Angiotensin II type 1a Receptors in the Subfornical Organ Modulate Neuroinflammation in the Hypothalamic Paraventricular Nucleus in Heart Failure Rats

Abstract

Inflammation in the hypothalamic paraventricular nucleus (PVN) contributes to neurohumoral excitation and its adverse consequences in systolic heart failure (HF). The stimuli that trigger inflammation in the PVN in HF are not well understood. Angiotensin II (AngII) has pro-inflammatory effects, and circulating levels of AngII increase in HF. The subfornical organ (SFO), a circumventricular structure that lacks an effective blood-brain barrier and senses circulating AngII, contains PVN-projecting neurons. We hypothesized that activation of AngII type 1a receptors (AT_1aR) in the SFO induces neuroinflammation downstream in the PVN. Male rats received SFO microinjections of an adeno-associated virus carrying shRNA for AT_1aR, a scrambled shRNA, or vehicle. One week later, some rats were euthanized to confirm the transfection potential and knockdown efficiency of the shRNA. Others underwent coronary ligation to induce HF or a sham coronary ligation (Sham). Four weeks later, HF rats that received the scrambled shRNA had increased mRNA in SFO and PVN for AT_1aR, inflammatory mediators and indicators of neuronal and glial activation, increased plasma AngII, tumor necrosis factor-α, norepinephrine and arginine vasopressin, and impaired cardiac function, compared with Sham rats that received scrambled shRNA. The central abnormalities were ameliorated in HF rats that received AT_1aR shRNA, along with plasma norepineprine and vasopressin. Sham rats that received AT_1aR shRNA had reduced SFO AT_1aR mRNA but no other changes compared with Sham rats that received scrambled shRNA. The results suggest that activation of AT_1aR in the SFO upregulates the neuroinflammation in the PVN that contributes to neurohumoral excitation in HF.

INTRODUCTION

Increased renin-angiotensin system (RAS) activity and inflammation in cardiovascular-related regions of the central nervous system contribute to the overactivity of neurohumoral systems that promote volume retention, cardiac remodeling and serious cardiac arrhythmias in systolic heart failure (HF). A primary central nervous system site involved in this process is the hypothalamic paraventricular nucleus (PVN), which contains both presympathetic and neuroendocrine neurons (Ferguson et al., 2008). Interventions that reduce RAS activity and inflammation in the PVN are uniformly successful in reducing sympathetic excitation and improving indices of volume regulation and cardiac hemodynamics in rats with HF (Zhang et al., 1999; Francis et al., 2001; Francis et al., 2004; Guggilam et al., 2008; Kang et al., 2008a; Kang et al., 2008b; Kang et al., 2010; Yu et al., 2012; Huang et al., 2014). However, the mechanisms upregulating the activity of these two excitatory neurochemical systems in the PVN in HF are still poorly understood.

The present study sought to determine whether increased RAS activity in the subfornical organ (SFO) - a forebrain circumventricular organ that lacks an effective blood-brain barrier, senses circulating humoral factors in HF (McKinley et al.; 2003, Ferguson, 2014) and projects directly to the PVN (Li and Ferguson, 1993; Kawano and Masuko, 2010) - contributes to the inflammatory response in the PVN in HF. In HF, angiotensin II (AngII) type 1 receptors (AT₁R) are upregulated in the SFO (Tan et al., 2004; Wei et al., 2008b), which is exposed to increased plasma levels of angiotensin II (AngII) in that setting (Huang and Leenen, 2009a; Wang et al., 2014). Previous studies have shown that a slow-pressor infusion of AngII upregulates the expression of tumor necrosis factor (TNF)-α, interleukin (IL)-1β, IL-6 and cyclooxygenase-2 (COX-2) in the PVN of normal rats (Yu et al., 2013b), and that chronic intracerebroventricular infusion of the AT₁R blocker losartan significantly reduces the expression of the TNF-α, IL-1β and IL-6 in the PVN of HF rats (Kang et al., 2008a), though neither of those studies considered what role AT₁R in the SFO might play in upregulating the expression of these neuroinflammatory mediators in the PVN.

To address this question, we used an adeno-associated viral (AAV) vector carrying an shRNA specific for AT_1aR, the AT₁R subtype that mediates the effects of AngII in the SFO and other cardiovascular autonomic regions of the brain (Lenkei et al., 1997). We microinjected the AT_1aR shRNA into the SFO prior to the induction of HF and measured the effects of AT_1aR knockdown in the SFO on mRNA for TNF-α, IL-1β, COX-2 and markers of neuronal and glial activation in the SFO and PVN, on plasma AngII, TNF-α, norepinephrine (NE), arginine vasopressin (AVP), and on indices of cardiac remodeling and hemodynamics. The results reveal that activation of AT_1aR in the SFO of HF rats promotes the expression of the inflammatory mediators in the PVN that contribute to neurohumoral excitation in HF.

EXPERIMENTAL PROCEDURES

Animals

Adult male Sprague-Dawley rats weighing 250–300 g were purchased from Envigo/Harlan (Indianapolis, IN). Animals were housed in temperature- (23 ± 2°C) and light-controlled animal care facility, and standard rat chow and water were given ad libitum. Experiments were performed in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. All procedures were approved by the Institutional Animal Care and Use Committee of the University of Iowa. All efforts were made to minimize the number of animals used and their suffering.

Experimental Protocols

Sixty-seven rats underwent SFO microinjection of an AAV vector carrying green fluorescent protein (GFP) and either shRNA against AT_1aR (AAV-AT_1aR shRNA, n = 27) or a scrambled control shRNA (AAV-CON shRNA, n = 31), or vehicle (Veh, artificial cerebrospinal fluid, n = 9).

One week later, some rats that had received SFO microinjection of Veh (n=9), AAV-AT_1aR shRNA (n = 9) or AAV-CON shRNA (n = 9) were euthanized to verify the transfection potential (by immunofluorescent imaging, n=3/group) and the knockdown efficiency (by RT-PCR, n=6/group) of the AAV-AT_1aR shRNA.

