No evidence for brain renin-angiotensin system activation during DOCA-salt hypertension.

Brain renin-angiotensin system (RAS) activation is thought to mediate deoxycorticosterone acetate (DOCA)-salt hypertension, an animal model for human primary hyperaldosteronism. Here, we determined whether brainstem angiotensin II is generated from locally synthesized angiotensinogen and mediates DOCA-salt hypertension. To this end, chronic DOCA-salt-hypertensive rats were treated with liver-directed siRNA targeted to angiotensinogen, the angiotensin II type 1 receptor antagonist valsartan, or the mineralocorticoid receptor antagonist spironolactone (n=6-8/group). We quantified circulating angiotensinogen and renin by enzyme-kinetic assay, tissue angiotensinogen by western blotting, and angiotensin metabolites by LC-MS/MS. In rats without DOCA-salt, circulating angiotensin II was detected in all rats, whereas brainstem angiotensin II was detected in 5 out of 7 rats. DOCA-salt increased mean arterial pressure by 19±1 mmHg and suppressed circulating renin and angiotensin II by >90%, while brainstem angiotensin II became undetectable in 5 out of 7 rats (<6 fmol/g). Gene silencing of liver angiotensinogen using siRNA lowered circulating angiotensinogen by 97±0.3%, and made brainstem angiotensin II undetectable in all rats ( P <0.05 vs. non-DOCA-salt), although brainstem angiotensinogen remained intact. As expected for this model, neither siRNA nor valsartan attenuated the hypertensive response to DOCA-salt, whereas spironolactone normalized blood pressure and restored brain angiotensin II together with circulating renin and angiotensin II. In conclusion, despite local synthesis of angiotensinogen in the brain, brain angiotensin II depended on circulating angiotensinogen. That DOCA-salt suppressed circulating and brain angiotensin II in parallel, while spironolactone simultaneously increased brain angiotensin II and lowered blood pressure, indicates that DOCA-salt hypertension is not mediated by brain RAS activation.


Introduction
Hypertension is the leading risk factor for death and disability worldwide. 1 End-organ damage caused by hypertension can be effectively prevented by blocking the reninangiotensin system (RAS), even when circulating RAS activity is low. [2][3][4] The underlying concept is that in these organs angiotensin (Ang) II generation relies on the local synthesis of angiotensinogen (AGT), i.e., occurs independently of AGT synthesis in the liver. 5 Indeed, even though liver-derived AGT is the main -if not the only -source of circulating Ang II 6 , gene expression of Agt also occurs in brain, kidney and adipose tissue. 7,8 Yet when the Agt gene was deleted selectively in hepatocytes, it was found that Ang II generation in kidney and adipose tissue did depend on the uptake of liver-derived AGT from the circulation. 6,9,10 To what degree this is also true for the brain is unknown. The brain RAS is thought to be activated by the combination of deoxycorticosterone acetate and excess dietary salt (DOCA-salt). The DOCA-salt rat represents a model for human primary hyperaldosteronism in which the circulating RAS is suppressed by volumedependent hypertension. That intracerebroventricular, but not intravenous, administration of RAS blockers partially reversed the hypertensive response to DOCA-salt has supported the concept of selective brain RAS activation. [11][12][13] The underlying assumption is that the blood-brain barrier prevents the diffusion of circulating AGT and Ang II into the brain, so that any Ang II in the brain must be derived from locally synthesized AGT. However, it is widely accepted that high blood pressure disrupts the blood-brain barrier, thereby allowing circulating Ang II access to the brain. 14,15 This perspective offers an alternative explanation as to why intracerebroventricular administration of an Ang II type 1 (AT 1 ) receptor blocker (ARB) attenuated hypertension better than its intravenous application. 16 An additional argument against selective brain RAS activation by DOCA-salt is that the brain lacks renin. Originally, it was proposed that intracellular renin -the isoform of renin expressed in the brain 17 -cleaves brain AGT. Yet, surprisingly, its deletion increased blood pressure and made mice more susceptible to Ang II-induced organ damage. 18 Consequently, intracellular renin is now believed to play a protective, RAS-suppressing role, without acting on AGT. 18 Furthermore, brain renin levels in DOCA-salt-hypertensive mice decreased in parallel with circulating renin. 19 In fact, brain renin levels were so low that they likely represented renin in trapped blood. 19 The same is true for prorenin 19 , which opposes Downloaded from http://portlandpress.com/clinsci/article-pdf/doi/10.1042/CS20201239/901276/cs-2020-1239.pdf by guest on 12 January 2021 Clinical Science. This is an Accepted Manuscript. You are encouraged to use the Version of Record that, when published, will replace this version. The most up-to-date-version is available at https://doi.org/10.1042/CS20201239 the concept that prorenin's interaction with the (pro)renin receptor underlies brain RAS activation in DOCA-salt treated animals.
In the current study, we set out to determine the contribution of liver-derived AGT to brain RAS activation in chronic DOCA-salt-hypertensive rats, making use of liver-targeted small interfering RNA (siRNA) to selectively silence hepatic Agt. 20 We focused on the brainstem as a proxy for brain RAS activation, because this region is fully sequestered from the circulation by the blood brain barrier and it contains the highest levels of renin in the brain. 19 To simultaneously delineate the role of blood pressure and AT 1 receptor-mediated uptake of circulating Ang II into the brain, we also evaluated DOCA-salt rats treated with the mineralocorticoid receptor antagonist spironolactone (which normalizes blood pressure in this model), the ARB valsartan, or their combination. We found that, in contrast to AGT in the kidney or adipose tissue, AGT in the brain did not depend on uptake of liver-derived AGT from the circulation. Yet, DOCA-salt suppressed levels of brain Ang II in parallel with those in blood, kidney and adipose tissue. Thus, selective brain RAS activation by DOCA-salt did not occur, while a role for brain-derived AGT still needs to be demonstrated.

