A total of 23 human brains samples from autopsies were collected at the Tulane Medical Center between May 2011 and August 2015. Brain tissues were collected within 5 h of death. The PVN and RVLM brain regions were confirmed by microscopic examination using hematoxylin and eosin (H&E) staining, with reference to a human brain atlas (Ding et al., 2016). For the RVLM, we used an anti-tyrosine hydroxylase (TH) antibody to identify the C1 group of catecholaminergic neurons as an additional landmark. In the final analysis, the PVN could be identified in 11 of the 23 brain specimens (four NTNs, seven HTNs), and the RLVM could be identified in nine specimens (four NTNs, five HTNs). An example of human PVN identification is presented in Figure 1, which shows a schematic of the human PVN (Figure 1A) with H&E staining and immunohistochemical labeling of the PRR using 3,3'-diaminobenzidine (DAB) as a substrate (Figures 1B,C). Figure 2 shows a schematic illustration of the human RVLM (Figure 2A), with H&E staining and immunohistochemical labeling of the PRR and TH (Figures 2B–D). Brain tissues with no identifiable PVN or RVLM were excluded from our final analysis. All clinical data, including patient history and diagnosis, were collected, documented, and de-identified. BP used for analyses, recorded in study participants during standard care, was obtained from patient charts. The designations of hypertension and normotension used in this study are based on the individual clinical diagnosis of each human subject according to the retrospective clinical record. The next of kin consented to all autopsies and tissue collection through a standard medical autopsy protocol. This study was approved by the Institutional Review Board and the Research Integrity Offices at the University of Nevada, Reno, and complies with human research project oversight by the Institutional Review Board.

Human brain tissues containing the RVLM and PVN were dissected, fixed in 10% formalin, and embedded in paraffin. Paraffin-embedded tissues were sectioned at 5 μm thickness and mounted onto slides. Slide-mounted sections were rehydrated by two washes with xylene and two washes with 100% ethanol (10 min each) followed by washes with a graded ethanol series (95 to 80% and 80 to 70%, each for 3 min). After washing with phosphate-buffered saline (PBS; Fisher Scientific, Fair Lawn, NJ, United States) for 10 min, slides were incubated in hematoxylin for 1 min and then in eosin for 30 s; after each staining step, slides were washed in running tap water until the water ran clear. Finally, slides were dehydrated by washing with a graded ethanol series (50 to 70%, 70 to 80%, 80 to 95%, and 95 to 100%, each for 1 min).

Bright-field stitched images were captured using an optical microscope (model BZ-X710; Keyence, Elmwood Park, NJ, United States). The entire section was imaged so as to accurately identify PVN and RVLM regions by comparison with a human brain atlas (Ding et al., 2016).

Paraffin-embedded PVN and RVLM tissues were sectioned at 5 μm thickness and mounted onto slides, as described previously (Cooper et al., 2018). Briefly, the tissue was deparaffinized by heating slides in an oven at 60°C for 1 h. Sections were rehydrated by washing twice with xylene for 10 min each, followed by washing with a graded ethanol series (100 to 95% to 70 to 50 to 30%, 2 min each), and then with distilled water (dH2O) for 2 min. Antigen retrieval was performed by steaming slides in sodium citrate buffer (10 mM, pH 6.0) for 45 min. Non-specific binding was blocked by incubating with 10% goat serum for 2 h at room temperature. PVN containing sections were incubated with rabbit anti-PRR primary antibody (ab264763; Abcam, Cambridge, United Kingdom, antibody registry ID: RRID: AB_287695 at http://antibodyregistry.org), diluted 1:100 in PBS containing 0.3% Triton X-100 (PBST). PVN sections were additionally incubated overnight with mouse anti-HuC/HuD monoclonal antibody (A21271; Thermo Fisher Scientific, Waltham, MA, United States), diluted to 1:50 in PBST or chicken anti-glial fibrillary acidic protein (GFAP; ab4674; Abcam), and diluted to 1:1,000 in PBST under the same conditions. Thereafter, slides were incubated at room temperature for 2 h with Alexa Fluor 594‐ or Alexa Fluor 488-conjugated goat anti-rabbit (1:1,000), goat anti-mouse (1:1,000) or goat anti-chicken (1:1,000) secondary antibody, as appropriate. After staining with 4',6‐ diamidino-2-phenylindole (DAPI) for 5 min, all sections were incubated in TrueBlack Lipofuscin Autofluorescence Quencher (23007, Biotium Inc., Fermont, CA, United States) for 1 min to reduce the autofluorescence from aged human brain tissue. All sections were coverslip-mounted with Vectashield hard-set mounting medium (Vector Laboratories, Burlingame, CA, United States).

