Estrogen receptor beta activity contributes to both tumor necrosis factor alpha expression in the hypothalamic paraventricular nucleus and the resistance to hypertension following angiotensin II in female mice

Abstract

Sex differences in the sensitivity to hypertension and inflammatory processes are well characterized but insufficiently understood. In male mice, tumor necrosis factor alpha (TNFα) in the hypothalamic paraventricular nucleus (PVN) contributes to hypertension following slow-pressor angiotensin II (AngII) infusion. However, the role of PVN TNFα in the response to AngII in female mice is unknown. Using a combination of in situ hybridization, high-resolution electron microscopic immunohistochemistry, spatial-temporal gene silencing, and dihydroethidium microfluorography we investigated the influence of AngII on both blood pressure and PVN TNFα signaling in female mice. We found that chronic (14-day) infusion of AngII in female mice did not impact blood pressure, TNFα levels, the expression of the TNFα type 1 receptor (TNFR1), or the subcellular distribution of TNFR1 in the PVN. However, it was shown that blockade of estrogen receptor ß (ERß), a major hypothalamic estrogen receptor, was accompanied by both elevated PVN TNFα and hypertension following AngII. Further, AngII hypertension following ERß blockade was attenuated by inhibiting PVN TNFα signaling by local TNFR1 silencing. It was also shown that ERß blockade in isolated PVN-spinal cord projection neurons (i.e. sympathoexcitatory) heightened TNFα-induced production of NADPH oxidase (NOX2)-mediated reactive oxygen species, molecules that may play a key role in mediating the effect of TNFα in hypertension. These results indicate that ERß contributes to the reduced sensitivity of female mice to hypothalamic inflammatory cytokine signaling and hypertension in response to AngII.

1. INTRODUCTION

Globally, hypertension is a leading risk factor for increased morbidity and mortality associated with diseases of the heart, blood vessels, kidney, and brain (Zhou et al 2021). The most common form of hypertension, essential or primary hypertension, is associated with a gradual rise in blood pressure and an increase in sympathetic activation (Grassi 2021), suggesting a significant neurogenic component. An established model of hypertension achieved by continuous systemic administration of a subpressor dose of angiotensin II [AngII; (Dickinson & Lawrence 1963)] mimics these key features of hypertension (Lerman et al 2019), including activation of critical sympathetic regulating circuits involving the paraventricular nucleus (PVN) of the hypothalamus (Bains & Ferguson 1995, Llewellyn et al 2012).

In male mice, inflammatory cytokine signaling in the PVN has been shown to play an important role in hypertension induced by AngII (Shi et al 2010) as well as other hypertensive challenges (Dai et al 2015, Dange et al 2015, Sriramula et al 2013). Tumor necrosis factor alpha (TNFα), a glial cell-derived cytokine, has been reported to contribute to both increased sympathetic activation (Shi et al 2011) and hypertension (Sriramula et al 2013) through actions in the PVN. Further, there is evidence that TNFα signaling in the PVN heightens sympathetic nervous system activation through a mechanism involving excitatory neuronal signaling (Mourão et al 2021) and also plays a role in the pressor response to AngII mediated by the TNFα type 1 receptor (TNFR1) in males (Woods et al 2021).

Significantly, there is an important sex divergence in the susceptibility to hypertension (Reckelhoff 2018) and inflammation (Gillis & Sullivan 2016). Women are protected from hypertension at reproductive age but become increasingly vulnerable at the onset of menopause (Van Kempen et al 2016a). Compared to males, female rodents also are less sensitive to various forms of experimental hypertension including the AngII model (Girouard et al 2009, Marques-Lopes et al 2017, Xue et al 2005). However, the role of TNFα in blood pressure regulation in females, particularly in response to AngII, is unclear.

The mechanisms subserving female protection from hypertension are not fully understood, however, gonadal hormones may play a significant role. For example, following ovariectomy mice show a greater increase in blood pressure in response to AngII (Xue et al 2005). Additionally, in a model of chemical accelerated ovarian failure, which mimics the hormonal changes seen in human menopause, female mice also respond with an increase in blood pressure following AngII (Marques-Lopes et al 2017, Milner et al 2021, Van Kempen et al 2016b). Additionally, it has been shown that estrogen receptor beta (ERß), the major estrogen receptor in the PVN (Contoreggi et al 2021), plays an important role in AngII-dependent blood pressure regulation in females (Marques-Lopes et al 2017, Milner et al 2021, Ovalles et al 2019). However, the role of ERß in TNFα-mediated signaling, particularly in the context of blood pressure regulation, in females is not clear.

Inflammatory cytokines have been shown to influence the production of free radicals, which also play a critical role in hypertension in males (Coleman et al 2013, Kang et al 2008, Kang et al 2009, Oliveira-Sales et al 2009). The canonical NADPH oxidase (NOX2) is a significant generator of brain reactive oxygen species (ROS) implicated in hypertension (Zimmerman & Davisson 2004), particularly via actions in the PVN (Coleman et al 2013, Wang et al 2013). However, little is known about the role of TNFα in ROS signaling in the PVN of female mice.

Despite the significant role of PVN TNFα signaling in AngII hypertension in male mice, little is known about the role of this inflammatory cytokine in the context of blood pressure control in females. In the present study, a combination of molecular, high resolution anatomical, gene silencing, and dihydroethidium (DHE) microfluorographic approaches were used to examine the role of PVN TNFα signaling in hypertension in young adult female mice.

2. METHODS

2.1. Subjects.

The experimental subjects were adult wild-type female (N=91) and male (N=3) C57BL/6 mice, and female TNFR1 knockout mice (N=3) on the C57BL/6 background. Wild-type animals were bred and maintained in a vivarium located at Weill Cornell Medicine, and knockout mice were purchased from The Jackson Laboratory (Cat. No. 3242; Bar Harbor, ME). All mice were housed 2–5 per cage, with unlimited access to rodent chow and water, and maintained on a 12-hr light/dark cycle. All experiments were approved by the Institutional Animal Care and Use Committees at Weill Cornell Medicine in accordance with guidelines established by the National Institutes of Health Guide for the Care and Use of Laboratory Animals. All efforts were made to minimize the number of animals used and their suffering.

2.2. Slow-pressor response to AngII.

Female mice have been shown to have reduced sensitivity to the pressor effects of AngII, and we were interested in whether blocking ERß would impact blood pressure after AngII. Although tail-cuff plethysmography may be insensitive to subtle changes in systolic blood pressure (SBP), this method is able to detect large changes in pressure (> 15 mmHg systolic pressure) as described (Gross & Luft 2003). Indeed, the tail-cuff method is a sensitive and reliable approach for detecting large between-group changes in SBP in female mice in a model of accelerated ovarian failure (Marques-Lopes et al 2017, Marques-Lopes et al 2014). Further, as the immunohistochemical methods used in these studies require aortic arch perfusion fixation (Milner et al 2011a), the tail-cuff approach is preferred given that it is noninvasive and does not compromise the carotid artery, unlike telemetric blood pressure recording that requires vascular catheterization (Butz & Davisson 2001). The blood pressure recording protocol used strict handling and training steps to minimize stress (Woods et al 2021). Further, mice were sacrificed one day after the last blood pressure recording to minimize any impact of stress on in situ hybridization and immunohistochemical measures (Marques-Lopes et al 2017, Marques-Lopes et al 2014).