The remaining rats that had received AAV-AT_1aR shRNA or AAV-CON shRNA underwent either coronary artery ligation (CL) to induce HF or a sham operation (Sham), resulting in 4 experimental groups: HF+AAV-AT_1aR shRNA (n = 11), HF+AAV-CON shRNA (n = 15), Sham+AAV-AT_1aR shRNA (n = 7), Sham+AAV-CON shRNA (n = 7).

Left ventricular function was evaluated by echocardiogram within 24 hours of CL or sham operation. Four weeks later, a second echocardiogram was performed to determine treatment effects. At the end of the 5-week protocol, rats from each treatment group were anesthetized for invasive assessment of cardiac hemodynamics and then euthanized while still under anesthesia to collect brain and blood for real-time PCR and ELISA. The heart and lungs were also harvested and weighed to assess peripheral indicators of HF.

Six rats died within 24 hours of coronary ligation (HF+AAV-AT_1aR shRNA: n = 2; HF+AAV-CON shRNA: n = 4). Three more rats died between 24 hours and the end of the protocol (HF+AAV-AT_1aR shRNA: n = 1; HF+AAV-CON shRNA: n = 2). Five rats with a small myocardial infarction on initial echocardiogram (ischemic zone, as defined below, ≤ 30%) were excluded from further study (HF+AAV-AT_1aR shRNA: n = 2; HF+AAV-CON shRNA: n = 3). No Sham rats died prior to completing the protocol. The final animal numbers that were used for data analysis were as follows: n = 7 rats for Sham+AAV-CON shRNA and Sham+AAV-AT1aR shRNA; n = 6 rats for HF+AAV-CON shRNA and HF+AAV-AT1aR shRNA.

SFO Microinjections

SFO microinjection was performed as previously described (Moreau et al., 2012; Wei et al., 2015). Briefly, rats were anesthetized (ketamine 100 mg/kg + xylazine 10 mg/kg IP) and positioned in a stereotaxic apparatus (Kopf Instruments; Tujunga, CA, USA). A longitudinal skin incision was made to expose the skull, and a small hole was drilled at stereotaxic coordinates 0.8 mm caudal and 1.7 mm lateral to bregma. A 30-gauge guide cannula was inserted and advanced at a 25° angle to the vertical meridian to a position 5 mm ventral to the cortical surface, approaching the SFO at an angle to avoid the sagittal sinus. A 35-gauge injection cannula connected via calibrated polyethylene tubing to a 1-μl Hamilton microsyringe was inserted into the guide cannula and extended 0.5 mm beyond the tip of the guide cannula. AT_1aR shRNA (0.3 μl of 1.1 × 10¹² genomic particles/ml) or CON shRNA AAV vector, both carrying GFP (Genedetect, Sarasota, FL, USA), was then injected into the SFO over 15–30 s. The injection cannula was left at the injection site for additional 5 min before removing from the brain. The scalp incision was closed with sutures. The dose of AAV-AT_1aR shRNA used to knock down AT_1aR gene expression in the SFO in rats was derived from a previous study (Walch et al., 2014). Buprenorphine (0.03 mg/kg, SC) was administered immediately postoperatively, and then every 12 hours for 48 hours.

Induction of HF

CL to induce HF, or sham CL, was performed as described before (Yu et al., 2012). Briefly, rats were anesthetized with ketamine and xylazine (100 mg/kg and10 mg/kg, respectively, IP), intubated and placed on a ventilator. Under sterile conditions, the chest wall and pericardium were opened and the left coronary artery was tied off proximally to induce a large myocardial infarction. Sham rats underwent the same operative procedures without occlusion of the left coronary artery. The chest was closed and the lungs were re-inflated. Buprenorphine (0.03 mg/kg, SC) was administered immediately postoperatively and then every 12 hours for 48 hours.

Assessment of Severity of HF

Echocardiography

Rats that had undergone CL or sham CL were sedated with ketamine (60 mg/kg IP) and echocardiography was carried out to assess the ischemic zone as a percent of the left ventricular (LV) circumference (% IZ), LV ejection fraction (LVEF), LV end-diastolic volume (LVEDV) and LV volume/mass ratio, as described before (Yu et al., 2012).

Cardiac Hemodynamics

Rats that had undergone CL or sham CL were anesthetized with urethane (1.5 g/kg, IP). Systolic blood pressure (SBP), diastolic blood pressure (DBP), left ventricle end-diastolic pressure (LVEDP), LV peak systolic pressure (LVPSP), and maximum rate of rise of LV pressure (LV dP/dt_max) were assessed by a Millar catheter that was inserted via the right carotid artery into the LV, as described previously (Yu et al., 2012). HR (beats/min) was calculated as the frequency of arterial pressure pulses.

Anatomical Measures

Wet lung weight, LV weight and right ventricular (RV) weight, with respect to body weight (BW), were measured as indices of pulmonary congestion and cardiac remodeling, two indicators of the severity of HF.

Immunofluorescent Studies

Rats for immunohistochemical studies were transcardially perfused with 4% paraformaldehyde while deeply anesthetized with urethane (1.5 g/kg, IP). Brains were removed and post-fixed in 4% paraformaldehyde for 24 h at 4°C followed by cryoprotection in 30% sucrose for 48 h at 4°C. Brains were frozen in OCT compound on dry ice and coronal sections (20 μm) of target tissues were cut using a cryostat and stored at −80 °C. GFP fluorescence was visualized using a confocal laser-scanning microscope (Zeiss LSM 710, Carl Zeiss, Inc, Oberkochen, Germany).

Molecular Studies

Rats for molecular studies were decapitated while deeply anesthetized with urethane (1.5 g/kg, IP). Trunk blood was collected for measurement of plasma AngII, TNF-α, NE and AVP levels using commercial ELISA kits (Cloud-Clone Corp, Katy, TX, USA; R &D Systems, Inc, Minneapolis, MN, USA; Labor Diagnostika Nord GmbH & Co KG, Nordhorn, Germany; Enzo Life Sciences, Inc, Plymouth Meeting, PA, USA, respectively), according to the manufacturer’s instruction.