Rostral ventrolateral medulla sections were incubated with rabbit anti-PRR primary antibody (ab264763; Abcam) and diluted to 1:100 in PBST. RVLM sections were concomitantly incubated overnight with mouse anti-HuC/HuD monoclonal antibody (A21271; Thermo Fisher Scientific), diluted to 1:50 dilution in PBST, chicken anti-GFAP (ab4674; Abcam), and diluted to 1:1,000 in PBST. Thereafter, sections were incubated at room temperature for 2 h with Alexa Fluor 594‐ or Alexa Fluor 488-conjugated goat anti-rabbit (1:1,000), goat anti-chicken (1:1,000), or goat anti-mouse (1:1,000) secondary antibody, as appropriate. Sections were further stained with DAPI and TrueBlack following the same steps as described above for PVN sections. The specificity of all commercial antibodies used in this study was validated by negative control sections labeled without addition of primary antibodies, and data are presented in the respective immune-labeling figures. Images were captured using an epifluorescence microscope (model BZ-X710; Keyence).

Paraventricular nucleus of the hypothalamus and RVLM tissues were fixed, paraffin-embedded, sectioned, and rehydrated, followed by antigen retrieval using Na-citrate buffer, as described above. Slides were then washed once with PBS for 5 min, placed in 3% H2O2 for 5 min, and washed in PBS for 5 min. Blockade of nonspecific binding and permeabilization was performed simultaneously by incubating with 10% goat serum in PBS containing 0.3% Triton X-100 for 1 h. For RVLM only, sections were incubated for 48 h at 4°C with the same rabbit anti-PRR antibody (diluted 1:50 in PBST containing 2% goat serum) or rabbit anti-TH polyclonal antibody (AB152; Millipore Sigma, Massachusetts, United States), diluted to 1:100 in PBST containing 2% goat serum. After a 10-min wash with PBS, sections were incubated with diluted, biotinylated rabbit IgG secondary antibody, a component of the Vectastain ABC HRP Peroxidase Kit (PK-4001; Vector Laboratories), for 30 min, followed by a 5-min wash in PBS. All slides were incubated at the same time for 2 min in a working solution of DAB Peroxidase Substrate (SK-4100; Vector Laboratories) containing 250 ml of dH2O, 4.2 ml of buffer stock solution, 4 ml of H2O2 solution, and 5 ml of DAB stock solution. The reaction was stopped by rinsing slides with tap water for 5 min, after which slides were coverslip-mounted using a xylenebased hard-set Cytoseal XYL mounting medium (Richard-Allen Scientific, Kalamazoo, MI, United States). All immunoreactions with DAB substrate, including no-primary antibody and pre-immune serum controls, were performed at the same time in the same DAB substrate developing solution.

Bright-field images were captured with an epifluorescence microscope (model BZ-X710; Keyence) using the same settings. Once the PVN or RVLM region was identified in each section, at least three to four images were captured from each section. The relative intensity of each image was analyzed using ImageJ software (version: 2.0.0-rc-68/1.52 h) and is presented as mean intensity in arbitrary units (AUs). Within each subject, the mean intensity of all images was averaged and presented as a single “N” value. All experiments and data analyses were performed independently in a blinded fashion by two investigators.

The significance of differences in sBP, diastolic BP (dBP), age, body mass index (BMI), and PRR immunoreactivity between HTN and NTN subjects was assessed with a two-tailed, unpaired Student’s t-test using Prism 8 software (GraphPad, La Jolla, CA, United States). Pearson correlation and least squares linear regression were used to determine correlation and regression coefficients between variables. Differences were considered statistically significant at p < 0.05. Multiple linear regressions were performed using SAS 9.4. All statistical tests were two-tailed, and α was set at 0.05. Data are reported as means ± SEs.

