Caspase lesions of PVN-projecting MnPO neurons block the sustained component of CIH-induced hypertension in adult male rats

Abstract

Obstructive sleep apnea is characterized by interrupted breathing that leads to cardiovascular sequelae including chronic hypertension that can persist into the waking hours. Chronic intermittent hypoxia (CIH), which models the hypoxemia associated with sleep apnea, is sufficient to cause a sustained increase in blood pressure that involves the central nervous system. The median preoptic nucleus (MnPO) is an integrative forebrain region that contributes to blood pressure regulation and neurogenic hypertension. The MnPO projects to the paraventricular nucleus (PVN), a preautonomic region. We hypothesized that pathway-specific lesions of the projection from the MnPO to the PVN would attenuate the sustained component of chronic intermittent hypoxia-induced hypertension. Adult male Sprague-Dawley rats (250–300 g) were anesthetized with isoflurane and stereotaxically injected bilaterally in the PVN with a retrograde Cre-containing adeno-associated virus (AAV; AAV9.CMV.HI.eGFP-Cre.WPRE.SV40) and injected in the MnPO with caspase-3 (AAV5-flex-taCasp3-TEVp) or control virus (AAV5-hSyn-DIO-mCherry). Three weeks after the injections the rats were exposed to a 7-day intermittent hypoxia protocol. During chronic intermittent hypoxia, controls developed a diurnal hypertension that was blunted in rats with caspase lesions. Brain tissue processed for FosB immunohistochemistry showed decreased staining with caspase-induced lesions of MnPO and downstream autonomic-regulating nuclei. Chronic intermittent hypoxia significantly increased plasma levels of advanced oxidative protein products in controls, but this increase was blocked in caspase-lesioned rats. The results indicate that PVN-projecting MnPO neurons play a significant role in blood pressure regulation in the development of persistent chronic intermittent hypoxia hypertension.

NEW & NOTEWORTHY Chronic intermittent hypoxia associated with obstructive sleep apnea increases oxidative stress and leads to chronic hypertension. Sustained hypertension may be mediated by angiotensin II-induced neural plasticity of excitatory median preoptic neurons in the forebrain that project to the paraventricular nucleus of the hypothalamus. Selective caspase lesions of these neurons interrupt the drive for sustained hypertension and cause a reduction in circulating oxidative protein products. This indicates that a functional connection between the forebrain and hypothalamus is necessary to drive diurnal hypertension associated with intermittent hypoxia. These results provide new information about central mechanisms that may contribute to neurogenic hypertension.

INTRODUCTION

Obstructive sleep apnea (OSA) is characterized by interruptions in breathing that are sufficient to cause significant arterial hypoxemia (57). Intermittent hypoxia and hypercapnia associated with OSA contribute to elevated sympathetic nerve activity that leads to hypertension and other cardiovascular morbidities (22). As OSA progresses, the hypertension observed during the intermittent hypoxic sleep cycle begins to manifest into the normoxic waking hours leading to the development of sustained systemic hypertension and cardiovascular disease (13, 47).

To understand the early events in the pathogenesis of hypertension associated with OSA, a rodent model of chronic intermittent hypoxia (CIH) was designed to mimic the intermittent hypoxemia, hypertension, and other cardiovascular comorbidities related to OSA (13, 16, 18, 47). Several mechanisms have been linked to hypertension associated with CIH such as changes in chemoreceptor reflex sensitivity (18, 41), baroreceptor reflex function (48), altered cardiorespiratory coupling (58), and catecholamine neurons in the nucleus of the solitary tract in the hindbrain (3). However, additional mechanisms, including the median preoptic nucleus (MnPO) in the forebrain (9, 15, 46), may contribute to the sustained component of CIH hypertension, which occurs in normoxic components of the diurnal cycle.

During the initiation of CIH hypertension, chemoreceptors become sensitized, which increases sympathetic outflow and plasma angiotensin (16, 17, 19, 32, 47). Our working hypothesis is that increased circulating angiotensin II (ANG II) activates the forebrain, thereby increasing sympathetic drive, creating a vicious cycle (35, 47). Peripheral ANG II could be sensed by forebrain circumventricular organs (CVOs)—the subfornical organ (SFO) and organum vasculosum of the lamina terminalis (OVLT)—through angiotensin type 1a receptors (AT_1aRs) and relayed to downstream sympathoregulatory regions, including the MnPO (47).

The MnPO lies in the anteroventral third ventricle (AV3V) region of the forebrain and plays an important role in receiving and integrating afferent signals from the CVOs that lie just dorsal and ventral to the MnPO on the AV3V, as well as propagating information to the hypothalamus (2, 24, 33, 39, 44). Neurons in the MnPO are involved in thermoregulation, osmoregulation, sleep, body fluid balance, and autonomic regulation (2, 7, 24, 33). Recent studies have found that an intact MnPO is necessary for the development of CIH-induced hypertension (9, 45), as well as ANG II-induced hypertension (38, 39). The MnPO has projections to the paraventricular nucleus (PVN) of the hypothalamus (30, 33, 35, 47). The PVN contains preautonomic neurons involved in blood pressure regulation as well as other groups of cells that contribute to neuroendocrine function, ingestive behavior, and metabolism that are still not yet fully understood (5, 45, 47). The PVN has been implicated as a major site in the pathogenesis and maintenance of hypertension, as it is a key area for integrating central and peripheral stimuli to regulate blood pressure (5, 8). Through its efferent projections, the PVN can influence activity in sympathoregulatory regions of the spinal cord and hindbrain, specifically the intermediolateral column of the spinal cord and the rostral ventrolateral medulla (RVLM).

The MnPO influences sympathetic responses through a glutamatergic input to the PVN influencing sympathetic vasoconstrictor pathways (30, 33) that contributes to hypertension dependent on elevated sympathetic tone and the PVN (35, 45). Along with other adaptations, activation of the AV3V could contribute to increased sympathetic outflow during normoxia, resulting in sustained hypertension (13, 20, 26, 34, 35). Our work suggests that the MnPO is necessary for hypertension produced by CIH (9, 45), whereas other studies indicate that PVN-projecting MnPO neurons have an excitatory effect on PVN activity (30). The role of MnPO-PVN pathway neurons that contribute to CIH-induced hypertension, however, has not been elucidated.

We hypothesized that a pathway-specific lesion of the MnPO-PVN projection would attenuate the hypertension that persists into the normoxic period, or the sustained component, of CIH-induced hypertension. A caspase-3 lesion approach was used to selectively eliminate MnPO neurons projecting to the PVN. Each rat was injected bilaterally in the PVN with a retrograde adeno-associated virus (rAAV) containing Cre recombinase. This caused Cre expression in all neurons that project to the PVN at the injection site in the MnPO. The MnPO was injected with an adeno-associated virus (AAV) containing caspase-3 and tobacco etch virus protease with a flex promotor, which allows the constructs to be translated only in cells expressing Cre (56). This approach results in apoptosis of MnPO neurons projecting to the PVN. The study used a 7-day CIH protocol to study the early events involved in the initiation and maintenance of CIH hypertension, without potential confounding effects of end-organ damage.

METHODS

Ethical Approval

The experimental procedures involving animals adhered to the standards of the National Institutes of Health’s Guide for the Care and Use of Laboratory Animals and were approved by the University of North Texas Health Science Center Institutional Animal Care and Use Committee.