Before blood pressure recording and pump implantation, mice were acclimated to the blood pressure recording room and instrumentation for 1-week. Non-invasive plethysmography was used to record SBP in awake mice using a Hatteras MC-4000 tail-cuff blood pressure system (Cary, NC). A total of 10–20 blood pressure measurements were recorded at baseline and following pump implantation in each recording session. For each mouse, blood pressure values were averaged during each session, and these values were averaged across animals to produce group mean SBP values at each timepoint across treatments.

After baseline SBP recording, mice were prepared for AngII infusion. Prior to implantation, mini-pumps (ALZET^®, Cupertino, CA) were filled with either vehicle (0.1% bovine serum albumin [BSA] in saline, termed “Sal”) or AngII dissolved in Sal (600ng/kg⁻¹/min⁻¹). Mice were surgically implanted with osmotic mini-pumps under isoflurane anesthesia (2–4%). Pumps were implanted subcutaneously in the dorsal flank region as described (Capone et al 2012, Coleman et al 2013). Sal or AngII were delivered over 14 days. Measurements of SBP were made before (baseline), and at 2 to 3-day intervals after minipump implantation. For in situ hybridization and immunohistochemical studies, mice were tested on day 13 and euthanized on day 14 to minimize the possibility of handling stress impacting gene/protein expression or receptor localization (Marques-Lopes et al 2015). To block ERß, mice were administered the antagonist 4-[2-Phenyl-5,7-bis(trifluoromethyl)pyrazolo[1,5-a]pyrimidin-3-yl]phenol (PHTPP; 3 mg/kg in sesame oil, s.c.) daily. Prior reports demonstrate that systemic administration of PHTPP produces effects consistent with central actions. Indeed, PHTPP has been shown to block the antidepressant-like effects of estrogenic compounds (Valdés-Sustaita et al 2021) and suppress hippocampal markers of synaptic plasticity (Xing et al 2018). Significantly, in ovariectomized female rats, systemic PHTPP injection has also been shown to modulate PVN NADPH-diaphorase expression (Grassi et al 2013).

2.3. Estrous cycle determination.

Immediately prior to anesthesia, estrous cycle stage was assessed by vaginal smear cytology (Turner & Bagnara 1971). Female mice used in this study were determined to be in estrus or diestrus at the time of sacrifice.

2.4. In situ hybridization.

At the conclusion of the experiment, mice were deeply anesthetized with sodium pentobarbital (150 mg/kg, i.p.). Brains were rapidly fixed by aortic arch perfusion. For this, animals were perfused sequentially with 5 ml of 1000 units/ml of heparin in 0.9% saline, and then by 35 ml of 4% paraformaldehyde (PFA) in 0.1 M phosphate buffer (PB, pH 7.4). All perfusion fluid was delivered at a flow rate of 20 ml/minute. Following dissection from the skull, brains were post-fixed in 4% PFA in 15% sucrose in PB overnight and then incubated for 1–2 days in a cryoprotectant solution consisting of 30% sucrose in PB. After sinking, brains were frozen and cryosectioned (30 μm) on a cryostat. Collected coronal brain sections were placed in cryoprotectant solution (30% sucrose, 30% ethylene glycol in PB) and stored at −20°C until in situ processing.

Forebrain sections containing the caudal PVN subdivision [1.0–1.2 mm caudal to bregma (Paxinos & Franklin 2000)], an area containing a low incidence of neuroendocrine neurons but a high density of preautonomic neurons projecting to the spinal cord (Biag et al 2012, Contoreggi et al 2021), were selected for analysis. After sections were selected, they were punch coded and mounted on a single glass slide and run in tandem. This was done to ensure that all brain sections were exposed to identical labeling conditions. To measure TNFR1 mRNA, tissue was processed using the RNAscope^® 2.0 HD Detection Kit (ACD, Cat. No. 310035) as described (Johnson et al 2021). At the start of tissue processing, slides were dehydrated overnight and then heated at 60°C for 1-hr. Sections were then quenched and incubated in target retrieval solution and protease. This was followed by hybridization of sections for 2-hours at 40°C with probe (mouse TNFR1 target region 331–1345; Mm-Tnfrsf1a Cat No. 426541). In addition, control sections were hybridized with the positive control probe Ppib (Cat No. 313911) and other sections were run with the bacterial negative control probe DapB (Cat No. 310043). The chromogen 3, 3’-diaminobenzidine (DAB) was used to detect hybridization signal, followed by counterstaining with blue hematoxylin.

An Eclipse Nikon 80i microscope (Melville, NY) connected to a Micropublisher 5.0 digital camera (Q-imaging) and IP Lab software was used to take images of the PVN (20x). For quantification, the number of black particles overlying each blue Nissl-stained cell were made using Halo^® Software (Indica Labs, Albuquerque, New Mexico) as per ACD recommendations (indicalab.com/wpcontent/uploads/2018/04/MK_51_103_RNAScope_data_analysis_guide_RevB.pdf). The number of particles per cell in the PVN were counted for each animal and these were averaged across treatments. In addition, the number of labeled cells were also counted in the PVN of each animal, and these were also averaged across treatments.

A total of 12 mice were used to examine the effect of AngII on TNFR1 gene expression and another 12 mice were used to investigate the impact of the

2.5. Primary Antisera.

A rabbit polyclonal antiserum (Abcam, Cat. No. AB223352) was used to label TNFR1 (Glass et al 2017). To label green fluorescent protein (GFP), a chicken polyclonal antiserum (Aves Labs, Cat. No. GFP-1020) was used (Marques-Lopes et al 2015). A rabbit polyclonal antibody (Invitrogen, Cat. No. PA5–19810) was used to label TNFα (Potor et al 2021). A rabbit polyclonal antiserum (Millipore, Cat. No. ABN1480) was used to label the NOX2 organizing protein p47^phox (Coleman et al. 2013). To label p47^phox phosphorylated at serine 304 (Ser304), a rabbit polyclonal (Invitrogen, Cat. No. PA5–36773) antiserum was used (Xu et al 2021).

2.6. Immunohistochemistry for light microscopy.

Light microscopic immunohistochemistry was performed as described (Woods et al 2021). The mice were deeply anesthetized with sodium pentobarbital and their brains were fixed by transcardial perfusion with 4% PFA in PB. Brains were post-fixed in 4% PFA in PB overnight followed by vibratome sectioning (30–40 μm). To minimize nonspecific labeling, sections were incubated in 0.5% BSA in 0.1 M TBS for 30-minutes. Sections were then incubated for 48-hrs in rabbit anti-TNFR1 (1:100), anti-TNFα (1:500 in 0.1% Triton X-100), anti-p47^phox (1:100), anti-47^phox Ser304 (1:100), or chicken anti-GFP (1:1000) primary antisera in 0.1% BSA. Tissue was then incubated for 30-min in anti-rabbit or anti-chicken IgG conjugated to biotin for 30-minutes, followed by incubation in avidin-biotin complex at half the manufacturer’s recommended dilution in TBS for 30-min. Sections were then incubated in DAB (Millipore Sigma, Milwaukee, MI) and H₂O₂ in TBS for 6-min. All incubations were separated by washes in TBS followed by PB. After the final washes, forebrain sections were mounted on 1% gelatin-coated glass slides in 0.05 M PB. They were then dehydrated in an ascending series of alcohol and xylene, followed by covering by a glass coverslip in DPX (Sigma-Aldrich, St. Louis, MO). Labeled brain sections were then examined using a Nikon light microscope.

