“Estradiol and 3β-diol protect female cortical astrocytes by regulating connexin 43 Gap Junctions”

Abstract

While estrogens have been described to protect or preserve neuronal function in the face of insults such as oxidative stress, the prevailing mechanistic model would suggest that these steroids exert direct effects on the neurons. However, there is growing evidence that glial cells, such as astrocytes, are key cellular mediators of protection. Noting that connexin 43 (Cx43), a protein highly expressed in astrocytes, plays a key role in mediating inter-cellular communication, we hypothesized that Cx43 is a target of estradiol (E2), and the estrogenic metabolite of DHT, 3β-diol. Additionally, we sought to determine if either or both of these hormones attenuate oxidative stress-induced cytotoxicity by eliciting a reduction in Cx43 expression or inhibition of Cx43 channel permeability. Using primary cortical astrocytes, we found that E2 and 3β-diol were each protective against the mixed metabolic/oxidative insult, iodoacetic acid (IAA).

Moreover, these effects were blocked by estrogen receptor antagonists. However, E2 and 3β-diol did not alter Cx43 mRNA levels in astrocytes but did inhibit IAA-induced Cx43 gap junction opening/permeability. Taken together, these data implicate astrocyte Cx43 gap junction as an understudied mediator of the cytoprotective effects of estrogens in the brain.

Given the wide breadth of disease states associated with Cx43 function/dysfunction, further understanding the relationship between gonadal steroids and Cx43 channels may contribute to a better understanding of the biological basis for sex differences in various diseases.

1. Introduction

Astrocytes play important roles in regulating central nervous system (CNS) function. Most often, they are recognized as regulating the chemical and ionic composition of the extracellular milieu and protecting against detrimental mechanisms that include excitotoxicity and oxidative stress (1–3). Additionally, it is important to recognize that these cells also play key roles in facilitating intercellular communication by enabling metabolite shuttling between cells (4,5). However, the mechanisms by which these functions are regulated remain to be fully understood.

Connexin 43 (Cx43), a protein highly expressed in astrocytes (6), is recognized as a key component of the intercellular communication (3,5,7–9). Cx43 channels exist as either gap junctions or hemichannels (10). Whereas a number of studies published to date have focused on the role of Cx43 hemichannels, through the use of the Cx43 hemichannel blocker-Gap19 (1,11–15), some findings support Cx43 gap junctions as relevant contributors to optimal astrocyte function, as well as playing indirect roles in regulating neuron function, as demonstrated through th use of the Cx43 gap junction blocker, Gap26 (16–18). Gap26 is a Cx43 mimetic peptide, consisting of amino acid residues 63-75 of the extracellular loop 1 of connexin 43 (Cx43), that inhibits Cx43 hemichannels and/or gap junctions (19–21).

Cx43 channel function is regulated by three variables: The Cx43 expression (22), the regulation of posttranslational modifications to Cx43 (such as the phosphorylation (23)), and the Cx43 channel permeability (24,25). Androgens and estrogens have been reported to regulate Cx43 channel expression, phosphorylation, and permeability in the peripheral tissue (26–30). Building off of a previous study that demonstrated estrogen and progesterone regulation of Cx43 function in the preoptic area (POA) of the hypothalamus (31), we set out to determine whether estrogens or androgens regulate cerebral cortical astrocyte Cx43.It is also worth noting that no information currently exists, in any tissue, on whether estrogens or androgens regulate Cx43 gap junction permeability. Only one study has assessed the effect of E2 on Cx43 phosphorylation, and the resulting effects on tissue susceptibility to damage following cytotoxic insult (30). Collectively, this information highlights the current gap in knowledge on the role of Cx43 gap junctions and hemichannels in mediating steroid hormone action in the CNS.

While the aromatization of testosterone to estradiol (E2) is a well-studied pathway by which androgens (testosterone, in particular) can elicit its effects through estrogen receptors, the recognition of 5α-androstane-3β,17β-diol (also known as 3β-diol) as a metabolite mediator of the effects of another major androgen, dihydrotestosterone (DHT) has only more recently been appreciated. 3β-diol, a member of the androstanediol family, is a 3β-hydroxylated metabolite of DHT, that can be produced in brain cells that express 5-α-reductase (for the conversion of testosterone to DHT) and 3-β-hydroxysteroid hydroxysteroid (32–34). The effects of 3β-diol is largely predicted by the expression and function of ERβ, given that 3β-diol preferentially binds to ERβ. The expression patterns and known biology of ERβ in different tissues has helped predict the physiology of 3β-diol. For example, the work of Weihua et al., (35,36) inferred a role of 3β-diol, through ERβ, in prostate tissue. Further, the 3β-diol has been described to have important effects on the HPA axis and potentially, neuroprotection in both males and females (reviewed in (37)).