Animals

Adult male Sprague-Dawley rats (250–300 g body wt; Charles River) were individually housed in a temperature-controlled (25°C) room on a 12-h:12-h light-dark cycle with light onset at 7:00 am. Food and water were available ad libitum except on the day of perfusions. Rats were weighed daily, and their food and water intake was monitored.

Microinjection Surgeries

Rats were anesthetized with 2% isoflurane and received stereotaxic microinjections of the rAAV containing Cre [AAV9.CMV.HI.eGFP-Cre.WPRE.SV40; Penn Vector Core; 1.3 × 10¹³ viral particles (vm)/mL] bilaterally into the PVN (−1.8 mm anterior, ±0.4 mm lateral, and 7.6 mm ventral from bregma; 37). In the same surgery, caspase-3 (AAV5-flex-taCasp3-TEVp; 2.1 × 10¹² vg/mL) or control (AAV5-hSyn-DIO-mCherry; 3.3 × 10¹² vg/mL) virus (Vector Core at the University of North Carolina at Chapel Hill) was injected into the MnPO (microinjector angled at 8° from medial to lateral to avoid the septum, at coordinates of 0.9 mm lateral and 6.7 mm ventral from bregma; 37). Burr holes were then drilled in the skull for each measured injection site. In the MnPO, 200–300 nL of AAV were delivered at a rate of 200 nL/min. For the PVN, each side was injected with 200 nL with a rate of 200 nL/min. Injections were made with 30-gauge hypodermic tubing that was connected to a Hamilton 5-µL syringe (No. 84851; Hamilton, Reno, NV) by calibrated polyethylene tubing that was used to determine the injection volume. Tubing was clamped to a standard electrode holder (model 1770; David Kopf Instruments, Tujunga, CA) connected to an electrode manipulator (model 920; David Kopf Instruments) mounted on a standard stereotaxic instrument (model 900LS; David Kopf Instruments). Separate injector systems were used to make the MnPO and PVN injections to prevent contamination of the AAVs. At each site, the injector remained in place for 5 min to allow for absorption and then was slowly withdrawn. Gel foam was placed into the hole in the skull. Absorbable antibiotic suture was used to close the incision and minimize postsurgical infection. Each rat was given carprofen (1 mg, Rimadyl; Bio-Serv) orally to minimize pain following surgery. Rats were allowed to recover for 1 wk before radiotelemetry surgery and for 3 wk before beginning the CIH protocol.

Radiotelemetry Implantation

Rats were anesthetized as before with 2% isoflurane 1 wk after microinjection surgeries. Rats were implanted with an abdominal aortic catheter attached to an HD-S10 radiotelemetry transmitter (Data Sciences International, St. Paul, MN) to continuously record hemodynamic measurements (Dataquest IV; Data Sciences International). The transmitter was secured to the abdominal wall using prolene suture and remained in the abdominal cavity through the duration of the experiment. Two weeks were allowed for recovery from surgery. Blood pressure measurements obtained during a 10-s sampling period (500 Hz) were averaged and recorded every 5 min.

CIH Protocol

After 2 wk of postsurgical recovery from microinjections and 1 wk recovery from telemetry implantation, rats individually housed in their home cages were relocated to custom-built plexiglass chambers for 2 days of acclimation and 5 days of baseline recording at normoxia (21% O₂). After this 7-day period, rats were exposed to CIH for 8 h/day (8:00 am to 4:00 pm) for 7 days. The O₂ concentration in the chambers was regulated using custom-built user-controlled timers that separately switched the flow of room air and nitrogen into each chamber. The O₂ concentration was continuously monitored using O₂ sensors (KE-25; Kent Scientific, Torrington, CT) and recorded using Spike2 software (v.5.07; Cambridge Electronic Design, Cambridge, UK). Flow rates of room air and nitrogen to each chamber were controlled separately using individual flowmeters, as previously described (26). The CIH protocol consisted of 6-min cycles, with 3 min of hypoxia (10% O₂) using nitrogen infusion and 3 min of normoxia (21% O₂) using reinfusion of room air, as previously described (9, 15, 26). These cycles repeated for 8 h/day for 7 days, resulting in 80 cycles of hypoxic exposure per day. During the remaining 16 h of the day, the chambers were open to normoxic room air (21% O₂). Controls were placed in identical chambers within the same room but only exposed to normoxic room air (21% O₂). After the 7-day CIH protocol, on the morning of the eighth day, all animals were euthanized.

Perfusions: Tissue and Body Fluid Collection

On the morning of day 8, after the 7-day CIH protocol terminated, animals were anesthetized using 100 mg/kg ip inactin (Sigma-Aldrich). Blood was collected from the left ventricle and transferred into an EDTA vacutainer containing heparin immediately preceding the perfusion, to measure advanced oxidative protein products (AOPPs). Blood was also transferred into a separate tube to measure hematocrit and plasma osmolality using a vapor pressure osmometer (Wescor, Logan, UT) as previously described (23). Rats were transcardially flushed first with phosphate-buffered saline solution (PBS, ~100 mL) followed by 4% paraformaldehyde (PFA, 300 mL). Each brain was kept in 4% PFA overnight before being placed in 30% sucrose in PBS.

Tissue Processing

Immunohistochemistry.

Forty-micrometer coronal sections of each previously perfused brain were cut using a cryostat. Three sets of serial sections were collected in cryoprotectant and stored at −20°C until they were processed for immunohistochemistry. One set of sections containing the MnPO and PVN were mounted directly on slides. The other sections were processed for immunohistochemistry as described below.

fosb.

Separate sets of serial sections from brains injected with caspase-3 or control virus were stained for FosB [1:1,000, sc-48-G, goat polyclonal anti-Fos antibody, Santa Cruz Biotechnology; Research Resource Identifier (RRID): AB_631516], which has been previously validated (15, 25, 46). After 48 h, sections were washed using PBS and transferred to a secondary antibody (BA-9500, biotinylated anti-goat; Vector Laboratories; RRID: AB_2336123) for 3,3′-diaminobenzidine (DAB) reaction and labeling (15, 46). After the DAB reaction, the sections were washed, placed in the primary antibody [1:500, DSHB-GFP3-1F5, green fluorescent protein (GFP) anti-mouse primary antibody; Developmental Studies Hybridoma Bank, University of Iowa; RRID: AB_2617424], and incubated for an additional 48 h followed by incubation with a Cy2-conjugated anti-mouse antibody (715-225-151; Jackson ImmunoResearch; RRID: AB_2340827) for 4 to 5 h (46). The sections were then mounted on gelatin-coated slides, dried, and coverslipped with Permount for imaging.

nitric oxide synthase 1 and glial fibrillary acidic protein.