The characteristics of NTN and HTN study subjects are detailed in Tables 1, 2. All HTN subjects were clinically diagnosed as HTN with a history of taking anti-HTN medication, and all NTN subjects had a record of normal BP with no history of diagnosed hypertension. For PVN samples, study subjects in HTN and NTN groups were similar in terms of age (p = 0.345), body weight (p = 0.428), height (p = 0.313), and BMI (p = 0.899). There were similarly no significant differences in age (p = 0.201), body weight (p = 0.384), height (p = 0.708), or BMI (p = 0.560) among subjects from whom RVLM samples were obtained. HTN subjects from whom PVN samples were acquired had higher sBP (p = 0.001) and an increased heart rate (HR; p = 0.047) compared with NTN subjects. Similarly, HTN subjects from whom RVLM samples were obtained had higher sBP (p = 0.01) than NTN subjects; however, there was no difference in HR between groups. For both RVLM and PVN sample sets, there was no difference in dBP between HTN and NTN subjects.

It is well-established that the PVN is a critical networking center for the angiotensinergic neural circuit in the central regulation of BP that relays excitatory pathways to the RVLM (Chen and Toney, 2003; Li et al., 2003; Cato and Toney, 2005). We recently reported that, in the human SFO, the PRR is expressed in neurons but not in astrocytes (Cooper et al., 2018). Here, we performed dual immunofluorescence labeling of the PVN for the PRR and markers of neurons or astrocytes. Within this region, we found that the PRR was colocalized with the neuronal marker HuC/HuD (Figure 3A) but not with the astrocyte marker GFAP (Figure 3B); no labeling was observed in controls (Figure 3C). These data indicate that the PRR is mostly expressed in neurons and not astrocytes in the human PVN.

The RVLM contains sympathetic neurons that control tonic and phasic arterial pressure and receives hypothalamic afferent inputs from the PVN (Dampney, 1994; Tagawa and Dampney, 1999). Similar to the PRR expression pattern in the PVN, we found that, in the RVLM, the PRR was expressed mostly in neurons (Figure 4A) and not astrocytes (Figure 4B); again, no labeling was observed in controls (Figure 4C).

With aging, human brain tissue develops a tendency to exhibit strong autofluorescence due to lipofuscin accumulation (Goyal, 1982), as we also observed in our human brain samples. Thus, to semi-quantitatively determine PRR expression levels in the PVN and RVLM, we used immunohistochemistry, presenting staining results as relative intensity in AU. Figure 5 (top panel) shows representative images of immunohistochemical labeling of the PRR in the PVN of NTN and HTN subjects, together with negative control staining without primary antibody. PRR immunoreactivity was significantly higher in HTN subjects compared with NTN subjects (p = 0.022), as shown in Figure 5D. To ensure the changes in PRR immunoreactivity in this brain region are specific, we also performed additional analysis of PRR immunoreactivity lateral outside the PVN and the background intensity (Figures 5E,F); we observed no difference between the NTN and HTN subjects. Figures 6A–C show representative images of PRR immunoreactivity in the RVLM of NTN and HTN subjects. We found that HTN subjects exhibited significantly higher PRR immunoreactivity compared with NTN subjects (p = 0.018; Figure 6D). Similarly, we analyzed the PRR immunoreactivity lateral outside the RVLM and the background intensity (Figures 6E,F); we observed no difference between the NTN and HTN subjects. The data indicate elevation in the PRR immunoreactivity in both the PVN and RVLM.

To examine whether PRR immunoreactivity is correlated with BP, we performed a linear regression using Pearson correlation coefficient and least-square correlation post analyses. We found that PRR immunoreactivity in the PVN was significantly correlated with sBP (R ² = 0.6435, p = 0.0030) but not with dBP (R ² = 0.05243, p = 0.4982; Figures 7A,B). Because age, BMI, and sex are well-established factors that affect BP in humans, we also performed multiple regressions to investigate possible associations of BMI or age with PRR immunoreactivity in our study subjects. We found no significant correlation between PRR immunoreactivity in the PVN and BMI or age (Figures 7C,D). Although, there was a trend toward a positive correlation between PRR immunoreactivity and sBP in the RVLM (R ² = 0.4416, p = 0.0509), this difference did not reach statistical significance (Figure 8A). There was no significant correlation between PRR immunoreactivity in the RVLM and dBP, BMI, or age (Figures 8B–D). Although not significant, the linear regression analysis showed a trend toward a positive correlation between PRR and age (Figures 7D, 8D). To further determine whether age, sex, and BMI affect the relationship between PRR and sBP, we performed a multiple regression analysis between PRR and sBP controlling for age, sex, and BMI. As showed in Table 3, PRR level in the PVN remained positively correlated with sBP (p = 0.015) after controlling for these parameters. In addition, there was no significant contribution of age, sex, or BMI to PRR levels in our subjects.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.