Given our recent observation that E2 and 3β-diol can elicit protection of astrocytes and that 3β-diol activates several of the same estrogen receptor (ER)-dependent mechanisms activated by E2 (37–42), we proposed that both E2 and 3β-diol protect astrocytes from metabolic/oxidative insult through a common estrogen-receptor mediated mechanism, and do so by modulating Cx43 gap junctions through Gap26. Our data demonstrate that E2 and 3β-diol inhibit Cx43 gap junction permeability in cortical astrocytes and that this inhibition was correlated with their cytoprotective efficacy. As such, these studies demonstrate for the first time that cortical astrocyte Cx43 is a relevant target of estrogens in the brain.

2. Materials and Methods

2.1. Tissue Acquisition and Culture

Primary cortical astrocytes were derived from female postnatal day 2 (P2) C57Bl/6 mice as described previously with some modifications (43,44). Additionally, we also generated primary cutures of astrocytes from P2 mouse cerebral cortex obtained from a commercial source (BrainBits LLC., Springfield, IL). The data were consistent between the two sources of tissue, and as such, and served to further validate the data. Briefly, pups were anesthetized by hypothermia, followed by cardiac puncture as the method of euthanasia. Following craniotomy, the brain was removed and microdissected to isolate the cerebral cortex. The resulting cortical tissue was then mechanically and enzymatically dissociated into individual cells that were centrifuged and filtered (to remove cellular debris), and subsequently plated onto 75 cm² noncoated flasks at a plating equivalent of two brains worth of tissue per 96 well plate (for viability and permeability assessments) or 6 well plates for obtaining RNA and protein. Once confluent, the cells were then subcultured and plated in either 96-well black bottom plates (noncoated, Thermo Fisher Scientific, Rochester, NY; cat# 165305) for cell viability assays and hemichannel/gap junction permeability assays, or clear bottom 6-well plates (noncoated, VWR North American, Radnor, PA; cat# 10062-892), for mRNA and protein assessments. Cells from a single confluent 75 cm² flask were sufficient to plate either one 96-well or one 6-well plate. For each experiment (biological replicate), primary cortical astrocytes derived from two brains, and were used to produce three technical replicates for in vitro assays. Astrocytes were cultured in astrocyte media (Dulbecco’s Modified Eagle Media containing 110 mg/L Sodium Pyruvate, Gibco Life Technologies, Grand Island, NY; cat# 11995-065) with 10% v/v Fetal Bovine Serum (Atlanta Biologicals, Flowery Beach, GA; cat# S11550) and 1% v/v PenStrep 10,000 units/mL (HyClone, South Logan, UT cat# SV30010). 24 hrs prior to treatment of the astrocytes with the respective treatment (e.g., steroid), the cultures were switched to media that contained 10% v/v charcoal-stripped Fetal Bovine Serum, FBS (R&D Systems, Inc., Minneapolis, MN; cat# S11650) to reduce the potential confounding effect of high levels of steroid hormones in media containing “regular” FBS. Since studies have reported that the presence or absence of phenol red makes no difference to the demonstration of estrogenic effects (45,46) and our previous studies also revealed the data having no estrogenic effects of phenol red, we used phenol red contained culture media.

The purity of astrocytes was verified by counter-staining of astrocytes using antibody staining against dendrite. All experimental procedures involving animals were approved by the UNT Health Science Center’s Institutional Animal Care and Use Committee (IACUC).