Brains were removed and the SFO and PVN regions, including small amounts of surrounding tissues, were punched using a 15-gauge needle stub (inner diameter: 1.5 mm). The total RNA was isolated with TRI Reagent (Molecular Research Center, Inc, Cincinnati, OH, USA). mRNA expression of AT_1aR, TNF-α, TNF-α receptor 1 (TNFR1), IL-1β and COX-2, the immediate early gene product c-Fos, glial fibrillary acidic protein (GFAP) and cluster of differentiation 68 (CD68) were analyzed with SYRB Green real-time PCR following reverse transcription of total RNA, as previously described (Yu et al., 2013b). The sequences for the primers are presented in Table 1. Real-time PCR was performed using the ABI prism 7000 Sequence Detection System (Applied Biosystems, Carlsbad, CA, USA). The values were corrected by GAPDH and were calculated using the formula x=2^−(ΔΔCt) (Livak and Schmittgen, 2001). All mRNA data were expressed as a fold change relative to Veh or Sham+AAV-CON shRNA.

Table 1.

Sequences for primers

Gene	Primers	Sequences
AT1aR	Forward primer:	5′-ACTCACAGCAACCCTCCAAG-3′
	Reverse primer:	5′-ATCACCACCAAGCTGTTTCC-3′
TNF-α	Forward primer:	5′-CCTTATCTACTCCCAGGTTCTC-3′
	Reverse primer:	5′-TTTCTCCTGGTATGAATGGC-3′
TNFR1	Forward primer:	5′-GGTTCCTTTGTGGCACTTGGT-3′
	Reverse primer:	5′-CTCTTGGTGACCGGGAGAAG-3′
IL-1β	Forward primer:	5′-CGACAGAATCTAGTTGTCC-3′
	Reverse primer:	5′-TCATAAACACTCTCATCCACAC-3′
COX-2	Forward primer:	5′-AAGGGAGTCTGGAACATTGTGAAC-3′
	Reverse primer:	5′-CAAATGTGATCTGGACGTCAACA-3′
GAPDH	Forward primer:	5′-AAGGTCATCCCAGAGCTGAA-3′
	Reverse primer:	5′-ATGTAGGCCATGAGGTCCAC-3′

AT_1aR: angiotensin II type 1a receptor; TNF-α: tumor necrosis factor-α; TNFR1: TNF-α receptor 1; IL-1β:interleukin-1β; COX-2: cyclooxygenase-2; GAPDH: glyceraldehyde-3-phosphate dehydrogenase.

Statistical Analysis

All data are expressed as mean±SEM. The significance of differences in mean values was analyzed using one-way or two-way ANOVA followed by post hoc multiple comparison tests. P<0.05 was considered statistically significant.

RESULTS

Validation of AT_1aR Knockdown in SFO

One week after SFO microinjection, GFP immunoreactivity was present throughout the SFO of rats treated with either AAV-AT_1aR shRNA or AAV-CON shRNA (Figure 1). No GFP was found in PVN. SFO microinjection of AAV-AT_1aR shRNA significantly reduced AT_1aR mRNA in the SFO compared with SFO microinjection of VEH or AAV-CON shRNA, but had no effect on AT_1aR mRNA in the PVN. AAV-CON shRNA did not alter AT_1aR mRNA expression in either SFO or PVN. These observations confirmed the accuracy of the microinjection technique and the effectiveness of gene-knockdown in the SFO with AT_1aR shRNA.

Figure 1. Accuracy and selectivity of angiotensin II type 1a receptor (AT_1aR) knockdown in subfornical organ (SFO).

A: Representative digital images showing green fluorescent protein (GFP) expression in the SFO and the hypothalamic paraventricular nucleus (PVN) in rats (n=3/group) one week after SFO microinjection of vehicle (Veh, left column), an adeno-associated viral (AAV) vector carrying GFP and AT_1aR shRNA (middle column) or scrambled (CON) shRNA (right column). In rats receiving AAV-AT_1aR shRNA or AAV-CON shRNA, GFP expression was mainly observed in the SFO, with little expression in the surrounding tissue and none in the PVN. Scale bar: 100 μm. B: mRNA expression of AT_1aR in the SFO and the PVN in rats one week after SFO microinjection of Veh, AAV-AT_1aR shRNA or AAV-CON shRNA. Data were analyzed with one-way ANOVA followed by multiple-comparison tests. Values are mean ± SEM (n = 6/group) and expressed as a fold change compared to Veh. Individual data points for each animal are shown. *p<0.05 vs Veh or AAV-CON shRNA.

Effect of SFO AT_1aR Knockdown on Expression of Inflammatory Mediators

Four weeks after CL, HF+AAV-CON shRNA rats exhibited markedly increased mRNA expression of AT_1aR (Figure 2), and of the inflammatory mediators TNF-α, TNFR1, IL-1β and COX-2 (Figure 3) in both SFO and PVN, compared with Sham+AAV-CON shRNA rats. Compared with HF+AAV-CON shRNA rats, HF+AAV-AT_1aR shRNA rats had significantly reduced mRNA expression of AT_1aR not only in the SFO but also in the PVN. In addition, the mRNA expression of TNF-α, TNFR1, IL-1β and COX-2 was reduced in both SFO and PVN in the HF+AAV-AT_1aR shRNA rats. Notably, Sham+AAV-AT_1aR shRNA rats also had significantly reduced mRNA expression of AT_1aR in the SFO, but no differences in mRNA expression of TNF-α, TNFR1, IL-1β and COX-2 were observed in SFO or PVN between Sham+AAV-AT_1aR shRNA and Sham+AAV-CON shRNA rats.

Figure 2. Effect of angiotensin II type 1a receptor (AT_1aR) knockdown in the subfornical organ (SFO) on AT1aR gene expression in the SFO and the hypothalamic paraventricular nucleus (PVN).