2.7. Light microscopic densitometry.

Densitometric quantification within the PVN was performed using previously described methods (Milner et al 2022, Pierce et al 2014). Forebrain sections were photographed using a microscope (Nikon Eclipse 80i) interfaced to a digital camera (Micropublisher 5.0, Q-imaging, BC, Canada) and IP Lab software (Scanalytics IPLab, RRID: SCR_002775). The region of interest (ROI) was highlighted, and the average pixel density determined using ImageJ64 (Image J, RRID:SCR_003070) software. To control for variations in illumination between images as well as background labeling, the pixel density of a region lacking labeling (white matter) was calculated and subtracted from ROI measurements. It has been previously reported that there is a strong linear correlation between average pixel density and actual transmittance demonstrating the accuracy of the technique (Pierce et al 2014). Density measures were determined for each mouse, and these were averaged across treatments. Data are presented as a percent of the control (i.e., Sal) condition.

2.8. Tissue preparation and immunoelectron microscopy procedures.

Caudal PVN sections were prepared for immunohistochemistry using previously described methods (Milner et al 2011b). Mice were deeply anesthetized with sodium pentobarbital (150 mg/kg, i.p.) and then perfused intracardially with: (a) 15 ml of 1000 units/ml of heparin in 0.9% saline, and (b) 40 ml of a mixture of 3.75% acrolein/ 2% PFA in 0.1 M PB. Brains were removed and then post-fixed in a solution of 2% PFA for a duration of 30-minutes. Forebrains were sectioned coronally through the rostro-caudal extent of the PVN (40 μm) with a VT1000X vibratome (Leica Microsystems) and stored at −20^oC in cryoprotectant until immunocytochemical processing.

Fixed brain sections were processed for immunohistochemical labeling and detection of TNFR1 using an immunogold-silver (IGS) labeling method (Milner et al 2011b). After punch coding and pooling brain sections into a single vial, excess aldehydes were removed by incubating tissue in 1.0% sodium borohydride in PB. Sections were then blocked by a 30-minute incubation in 0.5% bovine serum albumin (BSA) in 0.1 M Tris-buffered saline (TBS). Brain sections were then incubated for 48-hours in a primary rabbit anti-TNFR1 antiserum in 0.1% BSA (1:100). In preparation for IGS labeling, brain tissue was incubated for 10-minutes in a blocking solution consisting of 0.8% BSA and 0.1% gelatin in in 0.01 M phosphate-buffered saline (PBS; pH 7.4). This was followed by 2-hr incubation in anti-rabbit 1 nm gold particle-conjugated IgG [1:50, Electron Microscopy Sciences (EMS), Fort Washington, PA] diluted in the blocking solution. Sections were next placed in 2% glutaraldehyde in PBS for 10-minutes. Enlargement of nano-gold particles was performed with the IntenSE-M silver-enhancement kit during a 6-min incubation (GE Healthcare, Waukesha, WI).

2.9. Electron Microscopy.

In preparation for electron microscopy, immunohistochemically processed forebrain tissue sections were post-fixed in 2% osmium tetroxide (EMS, Fort Washington, PA) in PB for 1-hr. After hardening, brain sections were flat embedded in a matrix of EMbed 812 [Electron microscopy Sciences (EMS)] and covered in Aclar.

Prior to ultrathin sectioning, a trapezoidal section of the PVN was cut, then mounted and fixed on an Epon block. Using a diamond knife and ultramicrotome (Ultratome, NOVA, LKB, Bromma, Sweden), the surface of the PVN sample was then cut in ultrathin sections (60–80 nm), which were collected on 400-mesh, thin-bar copper grids (EMS). Samples were counterstained with uranyl acetate and Reynold’s lead citrate (Milner et al 2011b). These sections were examined with a transmission electron microscope (Tecnai 12 BioTwin, FEI, Hillsboro, OR). The electron microscope was interfaced to a digital camera (Advantage HR/HR-B CCD Camera System, Advanced Microscopy Techniques, Danvers, MA) that allowed for the collection of digital micrographs from the analyzed ultrathin sections.

2.10. Ultrastructural analysis.

The analysis of ultrathin PVN sections was conducted in the zone between the embedment matrix and the tissue surface to ensure sampling was conducted in an area of consistent reagent penetrance. Profile identification was guided by the use of well-established ultrastructural criteria for neuronal and glial elements (Peters et al 1991). Profiles were considered of dendritic origin if they showed microtubule arrays, vesicular/endomembranous organelles, and/or postsynaptic contacts. Profiles of at least 0.2 μm in diameter also containing small synaptic vesicles were considered to be from axon terminals. If profiles were less than 0.2 μm without small synaptic vesicles, they were categorized as unmyelinated axons. Profiles lacking regular shape, devoid of organelles and/or presenting filaments were categorized as glial in origin.

Assessing the distribution of IGS labeling for TNFR1 in subcellular locations in dendritic profiles was performed using previously described methods (Glass et al 2015, Marques-Lopes et al 2016). Sampling was conducted from 30 randomly selected fields from each mouse in each treatment group. This amounted to a total of 15,606 μm² of tissue equally sampled from both groups. Quantitative morphological analysis of profile surface area and cross-sectional area was performed with the aid of Microcomputer Imaging Device software (MCID Imaging Research Inc., Ont., Canada; RRID:SCR_014278). The subcellular location of IGS particles was apportioned to the plasma membrane or the cytoplasm. Density ratios were determined by dividing the number of particles by profile size. These were calculated for each profile and then averaged for each animal in each treatment.

2.11. Virus administration.

Bilateral microinjections of viral vectors into the PVN were performed using previously described procedures for stereotaxic surgery (Glass et al 2015). Under isoflurane anesthesia, vectors were microinjected into the mouse PVN at a volume of ~100 nl (1.2×10¹² GC/ml) per side. Coordinates were ~1.0 mm posterior and 0.2 mm lateral to bregma, at a depth of 4.8 mm (Paxinos & Franklin 2000). Microinjections were made using a glass pipette connected to a picospritzer (Picospritzer II, General Valve Corp., Fairfield, NJ). After surgery mice were allowed to recover in their home cages for 21-days.

Silencing of TNFR1 was performed using a recombinant adeno-associated viral vector (AVV) expressing a mouse TNFR1 short hairpin RNA (AAV-TNFR1 shRNA). Expression of the targeting sequence (5’-CCTCGTGCTTTCCAAGATGAA-3’; Vector Biolabs, Cat. No. shRNA 274706) was driven by a U6 promoter. An enhanced GFP reporter was driven by a CMV promoter. In control mice, an AAV-GFP vector (Vector Biolabs, Cat. No. 7004) was microinjected into the PVN. The shRNA vector was validated for ~90% knockdown of mRNA in NIH/3T3 cells.

2.12. Retrograde labeling of PVN-spinal cord projection neurons.

Immediately prior to surgery, mice were anesthetized with 87.5 mg/kg ketamine and 12.5 mg/kg xylazine (i.p.) or isoflurane (2–4%). Using a surgical microscope, mouse spinal cords were exposed at the T₂-T₄ level as described (Wang et al 2013). Then, using a Hamilton syringe, rhodamine-labeled fluorescent microspheres (~0.5 μl; FluoSpheres, 0.04 μm; Molecular Probes, Eugene, OR) were manually injected (~5-min) into the intermediolateral nucleus, and the needle left in place for ~3-min before removal. Mice were left to recover at least 7-days prior to experimental use.