2.2. Pharmacologic Treatments

17-β estradiol (E2, Sigma, St. Louis, MO; cat# E2758), dihydrotestosterone (DHT, Steraloids, Newport, RI; cat# A2570-000), and 5α-androstane-3β, 17β-diol (3β-diol, Steraloids, Newport, RI; cat# A1220-000) were dissolved in DMSO and applied to the cells at a final concentration of 100 nM (where the final concentration of DMSO was < 0.1%). This concentration was selected based on concentration-response and time course optimizations that indicated 100 nM E2 induced significant protection against iodoacetic acid-induced cytotoxicity. Moreover, this concentration is consistent with published reports in which neuro/cytoprotection was evaluated. Iodoacetic acid (IAA, final concentration 25 μM, Sigma, St. Louis, MO; cat# I4386), was dissolved in de-ionized water, sterile filtered, aliquoted, and used at the final concentrations stated in the Results and Figure Legends. The concentration of IAA selected for our experiments was based on concentration-response optimization experiments to identify the concentration of IAA that reduced confluent astrocyte viability, after 6 hours of treatment, by approximately 30 percent (EC₃₀). Ethidium bromide (EtBr, VWR North American, Radnor, PA; cat# X328) and Lucifer yellow (Thermofisher Scientific, Pleasanton, CA; cat# L453) were prepared and aliquoted based on the manufacturer’s recommendation. ER antagonists ICI-182,780, (500 nM, cat# 1047/1), MPP Dihydrochloride, (100 nM, cat# 1991/10), PHTPP (500 nM, cat# 2662/10), the Connexin43 hemichannel blocker, TAT-GAP19 (10 μM, cat# 6227), the Panx1 blocker, 10Panx (25 μM, cat# 3348), and the Connecxin43 gap junction blocker, GAP26 (50 μM, cat# 1950), were purchased from Tocris Bioscience (Minneapolis, MN). The final concentrations used for these compounds were based on each compound’s respective EC₅₀ or IC₅₀.

2.3. RTPCR

Semi-quantitative, real-time PCR was conducted using 20 ng of template per reaction to assess treatment-induced changes in astrocyte mRNA expression of Cx43 (primer: Mm01179639_s1 from Thermofisher Scientific, Pleasanton, CA), GAP26 (primer: Mm00433610_s1 from Thermofisher Scientific, Pleasanton, CA) and GAPDH (primer: Mm99999915_g1 from Thermofisher Scientific, Pleasanton, CA), as the “housekeeping” gene control. RNA was extracted from astrocytes using the RNeasy Mini Kit (Qiagen, Germantown, MD cat# 74106), converted to cDNA using the High-Capacity cDNA Reverse Transcription Kit (Thermofisher Scientific, Pleasanton, CA cat# 4368814), and analyzed on an Applied Biosystems 7300 Real-Time PCR System Thermocycler (Foster City, CA). Target mRNA levels were evaluated in triplicate and averaged to obtain average cycle threshold (Ct) values for each sample’s Gap26, Cx43, and GAPDH expression. Fold change values were calculated using the 2^−ΔΔCt method (47), followed by transforming the data to the percent of DMSO control (serving as the experimental reference group). The resulting averages were compared for statistical significance by One-Way ANOVA followed by Tukey’s post hoc analysis to determine statistical significance between treatment groups in GraphPad Prism (San Diego, CA) software (Version 5).

2.4. Western Blot

Western blot was conducted, as previously described (48), to assess relative changes in Cx43 protein expression following treatment with E2, DHT, and 3β-diol (all 100 nM, 3 and 18 hours vehicle or steroid treatment). In brief, astrocyte cultures were lysed in lysis buffer containing protease and phosphatase inhibitors, including 50 mm Tris, pH 7.4, 150 mm NaCl, 1 mm EGTA, 1 mm Na₃VO₄, and 100 mm NaF plus 5 μm ZnCl₂, 10% glycerol, and 1% Triton X-100 and 1 mm PMSF (49). After homogenization, samples were centrifuged at 100,000 x g for 30 min at 4°C. 40 μg of the total protein supernatant by sample (quantified using the Bio-Rad DC assay (Rio-Rad, Hercules, CA; cat# 5000116)) per lane were separated by SDS-PAGE and subsequently transferred onto polyvinylidene fluoride membranes (PVDF, Bio-Rad Laboratories) by electroblotting. Membranes were blocked with 3% bovine serum albumin (BSA) in tris-buffered saline containing 0.2% Tween 20 (TBS-T) for 1 hour at room temperature, followed by overnight incubations (at 4°C) with the following antibodies: Cx43 / GJA1 rabbit polyclonal antibody (Abcam, Cambridge, MA; cat# ab11370, RRID: AB_297976; 1:5000 in TBST), GAPDH rabbit polyclonal antibody (Cell Signaling, Danvers, MA; cat# 2118, RRID: AB_561053; 1:1000 in TBST), and HRP-conjugated mouse anti-rabbit monoclonal secondary antibody, light chain only (Jackson ImmunoResearch, West Grove, PA; cat# 211-032-171, RRID: AB_2339149; 1: 10,000 in 3% v/v BSA in TBST). The ECL detection system (Thermofisher Scientific) was used, and chemiluminescence intensity was detected and band intensity visualized using an image analyzer (Alpha Innotech, San Leandro, CA). Densitometric analyses were conducted using ImageJ (National Institutes of Health) software. Each data point was derived from at least three independent replicates. Between-group differences were assessed using One-Way ANOVA followed by Tukey post hoc analyses in GraphPad Prism Version 5 software.