Other sections from subsets of each treatment group were stained separately for nitric oxide synthase 1 (NOS1; 1:1,000, sc5302; Santa Cruz Biotechnology; RRID: AB_626757) to determine MnPO-PVN-projecting neuron phenotype (28) or glial fibrillary acidic protein (GFAP; 1:500, G3893; Sigma-Aldrich; RRID: AB_477010) to ensure that only neurons were affected by caspase-3-induced apoptosis (46, 53). For these anatomical studies, sections came from rats injected with just the AAV9-GFP retrograde into the PVN or rats injected with the AAV9-GFP retrograde into the PVN and caspase-3 into the MnPO. After 48 h, sections were washed using PBS and transferred to a Cy3-conjugated anti-mouse secondary antibody (711-165-151; Jackson ImmunoResearch; RRID: AB_2315777). The sections were then mounted on gelatin-coated slides, dried, and coverslipped with ProLong Diamond mounting medium for imaging. All antibodies were diluted to working concentration in PBS diluent (0.25% Triton, 3% horse serum, and 96.75% PBS).

In situ hybridization.

In situ hybridization (ISH) experiments were performed to characterize the neuronal phenotype of PVN-projecting MnPO neurons transfected by the AAV. After the CIH protocol was terminated, separate groups of rats from those used for immunohistochemistry were anesthetized using 100 mg/kg ip inactin (Sigma-Aldrich) and transcardially flushed first with RNase-free PBS and then perfused using 4% PFA. Brains were dehydrated in RNase-free 30% sucrose. Twenty-micrometer coronal sections of each brain were cut using a cryostat (Leica). Four to six sets of serial MnPO sections were collected in RNase-free PBS, mounted onto Superfrost Plus Gold microscope slides (Thermo Fisher Scientific, Waltham, MA), left at room temperature for ~2 h, and then stored at −80°C until used for ISH experiments. All reagents used for ISH experiments were purchased from Advanced Cell Diagnostics (Newark, CA). Experiments were performed using a previously established protocol (12, 49, 54). ISH was performed for vesicular glutamate transporter 2 (vGLUT2; Cat. No. 317011) to study the phenotype of PVN-projecting MnPO neurons in rats only injected in the PVN with AAV9-GFP virus. ISH for vGLUT2 was performed using a fluorescence assay (Cat. No. 320293). Experimental protocols were performed according to the manufacturer’s instructions described in RNAscope Fluorescent Multiplex Kit User Manual Part 2 (Cat. No. 320293; Advanced Cell Diagnostics).

Imaging and Quantification

All imaging and quantification was conducted by investigators blind to experimental conditions of the individual animals. Tissue sections were imaged with a microscope equipped for epifluorescence (Olympus BX41; Olympus, Center Valley, PA) with a digital camera (Olympus DP70). Cell counting and image analysis were performed using National Institutes of Health ImageJ software (v.1.49). For sections stained for FosB, tissue was visualized using bright-field illumination. FosB counts were calculated in ImageJ as previously described (44, 46). For sections stained for either vGLUT2 or NOS1, images were taken to identify AAV9-GFP expression at emission and excitation frequencies of 493 and 518 nm, respectively, and vGLUT2 or NOS1 expression at emission and excitation frequencies of 550 and 568 nm, respectively. The separate images were merged in ImageJ to determine raw cell counts for colocalization. Bright-field images were inverted using ImageJ and merged with images of GFP fluorescence to determine colocalization. Sections stained for GFAP were imaged for AAV9-GFP expression and GFAP expression. GFAP expression was then quantified in ImageJ using a densitometric approach. Sections from control rats that were not injected with virus (n = 4 rats) and stained only with Cy3 anti-mouse secondary antibody were processed according to the same immunohistochemistry protocol. Images were taken of these sections using the same exposure settings and were converted to 8-bit grayscale in ImageJ. In ImageJ, the threshold of each image was uniformly adjusted so that integrated density was as close to zero as possible in control sections to subtract out background. Once this threshold was set, sections stained for GFAP were analyzed using the same threshold to measure integrated density. Integrated density of GFAP fluorescence in the MnPO was then compared between rats injected with only AAV9-GFP in the PVN and rats injected with AAV9-GFP in the PVN and caspase-3 in the MnPO. This approach allowed us to determine whether caspase-3 was significantly decreasing GFAP fluorescence density.

Injection Verification

For the PVN injection sites, GFP expression from AAV9 was examined in the PVN and in adjacent regions such as the lateral hypothalamus, the perifornical region, and the third ventricle. Fluorescence images containing the injection sites of AAV9-GFP in all rats were taken under ×1.25 magnification using a Leica DM6 B upright microscope. Images (1,296 × 966 pixels) were aligned to the reference brain section image depicted in the rat brain atlas (37). The heat map was created using the matplotlib package in Python 3.7.1. To create the heat map, the fluorescence images were converted to 8-bit grayscale images. The pixel gray values (0–255) were measured for all the images. For each of the 1,296 × 966 pixels, the average pixel gray value of all rats was calculated. The newly generated 1,296 × 966 pixels were used to create a heat map representing the average AAV9-GFP levels in the PVN and the surrounding areas among all rats. The heat map was then overlaid on the reference brain section image to estimate the distribution of AAV9-GFP around the injection site. Rats with injections that were not confined to the PVN or did not include the PVN were not included in the study.

For the MnPO injection sites and lesions, the numbers of GFP-positive cells in the OVLT, MnPO, and SFO were counted. Counts from multiple sections were averaged for each individual. All regions were defined using the stereotaxic atlas of Paxinos and Watson (37). The anterior border of the RVLM was defined by the caudal pole of the facial nucleus, and the rostral hypoglossal nucleus was used for the posterior border (10, 42), which corresponds to −11.60 to −13.68 posterior to bregma (37).

Rats that did not have successful injections into the MnPO or both PVNs were excluded from anatomical studies. Those that received the caspase virus and were exposed to CIH but had unsuccessful microinjections were used, however, in hemodynamic and blood component analyses (CASP CIH MISS) to serve as an additional control.

Advanced Oxidative Protein Product Assay

Plasma oxidative stress was measured using Cell Biolabs’ OxiSelect AOPP Assay Kit (STA-318) and performed using a previously described protocol based on the manufacturer’s instructions (50, 51). The AOPP assay kit provides the ability to measure the concentration of total oxidized protein in a sample (μM) by reacting with chloramine to initiate a color change. Samples are read at a wavelength of 340 nm, and concentration is calculated by comparing with the predetermined chloramine standard curve. Assay results are reported as a percentage of control [individual value/(average of normoxic control values) × 100], as previously described (50).

Exclusion Criterion

Rats with unsuccessful AAV bilateral PVN or MnPO injection were eliminated from analysis (n = 11 rats). Rats exposed to CIH with successful AAV bilateral PVN but unsuccessful MnPO injection of caspase (referred to as “miss”) were used as an additional control only for hemodynamic and AOPP analyses (n = 6 rats). Rats with successful AAV injections but that were instrumented with faulty telemetry or whose hemodynamic recordings had technical problems were only included in neuroanatomical studies (n = 6 rats).