AT_1aR mRNA expression in SFO and the PVN 4 weeks after coronary artery ligation to induce heart failure (HF) or sham operation (Sham) and 5 weeks after SFO microinjection of AAV-AT_1aR shRNA or AAV-CON shRNA. Data were analyzed with two-way ANOVA followed by multiple-comparison tests. Values are mean ± SEM (n = 6–7 for each group) and expressed as a fold change compared to Sham+AAV-CON shRNA. Individual data points for each animal are shown. *p<0.05 vs Sham+AAV-CON shRNA, †p<0.05 HF+AAV-AT_1aR shRNA vs HF+AAV-CON shRNA.

Figure 3. Effect of angiotensin II type 1a receptor (AT_1aR) knockdown in the subfornical organ (SFO) on gene expression of inflammatory mediators in the SFO and the hypothalamic paraventricular nucleus (PVN).

mRNA expression for tumor necrosis factor – alpha (TNF-α, A), TNF-α receptor 1 (TNFR1, B), interleukin-1 beta (IL-1β, C) and cyclooxygenase-2 (COX-2, D) in the SFO and the PVN 4 weeks after coronary artery ligation to induce heart failure (HF) or sham operation (Sham) and 5 weeks after SFO microinjection of AAV-AT_1aR shRNA or AAV-CON shRNA. Data were analyzed with two-way ANOVA followed by multiple-comparison tests. Values are mean ± SEM (n = 6–7 for each group) and expressed as a fold change compared to Sham+AAV-CON shRNA. *p<0.05 vs Sham+AAV-CON shRNA, †p<0.05 HF+AAV-AT_1aR shRNA vs HF+AAV-CON shRNA.

Effect of SFO AT_1aR Knockdown on Glial Markers

Compared with Sham+AAV-CON shRNA rats, HF+AAV-CON shRNA rats had higher mRNA expression of the microglia-specific marker CD68 and the astrocyte-specific marker GFAP in both SFO and PVN (Figure 4). Compared with HF+AAV-CON shRNA rats, HF+AAV-AT_1aR shRNA rats had significantly lower mRNA expression of CD68 and GFAP in both SFO and PVN. mRNA expression of CD68 and GFAP in SFO and PVN of Sham+AAV-AT_1aR shRNA rats was not different from that in Sham+AAV-CON shRNA rats.

Figure 4. Effect of angiotensin II type 1a receptor (AT_1aR) knockdown in the subfornical organ (SFO) on indicators of glial activity in the SFO and the hypothalamic paraventricular nucleus (PVN).

mRNA expression of cluster of differentiation 68 (CD68, a microglia marker) and glial fibrillary acidic protein (GFAP, an astrocyte marker) in the SFO and the PVN 4 weeks after coronary artery ligation to induce heart failure (HF) or sham operation (Sham) and 5 weeks after SFO microinjection of AAV-AT_1aR shRNA or AAV-CON shRNA. Data were analyzed with two-way ANOVA followed by multiple-comparison tests. Values are mean ± SEM (n = 6–7 for each group) and expressed as a fold change compared to Sham+AAV-CON shRNA. *p<0.05 vs Sham+AAV-CON shRNA, †p<0.05 HF+AAV-AT_1aR shRNA vs HF+AAV-CON shRNA.

Effect of SFO AT_1aR Knockdown on Neuronal Excitation

Compared with Sham+AAV-CON shRNA rats, HF+AAV-CON shRNA rats exhibited higher mRNA expression of c-Fos (an indicator of neuronal excitation) in SFO and PVN (Figure 5). The c-Fos expression in SFO and PVN was significantly decreased in HF+AAV-AT_1aR shRNA rats, compared with HF+AAV-CON shRNA rats. There was no difference in mRNA expression of c-Fos in SFO or PVN between Sham groups.

Figure 5. Effect of angiotensin II type 1a receptor (AT_1aR) knockdown in the subfornical organ (SFO) on neuronal excitation in the SFO and the hypothalamic paraventricular nucleus (PVN).

mRNA expression of c-Fos, an indicator of neuronal excitation, in the SFO and the PVN 4 weeks after coronary artery ligation to induce heart failure (HF) or sham operation (Sham) and 5 weeks after SFO microinjection of AAV-AT_1aR shRNA or AAV-CON shRNA. Data were analyzed with two-way ANOVA followed by multiple-comparison tests. Values are mean ± SEM (n = 6–7 for each group). Values are expressed as a fold change compared to Sham+AAV-CON shRNA. *p<0.05 vs Sham+AAV-CON shRNA, †p<0.05 HF+AAV-AT_1aR shRNA vs HF+AAV-CON shRNA.

Effect of SFO AT_1aR Knockdown on Humoral Responses to HF

Compared with Sham+AAV-CON shRNA rats, HF+AAV-CON shRNA rats exhibited markedly higher plasma levels of TNF-α, AngII, AVP and NE (Figure 6). HF+AAV-AT_1aR shRNA rats had significantly decreased plasma NE and AVP levels compared with HF+AAV-CON shRNA rats, but these levels remained higher than those in the Sham groups in which there were no differences in plasma NE and AVP levels. In contrast, plasma AngII and TNF-α levels remained elevated in the HF+AAV-AT_1aR shRNA rats, similar to levels in the HF+AAV-CON shRNA rats.

Figure 6. Effect of angiotensin II type 1a receptor (AT_1aR) knockdown in the subfornical organ (SFO) on humoral factors.

Plasma levels of angiotensin II (AngII), tumor necrosis factor-α (TNF-α) norepinephrine (NE) and arginine vasopressin (AVP) 4 weeks after coronary artery ligation to induce heart failure (HF) or sham operation (Sham) and 5 weeks after SFO microinjection of AAV-AT_1aR shRNA or AAV-CON shRNA. Data were analyzed with two-way ANOVA followed by multiple-comparison tests. Values are mean ± SEM (n = 6–7 for each group). *p<0.05 vs Sham+AAV-CON shRNA, †p<0.05 HF+AAV-AT_1aR shRNA vs HF+AAV-CON shRNA.