2.5. MTT Viability Assay

Methyltetrazolium (MTT, Bio Vision, Milpitas, CA; cat# 2809) viability assay was conducted in a 96-well plate format, with 8 wells allocated per treatment group. Iodoacetic acid (IAA) was chosen as the metabolic/oxidative insult because it simulates age-associated metabolic stress and oxidative load (50,51), as well as opening Cx43 hemichannels without similar effects at Cx43 gap junctions (52). Preliminary Iodoacetic acid (IAA) concentration-response assays revealed that 25 μM IAA for 6 hours was sufficient to reduce cell viability by 50%. Steroid-treated groups were treated for 18 hours prior to IAA treatment. The resulting treatment/insult-induced changes in MTT colorimetric absorbance (at 595 nm) were detected on a FlexStation 3 Microplate Reader, (Molecular Devices, Sunnyvale, CA). Results for 8 wells were used to obtain the treatment group average and variability per individual plate. Each experiment run (i.e., an experiment run at different times, where cells were derived from different animals) served as a single biological replicate (“n”). Each data point represents the average from at least three independent replicates. Between-group differences were assessed using One-Way ANOVA followed by Tukey post hoc analyses in GraphPad Prism Version 9 software.

2.6. Ethidium Bromide and Lucifer Yellow Dye Uptake Assay to Assess Hemichannel/Gap Junction Permeability

The Ethidium Bromide and Lucifer Yellow Dye Uptake Assay were conducted under Ca²⁺-free conditions described previously and adapted for use with a 96-wells fluorescence plate reader. Both hemichannel and gap junction permeability in cultured astrocytes were induced by exposing the cells to a Ca²⁺-free solution. In order to assess the hemichannel permeability, ethidium bromide was applied at a final concentration of 4 μM and incubated at 37°C for 10 minutes. For the purposes of checking the permeability of gap junctions, lucifer yellow was applied at a final concentration of 2 mM and incubated at room temperature for 20 minutes. Cells were washed with Hank’s balanced salt solution (HBSS, Corning Cellgro, Manassas, VA; Ref: 21-022-CM) supplemented with 1.2 mM CaCl₂ (HBSS-Ca²⁺). Treatment-induced changes in fluorescence were then assessed on a Flexstation 3 Microplate Reader, using 520 nm excitation/610 nm emission for ethidium bromide and 428 nm excitation/540 nm emission for lucifer yellow. Resulting between-group differences were assessed using One-Way ANOVA followed by Tukey post hoc analyses in GraphPad Prism Version 5 software.

3. Results

3.1. E2 and 3β-diol each protect against metabolic and oxidative stress in cortical astrocytes.

The cytoprotective effects of these three steroid hormones against the metabolic and pro-oxidative stressor, iodoacetic acid (IAA), were evaluated. Figure 1 shows that 18 hours of pretreatment with equimolar (100 nM) concentrations of both E2 and 3β-diol, but not DHT, protected primary cortical astrocytes against IAA-induced cell death. The protective effects of E2 are known to be at least partially mediated by estrogen receptors (in addition to estrogens’ antioxidant properties) (53). To validate the involvement of estrogen receptors in the protective are known to be at least partially mediated by estrogen receptors (in addition to estrogens’ effect of 3β-diol against IAA, we performed the cytoprotective assay in the presence or absence of ER antagonists that included ICI 182,780, an antagonist for both ER-alpha and ER-beta; MPP, an ER-alpha antagonist; and PHTPP, an ER-beta antagonist (Figure 2). Since both ICI and PHTPP, but not MPP, inhibited the protective effect of 3β-diol against IAA-induced cell death, we interpreted that the protective effects of 3β-diol were mediated by estrogen receptor-β (Figure 2).