Statistics

Student’s t-tests were performed when comparing AAV9-GFP labeling in control versus caspase virus-injected rats for vGLUT2 and NOS1 analysis. One-way ANOVA was used for comparing plasma measurements (e.g., hematocrit, plasma osmolality, and AOPP levels), neuronal phenotyping dependent on CIH exposure and virus (e.g., FosB and AAV9-GFP labeling), body temperature (BT), and body weight changes. Baseline data for telemetry recordings were analyzed for between-group differences using two-way mixed-effects ANOVA. The baseline measurements were averaged for each rat, and within-group differences for each condition were analyzed using separate repeated measures (RM) ANOVA followed by Holm-Sidak post hoc test to compare against baseline conditions. Effects of CIH on mean, systolic, and diastolic arterial pressure (MAP, SAP, and DAP, respectively), heart rate (HR), and BT during the dark phase (6:00 pm to 6:00 am) and during the 8-h CIH exposure period (8:00 am to 4:00 pm) were analyzed separately by two-way RM ANOVA. Post hoc analysis of between-group differences was performed using the Student-Newman-Keuls (SNK) test. Statistical significance is defined at an α-level of 0.05, and exact P values are reported. Values are reported as means ± SE. All statistics were performed using SigmaPlot v.12.0 (Systat Software, San Jose, CA).

RESULTS

MnPO Neurons Projecting to the PVN Are Excitatory

In situ hybridization experiments for vGLUT2 were performed to determine whether the AAV retrograde virus injected into the PVN was labeling putative excitatory MnPO neurons (Fig. 1A). The results indicate that 89.3 ± 1.4%, or 205 ± 29 of 230 ± 34, of the labeled MnPO neurons (n = 5 rats, 2 to 3 MnPO sections per rat) that project to the PVN (green) are glutamatergic (red). Caspase-3 lesions significantly reduced this number to 21.4 ± 1.1%, or 15 ± 3 of 75 ± 12, of the labeled MnPO neurons [n = 5 rats, 2 to 3 MnPO sections per rat; t(8) = 38.074, P < 0.001, Student’s t-test].

Fig. 1.

Median preoptic nucleus (MnPO) neurons that project to the paraventricular nucleus are mainly glutamatergic and nitric oxide synthase 1 (NOS1) positive. A, left to right: representative single-channel adeno-associated virus 9-green fluorescent protein (AAV9-GFP) labeling, vesicular glutamate transporter 2 (vGLUT2) labeling, and merged image of vGLUT2 in situ hybridization (red) and AAV9-GFP retrograde labeling (green) in the MnPO of AAV9-only injected without (row at top; n = 5 rats) or with caspase (CASP + AAV9, row at bottom; n = 5). B, left to right: representative single-channel AAV9-GFP labeling, NOS1 labeling, and merged image of NOS1 staining (red) and AAV9-GFP retrograde labeling (green) in the MnPO of AAV9-only injected (row at top; n = 4) or CASP + AAV9 (row at bottom; n = 4). White arrows indicate double-labeled neurons. Here, ac, anterior commissure; dMnPO, dorsal MnPO. Scale bars, 100 µM.

Immunohistochemistry was performed on a separate set of sections containing the MnPO (n = 4 rats, 2 to 3 MnPO sections per rat) for NOS1 to further identify the phenotype of MnPO neurons that project to the PVN (Fig. 1B). The results indicate that 42.3 ± 3.0%, or 97 ± 13 of 229 ± 30, of MnPO neurons that project to the PVN (green) are also NOS1 positive (red). In rats with caspase-3 lesions (n = 4 rats, 2 to 3 MnPO sections per rat), the number of NOS1-positive cells significantly decreased by 30% to 30.1 ± 2.5%, or 33 ± 11 of 105 ± 36, labeled MnPO neurons [t(6) = −3.112, P = 0.021, Student’s t-test].

A cytomegalovirus (CMV) promotor in the Cre-expressing AAV9-GFP retrograde virus was utilized to increase transduction rate and viral spread to the MnPO neurons that project to the PVN (55). To validate that we were specifically affecting neuronal populations, we performed densitometric analysis of GFAP and calculated the integrated density in the MnPO of rats injected with or without the caspase-3 virus (Fig. 2). We found no changes in GFAP expression, suggesting that the AAV5-caspase-3 flex viral construct-induced apoptosis did not affect astrocytes [t(7) = 0.303, P = 0.771, Student’s t-test].

Fig. 2.

Astrocytes were unaffected by caspase lesions in the median preoptic nucleus (MnPO). Representative merged images of glial fibrillary acidic protein (red) and adeno-associated virus 9 (AAV9)-green fluorescent protein (green) in the MnPO injected without (left bar in panel at bottom; n = 4 rats) or with caspase-3 (Casp-3) virus (right bar in panel at bottom; n = 5) and integrated density for each group. Scale bars, 100 µm. No significant difference between groups; Student’s t-test. Data are expressed as means ± SE.

Injection Verification

Immunohistochemistry was performed after the 7-day CIH protocol for AAV retrograde to verify the injection sites and caspase lesions (n = 6 rats for each group). The PVN was analyzed to verify bilateral retrograde AAV injections (Fig. 3). GFP-positive nuclei were primarily observed in the PVN with minimal bleed outside of the region of interest. GFP-positive nuclei in the MnPO were assessed to verify the caspase lesions (Fig. 4A). GFP-positive nuclei in the OVLT and SFO were also analyzed to verify that caspase lesions were isolated to the MnPO (Fig. 4A). Overall, the numbers of GFP-positive nuclei in the MnPO of caspase-injected rats were significantly reduced compared with rats injected with control virus [caspase, 101 ± 7; control, 270 ± 8; Fig. 4C, t(10) = 33.303, P < 0.001, Student’s t-test; Fig. 4B]. There were no significant differences for GFP-positive nuclei in OVLT, SFO, or PVN (Fig. 4B).

Fig. 3.

Retrograde adeno-associated virus 9 (AAV9) injections in the paraventricular nucleus (PVN). Heat map containing the injection sites of AAV9-green fluorescent protein (GFP) in the rostral (A), medial (B), and caudal PVN (C) and surrounding areas overlaid onto the reference brain section image (37) to estimate the distribution of AAV9-GFP around the injection site. In the corresponding intensity scale, red indicates maximal GFP, and blue indicates minimal GFP.

Fig. 4.

Caspase lesions significantly decrease retrograde labeling in the median preoptic nucleus (MnPO). A: representative images of retrograde adeno-associated virus (rAAV) retrograde labeling [green fluorescent protein (GFP)] in the organum vasculosum of the lamina terminalis (OVLT, top), subfornical organ (SFO, top middle), MnPO (bottom middle), and paraventricular nucleus (PVN, bottom) of animals injected with the control (CTRL, left) or caspase-3 (CASP, right) virus. B: rats injected with the CASP virus had significantly less AAV retrograde labeling in the MnPO only than CTRL (*Student’s t-test, P < 0.001 compared with CTRL); n = 6 rats for each group. Here, ac, anterior commissure; dMnPO, dorsal MnPO; 3V, third ventricle. Scale bars, 100 µM. Data are expressed as means ± SE.

Caspase-Induced Inhibition of PVN-Projecting MnPO Neurons Blocks CIH-Induced Increases in MAP, SAP, DAP, and HR

To determine whether CIH-induced hypertension could be blunted by selectively eliminating MnPO neurons that project to the PVN, hemodynamic measurements were assessed during a 5-day baseline period and 7 days of CIH for each rat in each group (n = 6–9 rats). During the baseline period, there were no significant differences among treatment groups for baseline MAP, SAP, DAP, HR, or BT (Table 1). This suggests that MnPO neurons that project to the PVN do not significantly contribute to basal MAP, SAP, DAP, HR, or BT.

Table 1.