Effects of SFO AT_1aR Knockdown on Echocardiographic Indicators of HF

Echocardiography performed within 24 hours of CL showed that the infarct sizes and degrees of cardiac dysfunction, as indicated by LVEF and LVEDV, were comparable between HF+AAV-AT_1aR shRNA and HF+AAV-CON shRNA rats (Table 2). Four weeks after CL, %IZ remained similar in the two HF groups. LVEDV had increased further in HF+AAV-CON shRNA rats, but not in the HF+AAV-AT_1aR shRNA rats though a non-significant trend in that direction was observed. Echocardiographic indices of cardiac function were comparable between the two Sham groups.

Table 2.

Echocardiographic Measurements

Variables at Baseline	Sham+AAV-Con shRNA	Sham+AAV-AT1aR shRNA	HF+AAV-Con shRNA	HF+AAV-AT1aR shRNA
LVEDV (ml)	0.49 ± 0.04	0.48 ± 0.04	0.78 ± 0.06*	0.77 ± 0.04*
LV Vol/Mass (μl/mg)	0.53 ± 0.05	0.55 ± 0.05	1.17 ± 0.12*	1.20 ± 0.10*
LVEF	0.72 ± 0.03	0.73 ± 0.04	0.35 ± 0.06*	0.36 ± 0.03*
%IZ	----	----	41 ± 3*	42 ± 4*
Variables at week 4
LVEDV (ml)	0.52 ± 0.05	0.51 ± 0.02	1.02 ± 0.06*§	0.99 ± 0.08*
LV Vol/Mass (μl/mg)	0.62 ± 0.06	0.64 ± 0.04	1.41 ± 0.11*	1.25 ± 0.12*
LVEF	0.73 ± 0.05	0.76 ± 0.03	0.31 ± 0.04*	0.34 ± 0.03*
%IZ	---	---	43 ± 2*	40 ± 2*

LVEDV: left ventricular end-diastolic volume; LVEF: left ventricular ejection fraction; %IZ: ischemic zone as a percent of left ventricular circumference. Data were analyzed with two-way ANOVA followed by multiple-comparison tests. Values are expressed as mean ± SEM (n = 6–7 for each group)..

^*P<0.05 versus Sham+AAV-Con shRNA at same week.

^§P<0.05 HF+AAV-Con shRNA at 4 weeks versus 24 hours.

Effects of SFO AT_1aR Knockdown on Hemodynamic and Anatomical Indicators of HF

Hemodynamic studies (Table 3) revealed that SBP, LVPSP, and LV dP/dt_max were lower and LVEDP was higher in HF+AAV-CON shRNA rats compared with Sham+AAV-CON shRNA rats. HF+AAV-AT_1aR shRNA rats had significantly lower LVEDP and higher LV dP/dt_max than HF+AAV-CON shRNA rats, but SBP remained low. Of note, SFO AT_1aR knockdown in Sham rats did not alter any of the hemodynamic parameters. HR and DBP were comparable across the experimental groups.

Table 3.

Anatomical and hemodynamic measurements

Variables at 4 weeks	Sham+AAV-Con shRNA	Sham+AAV-AT1aR shRNA	HF+AAV-Con shRNA	HF+AAV-AT1aR shRNA
BW (g)	408 ± 5	406 ± 13	409 ± 5	405 ± 12
LV/BW (mg/g)	2.19 ± 0.06	2.31 ± 0.09	2.15 ± 0.09	2.16 ± 0.07
RV/BW (mg/g)	0.52 ± 0.01	0.56 ± 0.01	1.05 ± 0.06*	0.81 ± 0.04* †
Lung/BW (mg/g)	3.58 ± 0.11	3.52 ± 0.19	9.78 ± 0.57*	6.62 ± 0.52*†
HR (beats/min)	359 ± 7	353 ± 8	358 ± 6	352 ± 7
SBP (mmHg)	126 ± 2	124 ± 3	109 ± 3*	111 ± 2*
DBP (mmHg)	90 ± 2	98 ± 3	85 ± 2	86 ± 2
LVPSP (mmHg)	122 ± 2	120 ± 2	106 ± 3*	109 ± 2*
LVEDP (mmHg)	3.4 ± 0.2	3.2 ± 0.3	15.1 ± 0.7*	9.9 ± 0.8*†
LV dP/dtmax (mmHg/s)	9399 ± 181	9518 ± 213	6405 ± 165*	7541 ± 201*†

BW: body weight; LV: left ventricular; RV: right ventricular; HR: heart rate; SBP: systolic blood pressure; DBP: diastolic blood pressure; LVPSP: LV peak systolic pressure; LVEDP: LV end-diastolic pressure; dP/dt_max: maximum rate of rise of LV pressure. Data were analyzed with two-way ANOVA followed by multiple-comparison tests. Values are expressed as mean ± SEM (n = 6–7 for each group).

^*P<0.05 versus Sham+AAV-Con shRNA,

^†P<0.05 HF+AAV-AT₁aR shRNA versus HF+AAV-Con shRNA.

Four weeks after CL, RV/body weight (BW) and wet lung/BW ratios were substantially higher in HF+AAV-CON shRNA rats than Sham+AAV-CON shRNA rats (Table 3). The wet lung/BW and RV/BW ratios were significantly lower in the HF+AAV-AT_1aR shRNA group than the HF+AAV-CON shRNA group. SFO AT_1aR knockdown had no effect on RV/BW and wet lung/BW ratios in Sham rats. BW and LV/BW were comparable across the experimental groups.

DISCUSSION

Previous studies have established that central neuroinflammation (Kang et al., 2008b; Yu et al., 2012; Yu et al., 2017), and in particular neuroinflammation in the PVN (Kang et al., 2009, Kang et al., 2010), plays an important role in the autonomic dysregulation of HF. Rats with HF following myocardial infarction have increased glial activation (Yu et al., 2010) and increased pro-inflammatory cytokine (PIC) levels in the PVN (Yu et al., 2010; Yu et al., 2012), and maneuvers that block or counter the actions of these inflammatory mediators in the PVN reduce sympathetic activity and improve cardiac hemodynamics (Kang et al., 2009; Yu et al., 2012). However, the mechanisms that upregulate inflammation in the PVN are still poorly understood.

