Estrogen Protects Neurotransmission Transcriptome During Status Epilepticus

Abstract

Women with epilepsy commonly have premature onset of menopause. The decrease in estrogen levels is associated with increased occurrence of neurodegenerative processes and cognitive decline. Previously, we found that estradiol (E2) replacement in ovariectomized (OVX) female rats significantly reduced the seizure-related damage in the sensitive hilar region of hippocampal dentate gyrus (DG). However, the complex mechanisms by which E2 empowers the genomic fabrics of neurotransmission to resist damaging effects of status epilepticus (SE) are still unclear. We determined the protective effects of the estradiol replacement against kainic acid-induced SE-associated transcriptomic alterations in the DG of OVX rats. Without E2 replacement, SE altered expression of 44% of the DG genes. SE affected all major functional pathways, including apoptosis (61%), Alzheimer's disease (47%), cell cycle (59%), long-term potentiation (62%), and depression (55%), as well as synaptic vesicle cycle (62%), glutamatergic (53%), GABAergic (49%), cholinergic (52%), dopaminergic (55%), and serotonergic (49%) neurotransmission. However, in rats with E2 replacement the percentage of significantly affected genes after SE was reduced to the average 11% (from 8% for apoptosis to 32% for GABAergic synapse). Interestingly, while SE down-regulated most of the synaptic receptor genes in oil-injected females it had little effect on these receptors after E2-replacement. Our novel Pathway Protection analysis indicated that the E2-replacement prevented SE-related damage from 50% for GABA to 75% for dopaminergic transmission. The 15% synergistic expression between genes involved in estrogen signaling (ESG) and neurotransmission explains why low E2 levels result in down-regulation of neurotransmission. Interestingly, in animals with E2-replacement, SE switched 131 synergistically expressed ESG-neurotransmission gene pairs into antagonistically expressed gene pairs. Thus, the ESG pathway acts like a buffer against SE-induced alteration of neurotransmission that may contribute to the E2-mediated maintenance of brain function after the SE injury in postmenopausal women. We also show that the long-term potentiation is lost in OVX rats following SE but not in those with E2 replacement. The electrophysiological findings in OVX female rats with SE are corroborated by the high percentage of long-term potentiation regulated genes (62%) in oil-injected while only 13% of genes were regulated following SE in E2-replaced rats.

Keywords: GABAergic synapse; beta-estradiol; cholinergic synapse; dentate gyrus; dopaminergic synapse; glutamatergic synapse; jun oncogene; serotonergic synapse.

Figure 1

(A) Experimental design. OVX, ovariectomy; KA, kainic acid; saline, KA vehicle. Animals were injected with oil or 2 μg of β-estradiol benzoate for 4 days subcutaneously. (B,E) Percentage of the regulated genes within the entire transcriptome (ALL) and included in the analyzed pathways in CES and COS groups with respect to CEN, respectively CON groups. ALZ, Alzheimer's disease, APO, apoptosis, CC, cell cycle, LTD, long-term depression, LTP, long-term potentiation, SVC, synaptic vesicle cycle, GLU, glutamatergic synapse, GABA, GABAergic synapse, ACH, cholinergic synapse, DA, dopaminergic synapse, 5HT, serotonergic synapse. Note the massive effect of SE on the gene expression. (C,F) WPR scores for selected pathways. In (B,C,E,F) lower is better (less damaging). (D,G) Protection against transcriptomic regulation in E2-replaced animals quantified by both TPR (percent reduction of the number of significantly regulated genes) and PPR (percent reduction of WPR). Note that DA had the largest (better) protection as quantified by PPR.

Figure 2

Power of the flexible fold-change cut-off. (A) Distribution of the fold-change cut-offs for the COS vs. CON comparison. COS, expression ratio in the comparison COS vs. CON, CES, expression ratio in the comparison CES vs. CEN. When all 12,710 quantified unigenes are considered, CUT values range from 1.055 for Pcyt1a to 4.062 for Scgb1c1, with the average 1.451. (B) Distribution of the fold-change cut-offs for the CES vs. CEN comparison. When all 12,710 quantified unigenes are considered, CUT values range from 1.109 for Mapk1ip1l to 3.877 for Ttr, with the average 1.633. (C) “False negatives” and “false positive” regulation hits from the analyzed functional pathways. Note that expression ratio 2.226 was not enough for Chrna7 to be considered as up-regulated in COS vs. CON comparison (CUT = 2.297) but 1.986 was enough in CES vs. CEN (CUT = 1.873).