2.13. DHE microfluorography.

Levels of ROS were measured by DHE in isolated PVN cells from untreated mice or in projection neurons from mice microinjected with retrograde tracer into the spinal cord. Cells were enzymatically dissociated from PVN slices using 0.02% protease + 0.02% thermolysin. Isolated cells were mounted on a perfusion glass-bottom dish placed on the stage of an inverted microscope (Nikon E300, Japan). Cells were loaded with DHE (2 μmol/L) for 30-minutes in Mg²⁺-free buffer composed of (in mmol/L): NaCl 121, KCl 5, CaCl₂ 1.8, glycine 0.01, Na-pyruvate 1, glucose 20, NaCO₃ 26, NaH₂PO₄ 3, pH = 7.35. Using a Nikon E300 microscope equipped with a CCD camera, DHE was visualized with a Bromide-ethidium filter, and a FITC filter was used to identify rhodamine-labeled neurons. Time-resolved fluorescence was measured every 30-sec with an exposure time of 100-ms using image analysis software (RRID: SCR_002775, Scanalytics, Fairfax, VA). Recordings were started after stable baseline fluorescence readings were achieved. Baseline ROS-dependent fluorescence was expressed as relative fluorescence units. Changes in ROS fluorescence induced by TNFα (unlabeled cells: 1–100 ng/ml; labeled cells: 10 ng/ml) were expressed as the ratio of Ft/Fo, where Fo is the baseline fluorescence before application of drugs, and Ft is fluorescence in the same cell after TNFα application (Wang et al 2006). All background fluorescence was subtracted from the readings. In some experiments, cells were pretreated with PHTPP in DMSO, a polypeptide inhibitor of the NADPH oxidase gp91^phox docking sequence (gp91-ds) [H]GGGGCSTRIRRQL[NH2] (1 μmol/l), or scrambled peptide (gp91-sc; 1 μmol/l). To verify the stability of the preparation, parallel time control experiments in which drugs were not perfused were performed.

2.14. Outline of experiments

Experiment 1.

A total of 12 mice (6 Sal and 6 AngII) were used to examine the impact of AngII on PVN TNFR1 gene expression by in situ hybridization and TNFR1 and TNFα protein by immunohistochemistry in 2 sections per mouse.

Experiment 2.

A total of 6 mice (3 Sal and 3 AngII) were used to examine the effect of AngII on the subcellular localization of TNFR1 in dendritic profiles of PVN neurons by electron microscopy in 1 ultrathin section per mouse.

Experiment 3.

A total of 12 mice (6 AAV-GFP, 6 AAV-shRNA) were used to investigate the influence of TNFR1 silencing on blood pressure basally and after AngII.

Experiment 4.

To examine the impact of ERß blockade on blood pressure following AngII infusion, female mice were co-administered AngII along with vehicle (N=8) or the ERß antagonist PHTPP (N=8). Additionally, other groups of mice were also treated with Sal and vehicle (N=6) or PHTPP (N=6). Light microscopic in situ and immunohistochemical experiments for TNFR1 were conducted in animals infused with AngII and treated with PHTPP or vehicle (N=5, 2 sections per mouse), or in the case of TNFα in all groups (N=5, 2 sections per mouse). Electron microscopic analysis of TNFR1 was performed in animals infused with AngII and injected with PHTPP or vehicle (N=3 per group, 1 ultrathin section per mouse). To examine the role of PVN TNFR1 in hypertension following AngII and ERß blockade, mice were microinjected with control (N=6) or AAV-shRNA (N=6) vectors.

Experiment 5.

To investigate the influence of TNFα on free radicals, TNFα-induced ROS production was first measured in isolated PVN cells from female mice (N=3, 57 cells). Next, the effect of ERß blockade on TNFα-induced ROS production was further investigated in spinally-projecting neurons. The impact of TNFα and ERß blockade was examined in isolated PVN projection neurons treated with vehicle and vehicle, TNFα and vehicle, vehicle and PHTPP, or TNFα and PHTPP (N=3 per condition, 15–20 cells per condition). The role of TNFR1 in TNFα-induced ROS was further tested in isolated PVN projection neurons from TNFR1 KO mice (N=3, 10 cells). Additionally, the involvement of NOX2 in TNFα-induced ROS production was examined in isolated projection neurons treated with the NOX2 inhibitor gp91ds (N=3, 31 cells) or scrambled peptide (N=3, 12 cells). The impact of AngII on TNFα-induced ROS in PVN projection neurons was also investigated in mice infused with 14-day AngII (N=3, 16 cells). As a comparison with females, TNFα-induced ROS was also investigated in PVN-spinal cord projection neurons from male mice (N=3, 23 cells).

2.15. Drugs and reagents.

AngII (Cat. No. A9525), thermolysin, and TNFα (Cat. No. T7539) were purchased from Sigma. PHTPP (Cat. No. 2662/10) was purchased from Tocris. The gp91^phox peptide inhibitor gp-91ds-tat (Cat. No. 14425–01) and scrambled peptide (Cat. No. 14426–01) were synthesized by Bio-Synthesis (Lewisville, TX). DHE was purchased from Thermo Fisher (Cat. No. 11347).

2.16. Statistical analyses.

Analysis of the SBP results and ROS dose-response data were performed using repeated measures ANOVA followed by post-hoc testing (Tukey’s test). The in situ hybridization and immunohistochemical results were analyzed by t-tests. ROS studies in identified PVN neurons were analyzed by ANOVA. Prior to analysis distributions were tested for both normality and equality of variance. When distributions did not express equality of variance data were analyzed by Welch’s t-test or Welch’s ANOVA using the Dunnett test for multiple comparisons for post-hoc testing. Prism 8 (GraphPad) software was used to conduct statistical analyses.

2.17. Image preparation.

Contrast, sharpness, and/or brightness were adjusted in light and electron micrographs using Photoshop 11 software. Figures were constructed by importing images into PowerPoint. Prism 8 software was used to produce the graphical figures (GraphPad Software, La Jolla, CA).

3. RESULTS

3.1. Effect of AngII infusion on blood pressure and TNFR1 expression in PVN of female mice

To investigate the effect of AngII on blood pressure and TNFR1 expression in the PVN of female mice, animals were infused with either Sal (N = 6) or AngII (N = 6) for 14-days. There was no effect of Sal or AngII [F(1, 40) = 1.7, p > 0.2], session [F(4, 40) = 1.7, p > 0.16], or a treatment by session interaction [F(4, 40) = 1.1, p > 0.3, repeated measures ANOVA] with respect to SBP (Fig. 1A).

Fig. 1. Blood pressure and PVN TNFR1 mRNA are unaltered by AngII in female mice.