The Subfornical Organ as a Nidus of Inflammation in Heart Failure

Recent studies have demonstrated that activation of neurons in the SFO, which is exposed to humoral factors and has direct projections to the PVN, can influence the inflammatory neurochemical milieu of the PVN. Blood-borne IL-1β increases the expression of inflammatory elements in the PVN in normal rats, an effect that can be significantly reduced by lesioning the SFO (Wei et al., 2017), and TNFR1 receptor knockdown in the SFO of HF rats reduces the expression of inflammatory elements in the PVN (Yu et al., 2017). Indirect evidence suggesting a similar effect of AngII acting upon AT_1aR in the SFO comes from studies in which a slow-pressor dose of AngII induced upregulation of inflammatory mediators in the PVN of normal rats (Yu et al., 2013b). The present study demonstrates directly that activation of the AT_1a receptor in the SFO contributes to the inflammatory response in the PVN of HF rats. In this study, a partial knockdown of AT_1aR expression in the SFO sufficient to prevent the increase in AT₁R expression that normally occurs in the SFO of HF rats (Tan et al., 2004) resulted in lower levels of mRNA for TNF-α and IL-1β, TNFR1 and COX-2 in the PVN. CD68 and GFAP mRNA levels were also lower, suggesting that microglia and astrocytes, the major sources of PIC production in the brain (Ransohoff and Brown, 2012; Gimsa et al., 2013), were less active in the PVN after a reduction in AT_1aR expression in the SFO.

“Cross-talk” between Inflammation and Renin-Angiotensin System Activity

The relationship between AngII and neuroinflammation has long been recognized (Dandona et al., 2007; Villapol and Saavedra, 2015), with circumventricular organs that have high densities of AT₁R proposed as one potential site at which inhibition of RAS activity may reduce inflammation in the central nervous system (Villapol and Saavedra, 2015). The present demonstration of a RAS-mediated inflammatory response in the SFO, taken together with the recent demonstration of a TNF-α-mediated RAS response in the SFO of HF rats (Yu et al., 2017), suggests a level of “cross-talk” between PICs and the RAS at the SFO level comparable to that which has been described in PVN (Kang et al., 2008a; Kang et al., 2010) and raises the question of what neurochemical stimulus drives SFO to PVN projection neurons to upregulate inflammation in the PVN. Might the effect of TNF-α activation of TNFR1 in the SFO on inflammation in the PVN be mediated by TNF-α-induced increases in RAS activity in the SFO? That possibility is supported by studies in normal rats demonstrating that microinjection of TNF-α and IL-1β into the SFO upregulates the expression of angiotensin converting enzyme (ACE) and AT₁R in the SFO, and that the sympathoexcitatory effects elicited by those inflammatory agonists can be significantly attenuated by blocking RAS activity in the SFO with an ACE inhibitor or an AT₁R blocker (Wei et al., 2015). On the other hand, might the effects of AngII activation of AT_1aR in the SFO on inflammation in the PVN be driven by AngII-induced increased inflammation in the SFO as manifest by increased TNF-α, IL-1β or COX-2 activity (with prostaglandin E2 production)? That possibility is supported by in vitro studies showing that PICs can directly activate SFO neurons (Desson and Ferguson, 2003; Simpson and Ferguson, 2017). Likely, inflammation and RAS activity both contribute, along with molecular signaling mechanisms downstream of AT_1aR and TNFR1 activation, e.g., generation of reactive oxygen species or activation of mitogen activated protein kinases. Further studies will be needed to decipher the nature of the stimulus driving the SFO to PVN pro-inflammatory pathway in HF. Such studies would require identification of the phenotypes of SFO neurons projecting to specific regions of PVN, the receptors expressed by those neurons, the neurotransmitters those SFO neurons release at the PVN level, and the effects of those transmitters on glial and neuronal elements in the PVN that produce the inflammatory mediators.

Similarly, since knockdown of AT_1aR activity in the SFO ultimately resulted in reduced expression of AT_1aR as well as inflammatory mediators in the PVN, the observed improvements in the peripheral manifestations of HF cannot be attributed entirely to the observed reduction in inflammatory mediators in the PVN. While AT₁ receptors are reportedly not expressed on presympathetic or vasopressinergic neurons in the PVN (Oldfield et al., 2001; de Kloet et al., 2017), at least under normal conditions, there is ample evidence that AngII acts upon AT₁ receptors in the PVN to increase the activity of pre-autonomic and magnocellular neurons (Ferguson, 2009). Presynaptic influences of AT₁ receptors on fibers in close proximity to these neurons may be implicated (de Kloet et al., 2017). Notably, however, the cellular localization of the AT₁ receptors that are upregulated in the PVN in HF (Tan et al., 2004) has not been determined. The relative contributions of reduced inflammation and RAS activity in the PVN were not determined in this study, and previous studies have demonstrated that intervening in either pathway within the PVN can ameliorate sympathetic excitation in HF (Kang et al., 2010; Yu et al., 2013a).

Nevertheless, it is clear from this study that preventing the increase in AT_1aR expression in the SFO had beneficial cardiopulmonary effects, likely due at least in part to reductions in plasma NE and AVP and their vascular and renal effects on preload (i.e., reduced renal sodium and water reabsorption) and afterload (i.e., reduced vasoconstriction). The SFO has direct projections to parvocellular and magnocellular regions of the PVN (Kawano and Masuko, 2010), and the reductions in plasma NE and AVP suggest that both parvocellular presympathetic and magnocellular vasopressinergic PVN neurons were affected – directly or indirectly - by AT_1aR knockdown in the SFO. Notably, however, the plasma levels of AngII and TNF-α were unaffected, suggesting that the blood-borne stimuli to AT_1aR and TNFR1 in the SFO persisted despite AT_1aR knockdown. Of interest, others have reported that a chronic PVN infusion of the AT₁R blocker losartan reduced circulating cytokines as well as NE in this same model of ischemia-induced HF (Yu et al., 2013a).