Figure 3

Electrophysiology and transcriptomics of the long-term potentiation (LTP). (A,B) Medial perforant path LTP in the DG in OVX rats is lost following SE but preserved in E2-replaced animals. LTP induced by TBS following SE (filled circles) compared to rats with no seizures (open circles). An arrow marks a moment when the TBS was administered. In slices from oil-injected rats exposed to SE (COS; n = 5), TBS-induced LTP was severely reduced compared to LTP from rats with no SE (CON; n = 10). In contrast, the magnitude of TBS-induced LTP in slices from E2-replaced OVX rats following SE (CES; n = 5) was not different from LTP in slices from rats with no seizure experience (CEN; n = 10). (C,D) Dramatic regulation of LTP associated genes by SE in OVX rats compared to their non-SE counterparts (COS vs. CON) is in contrast to protection against the transcriptomic regulation of LTP associated genes in E2 replaced animals (CES vs. CEN). As specified in the “COLOR CODE” inset table, red/green/yellow background of the gene symbol indicates up-/down-/no regulation, while the white background indicates that expression of that gene was not quantified. Note the significantly lower number of regulated genes in rats with E2 replacement.

Figure 4

Regulation of the glutamatergic synapse pathway by SE in OVX rats with (CES)/without (COS) E2-replacement with respect to their non-SE counterparts (CEN and CON). Note the large number of down-regulated genes by SE in oil injected rats [e.g., metabotropic glutamate receptor 3, Grm3 in (A)] whose expression was not affected by SE in rats with E2 replacement (B).

Figure 5

Regulation of the GABAergic synapse pathway by SE in OVX rats with (CES)/without (COS) E2-replacement with respect to their non-SE counterparts (CEN and CON). Note large number of genes down-regulated by SE in oil-injected animals [e.g., GABA B receptor 1, Gabbr1 in (A)] whose expression was not affected by SE in rats with E2 replacement (B).

Figure 6

Regulation of the dopaminergic synapse pathway by SE in OVX rats with (CES)/without (COS) E2-replacement with respect to their non-SE counterparts (CEN and CON). Note the number of genes down-regulated by SE in oil-injected animals [e.g., DOPA decarboxylase, Ddc in (A)] whose expression was not affected by SE in rats with E2 replacement (B).

Figure 7

Regulation of the cholinergic synapse pathway by SE in OVX rats with (CES)/without (COS) E2-replacement with respect to their non-SE counterparts (CEN and CON). Again, note the large number of genes down-regulated by SE in oil-injected animals [e.g., acetylcholinesterase, Ache in (A)] whose expression was not affected by SE in rats with E2 replacement (B).

Figure 8

Regulation of the serotonergic synapse pathway by SE in OVX with (CES)/without (COS) E2-replacement with respect to their non-SE counterparts (CEN and CON). Note the number of down-regulated genes by SE in oil-injected animals [e.g. serotonin receptor 5B, Htr5b in (A)] whose expression was not affected by SE in rats with E2 replacement (B).

Figure 9

Regulation of the synaptic vesicle cycle pathway by SE in OVX with (CES)/without (COS) E2-replacement with respect to their non-SE counterparts (CEN and CON). Note that in oil-injected animals (A) the vesicle cycle was altered by SE in all phases but only few genes were affected in animals with E2 replacement (B).

Figure 10

Regulation of the estrogen signaling pathway by SE in OVX females with E2 replacement. Note that, in addition to up-regulating Adcy7, Adcy8, Creb1, Creb3l1, Gnai1, Gnai3, Kcnj3, Kras, Prkacb, and Sp1, SE turned on Calml3, Itpr3, Mmp9, Pik3r1, and Pik3r3. No gene of this pathway was found as down-regulated or turned off by SE.

Figure 11

The transcriptomic networks by which the estrogen-signaling (ESG) pathway genes regulate the glutamatergic synapse (GLU) genes through expression coordination with the common genes (CG) between the two pathways in OVX rats with E2-replacement. (A) Networks in the DG of non-SE animals. A red/blue line indicates that the linked genes are significantly (p-value < 0.05) synergistically/antagonistically expressed, respectively. Numbers on the left side represent the percentages of the synergistically (S)/antagonistically (A)/independently (I) expressed pairs that can be formed between the two groups of genes. (B) Networks in the DG of SE animals. Red/green/yellow background of the gene symbol indicates that that gene was up-/down-/not regulated by SE. Note that there is a substantial shift from synergistic to antagonistic expression between ESG and CG groups in SE animals.

Figure 12

The transcriptomic networks by which the estrogen-signaling (ESG) pathway genes regulate the cholinergic synapse (ACH) genes through expression coordination with the common genes (CG) between the two pathways in OVX rats with E2-replacement. (A) Networks in the DG of non-SE animals. (B) Networks in the DG of SE animals. Note the substantial shift from synergistic to antagonistic expression between ESG and CG groups in SE animals.

Figure 13

ES-neurotransmission gene pairs that switched their expression coordination from synergistic to antagonistic (white to blue bullets) or from antagonistic to synergistic (white to red bullets) following SE.

Figure 14

Expression correlation of Jun oncogene with some synapse membrane receptors in the DG of OVX rat female rats. (A) Oil-injected, non-SE rats; (B) Oil-injected SE rats; (C) E2-replaced, non-SE rats; (D) E2-replaced, SE rats. A continuous red/blue line indicates significant (P-value < 0.05) expression synergism/antagonism of Jun with the linked gene, an interrupted black line indicates significant (p-value < 0.05) expression independence, while missing lines mean that we have not enough evidence to categorize the expression coordination of Jun with that gene.