Figure 1: Estradiol (E2) and 3β-diol protect primary cortical astrocytes from IAA-induced cell death.

Primary cortical astrocytes were treated with E2, DHT, and 3β-diol for 18 hours prior to treatment with the oxidative/metabolic insult, IAA, for an additional 6 hours. Cell viability, as assessed by the MTT viability assay, revealed cell death following IAA, while E2 and 3β-diol were protective. Data were normalized to the vehicle-treated control and are presented as average ± SEM. Tukey’s post hoc analysis indicated statistical significance with p < 0.05. * = different from IAA alone with p < 0.05, n= 4 independent replicates.

Figure 2: 3β-diol mediated protection in astrocytes is estrogen receptor-dependent.

Primary cortical astrocytes were treated with 3β-diol with or without co-application of ER antagonists (ICI, an antagonist for both ER alpha and ER beta; MPP, ER alpha antagonist; PHTPP, ER beta antagonist) for 18 hours prior to treatment with the oxidative/metabolic insult, IAA, for an additional 6 hours. Cell viability was assessed by the MTT viability and revealed cell death following IAA, while 3β-diol protection was dependent upon estrogen receptors. Data were normalized to the vehicle-treated control, and are presented as average ± SEM. One-Way ANOVA results. Tukey’s post hoc analysis indicated statistical significance with p < 0.05. * = different from IAA alone with p < 0.05, # = different from indicated steroid + IAA, n= 4 independent replicates.

3.2. E2 and 3β-diol inhibited IAA-induced Cx43 Gap junction opening and were associated with steroid-induced protection against IAA toxicity.

Recently published data implicated the role of Cx43 hemichannel or gap junction as possible contributors to optimal astrocyte function when associated with Gap19 and Gap26, respectively (54,55). Moreover, like Cx43, Pannexin-1 channels are also widely expressed in immune cells and are involved in the regulation of inflammatory responses (56–59). To clarify the roles of these regulators of Cx43 function on estrogen-induced cytoprotection, we assessed the effects of the steroid hormones in the presence or absence of Gap26, a Cx43 gap junction inhibitor; TAT-Gap19, a hemichannel targeting inhibitor; and 10Panx, a pannexin channel inhibitor (Figure 3). While Gap26, which inhibits Cx43-containing gap junctions with little influence on hemichannels, enhanced the protective effects of E2 and 3β-diol against IAA-induced toxicity, TAT-Gap19, and 10Panx, didn’t produce a similar effect. Given the robust association between Cx43 gap junction opening (i.e., increased permeability) and cytotoxicity, we proposed that E2 and 3β-diol may exert protection against oxidative stress in astrocytes by attenuating IAA-induced Cx43 gap junction permeability.

Figure 3: Gap26, a Cx43 Gap junction inhibitor, enhanced the protective effect of E2 from IAA-induced cell death.

Oxidative stress was induced in cortical astrocytes by 6 hours of IAA exposure. While the pharmacological inhibitor of Cx43 Gap Junction, Gap26, enhanced the protection of the cells by E2 and 3β-diol from IAA-induced toxicity (B), TAT-Gap19, Cx43 hemichannel blocker (Al), and 10Panx, Panx receptor blocker (C), did not produce a parallel effect. In graphs, data were normalized to the vehicle-treated control, and are presented as average ± SEM. For graphs, Tukey’s post hoc analysis indicated statistical significance with p < 0.05. For graphs, * = different from vehicle+IAA with p < 0.05, n= 3 independent replicates. # = different from hormone+IAA with p < 0.05.