Average mean, systolic, and diastolic arterial pressure, heart rate, and body temperature during the light and dark phases during a 5-day baseline period before the 7-day CIH protocol

	CTRL NORM	CASP NORM	CTRL CIH	CASP CIH	CASP CIH MISS
n	9	6	8	9	6
MAP, mmHg
Light baseline	97.5 ± 1.5	99.5 ± 1.0	96.5 ± 1.2	98.5 ± 1.1	101.4 ± 2.0
Dark baseline	103.8 ± 2.0	104.1 ± 1.8	102.5 ± 1.4	104.2 ± 1.5	103.1 ± 2.4
SAP, mmHg
Light baseline	113.9 ± 2.7	117.0 ± 2.5	110.9 ± 1.6	117.3 ± 2.2	119.4 ± 3.6
Dark baseline	120.3 ± 3.2	121.3 ± 3.2	117.0 ± 1.4	123.4 ± 2.6	122.4 ± 4.1
DAP, mmHg
Light baseline	83.7 ± 1.2	85.3 ± 1.6	84.7 ± 1.1	83.9 ± 1.0	87.2 ± 1.4
Dark baseline	90.0 ± 1.7	90.2 ± 2.0	91.0 ± 1.6	89.8 ± 1.3	89.2 ± 1.4
Heart rate, beats/min
Light baseline	318.5 ± 8.1	321.8 ± 4.2	312.1 ± 5.0	311.2 ± 5.5	317.4 ± 6.8
Dark baseline	372.0 ± 10.0	375.2 ± 8.0	367.4 ± 9.2	364.7 ± 8.0	357.3 ± 11.9
Body temperature, °C
Light baseline	37.1 ± 0.0	37.2 ± 0.1	37.1 ± 0.0	37.2 ± 0.1	37.4 ± 0.1
Dark baseline	37.9 ± 0.01	37.9 ± 0.1	37.9 ± 0.1	38.0 ± 0.1	37.8 ± 0.1

Values are means ± SE; n = number of rats. Average mean, systolic, and diastolic arterial pressure (MAP, SAP, and DAP, respectively), heart rate, and body temperature during the light phase (8:00 am to 4:00 pm) and dark phase (7:00 pm to 7:00 am) during a 5-day baseline period before the 7-day chronic intermittent hypoxia (CIH) protocol. Control-injected rats were exposed to normoxia (CTRL NORM) or hypoxia (CTRL CIH); caspase-injected rats were exposed to normoxia (CASP NORM) or hypoxia (CASP CIH) or were exposed to hypoxia but had unsuccessful microinjections (CASP CIH MISS).

During CIH exposure from 8:00 am to 4:00 pm, the average daily changes in MAP varied significantly among the treatment groups [Fig. 5A, F(4,32) = 5.293, P = 0.002, 2-way RM ANOVA]. During the normoxic dark phase (6:00 pm to 6:00 am), significant differences were also detected between treatment groups [F(4,32) = 7.313, P < 0.001, 2-way RM ANOVA]. There was also a statistically significant interaction between day of exposure and treatment during the normoxic dark phase [Fig. 5A, F(28,295) = 1.713, P = 0.018, 2-way RM ANOVA].

Fig. 5.

Chronic intermittent hypoxia (CIH)-induced diurnal hypertension was significantly attenuated with caspase lesions of paraventricular nucleus-projecting median preoptic nucleus neurons. A: average daily change in mean arterial pressure (MAP) from baseline period recorded during CIH exposure from 8:00 am to 4:00 pm (left) and during the normoxic dark phase (right). Caspase lesions attenuated the increase in MAP during both CIH and the dark period [***compared with control-injected + CIH (CTRL CIH) and missed caspase-injected + CIH (CASP CIH MISS); **compared with normoxic controls]. B: daily average change in MAP for each group during CIH (open bars) and the normoxic dark phase (gray bars). Rats exposed to CIH showed significant increase in MAP except for those with caspase lesions (***compared with normoxic controls and CASP CIH during CIH; **compared with normoxic controls during CIH; #compared with normoxic controls and CASP CIH during the dark period). C: average daily changes in systolic arterial pressure (SAP) for the same groups. The results are the same as for MAP (***compared with normoxic controls and CASP CIH during CIH; **compared with normoxic controls during CIH; #compared with normoxic controls and CASP CIH during the dark period). D: average changes in diastolic arterial pressure (DAP) demonstrated results similar to those for MAP (***compared with normoxic controls and CASP CIH during CIH; **compared with normoxic controls during CIH; #compared with normoxic controls and CASP CIH during the dark period). E: average daily changes in heart rate (HR) observed during CIH and during the normoxic dark phase (**compared with normoxic controls during CIH; ‡compared with normoxic controls during the dark period). Groups: CTRL NORM, control-injected + normoxia (n = 9 rats); CTRL CIH, control-injected + CIH (n = 8); CASP NORM, caspase-injected + normoxia (n = 6); CASP CIH, caspase-injected + CIH (n = 9); CASP CIH MISS, missed caspase-injected + CIH (n = 6). Here, bpm, beats/min. Two-way mixed effects ANOVA and Student-Newman-Keuls tests; significance is detected at P < 0.05. Data are expressed as means ± SE.

When the daily changes in MAP were averaged over the 7 days of CIH, this pattern was clearer (Fig. 5B). During the CIH period, rats exposed to CIH and injected with the control virus (CTRL CIH) had significantly increased MAP compared with normoxic controls [vs. CTRL NORM, P = 0.013; vs. caspase-injected rats exposed to normoxia (CASP NORM), P = 0.009; SNK; Fig. 5B]. This is also true for the miss group exposed to CIH (CASP CIH MISS) compared with normoxic controls (vs. CTRL NORM, P = 0.020; vs. CASP NORM, P = 0.026; SNK).The rats exposed to CIH with successful caspase lesions of MnPO (CASP CIH) display increases in MAP that were significantly higher than the normoxic control groups but significantly lower than the other two groups exposed to CIH. Together, these results indicate that caspase lesions of PVN-projecting MnPO neurons attenuated the increase in blood pressure that occurs during intermittent hypoxia.

In contrast, caspase lesions of PVN-projecting MnPO neurons appeared to block the increase in MAP associated with CIH during the normoxic dark period. During the dark period (6:00 pm to 6:00 am), the average change in MAP recorded over the 7-day CIH protocol was significantly elevated in the CTRL CIH group (vs. CTRL NORM, P = 0.001; vs. CASP NORM, P = 0.005; SNK) and CASP CIH MISS (vs. CTRL NORM, P = 0.013; vs. CASP NORM, P = 0.018; SNK) compared with normoxic controls. With successful caspase injections, however, CIH-exposed rats did not display a significant increase in MAP (CASP CIH vs. CTRL CIH, P = 0.002; vs. CASP CIH MISS, P = 0.023; SNK) and maintained MAP comparable to that of normoxic controls (CASP CIH vs. CTRL NORM, P = 0.915; vs. CASP NORM, P = 0.915; SNK). These results indicate that diurnal increases in MAP produced by CIH in the control virus-injected group and the caspase miss group were significantly blocked by successful caspase lesions in the MnPO neurons that project to the PVN.