(A) Line graph showing systolic blood pressure (SBP) in female mice infused for 14-days with saline (Sal) or AngII. (B-E) Light micrographs illustrating TNFR1 mRNA in the caudal PVN of Sal-infused and AngII-infused mice at low magnification (top) and at a higher magnification (lower) in the areas shown in the dashed boxes. Cells showing TNFR1 particles are indicated by the arrowheads. (F) Dot plot showing TNFR1 mRNA presented as particles per cell is similar in Sal and AngII-treated animals. (G-J) Light micrographs showing TNFR1 immunolabeling in the PVN of Sal-infused and AngII-infused mice at low magnification (top) and at a higher magnification (lower) in the area shown in the dashed box. (K) Dot plot showing that immunolabeling density for TNFR1 is similar in Sal and AngII-treated animals. (L-O) Light micrographs showing TNFα immunolabeling in the PVN of Sal-infused and AngII-infused mice at low magnification (top) and at a higher magnification (lower) in the area shown in the dashed box. (P) Dot plot showing that immunolabeling density for TNFα is similar in Sal and AngII-treated animals. N = 6 mice/group. Scale bars: 100 μm in B, D, G, I, L, N and 20 μm in C, E, M, O and 40 μm in H, J. Data are presented as mean± SEM.

To investigate the effect AngII infusion on TNFR1 gene expression, forebrain sections containing the PVN were processed for in situ hybridization. Expression of TNFR1 was assessed in the caudal PVN, a region that contains a high concentration of sympathoexcitatory neurons and a region where TNFR1 is elevated following AngII hypertension in males (Woods et al 2021). Examples of TNFR1 mRNA in the PVN of Sal and AngII infused mice are presented in Figs. 1B–E. It was found that there was no difference in the number of TNFR1 probes per cell in Sal and AngII infused mice [t(10) = 0.3, p> 0.8, unpaired t-test; Fig. 1F).There was also no difference in the number of PVN cells with TNFR1 probe [Sal: 17.9± 6.6 versus AngII: 25± 4.6, t(10) = 0.9, p> 0.3; not shown]. In addition, TNFR1 protein labeling was measured by immunohistochemical densitometry. Examples of TNFR1 labeling are shown in Figs 1G–J. It was shown that TNFR1 labeling density was similar in Sal and AngII-infused mice [t(10) = 0.4, p> 0.9; Fig 1K]. Levels of TNFα were also assessed in the PVN of Sal and AngII-infused mice. Examples of TNFα labeling are shown in Figs 1L–O. It was found that Sal and AngII mice showed comparable densities of TNFα in the PVN [t(10) = 0.8, p> 0.4; Fig 1P].

Contrary to previous reports in males, the present results demonstrate that AngII infusion does not produce hypertension or result in changes in TNFR1 expression in female mice.

3.2. Influence of AngII infusion on the subcellular distribution of TNFR1 in PVN neurons of female mice

It has been previously reported that there is an increase in presumably functional plasma membrane TNFR1 in dendritic profiles of male hypertensive AngII-infused mice (Woods et al 2021). To examine if there were alterations in the plasma membrane availability of TNFR1 in PVN neurons of female mice, the subcellular location of TNFR1 in the PVN was examined following infusion with either Sal (N = 3) or AngII (N = 3) for 14-days. There was no effect of treatment on SBP [Fig. 2A; F(1, 4) = 1.3, p > 0.3]. In addition, there was neither a session [F(4, 16) = 1.9, p > 0.1], nor a treatment by session interaction [F(4, 16) = 1.1, p > 0.3, repeated measures ANOVA] with respect to SBP in female mice.

Fig. 2. Subcellular distribution of TNFR1 in PVN neurons is unaffected by AngII in female mice.

(A) Line graph illustrating SBP in Sal and AngII-infused mice. Both Sal and AngII-infused mice show similar SBP at all time points. (B-C) Electron micrographs showing immunogold-silver (IGS) labeling for TNFR1 (arrows) in dendritic profiles of PVN neurons from Sal and AngII-infused animals. (D-F) Dot plots showing the subcellular distribution of TNFR1 IGS labeling in Sal and AngII-treated mice. There are no differences in total (D), plasma membrane (E), or cytoplasmic (F) labeling in dendritic profiles of PVN neurons in Sal and AngII-infused animals. N=3 mice/group. Scale bars: 500 nm. Data are presented as mean± SEM.

Following SBP measurements, forebrain sections containing the PVN were processed for IGS labeling to assess total TNFR1 levels as well as the subcellular distribution of TNFR1 in dendritic processes of PVN neurons. Examples of TNFR1 labeling and TNFR1-labeled dendritic profiles of PVN neurons are shown in Figs. 2B, C. There was no difference in total TNFR1-IGS per unit area in Sal and Ang-infused mice [t(4) = 0.07, p > 0.9, unpaired t-test; Fig. 2D]. In addition, with respect to subcellular localization, there were no differences in plasma membrane TNFR1-IGS [t(4) = 0.05, p > 0.9, unpaired t-test; Fig. 2E] or cytoplasmic TNFR1-IGS [t(4) = 0.21, p > 0.8, unpaired t-test; Fig. 2F] in Sal and AngII-infused animals. There also were no treatment differences in morphological properties of TNFR1-IGS labeled dendritic profiles, including cross-sectional area [t(4) = 1.2, p> 0.2, unpaired t-test, not shown] and surface perimeter [t(4) = 1.35, p> 0.2, unpaired t-test, not shown].

Thus, in contrast to what has been reported in male mice, the present results demonstrate the AngII has no impact on the subcellular location of TNFR1 in female mice.

3.3. Impact of PVN TNFR1 silencing on basal blood pressure and blood pressure following AngII in female mice

To assess the effect of PVN TNFR1 silencing on basal blood pressure, SBP was measured in mice receiving bilateral AAV-TNFR1 shRNA (N = 6) or control vector (N= 6) pre-microinjection and 21-days following vector microinjection. There was no difference in SBP in either treatment group before and after microinjection [t(10) = 0.34, p > 0.7 unpaired t-test; Fig. 3A]. Mice were then implanted with osmotic minipumps containing AngII for 14-days. It was found that there was no difference in blood pressure following AngII in mice receiving the shRNA vector versus mice receiving the control vector [F(1, 10) = 0.03, p > 0.8; Fig. 3B]. In addition, there were no differences in SBP across recording sessions [F(5, 50) = 1.2, p > 0.3], nor was there a time by vector interaction [F(5, 50) = 0.7, p > 0.6, repeated measures ANOVA]. Expression of GFP in the PVN of a mouse receiving bilateral TNFR1 AAV-TNFR1 shRNA is shown in Fig. 3C. TNFR1was reduced in the PVN of AAV-TNFR1 shRNA treated mice (t(10) = 4.02, p< 0.003, unpaired t-test; Fig. 3D–F).

Fig. 3. Silencing TNFR1 in the PVN does not impact basal blood pressure or blood pressure following AngII in female mice.

(A) Histogram showing changes in SBP in PVN control and PVN TNFR1 knockdown mice at baseline and 21-days after vector microinjection. There are no changes in SBP in either group following vector administration. (B) Line graph showing SBP in both groups following 14-day infusion of AngII. Control and TNFR1 knockdown mice do not shown changes in SBP after AngII. (C) An example of GFP in the PVN of an AAV-TNFR1 shRNA microinjected mouse. (D, E) Light micrographs showing TNFR1 labeling in the PVN of control and TNFR1 shRNA injected mice. (F) Dot plot showing reduced TNFR1 in the PVN of mice receiving TNFR1 shRNA. N = 6 mice/group. Scale bars 200 μm (C), 100 μm (D, E): * p< 0.05, Data are presented as mean± SEM.

These results demonstrate that silencing TNFR1 in PVN neurons does not impact basal blood pressure or blood pressure following AngII infusion.