3.3. Hormone-induced changes in expression of Cx43 and Gap26 in primary astrocytes.

Next, we determined the relative expression of Cx43 and Gap26 in response to E2, DHT, and 3β-diol in primary cortical astrocytes. Using real-time PCR and Western blot to profile the mRNA and protein levels of Cx43, we found no significant changes in protein expression in response to either E2, DHT, or 3β-diol. However, all three hormones elicited an increase in the levels of mRNA for Gap26 (Figure 4). We hypothesized that E2, DHT, and 3β-diol would enhance the expression of Gap26 to induce neuroprotection against oxidative stress. Indeed, these steroids increased the expression of Gap26 mRNA, but not the total protein expression. Since hyperphosphorylation of Cx43 at either Tyr 265 or Ser 368 is associated with the channel closing (30), future studies are aimed at exploring whether the phosphorylation of these Gap26 amino acid residues in response to treatment with E2, DHT, and 3β-diol.

Figure 4: Effect of Estradiol (E2), DHT, and 3β-diol on Gap 26 and Cx43 expression.

A. Gap26 and Cx43 mRNAs expression was assessed following 18 hours of treatment with E2, DHT, 3β-diol using real-time PCR. E2, DHT, and 3β-diol didn’t alter Cx43 mRNA expression in primary cortical astrocytes. However, all three steroids significantly increased the Gap26 mRNA. Data were normalized to the vehicle-treated control and are presented as average ± SEM. B-D. E2, DHT, and 3β-diol failed to alter total Cx43 protein expression compared to baseline expression. Cx43 and GAPDH protein expression was assessed following 3 and 18 hrs treatment with E2 (B), DHT (Cl), and 3β-diol (D) using Western Blot. Neither E2, DHT, nor 3βdiol elicited a statistically significant change in Cx43 protein expression at either timepoint assessed. Data were normalized to the vehicle-treated control, and are presented as average ± SEM. For each graph, Tukey’s post hoc analysis indicated statistical significance with p < 0.05; * = different from DMSO vehicle control with p < 0.05. Treatment with IAA for 6 hours resulted in the significant flux of EtBr into the cells, serving as a surrogate indicator of Cx43 permeability.

3.4. E2 and 3β-diol inhibited IAA-induced cell death associated with Cx43 Gap junction opening.

While using the pharmacological regulators of Cx43 channels provided important insight into how E2 and 3β-diol engage Cx43 gap junctions to promote cell viability, we wanted to assess, more directly, the impact of these steroid hormones on Cx43 channel permeability. To this end, we used the ethidium bromide (EtBr) uptake assay (60) and lucifer yellow (Ly) uptake assay (61) to measure the permeability of hemichannels and gap junctions, respectively (Figure 5). Our data revealed that hemichannel permeability was not changed by IAA or hormone treatment. However, Ly permeability was increased significantly by IAA, and the enhancement of Ly permeability was diminished by all three hormones, strengthening our interpretation that the protective effects of E2 and 3β-diol against IAA-induced cell death were achieved by inhibiting Cx43-containing gap junctions.

Figure 5: Treatment of Primary Cortical Astrocytes with Iodoacetic acid (IAA) opens the Cx43 Gap junction.

Oxidative stress was induced in cortical astrocytes by 6 hours of IAA exposure. Treatment with IAA for 6 hours resulted in the flux of EtBr (hemichannel indicator) or lucifer yellow (gap junction indicator) into the cells, serving as surrogate indicators of Cx43 permeability. Whereas E2, DHT, and 3β-diol didn’t alter the increased permeability of Cx43 hemichannels elicited by IAA (left panel), all three steroids reduced the permeability of Cx43 gap junction elicited by IAA (right panel). Data were normalized to the vehicle-treated control, and are presented as average ± SEM. For each graph, Tukey’s post hoc analysis indicated statistical significance with p < 0.05. * = different from vehicle with p < 0.05, # = different from IAA+vehicle with p < 0.05.

3.5. Carbenoxolone enhances the neuroprotective role of E2 and 3β-diol against IAA-generated oxidative stress.

To complement the pharmacological strategy to dissect the role of gap junction in mediating the neuroprotective effect of E2 and 3β-diol against the IAA-generated cell death, carbenoxolone was used to determine whether inhibition of gap junction enhances the neuroprotective effect by E2 or 3β-diol. Our data show that the cytotoxicity associated with IAA correlated with an increase in Cx43 gap junction permeability, and that pharmacological inhibition of Cx43 gap junction permeability (confirmed by direct evaluation of Cx43 gap junction permeability) prevented the cytotoxic effects of IAA (Figure 6).

Figure 6: Carbenoxolone enhances the neuroprotective role of E2 and 3β-diol against the IAA-induced oxidative stress by blocking the Gap junction of Connexin 43.