Exposure to CIH and treatment significantly affected SAP (Fig. 5C) and DAP (Fig. 5D) with the same pattern as seen in the results for MAP. Significant differences were found between groups for changes in SAP during the CIH period [F(4,32) = 4.500, P = 0.005, 2-way RM ANOVA] and dark period [F(4,32) = 5.834, P = 0.001, 2-way RM ANOVA]. Results were consistent with those related to changes in MAP during the CIH period (CTRL CIH vs. CTRL NORM, P = 0.032; vs. CASP NORM, P = 0.011; CASP CIH MISS vs. CASP NORM, P = 0.026; SNK) and dark period (CTRL CIH vs. CTRL NORM, P = 0.005; vs. CASP NORM, P = 0.010; vs. CASP CIH, P = 0.006; SNK). Increases in SAP were maintained into the dark period only by the CTRL CIH and CASP CIH MISS groups, and these increases were significantly higher compared with normoxic controls and CASP CIH (P < 0.001 for each group, SNK).

Significant differences were found between groups for changes in DAP during the CIH period [F(4,32) = 3.737, P = 0.013, 2-way RM ANOVA] and dark period [F(4,32) = 4.513, P = 0.005, 2-way RM ANOVA]. This was due to significant differences in DAP changes during the dark period (CTRL CIH vs. CTRL NORM, P = 0.017; vs. CASP NORM, P = 0.044; vs. CASP CIH, P = 0.024; SNK). When analyzing overall average change in DAP, there was a significant effect of treatment between groups during the CIH period [F(4,32) = 42.035, P < 0.001, 1-way ANOVA] and normoxic dark period [F(4,32) = 48.214, P < 0.001, 1-way ANOVA]. These results were consistent with results from SAP. This was due to significant increases in DAP during CIH exposure (CTRL CIH, CASP CIH, CASP CIH MISS, P < 0.010 for each group compared with normoxic controls, SNK). Although DAP was significantly increased compared with controls, this increase was significantly decreased in CASP CIH versus CTRL CIH and versus CASP CIH MISS (P < 0.001, SNK). Increases in DAP were maintained into the dark period only by the CTRL CIH and CASP CIH MISS groups compared with normoxic controls and CASP CIH (P < 0.001 for each group, SNK).

No significant differences were found overall among any of the groups for daily changes in HR (Fig. 5E) during CIH [F(4,32) = 0.963, P = 0.441, 2-way RM ANOVA] or during the normoxic dark period [F(4,32) = 0.531, P = 0.714, 2-way RM ANOVA]. However, there were significant differences in averages measuring overall changes in HR during the CIH period [F(4,32) = 19.026, P < 0.001, 1-way ANOVA] and normoxic dark period [F(4,32) = 11.723, P < 0.001, 1-way ANOVA]. This was due to CIH exposure and was influenced by the caspase lesions (Fig. 5E).

CIH Causes Dysregulation in Core Body Temperature During the Normoxic Dark Period

No significant differences were found overall among any of the groups for daily changes in BT (Fig. 6) during CIH [F(4,32) = 0.398, P = 0.809, 2-way RM ANOVA] or during the normoxic dark period [F(4,32) = 0.757, P = 0.561, 2-way RM ANOVA]. However, there were significant differences in averages measuring overall changes in BT during the normoxic dark period [F(4,32) = 22.392, P < 0.001, 1-way ANOVA]. This was due to CIH exposure and was not dependent on whether groups were injected with the control or caspase virus (Fig. 6).

Fig. 6.

Chronic intermittent hypoxia (CIH) significantly affected changes in body temperature (BT). BT was decreased during CIH exposure (open bars) and increased during the normoxic dark period (gray bars) compared with normoxic controls. Groups: control-injected rats exposed to normoxia (CTRL NORM; n = 9 rats) or hypoxia (CTRL CIH; n = 8) and caspase-injected rats exposed to normoxia (CASP NORM; n = 6), hypoxia (CASP CIH; n = 9), or misses (CASP CIH MISS; n = 6). Data are expressed as means ± SE. ‡Significantly different from CTRL NORM (1-way ANOVA and Student-Newman-Keuls tests, P < 0.05).

CIH Exposure Significantly Increases Oxidative Stress, but This Effect is Blocked by Caspase Lesions

As previously reported (51), 7 days of CIH exposure significantly increased circulating AOPP [μM, Fig. 7, F(4,35) = 8.337, P < 0.001, 1-way ANOVA]. However, this was only the case for CIH-exposed rats injected with the control virus (CTRL CIH vs. CTRL NORM, P = 0.005; vs. CASP NORM, P = 0.005; SNK) and those that did not have successful caspase lesions (CASP CIH MISS vs. CTRL NORM, P = 0.002; vs. CASP NORM, P = 0.001; SNK). The increase in AOPP due to CIH exposure was blocked in the CIH-exposed caspase lesion group (CASP CIH vs. CTRL CIH, P = 0.003; vs. CASP CIH MISS, P < 0.001; SNK) to levels comparable to those of normoxic controls (CASP CIH vs. CTRL NORM, P = 0.681; vs. CASP NORM, P = 0.655; SNK).

Fig. 7.

Chronic intermittent hypoxia (CIH)-induced increases in plasma oxidative protein concentration were blocked with caspase lesions of paraventricular nucleus-projecting median preoptic nucleus neurons. Plasma advanced oxidative protein product (AOPP) concentrations at time of euthanize. Groups: control-injected rats exposed to normoxia (CTRL NORM) or hypoxia (CTRL CIH) and caspase-injected rats exposed to normoxia (CASP NORM), hypoxia (CASP CIH), or misses (CASP CIH MISS). Data are expressed as means ± SE. *Significantly different from CASP CIH and the normoxic control groups (1-way ANOVA and Student-Newman-Keuls tests, all P < 0.05; n = 6–9 rats).

Caspase Lesions Influence FosB Staining

In the SFO, FosB staining associated with CIH was not influenced by caspase lesion of MnPO-PVN neurons (Fig. 8D and Supplemental Fig. S1; Supplemental Material is available at https://doi.org/10.6084/m9.figshare.11343854). Following CIH, FosB staining in the SFO control and caspase virus-injected CIH rats (CTRL CIH and CASP CIH, respectively) increased compared with normoxic control groups [F(3,20) = 10.369, P < 0.001, 1-way ANOVA]. There was a trend for an effect of CIH on FosB staining in the OVLT [Fig. 8C and Supplemental Fig. S1; F(3,20) = 3.301, P = 0.041, 1-way ANOVA]; however, SNK did not detect significance between groups.

Fig. 8.

Caspase lesions of paraventricular nucleus-projecting median preoptic nucleus (MnPO) neurons block chronic intermittent hypoxia (CIH)-induced increases in FosB expression. A: representative images of FosB staining in the dorsal MnPO (dMnPO) of control-injected rats exposed to normoxia (CTRL NORM; n = 6 rats) or hypoxia (CTRL CIH; n = 6) and caspase-injected rats exposed to normoxia (CASP NORM; n = 6) or hypoxia (CASP CIH; n = 6). Scale bars, 100 µM. B: average FosB-positive nuclei in the MnPO. C–F: average FosB-positive nuclei in regions upstream [organum vasculosum of the lamina terminalis (OVLT, C) and subfornical organ (SFO, D)] and downstream [supraoptic nucleus (SON, E) and rostral ventral lateral medulla (RVLM, F)] of the MnPO. Data are expressed as means ± SE. **Significantly different from CTRL NORM and CASP NORM (1-way ANOVA and Student-Newman-Keuls tests, all P < 0.05). ***Significantly different from all other groups (1-way ANOVA and Student-Newman-Keuls tests, all P < 0.05).

