) 24-h retention (two-way RM ANOVA, seizure effect: F_(3,392) = 27.37, p < 0.0001; CS effect: F_(6,392) = 62.24, p < 0.0001; interaction: F_(18,392) = 2.038, p = 0.0075), and H) 48-h retention protocols (two-way RM ANOVA, seizure effect: F_(3,168) = 16.21, p < 0.0001; CS effect: F_(2,168) = 182.00, p < 0.0001; interaction: F_(6,168) = 5.11, p < 0.0001). Significance in figures is represented using a Tukey's multiple comparisons follow up test. Post-hoc analyses on fear retention graphs compare vehicle seizure to metyrapone seizure. Data are expressed as means ± SEM. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001; n = 10–15 per group. Graphics created with Biorender.com.

Fig. 4

Metyrapone rescues AEA levels and restores the seizure-induced imbalance of glutamate/GABA activity onto BLA pyramidal neurons. Repeated kindled seizures cause a downregulation in AEA content that is paralleled by an imbalance in both excitatory and inhibitory presynaptic activity within the BLA. A) Experimental schematic illustrating the timeline of eCB analysis and electrophysiological experiments. White filled arrows represent a 7-day break from experimentation. B/C) Kindled seizures cause a persistent drop in AEA but not 2-AG that can be fully restored using metyrapone. B) Anandamide (two-way ANOVA, seizure effect: F_(1,20) = 3.60, p = 0.26; drug effect: F_(1,20) = 0. 79, p = 0.38; interaction: F_(1,20) = 4.51, p = 0.046), and C) 2-Arachidonyl Glycerol (two-way ANOVA, seizure effect: F_(1,19) = 0.0021, p = 0.96; drug effect: F_(1,19) = 0.77, p = 0.39; interaction: F_(1,19) = 0.69, p = 0.42). D) Relationship of evoked action potential to increasing current steps was similar across all experimental groups (two-way RM ANOVA, seizure effect: F_(3,330) = 3.71, p = 0.012; current effect: F_(9,330) = 41.17, p < 0.0001; interaction: F_(27,330) = 0.21, p > 0.99); n = 9–12 cells, 8–10 rats. E) The resting membrane potential between all experimental groups was similar: Vehicle No Sz (−68.41 ± 1.29), Vehicle Sz (−68.44 ± 1.39), Metyrapone No Sz (−71.47 ± 1.86), Metyrapone Sz (−69.92 ± 1.90); n = 10–12 cells, 9–10 rats. F) I-V curves showing similar responses between all experimental groups (two-way RM ANOVA, seizure effect: F_(3,266) = 0.059, p = 0.98; current effect: F_(6,266) = 622.5, p < 0.0001; interaction: F_(18,266) = 3.011, p < 0.0001); n = 9–12 cells, 8–10 rats. (G/H) Repeated kindled seizures significantly increase excitatory presynaptic activity without changing the amplitude of the response. Metyrapone administration restored the seizure-induced increase in sEPSC frequency (G, n = 7–9 cells, 6–7 rats; two-way ANOVA, seizure effect: F_(1,28) = 9.732, p = 0.0048; drug effect: F_(1,28) = 1.64, p = 0.21; interaction: F_(1,28) = 7.51, p = 0.011) without any significant effect on sEPSC amplitude (H, n = 7–9 cells, 7 rats; two-way ANOVA, seizure effect: F_(1,28) = 0.0060, p = 0.94; drug effect: F_(1,28) = 0.0062, p = 0.94; interaction: F_(1,28) = 0.85, p = 0.036). (I/J) Repeated kindled seizures significantly reduces inhibitory presynaptic activity without changing the amplitude of the response. Metyrapone administration reversed the seizure-indued decrease in sIPSC frequency (I, n = 7–10 cells, 7–9 rats; two-way ANOVA, seizure effect: F_(1,29) = 13.39, p = 0.0010; drug effect: F_(1,29) = 7.92, p = 0.0087; interaction: F_(1,29) = 1.39, p = 0.25) without any significant effect on sIPSC amplitude (J, n = 8–12 cells, 7–11 rats; two-way ANOVA, seizure effect: F_(1,32) = 0.86, p = 0.36; drug effect: F_(1,32) = 0.044, p = 0.83; interaction: F_(1,32) = 1.55, p = 0.22). Significance in figures is represented using a Tukey's multiple comparisons follow up test. Data are expressed as means ± SEM. ∗p < 0.05, ∗∗p < 0.01. Graphics created with Biorender.com.