The E2-, DHT- or 3β-diol-associated inhibition of IAA-induced Cx43 gap junction permeability was associated with their effects on cell viability. Treatment of primary cortical astrocytes with IAA significantly reduced cell viability, while the pharmacological inhibitor of Cx43 gap junction, carbenoxolone, protected the cells from IAA-induced toxicity with steroids. In graphs, data were normalized to the vehicle-treated control, and are presented as average ± SEM. Tukey’s post hoc analysis indicated statistical significance with p < 0.05. For both graphs, * = different from Vehicle+IAA with p < 0.05, n= 3 independent replicates.

4. Discussion

The significant decline in circulating estrogen concentrations following the menopause has been correlated with a higher prevalence of such progressive neurodegenerative disorders as Alzheimer’s disease (62). However, the cellular and molecular mechanisms that contribute to this higher prevalence remain to be fully understood. Although healthy brain function requires the optimal function of both neurons and glia, the focus of most research in the brain with respect to neuroprotectant identification/development, including the gonadal steroid hormones, has been focused on neurons. Here, we focused on astrocytes since they represent a relatively understudied target of the protective effects of cytoprotectants like estrogens. Moreover, astrocytes not only play key roles in supporting homeostasis and protection of the brain, but they also express high levels of connexin 43 (Cx43), the focus of this study. Further, Cx43 does not appear to be expressed appreciably in the neurons (63–65). As such, this study served to expand our understanding of how gonadal steroids influence astrocyte viability through their action on cortical astrocyte Cx43.

Altered Cx43 function has been implicated in Alzheimer’s disease, and ischemic stroke pathology (8,66,67), and its function is also altered in those tissues affected following the menopausal transition, including the bone, heart, and reproductive tissue (26,30,68). Therefore, we postulated that studying how estrogens regulate Cx43 in the brain would not only advance our understanding of how estrogens exert their beneficial effects on the CNS but may shed light on the mechanism of action of steroid hormones in peripheral tissues as well.

Both Cx43 gap junctions and hemichannels contribute to astrocyte function, but the factors that regulate astrocyte Cx43 channels are not fully understood. The neurosteroids E2 and DHT, are known to regulate Cx43 gap junction expression and Cx43 channel opening in heart, bone, and reproductive tissue (26,28,68). However, E2 and DHT-mediated regulation of Cx43 channel expression and Cx43 channel opening in astrocytes have yet to be described. In addition, most studies of Cx43 fail to consider the separate contributions of Cx43-associated gap junctions and hemichannels. Another unique aspect of our study was to assess whether the cytoprotective effects of DHT or its influence on Cx43 were mediated by prior conversion of DHT to an estrogenic metabolite, such as 3β-diol. This is of particular importance in astrocytes since these cells appear to be devoid of the classical androgen receptor (69).

While we initially postulated that DHT, in a cell type that does not express appreciable levels of the androgen receptor (AR), could act through the estrogen receptor via its estrogen metabolite, 3β-diol, our data did not support this. Given the lack of protective effects of DHT in the cortical astrocytes, against the mixed metabolic/oxidative stress-inducing insult, IAA, we suggest one possibility where the expression of converting enzymes (e.g., 3β-hydroxysteroid dehydrogenase) may not have been present in sufficient abundance to catalyze the conversion of DHT to 3β-diol. Interestingly, cortical astrocytes treated with Gap26 protected against IAA-induced cytotoxicity (Figure 3B). Moreover, like E2 and 3β-diol, DHT also reduced the permeability of the Cx43 gap junction elicited by IAA, which was a bit surprising given that our lab has not found our cultured primary cortical astrocytes to express the classical AR. From these data, we postulate that DHT may induce the expression of Gap26, which in turn, reduces the permeability of the Cx43 gap junction. The effect might have been mediated by other androgen receptors such as the truncated AR, AR-22 (70), or the mAR (71). However, the effect on gap junction permeability may, in and of itself, not have been enough to protect the astrocytes against IAA-induced oxidative stress due to insufficient conversion enzymes, to include 3β hydroxysteroid dehydrogenase. Future studies, aimed at increasing the expression of requisite converting enzymes and assessing the effects of DHT on Gap26 associated with Cx43 gap junction, are planned. In addition, noting that our focus in this study were cerebral cortical astrocytes, there may be also value to determining if the results here extend to astrocytes derived from other brain regions.