Caspase lesions of MnPO influenced FosB staining associated with CIH in a region-specific manner (Fig. 8B). In the MnPO, CIH significantly increased FosB staining in rats injected with the control vector, and this increase was blocked by the caspase lesions [Fig. 8A, F(3,20) = 5.606, P = 0.006, 1-way ANOVA]. FosB staining in the MnPO of the CTRL CIH group was significantly increased compared with the other groups (Fig. 8B; vs. CTRL NORM, P = 0.007; vs. CASP NORM, P = 0.008; vs. CASP CIH, P = 0.018, all SNK). There were no significant differences between the CASP CIH group compared with either normoxic control group (vs. CTRL NORM, P = 0.829; vs. CASP NORM, P = 0.728). There was no effect of CIH or virus on FosB expression in the SON [Fig. 8E and Supplemental Fig. S1; F(3,20) = 1.497, P = 0.246, 1-way ANOVA].

Caspase lesions attenuated FosB expression in the PVN associated with CIH. In the PVN, FosB staining was significantly affected by CIH and caspase lesions [Fig. 9, A and B; F(3,20) = 9.88, P < 0.001, 1-way ANOVA]. The CTRL CIH group had significantly elevated FosB expression in the PVN compared with the CTRL NORM (P = 0.010, SNK), CASP NORM (P < 0.001, SNK), and CASP CIH (P = 0.001, SNK) groups. Subregions of the PVN were also analyzed (Fig. 8, C–G). There was a significant effect of CIH and virus on FosB expression in the medial parvocellular (MP) region [Fig. 9E; F(3,20) = 13.229, P < 0.001, 1-way ANOVA] and the lateral parvocellular (lp) region [Fig. 9G; F(3,20) = 3.429, P = 0.037]. The CTRL CIH group had significantly elevated FosB expression in both the MP of the PVN compared with the CTRL NORM (P < 0.001, SNK), CASP NORM (P < 0.001, SNK), and CASP CIH (P = 0.002, SNK) groups and the lp of the PVN compared with the CTRL NORM (P = 0.047, SNK), CASP NORM (P = 0.036, SNK), and CASP CIH (P = 0.047, SNK) groups.

Fig. 9.

Caspase lesions block chronic intermittent hypoxia (CIH)-induced increases in FosB expression in the paraventricular nucleus (PVN) and its autonomic-regulating subregions. A: representative images of FosB staining in the PVN of control-injected rats exposed to normoxia (CTRL NORM; n = 6 rats) or hypoxia (CTRL CIH; n = 6) and caspase-injected rats exposed to normoxia (CASP NORM; n = 6) or hypoxia (CASP CIH; n = 6). Scale bars, 100 µM. B–G: average total FosB-positive nuclei in the PVN (B) and its subregions, the posterior magnocellular (PM, C), dorsal parvocellular (DP, D), medial parvocellular (MP, E), ventral lateral parvocellular (vlp, F), and lateral parvocellular (lp, G) regions. 3V, third ventricle. Data are expressed as means ± SE. ***Significantly different from all other groups (1-way ANOVA and Student-Newman-Keuls tests, all P < 0.05).

The RVLM, a major sympathetic premotor region, was also analyzed for FosB staining. CIH and caspase lesions also had a significant effect on FosB expression in the RVLM [Fig. 8F and Supplemental Fig. S1, F(3,20) = 9.691, P < 0.001, 1-way ANOVA]. FosB expression was significantly elevated in the CTRL CIH group in the RVLM compared with the CTRL NORM (P = 0.001, SNK) and CASP NORM (P < 0.001, SNK) groups. This increase was again blocked by caspase lesions of MnPO in the CASP CIH (P = 0.005, SNK) group. Rats that did not have successful injections into the MnPO or PVN were excluded from anatomical studies.

DISCUSSION

The present studies tested the role of PVN-projecting MnPO neurons in the sustained component of CIH-induced hypertension. Previous studies conducted by our laboratory and others demonstrate that the MnPO is necessary in the development of neurogenic hypertension (9, 15, 31, 38, 39, 46). The PVN has also been shown to be important for the expression of CIH hypertension (11, 45). However, studies had not yet been conducted to demonstrate the contribution of the MnPO-to-PVN projection in the development of hypertension associated with CIH. Caspase-3 was therefore used to selectively lesion the PVN-projecting MnPO neurons and effectively blocked the sustained component of CIH hypertension. These lesions also prevented the increase in circulating oxidative by-products. Caspase-3 was the preferred viral construct over diphtheria toxin A or truncated BH3-interacting domain death agonist (tBid) because of its higher efficacy in committing cells to apoptosis and killing adult neurons in vivo, while minimizing toxicity to adjacent non-Cre-expressing cells (56). The lack of an apparent effect of GFAP staining would appear to support the specificity of this approach. It should be noted that this does not rule out an interaction between neurons and glia at the level of the MnPO that could contribute to CIH hypertension.

In the present study, 90% of PVN-projecting MnPO neurons were glutamatergic indicating they are involved in excitatory signaling. These results are consistent with those obtained by Stocker and Toney, which show that excitatory neurons that project from the MnPO to the PVN are involved in regulating blood pressure in response to osmotic challenges, ANG II, and baroreceptor input in the rat (52). Additionally, ~44% of PVN-projecting MnPO neurons were positive for NOS1. This is important as previous studies using PCR array analysis indicate that CIH exposure significantly increases NOS1 gene expression in the MnPO (9). Therefore, it is significant in this context that cell-autonomous apoptosis of MnPO neurons projecting to the PVN caused vGLUT2-positive MnPO neurons to be significantly reduced by ~75% and NOS1-expressing neurons to be significantly reduced by 30%.

Additionally, our laboratory has previously shown that using short hairpin RNA (shRNA) knockdown of AT_1aR in the MnPO during CIH exposure also blocks sustained hypertension (46). From these experiments, we found using retrograde labeling from the PVN to the MnPO that ~20% of PVN-projecting MnPO neurons were positive for the AT_1aR with in situ hybridization (46). We found in preliminary studies that caspase-induced apoptosis of PVN-projecting MnPO neurons reduced AT_1aR expression by ~90% (A. B. Marciante, L. A. Wang, and J. T. Cunningham, unpublished data). Therefore, we speculate that excitatory signaling, NOS1 expression, functional AT_1aRs, and a role for a brain renin-angiotensin system (RAS) in the MnPO may be key components in signaling to the PVN necessary for CIH hypertension and increased oxidative stress.