Glial Elements and AT_1aR-Mediated Inflammation

It is particularly interesting that knockdown of AT_1aR in the SFO of HF rats reduced the mRNA expression for CD68 and GFAP in both SFO and PVN of HF rats, suggesting that AT_1aR-mediated activation of these glial elements, whether directly or indirectly via their downstream products, plays an important role in AngII-induced central inflammation. While AT₁R reportedly are not expressed by glial cells under normal conditions (de Kloet et al., 2015), there is nevertheless evidence the AngII activation of AT₁R influences glial activity under stressful conditions. In vitro studies have demonstrated that inhibiting microglial activation with minocyline reduced AngII-dependent production of reactive oxygen species (ROS) in the PVN (Biancardi et al., 2016), and in vivo studies have demonstrated that intracerebroventricular infusion of minocycline reduced mean arterial pressure and TNF-α, IL-1β and IL-6 mRNA in the PVN of rats subjected to AngII-induced hypertension (Shi et al., 2010). AT₁R-mediated enhancement of LPS-induced microglial activation has been shown to evoke NF-κB activity (Miyoshi et al., 2008). AngII may also induce cytokine production by astrocytes (Danielyan et al., 2007; Gowrisankar and Clark, 2016), likely via an AT₁R/NF-κB/ROS pathway (Gowrisankar and Clark, 2016). A recent study (Stern et al., 2016) suggested that astrocytes in the PVN express AT₁R and play an important role in mediating the sympathoexcitatory effects of AngII on PVN neurons. Another (Isegawa et al., 2014) revealed that the increase in AT₁R mRNA in the brainstem of mice with ischemia-induced HF occurred predominantly in astrocytes, and that GFAP-specific AT₁R knockout HF mice had significantly lower urinary NE and reduced cardiac remodeling. These results suggest an important role for AngII activation of glial cells in the central inflammatory response that contributes to sympathetic excitation in HF.

The Subfornical Organ as a Potential Therapeutic Target in Heart Failure

An important implication of this study, in the context of our previous study regarding the effects of TNFR1 knockdown (Yu et al., 2017), is that the SFO may be a prime target for therapeutic interventions designed to lessen the pathophysiological impact of neurohumoral excitation in HF. It is particularly notable that the effect of this isolated intervention in the SFO closely approximates the effects we have previously observed with more global central interventions in sympathetic drive, e.g., with intracerebroventricular administration of drugs counteracting RAS and inflammatory mediators (Kang et al., 2008b; Yu et al., 2012). Thus, interventions targeting AT₁R in this one small region of the brain may have major therapeutic effects on the central autonomic pathways regulating neurohumoral excitation in HF. By virtue of its location outside the blood-brain barrier, its rich endowment with AT₁R and PIC receptors (McKinley et al., 2003; Wei et al., 2015; Wei et al., 2017), and its projections to PVN regions regulating sympathetic activity and AVP release (Kawano and Masuko, 2010), the SFO is well positioned to mediate the effects of drugs that might counter the adverse effects of blood-borne AngII and PICs in HF.

That anatomical advantage notwithstanding, the complex physiological interactions between RAS and inflammatory systems may complicate the therapeutic approach. Will intervention in only one of these systems suffice, or will the redundancy and overlap between these two systems require simultaneous inhibition of both? Or, might inhibition of a molecular mediator common to both RAS and inflammation be a more effective approach to reducing sympathetic and vasopressinergic overactivity in HF? Other work from our laboratory (Wei et al., 2008a; Yu et al., 2016) suggests that inhibition of mitogen-activated protein kinases (MAPK), and particularly of the extracellular signal-regulated kinases 1 and 2 (ERK1/2), holds promise as a therapeutic approach.

Central Influences on Cardiac Function in Heart Failure

AT_1aR knockdown in the SFO produced physiologically important improvements in LV dP/dt_max and LVEDP, but did not influence LV volumes, mass, or LVEF. Some (Francis et al., 2001; Francis et al., 2004; Guggilam et al., 2008; Kang et al., 2008a; Kang et al., 2008b) but not all (Kang et al., 2009; Kang et al., 2010; Huang et al., 2009b) previous studies have reported an analogous dissociation of the effects of central interventions upon measures of LV function after myocardial infarction. There are several possible explanations for the dissociation we observed between LV dP/dt_max and LVEF as indices of LV function. LV dP/dt_max occurs during isovolumic contraction, prior to the ejection phase of the cardiac cycle, so the hemodynamic and echocardiographic methods are actually measuring different phenomena. In addition, direct measurement of intracardiac pressures may be more sensitive than transthoracic echocardiography in detecting small changes in cardiac function. Finally, in the present study these two indices of contractility were obtained under different experimental conditions (i.e., different anesthetic agents).

The reductions in sympathetic nerve activity and plasma vasopressin levels likely contributed to the improvements in cardiac hemodynamics via their renal and vascular effects on preload and afterload. The changes in LV pressures were associated with significant reductions in lung/BW and RV/BW ratios, consistent with reduced pulmonary congestion and right ventricular remodeling.

Limitations of the Study

A simplistic interpretation of this study is that activation of AT_1aR within the SFO contributes to inflammation in the PVN in HF. However, the nature of the cellular elements and neurochemical mechanisms that produce these changes remain a mystery. Even the stimulus to the AT_1aR in the SFO is uncertain. Since plasma levels of AngII are increased in humans (Chatterjee, 2005) and in rats (Huang and Leenen, 2009a) with HF, the bloodstream seems a likely source of agonist, but the upregulation of angiotensin converting enzyme in the SFO in HF rats (Tan et al., 2004) makes local production of AngII an alternative (or perhaps additional) source. Within the SFO, a reasonable explanation for AT₁R-mediated upregulation of TNF-α, IL-1β and COX-2 might be AT_1aR activation of NF-κB, as suggested by a previous study of the PVN (Kang et al., 2008a).