3.4. Effect of ERß blockade on blood pressure during slow-pressor AngII infusion in female mice

Estrogen may contribute to the insensitivity of female mice to hypertension (Marques-Lopes et al 2017, Milner et al 2021). To investigate the role of ERß in blood pressure following AngII infusion, female mice were co-administered AngII along with vehicle (N=8) or the ERß antagonist PHTPP (N=8). Other groups of mice were treated with Sal and vehicle (N=6) or PHTPP (N=6). With respect to SBP, there were main effects of AngII [F(1, 120) = 18.6 p< 0.0001], PHTPP [F(1, 120) = 49.3, p< 0.0001], session [F(4, 120) = 3.5, p< 0.02, repeated measures ANOVA] and an AngII by PHTPP by session interaction [F(4, 120) = 2.4, p< 0.05, repeated measures ANOVA]. There were increases in blood pressure in AngII-PHTPP treated mice compared to baseline at days 9, 11, and 13, (p< 0.002–0.0001, Tukey’s test; Fig. 4A) and compared to AngII-Veh mice at days 9, 11, and 13 (p< 0.001–0.0001, Tukey’s test, Fig. 4A).

Fig. 4. Blockade of ERß during AngII-infusion increases blood pressure in female mice.

(A) Line graph showing SBP in mice treated with Sal and either Veh (Sal/Veh) or PHTPP (Sal/PHTPP), in addition to AngII and either Veh (AngII/Veh) or PHTPP (AngII/PHTPP). Mice co-treated with AngII and PHTPP showed elevated SBP compared to mice given Sal/Veh AngII/Veh, or Sal/PHTPP. N = 6–8 mice/group. * p<0.05 AngII/PHTPP days 0 versus 10, 12, 14; ^a p< 0.001 AngII/PHTPP vs AngII/Veh; ^b p< 0.003 AngII/PHTPP vs Veh/Veh; ^c p< 0.05 AngII/PHTPP vs Veh/PHTPP. (B, C) Dot plots showing TNFR1 mRNA and immunolabeling in the PVN of Ang/Veh and AngII/PHTPP mice. There were no differences in TNFR1 mRNA or labeling in the PVN of mice infused with AngII and either Veh or PHTPP. (D) Dot plot showing TNFα labeling in the PVN of mice infused with AngII and treated with Veh or PHTPP. The density of TNFα was elevated in AngII-infused mice treated with PHTPP versus Veh. * p< 0.05, Ang/Veh versus Ang/PHTPP. (E, F) Examples of TNFα labeling in the PVN of mice treated with Veh or PHTPP. (G) Dot plot showing a higher density of p47 phosphorylated at Ser304 in the PVN of mice treated with AngII and PHTPP compared to AngII and vehicle. (H, I) Light micrographs illustrating Ser304 labeling in the PVN of AngII/Veh (H) and AngII/PHTPP (I) treated mice. (J) Line graph showing SBP in mice receiving control or AAV-TNFR1 shRNA vectors and given AngII and PHTPP. Mice receiving the control vector along with AngII and PHTPP showed an increase in SBP. In contrast, animals microinjected with the TNFR1-shRNA vector did not show changes in SBP at any time point. N= 6/group. * p< 0.05 SBP Control versus TNFR1-shRNA at days 10 and 14. Data are presented as mean± SEM. Scale bar: 50 μm (E, F), 200 μm (H, I). Data are presented as mean± SEM.

To investigate the impact of ERß blockade on TNFR1 expression, forebrain sections containing the PVN were processed for in situ hybridization. There was no difference in the number of TNFR1 particles per cell in the PVN of mice infused with AngII and vehicle or PHTPP [t(8) = 0.79, p> 0.4, unpaired t-test; Fig. 4B]. In addition, TNFR1 and TNFα were assessed by immunohistochemical labeling. There was no difference in the density of TNFR1 in AngII/vehicle and AngII/PHTPP-infused mice [t(8) = 1.3, p> 0.2; Fig. 4C]. In contrast, TNFα levels were higher in AngII-infused mice treated with PHTPP compared to vehicle [309.6±30.7 versus 175.2± 22.3; t(8) = 3.5, p> 0.008; Fig. 4D–F], and also higher with respect to mice infused with Sal and PHTPP [200.6± 30.1; p< 0.05, not shown] or vehicle (97.4± 43.2; p< 0.05, not shown). The effect of ERß blockade on the distribution of TNFR1 in dendritic profiles of PVN neurons was also assessed by IGS electron microscopy. There was no difference with respect to total [t(4) 1.1, p> 0.3, unpaired t-test; not shown], cytoplasmic [t(4) = 0.5, p>0.6, unpaired t-test; not shown], or plasma membrane [t(4) = 0.8, p> 0.4, Unpaired t-test; not shown] IGS TNFR1 labeling in the PVN of mice infused with AngII and treated with either PHTPP or vehicle. Additionally, cross-sectional area [t(4) = 1.2, p> 0.2, unpaired t-test; not shown] and surface perimeter [t(4) = 1.35, p> 0.2, unpaired t-test; not shown] of labeled dendrites were comparable in both groups. These results demonstrate that blockade of ERß is associated with an increased susceptibility of female mice to AngII hypertension that is also accompanied by elevated levels of TNFα, but not TNFR1, in the PVN.

Given evidence that TNFα is associated with ROS production during hypertension (Guggilam et al 2011), we examined if there were changes in the p47^phox subunit whose phosphorylation is required for NOX2 activation (Koshkin et al 1996). Therefore, levels of total and phosphorylated p47^phox at the key Ser 304 residue (Vermot et al 2021) were measured in the PVN of female mice treated with vehicle or PHTPP and AngII. It was found that there was an increase in the density of p47^phox phosphorylated at Ser 304 in the PVN of PHTPP and AngII-infused mice [t(8) = 3.1, p< 0.02, unpaired t-test; Fig 4G, H, I]. The increased density was accompanied by an increase in the number of cells expressing Ser 304 [t(8) = 2.7, p< 0.03, not shown]. There was no difference in levels of total p47^phox density [t(8) = 0.8, p> 0.4, unpaired t-test; not shown]. These results indicate that there is an increase in PVN NOX2 activity following hypertension induced by AngII infusion and ERß blockade in female mice.

We assessed whether the increase in blood pressure observed in AngII-infused female mice after ERß blockade was dependent on TNFα signaling in the PVN. Mice first received PVN injections of AAV-TNFR1 shRNA (N= 6) or a control vector (N= 6). Then both groups of mice were infused with AngII and treated with PHTPP. There was a significant effect of vector [F(1, 40) = 82.3, p< 0.0001], session [F(4, 40) = 15.6, p< 0.0001], and a vector by session interaction [F(4, 40) = 19.0, p< 0.0001, repeated measures ANOVA] with respect to SBP. It was found that SBP was elevated in control-vector injected mice, but not AAV-TNFR1 shRNA-injected mice. SBP was lower in mice microinjected with TNFR1 shRNA compared to the control vector at days 10 and 14 (p< 0.05, Tukey’s test; Fig 4J). TNFR1 was reduced in the PVN of AAV-TNFR1 shRNA mice (100±18.1% versus 47±6.4%; t(6.2) = 2.7, p< 0.4, Welch’s t-test, not shown).