The 7-day CIH protocol used in previous studies (9, 15, 26) significantly increased FosB-positive nuclei in the SFO, MnPO, PVN, and RVLM. However, rats in the CASP CIH group had attenuated FosB staining comparable to that in the normoxic control groups in the MnPO. Similar effects were seen in the PVN and RVLM. Because of the caspase lesions of the PVN-projecting MnPO neurons, it is not surprising that FosB staining from CIH exposure would be significantly reduced in the MnPO and PVN. However, in the caspase-lesioned group exposed to CIH there was a significant reduction of FosB staining in the RVLM. This could be due to an interruption of signaling from the MnPO to the PVN that may influence the activity of presympathetic PVN neurons that project to sympathetic RVLM neurons consistent with the reduction in the pressor response to CIH. These observations are supported by the reduction in blood pressure in the present study, which is consistent with observations made by Menani et al., who suggested that the AV3V may serve as an “integrator” for autonomic control (34). In this model of CIH, the increase in blood pressure has been linked to the sympathetic nervous system on the basis of the results of experiments using ganglionic blockade (45). However, sympathetic activity was not directly measured in this study. In the present study, CIH was associated with statistically significant increases in both SAP and DAP, which can be indicative of elevated vascular tone. Lesions of PVN-projecting MnPO neurons prevented the increases in both SAP and DAP. The changes in FosB staining that are correlated to the reduction in the CIH pressor response also could be related to decreased sympathetic outflow, but additional experiments are required to directly test this hypothesis.

The RVLM contains a heterogeneous population of neurons including respiratory neurons that were likely affected by CIH (4, 21). Respiratory rate has been shown to be increased during CIH (15) and is potentially reflected in the FosB data shown for the CTRL CIH group, but not for the CASP CIH group. Since RVLM neurons were not phenotyped, we are only able to speculate; however, the role of forebrain structures in modulating respiratory centers is an interesting concept that should be explored in future studies.

CIH exposure significantly increased MAP in the CTRL CIH and CASP CIH MISS groups compared with the normoxic controls. Interestingly, this significant increase was not observed for the CASP CIH group on a day-to-day basis during the CIH exposure period, indicating that PVN-projecting MnPO neurons contribute to CIH hypertension during both the intermittent hypoxia and the sustained, normoxic phase. SAP and DAP were also analyzed to better assess cardiovascular health associated with changes in CIH (36a). This pattern was observed for SAP and DAP as well. Changes in HR appeared to be dependent on CIH rather than viral injection, and this is consistent with previous studies conducted in our laboratory (9, 15, 46).

The CIH-induced AOPP increase was blocked in the CASP CIH group and comparable to that of normoxic controls. These results suggest that oxidative stress associated with CIH hypertension was also mitigated by lesioning PVN-projecting MnPO neurons. We have previously observed that other manipulations attenuate CIH hypertension that are also associated with reduced AOPP (46). The hypoxemic environment enhances carotid body sensory activity, and the oxidative stress generated from the hypoxia-reoxygenation cycles causes carotid reflex sensitization and then mediates the effects of ANG II (6, 13). The high circulating concentration of plasma ANG II is thought to further increase oxidative stress in the carotid body fueling a viscous feed-forward redox stimulation that contributes to hypertension and other cardiovascular diseases (13, 14, 40).

In the context of these studies, we speculate that lesioning excitatory MnPO neurons (possibly the AT_1aR-positive ones) that project to the PVN interrupts signaling to presympathetic PVN neurons, therefore reducing the drive to autonomic regulatory regions in the hindbrain, which is consistent with the FosB data. This decreased signaling to autonomic regulatory regions in the hindbrain could be resulting from upstream lesions preventing the increase in sympathetic outflow and reduced RAS activation. Reduced RAS activation provides, then, a potential answer as to why we see the marked decrease in AOPP in the lesioned group exposed to CIH.

Additionally, increases in AOPP in this model could be related to factors that contribute to CIH hypertension or be caused by the RAS and may be potentiated by sustained hypertension or peripheral ANG II. Because we are blocking the sustained component of CIH-induced hypertension by lesioning the PVN-projecting MnPO neurons, the attenuated blood pressure could possibly be contributing to the decrease in plasma AOPP. Additional studies are needed to determine the relationship between CIH hypertension and elevated AOPP.

Interestingly, we observed CIH-mediated effects on thermoregulation that were not influenced by lesions. Although glutamatergic neurons in the MnPO have been shown to be involved in thermoregulation (1, 33), these may be a different population from those projecting to the presympathetic neurons of the PVN (2). Future studies should be focused on understanding how CIH influences body temperature possibly in the context of changes in metabolism.

When verifying injection sites, GFP-positive neurons were localized largely in the PVN. Although there appeared to be GFP-positive neurons in surrounding areas of the PVN, findings from the heat map analyses suggest that this was dorsal to the PVN in the nucleus reuniens of the thalamus and likely due to retracting the microinjector during surgery. Although studies suggest that the nucleus reuniens of the thalamus may modulate the PVN-RVLM pathway, we used a retrograde virus, eliminating interference (36). Additionally, studies have indicated that there is no projection from the nucleus reuniens of the thalamus to the MnPO (43). Therefore, we are confident in our assessment that these results are specific to MnPO-PVN neuronal knockdown.

OSA is a disease that can go unnoticed in patients for extensive lengths of time, until secondary symptoms become more prominent. Despite efforts in studying OSA, the pathogenesis of this disease is still not fully understood. Early in OSA pathophysiology, hypertension and oxidative stress can be identified (13). The 7-day CIH model used in these studies allows us to examine early adaptations and events that could contribute to the development of neurogenic hypertension. Previous studies have identified several mechanisms that could contribute to CIH hypertension, including long-term facilitation, chemoreflex sensitivity, baroreflex desensitization, and the peripheral RAS (20, 32, 35, 41, 47, 48). Other recent studies suggest that the brain RAS may also contribute to increased sympathetic outflow (15, 27, 47). This study demonstrates for the first time that the MnPO neurons that project to the PVN contribute to the sustained component of CIH-induced hypertension. The results also show that these MnPO neurons are primarily glutamatergic and some express NO synthase. Earlier work has linked the MnPO to the brain RAS in this form of hypertension. Virally mediated knockdown of AT_1aRs in the SFO attenuated CIH hypertension and decreased FosB staining in the MnPO. Similar results are observed when the same virally mediated knockdown strategy is applied to AT_1aRs (46) or angiotensin-converting enzyme (15) in the MnPO. Together, these results suggest the following mechanism for the sustained component of CIH hypertension. Chemoreceptor stimulation and elevated sympathetic tone would increase circulating ANG II, acting at the SFO, and activate ANG II receptors in the MnPO. This would result in the activation of an excitatory pathway from the MnPO to the PVN, which is necessary for maintaining CIH hypertension during normoxia. Eliminating PVN-projecting MnPO neurons interrupts this pathway attenuating the pressor response to intermittent hypoxia and allowing blood pressure to return to normal during normoxia. Additional experiments are necessary to determine how changes in gene expression of AT_1aRs and angiotensin-converting enzyme are related to function of the MnPO-to-PVN pathway and its contribution to CIH hypertension.

GRANTS

This work was supported by National Institutes of Health Grants P01-HL-088052 and T32-AG-020494.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

A.B.M. and J.T.C. conceived and designed research; A.B.M., L.A.W., and J.T.L. performed experiments; A.B.M., L.A.W., J.T.L., and J.T.C. analyzed data; A.B.M., L.A.W., and J.T.C. interpreted results of experiments; A.B.M. and L.A.W. prepared figures; A.B.M. and J.T.C. drafted manuscript; A.B.M., L.A.W., J.T.L., and J.T.C. edited and revised manuscript; A.B.M., L.A.W., J.T.L., and J.T.C. approved final version of manuscript.