To our knowledge, this was the first study to assess the effects of these steroids on Cx43 expression and Cx43 channel permeability in any brain tissue. Additionally, this was the first study to demonstrate that 3β-diol, can promote astrocyte viability through an estrogen receptor-mediated mechanism. Our findings that E2 and 3β-diol inhibited IAA-induced opening of Cx43 gap junction is therefore consistent with their cytoprotective effects. Furthermore, the protective efficacy of these steroids was equivalent to that of carbenoxolone, a compound that selectively inhibits Cx43 gap junction permeability, and as such, bolstered our overall conclusion that Cx43 gap junctions are important targets that mediate estrogen-induced protection of brain tissue. We acknowledge, however, that there is an opportunity to evaluate the efficacy of these steroid hormones at different times relative to the application of the stressor/insult. Such studies may inform whether the presence of the hormone helps prevent the damage-associated with oxidative/metabolic stress, or whether these hormones could be used as therapeutics to already stress/damaged cells or tissue. Such studies are part of our ongoing efforts to explore mechanisms underlying the effects of these hormones.

The broader implication of our studies is that Cx43 gap junctions are a potentially viable target to exploit in enhancing gonadal hormone-induced cytoprotection in certain neurodegenerative diseases. Supporting this interpretation are reports that the Cx43 gap junction blocker, carbenoxolone, inhibits the aggregation of Aβ42 peptide and destabilizes the Aβ42 aggregates (72). Furthermore, carbenoxolone prevents the Aβ42 oligomer–induced oxidative damage and anxiety-like behavior in rats (73). Moreover, a study reported that 3β-diol is a neurosteroid agent that mediates the impact of testosterone on neuronal activity and seizure susceptibility based on its capacity to modify GABA_a receptors (74). It may be especially crucial in men with epilepsy when androgen levels decrease during aging. Taken together, this information underscores astrocyte Cx43 gap junctions as an understudied, but important, therapeutic target for brain protection in males and females, both in vitro and in vivo. These effects on astrocytes do not, however, negate the critical importance of the inter-play between astrocytes and neurons such that astrocytes may be a critical cellular mediator of the neuroprotective effects of gonadal hormones. Indeed, we have shown recently that potential paracrine signaling between astrocytes and neurons influences neuronal viability (48). Future studies will determine if such paracrine signaling are potentially mediated by effects on Cx43 gap junction permeability in astrocytes.

As a first step to the study of steroid hormone action on Cx43 channels, and the relationship to such regulation of cell viability, we conducted our experiments under conditions where the timing of hormone administration occurred prior to insult, and thus, under healthy/non-insulted conditions. Under such conditions, our study demonstrated that estradiol and 3β-diol protected cortical astrocytes against subsequently applied oxidative/metabolic insult. However, it is unclear if steroid hormones will reveal the same protective effects in an unhealthy environment, including the steroid treatment with oxidative stress simultaneously or after. Further investigation will help determine if there is a healthy cell bias, as has been suggested by others (75).

Based on our findings, we propose that Cx43 channels play a role in regulating astrocyte viability. However, a more complete understanding of the contributions of Cx43 gap junctions is still warranted. Such enhanced understanding of these channels, along with their interplay with steroid hormone neurobiology, may facilitate the development of more precise therapeutics (i.e., enhancing cytoprotective therapeutic efficacy by selectively targeting Cx43 gap junctions in the target tissue) to enhance the beneficial actions of estrogens, and potentially androgens, in the brain. Furthermore, since alterations in Cx43 channel opening have been associated with a wide range of diseases that present significant epidemiological and economic burdens, including cancer (76), heart disease (28), and ischemic stroke (67), the impact of this research could extend to the treatment of these diseases as well.

Highlights (MCE-D-23-00140).

• Astrocytes are an important subpopulation of cells in the brain

• Estrogens, to include estradiol and 3β-diol, have brain protective effects that include effects on astrocytes

• Dihydrotestosterone, a member of a family of hormones called androgens, can have important effects on brain health through its metabolite, 3β-diol

• Cx43-containing channels on the membranes of astrocytes are novel targets of estrogens, the impact of which is to positively influence brain cell health and viability