How a reduction in AT₁R activity in the SFO might lead to reduced inflammation and AT₁R expression in the PVN is even less clear. SFO neurons project directly to the PVN to innervate presympathetic and neuroendocrine neurons (Kawano and Masuko, 2010) but, as mentioned above, the phenotypes of the SFO to PVN projection neurons that mediate the neurochemical changes in the PVN that we have observed have not been defined. Some SFO to PVN projection neurons are reported to release AngII (Li and Ferguson, 1993), others glutamate (Llewellyn et al., 2012). The mechanism(s) by which AT₁R knockdown in the SFO achieves a reduction in AT₁R downstream in the PVN of HF rats, along with reduced CD68, GFAP and c-Fos mRNA in the PVN, remains to be determined, but may be related to reduced activity of the angiotensinergic neurons projecting from SFO to PVN (Li and Ferguson, 1993). Notably, a previous study in rats subjected to AngII-induced hypertension reported a similar reduction in PVN AT_1aR mRNA expression following AT_1aR knockdown in the SFO (Wang et al., 2016).

Finally, the present study did not elucidate the mechanism(s) for the reduction in plasma AVP, which may have resulted from reductions in brain RAS activity (McKinley et al., 2004; de Kloet et al., 2015) or inflammation (Landgraf et al., 1995; Ferri et al., 2005; Kageyama and Suda, 2009), or both.

Several technical limitations deserve comment. The present study used mRNA levels as indicators of neurochemical changes in the excitatory milieu of the SFO and PVN. Although our previous studies have shown parallel changes in mRNA and protein (Yu et al., 2010; Yu et al., 2012; Yu et al., 2017), as might be expected in this chronic model of HF, it is possible that the pattern of mRNA changes might not be matched by similar changes in protein. Real-time PCR was also used instead of immunohistochemistry to assess neuronal excitation, as others have done (Kawasaki et al., 2009; Krause et al., 2011), precluding assessment of activity in the different regions of the PVN. However, humoral data supports the conclusion that presympathetic and vasopressinergic PVN neurons were affected by downregulation of AT_1aR in the SFO. In addition, the plasma levels of humoral factors were determined on trunk blood obtained under anesthesia at the time of euthanasia. Thus, while significant differences were noted between sham and HF rats and between HF rats treated with control shRNA and AT_1aR shRNA, the absolute values were undoubtedly exaggerated by stress.

Most previous studies of the role of AT₁R in the SFO have focused on the influences of AngII on sympathetic drive, volume regulation and arterial pressure, with an emphasis on AT₁R-mediated ROS production and other molecular mechanisms related to the pathogenesis of hypertension. For example, recent studies examining the influences of AT₁R in the SFO on the neurochemistry of the PVN have demonstrated that AT₁R knockdown in the SFO prevents AngII-induced ROS production in the PVN (Wang et al., 2016) and that microinjection of AngII into the SFO elicits a sympathetic response mediated by glutamate release in the PVN (Llewellyn et al., 2012). The present study extends our understanding of the prominent role of the SFO in cardiovascular disease by identifying a link between AT_1aR-mediated events in the SFO and inflammatory mechanisms in the PVN that are known to drive neurohumoral excitation in HF. HF rats in which the expected upregulation of AT_1aR in the SFO failed to occur because of pretreatment with AT_1aR shRNA had less expression of inflammatory markers in the SFO and in the PVN than rats treated with a control shRNA. These rats also had less peripheral evidence of HF, presumably due at least in part to a reduction in inflammatory mediators in the PVN, though a reduction in AT_1aR expression in the PVN also occurred and may have contributed. These findings extend our understanding of the complexity of RAS and inflammatory interactions in the central nervous system in HF. It appears likely that circulating AngII, acting upon AT_1aR at the SFO level, promotes the inflammatory state in central neural pathways that contributes to the augmented sympathetic and vasopressinergic activity in HF, with devastating consequences. The exposure of the SFO to blood-borne agents suggests that it may be an ideal target for pharmacological interventions designed to target these central mechanisms. At this point, because of the convergence of blood-borne AngII and PICs on the SFO and their interactions at that level, it is unclear whether blocking the activity of only one of these excitatory systems would be effective in reducing neurohumoral excitation in HF. A better understanding of the phenotypes of the SFO neurons involved and the neurochemical mechanisms affecting presympathetic and neuroendocrine PVN neurons will be required before an effective therapeutic approach can be designed.

Highlights.

• Central inflammation contributes to neurohumoral excitation in rats with ischemia-induced heart failure (HF)

• Angiotensin II induces inflammation in the subfornical organ (SFO) and the hypothalamic paraventricular nucleus (PVN)

• Preventing upregulation of angiotensin II type 1a receptor (AT_1aR) in the SFO in HF reduces inflammation in SFO and PVN

• Preventing upregulation of AT_1aR in the SFO in HF also reduces sympathetic excitation and vasopressin release

• Increased circulating angiotensin II likely contributes to neuroinflammation that drives neurohumoral excitation in HF

ABBREVIATIONS

AAV

adeno-associated vírus

ACE

angiotensin converting enzyme

AT₁R

angiotensin II type 1 receptors

AT_1aR

angiotensin II type 1a receptors

AngII

angiotensin II

AVP

arginine vasopressin

BW

body weight

CD68

cluster of differentiation 68

CL

left coronary artery ligation

CON

control

COX-2

cyclooxygenase-2

COX-1

cyclooxygenase-1

DBP

diastolic blood pressure

dP/dt_max

maximum rate of rise of LV pressure

GFAP

glial fibrillary acidic protein

GFP

green fluorescent protein

HF

systolic heart failure

HR

heart rate

IL-1β

interleukin-1 beta

IL-6

interleukin-6

%IZ

ischemic zone as a percent of left ventricular circumference

LPS

lipopolysaccaride

LV

left ventricle

LVEDP

left ventricular end diastolic pressure

LVEDV

left ventricular end diastolic volume

LVEF

left ventricular ejection fraction

LVPSP

left ventricular peak systolic pressure

NE

norepinephrine

NF-κB

nuclear factor κB

PGE₂

prostaglandin E2

PIC

proinflammatory cytokine

PVN

hypothalamic paraventricular nucleus

SFO

subfornical organ

TNF-α

tumor necrosis factor - alpha

TNFR1

tumor necrosis factor - alpha receptor 1

RAS

renin-angiotensin system

ROS

reactive oxygen species

RV

right ventricle

SBP

systolic blood pressure

shRNA

short hairpin RNA

Vol

volume