Fig. 5

The seizure-induced CORT response remains after glucocorticoid and mineralocorticoid receptor antagonism. Selective pharmacological blockade of GR and MR does not prevent the seizure-induced increase in corticosterone. A) Experimental schematic illustrating the timeline of corticosterone extraction/analysis experiments. White filled arrows represent a 7-day break from experimentation. B) Pre-seizure administration of CORT113176 significantly raised baseline CORT levels but did not reduce the rapid CORT increase in response to seizures (two-way RM ANOVA, seizure effect: F_(1,38) = 145.60, p < 0.0001; drug effect: F_(1,38) = 0.24, p = 0.63; interaction: F_(1,38) = 8.83, p = 0.0051). C) Pre-seizure administration of spironolactone significantly raised baseline CORT levels but did not reduce the rapid CORT increase in response to seizures (two-way RM ANOVA, seizure effect; F_(1,36) = 109.10, p < 0.0001; drug effect: F_(1,36) = 0.053, p = 0.82; interaction: F_(1,36) = 9.43, p = 0.0040. D) Co-administration of CORT113176 and spironolactone also raised CORT levels at baseline without changing the seizure-induced CORT response (two-way RM ANOVA, seizure effect: F_(1,31) = 127.00, p < 0.0001; drug effect: F_(1,31) = 0.53, p = 0.47; interaction: F_(1,31) = 6.08, p = 0.02). Overall ANOVA statistics are reported by comparing day 10 of each experimental group. Significance in figures is represented using a Tukey's multiple comparisons follow up test. Data are expressed as means ± SEM. ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001; n = 8–11 per group. Graphics created with Biorender.com.

Fig. 6

CORT113176 when co-administered with spironolactone prevents seizure-induced emotional dysfunction. CORT11376 and Spironolactone (but not CORT113176 alone or spironolactone alone) given together before each electrically kindled seizure reverted the chronic dysfunctions observed on the elevated plus maze and fear processing paradigms. A) Experimental schematic illustrating the timeline of elevated plus maze experiments. White filled arrows represent a 7-day break from experimentation. (B-D) Co-administration of CORT11376 and Spironolactone prevented the increase in seizure-induced anxious-like behaviours on the EPM. B) Effects of GR and MR antagonism and repeated kindled seizures on total distance travelled, (two-way ANOVA, seizure effect: F_(1,69) = 0.11, p = 0.74; drug effect: F_(3,69) = 0.41, p = 0.74; interaction: F_(3,69) = 0.65, p = 0.59), C) percentage of time spent in the open arms (two-way ANOVA, seizure effect; F_(1,69) = 10.19, p = 0.0021; drug effect: F_(3,69) = 3.51, p = 0.020; interaction: F_(3,69) = 1.85, p = 0.15), and D) percentage of open arm entries (two-way ANOVA, seizure effect: F_(1,69) = 2.66, p = 0.11; drug effect: F_(3,69) = 2.30, p = 0.085; interaction: F_(3,69) = 3.34, p = 0.024). E) Experimental schematic illustrating the timeline of auditory fear conditioning experiments. White filled arrows represent a 7-day break from experimentation. (F-H) Co-administration of CORT11376 and Spironolactone prevented the seizure-induced impairment of fear behaviours measured 24- and 48-h post-conditioning. F) Effects of GR and MR antagonism and repeated kindled seizures on freezing rates expressed as a percent during cued fear conditioning (two-way RM ANOVA, seizure effect: F_(7,328) = 0.078, p = 0.99; CS effect: F_(3,328) = 356.20, p < 0.0001; interaction: F_(21,328) = 0.22, p > 0.99). G) 24-h retention (two-way RM ANOVA, seizure effect: F_(7,574) = 28.27, p < 0.0001; CS effect: F_(6,574) = 187.80, p < 0.0001; interaction: F_(42,574) = 1.21, p = 0.18), and H) 48-h retention protocols (two-way RM ANOVA, seizure effect: F_(7,246) = 7.12, p < 0.0001; CS effect: F_(2,246) = 422.00, p < 0.0001; interaction: F_(14,246) = 2.11, p = 0.012). Significance in figures is represented using a Tukey's multiple comparisons follow up test. Post-hoc analyses on fear retention graphs compare vehicle seizure to CORT113176+Spironolactone seizure. Data are expressed as means ± SEM. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001; n = 8–18 per group. Graphics created with Biorender.com.