3.5. Impact of TNFα on ROS production in isolated PVN neurons from female mice

We further investigated the impact of TNFα on ROS production by DHE microfluorography in isolated PVN cells from female mice (N = 3, n = 57 cells). There was a concentration-dependent effect of TNFα on ROS-signal in PVN cells [F(5, 336) = 92.4, p< 0.0001 repeated measures ANOVA; p < 0.01–0.0001, Tukey’s test; Figs 5A–F).

Fig. 5. TNFα-induced ROS-signal is heightened during ERß inhibition in isolated PVN projection neurons from female mice.

(A) Isolated PVN cells from female mice show dose-dependent increases in ROS signal after TNFα. (B-F) Fluorescent micrographs illustrating ROS signal following 1 (B), 3 (C), 10 (D), 30 (E), and 100 ng/ml (F) TNFα. * p< 0.01, **** p< 0.0001, relative to vehicle. (G) Dot plot illustrating ROS-dependent fluorescence in isolated neurons from the PVN of female mice. Cells were treated with vehicle and vehicle (V/V), TNFα and vehicle (T/V; T: 10 ng/ml), vehicle and ERß antagonist PHTPP (V/P), or TNFα and PHTPP (T/P). The intensity of ROS-dependent fluorescence is elevated following application of TNFα (*p< 0.05 T/V versus V/V) and further elevated after TNFα and PHTPP (*** p= 0.0003 T/P versus T/V). The increase in ROS in response to TNFα and PHTPP is inhibited in isolated neurons from TNFR1 KO mice (KO; **** p< 0.0001 T/P KO versus T/P) and inhibition of the catalytic gp91 domain by gp91-ds-tat (gp91ds; #### p< 0.0001 T/P gp91ds versus T/P), but not scrambled peptide (gp91sc, p> 0.9). The production of ROS by TNFα and PHTPP is further heightened in PVN neurons from mice infused with AngII for 14 days (# p< 0.04 T/P AngII versus T/P). DHE signal following TNFα alone in females was lower when compared to males (M; ## p< 0.005 M T/V versus T/V) but ROS production was comparable to males in the presence of both TNFα and PHTPP (p> 0.9). Scale bar: 10 μm. Data are presented as mean± SEM.

Sex differences in ROS production associated with excitatory signaling in the PVN are dependent on ERß, the major estrogen receptor in the PVN (Milner et al 2021), however, the role of ERß in production of ROS following TNFα in females is unknown. To investigate the role of ERß in the production of ROS by TNFα in females, DHE fluorescence was measured in isolated PVN cells following pretreatment with the ERß antagonist PHTPP. To assess the sensitivity of spinally-projecting PVN neurons that play a critical role in hypertension (Contoreggi et al 2021), ROS production was measured in PVN neurons retrogradely labeled from the spinal cord. The dependency of ROS production on TNFR1 was further assessed in TNFR1 KO mice. In addition, to determine if ROS-signal was mediated by NADPH oxidase, isolated cells from female mice were treated with TNFα and PHTPP following NOX2 inhibition by gp91-ds-tat (1 μM), a peptide inhibitor of the catalytic gp91^phox subunit, or scrambled peptide. The effect of AngII was also examined in isolated neurons from mice infused with the peptide. To assess sex differences in cytokine-induced free radical production, ROS signal was measured in TNFα-treated isolated cells from male mice.

There was a significant effect of treatment on ROS-signal [F(8, 49.6) = 24.4, p< 0.0001, Welch’s ANOVA; Fig 5G]. TNFα (10 ng/ml) modestly elevated ROS in neurons (+11±3.1%; N = 3, n = 15 cells/group) from females (p<0.05, Dunnett test; Fig 5G). The addition of PHTPP (1 μM) to TNFα (N = 3, n = 22 cells) significantly elevated DHE fluorescence compared to vehicle (111.1 ± 3.1% versus 139.6 ± 4.5%, p = 0.0003, Dunnett test; Fig 5G). In addition, PHTPP alone (N = 3, n = 22 cells) did not result in an increase in ROS compared to vehicle (p> 0.9, Dunnett test; Fig 5G). The increase in TNFα-mediated ROS signal following PHTPP was inhibited in isolated neurons from TNFR1 knockout mice (p< 0.0001, Dunnett test; N = 3, n = 10 cells, Fig 5G), verifying that the TNFα-induced elevated radical production following ERß blockade was dependent on TNFR1 signaling. The increase in ROS by TNFα and PHTPP was reduced following NOX2 inhibition by gp91-ds-tat (1 μM; N = 3, n = 31 cells), a peptide inhibitor of the catalytic gp91 subunit (p< 0.0001, Dunnett test, Fig 5G], but not by scrambled peptide (N = 3, n = 12 cells) treatment (p> 0.9, Dunnett test Fig 5G), demonstrating the dependency of ROS production on NOX2. Further, there was an increase in DHE fluorescence in response to TNFα and ERß blockade in isolated neurons from mice infused with AngII compared to TNFα and ERß blockade alone (219.6 ± 21.5% versus 139.6 ± 4.5%p< 0.04, Dunnett test, Fig 5G). In addition, compared to PVN cells from females, ROS production in PVN neurons from males after TNFα was higher (111 ± 3.1% versus 132 ± 3.6%, p< 0.005, Dunnett test; N=3, n = 23 cells; Fig 5G). However, DHE-signal in response to TNFα and PHTPP in isolated PVN neurons from females was comparable to that produced by TNFα alone in cells from male mice [139.6 ± 4.5% versus 132 ± 3.6%, p> 0.9, Dunnett test; Fig 5G]. These results demonstrate that inhibition of ERß results in heightened NOX2-dependent ROS production in spinally-projecting PVN neurons in female mice and does so to a degree that is comparable to males.

4. DISCUSSION

In male mice, TNFα signaling in the PVN has been implicated in hypertension (Woods et al 2021). Although there are notable sex differences in the susceptibility to both inflammatory processes and hypertension (Gillis & Sullivan 2016), the role of PVN TNFα in blood pressure control in females is unclear. In the present study, we report that AngII infusion had no effect on blood pressure, in addition to TNFR1 mRNA, TNFR1 immunolabeling, TNFR1 subcellular localization, or TNFα immunoreactivity in the PVN of female mice. Additionally, silencing PVN TNFR1 did not impact SBP basally or following AngII-infusion. Significantly, administration of the ERß antagonist PHTPP was associated with sensitivity to the slow-pressor response to AngII in females. The increase in blood pressure was associated with an increase in PVN TNFα and phosphorylation of the p47^phox NOX2 organizing protein, but not with changes in TNFR1 expression or subcellular localization. Further, the susceptibility of female mice to AngII hypertension following ERß antagonist administration was inhibited by silencing of TNFR1 in the PVN. At the cellular level, blockade of ERß was associated with an increase in TNFα-induced ROS production in isolated PVN-spinal cord projection neurons that was heightened in AngII-infused female mice.

In female rodents, autonomic dysfunction associated with neurohumoral (Dai et al 2015), cardiac (Yu et al 2019), metabolic (Fouda et al 2020), and gestational (Issotina Zibrila et al 2021) factors is accompanied by increased TNFα levels in the PVN. In female rats, TNFα levels were reported to be elevated in the PVN following DOCA-salt hypertension, although compared to males they showed a reduced hypertensive response (Dai et al 2015). It has also been reported that in a model of heart failure, females show an increase in TNFα, but at levels that are lower (Yu et al 2019), or comparable to males (Najjar et al 2018). Other evidence indicates that increased sympathetic activation in diabetic female animals was associated with increased TNFα compared to males only when ovariectomized and given estrogen (Fouda et al 2020). These divergent results may be due to various methodological variables across these studies including the type of autonomic challenge, as well as metabolic or endocrine factors.

In the present study, we found that AngII administered at a dose that produces a slow-pressor response in males did not impact levels of TNFα in the PVN. Additionally, unlike the typical hypertensive response of males, female mice did not show an increase in blood pressure in response to AngII infusion. This finding is consistent with other reports demonstrating the reduced sensitivity of female mice to the hypertensive actions of AngII when administered using the slow-pressor model (Girouard et al 2009, Marques-Lopes et al 2017, Xue et al 2005).

The signaling actions of TNFα are critically mediated by activation of TNFR1 (Santello & Volterra 2012). Functional TNFR1 in the PVN has been shown to play an important role in the hypertensive response to slow-pressor AngII infusion in males (Woods et al 2021), however, the relationship between TNFR1 signaling and AngII infusion in females is not clear. In contrast to males, in the present study it was found that female mice did not show an increase in TNFR1 transcription in the PVN when infused with AngII. Further, after AngII infusion females did not show an increase in TNFR1 located on the plasma membrane. These results indicate that TNFR1 is not affected at transcriptional or posttranslational levels in the PVN during AngII infusion in female mice.

Estrogen has been shown to play an important role in the decreased sensitivity of females to hypertension. In rodents, states of low estrogen associated with ovariectomy (Xue et al 2013), aging (Marques-Lopes et al 2014), or the accelerated ovarian failure model of menopause (Marques-Lopes et al 2017, Ovalles et al 2019) have been shown to increase the susceptibility to AngII hypertension. Compared to ER alpha (ERα), ERß has preferential effects on suppression of TNFα signaling in the context of various models including airway (Ambhore et al 2019, Bhallamudi et al 2020) and aortic smooth muscle cells (Xing et al 2007), vascular endothelial cells (Fortini et al 2017), and neuroglial cells (Guo et al 2020). Further, it has also been shown that compared to ERα, ERß is more prominently expressed in the PVN (Contoreggi et al 2021, Oyola et al 2017), and our prior studies have also shown that ERß is critically involved in AngII hypertension in females (Milner et al 2021). We examined if inhibiting ERß, by infusion of the antagonist PHTPP, would affect hypothalamic TNFα signaling and blood pressure in females during AngII administration. It was shown that ERß blockade during AngII infusion was associated with an increase in blood pressure in female mice. However, it was found that there were no changes in TNFR1 mRNA or plasma membrane receptor localization in females rendered hypertensive following AngII and ERß blockade. However, there was an increase in TNFα in the PVN of PHTPP and AngII-treated mice. To investigate if TNFα signaling contributed to this hypertensive response, mice were co-administered AngII and PHTPP after PVN TNFR1 silencing. Following TNFR1 knockdown, it was found that mice did not show an increase in blood pressure when AngII was infused during ERß inhibition, demonstrating an important role for PVN TNFα signaling via TNFR1 in the elevation in blood pressure achieved by estrogen receptor blockade during AngII exposure.

Given that vascular and smooth muscle cells express ERß, and constitutive ERß knockout mice express hypertension (Zhu et al 2002), the present results using systemic PHTPP preclude definitive identification of a prohypertensive action of hypothalamic estrogen receptor blockade. However, there are compelling reasons supporting the involvement of the PVN as an important site for ERß inhibition in promoting AngII hypertension in female mice. Indeed, the PVN is a major site of brain ERß expression (Contoreggi et al 2021, Milner et al 2010). Moreover, this population of receptors has been previously implicated in AngII hypertension in females (Marques-Lopes et al 2017, Marques-Lopes et al 2014). Additionally, hypothalamic ERß gene deletion has been shown to significantly impact blood pressure and hypertension in 2–3 month old female mice (Milner et al 2021), actions that parallel the increase in blood pressure following AngII infusion and ERß inhibition seen here. Significantly, this contrasts with the hypertension seen in constitutive ERß knockout mice, which was only reported in mice of at least 6 months of age (Zhu et al 2002), older than the mice used here.

It has been reported that ERß contributes to sex differences in NOX2-mediated ROS production in the PVN (Milner et al 2021), however, the role of ERß in TNFα-induced ROS production in the PVN is unclear. The present study examined the role of ERß in NOX2 signaling in the PVN of female mice. The activity of the heteromeric NOX2 enzyme requires recruitment, organization, and membrane assembly of its constitutive subunits, a process that requires phosphorylation of the key organizing p47^phox subunit (Koshkin et al 1996, Vermot et al 2021). In the present study it was found that ERß blockade resulting in an increase in blood pressure during AngII infusion was also accompanied by an increase in levels of p47^phox phosphorylated at Ser304, but not total p47^phox. Further, it was shown that blockade of ERß in isolated sympathoexcitatory PVN-spinal cord projection neurons was associated with a heightening of TNFα-induced ROS production that was blocked by inhibiting the catalytic gp91^phox subunit or TNFR1 knockout. In addition, the increase in ROS production was further heightened in neurons from female mice receiving 14-day AngII infusion. These results point to ERß activation as a modulator of hypothalamic TNFα-mediated ROS signaling and the sensitivity to blood pressure changes during AngII administration in female mice.

5. CONCLUSION

In summary, contrary to previous reports in male mice, intact females do not show an increase in TNFα signaling in the PVN in response to AngII. However, following ERß blockade females show elevated PVN TNFα and hypertension in response to AngII that is reversed by PVN TNFR1 silencing. At the cellular level, the blockade of estrogen signaling at ERß was coupled to activation of TNFα-induced NOX2-mediated ROS production in PVN-spinal cord projection neurons. Although TNFα-induced ROS production in the absence of ERß blockade was lower in females compared to males, the magnitude of the increase in ROS following ERß blockade in females was comparable to what was seen in isolated neurons from males following TNFα stimulation alone. These results suggest that estrogen, acting at ERß in the PVN of females, promotes the suppression of signaling pathways, involving TNFα and NOX2-mediated reactive oxygen production, that normally stimulate AngII-mediated hypertension in males. When the ERß-dependent constraint on this pathway is removed by receptor blockade, the TNFα-NOX2 pathway induced by AngII then becomes active in the hypothalamus of females, which may contribute to their sensitivity to hypertension. Thus, reduced TNFα-induced ROS production in sympathoexcitatory neurons may provide a mechanism by which estrogen influences the susceptibility of females to hypertension in response to AngII.

ABBREVIATIONS

AAV

adeno-associated virus

AngII

angiotensin II

BSA

bovine serum albumin

CSF

cerebrospinal fluid

DAB

diaminobenzidine

DHE

dihydroethidium

GFP

green fluorescent protein

IGS

immunogold-silver

NOX

NADPH Oxidase

PFA

paraformaldehyde

PHTPP

4-[2-Phenyl-5,7-bis(trifluoromethyl)pyrazolo[1,5-a]pyrimidin-3-yl]phenol

PVN

paraventricular nucleus of the hypothalamus

ROS

reactive oxygen species

SBP

systolic blood pressure

shRNA

short hair-pin RNA

TNFα

tumor necrosis factor alpha

TNFR1

tumor necrosis factor alpha receptor 1