Animal studies
All studies were performed at the Erasmus MC (Rotterdam, The Netherlands) under the regulation and permission of the Animal Care Committee of the Erasmus MC (protocol number 17-870-01). Male, 14-week old Sprague Dawley rats (Janvier Labs, France) were housed in individual cages and maintained on a 12-h light-dark cycle with access to standard rat chow and tap water ad libitum. Radiotelemetry transmitters (HD-S10, Data Sciences International, St. Paul, USA) were implanted to continuously measure blood pressure, heart rate and activity. 21 Peri-operative analgesia consisted of buprenorphine (0.5 mg/kg s.c.) given 1h prior to surgery and 6 hours after surgery, followed by twice daily injections up until 2 days after surgery. After a two-week recovery period, hypertension was induced by subcutaneous implantation of a DOCA-pellet (200 mg; 60-day release; Innovative Research of America, Sarasota, Fl, USA) and by replacing the drinking water with saline (DOCA-salt).
The siRNA consisted of a chemically modified antisense strand with sequence UUGAUUUUUGCCCAGGAUAGCUC, hybridized with a chemically modified sense strand of sequence GCUAUCCUGGGCAAAAAUCAA. Oligonucleotides were synthesized as previously described. 22 To ensure selective and efficient delivery to hepatocytes, a triantennary Nacetylgalactosamine (GalNAc) -a high-affinity ligand for the hepatocyte-specific asialoglycoprotein receptor -was attached to the 3′ end of the sense strand. 20 Valsartan was dissolved in 1 mol/L sodium hydroxide (Sigma Aldrich; made up in saline). The solution was titrated back to a pH of ~7 with 10% hydrochloric acid (Sigma Aldrich) and delivered by an osmotic mini-pump (Alzet, Cupertino, CA, USA). Spironolactone was dissolved in sesame oil (Sigma Aldrich) and delivered by daily subcutaneous injection. To control for effects of these methods of delivery, all rats underwent a sham pump implantation and received daily injections of equivalent volumes of sesame oil. For biochemical measurements, we collected 24-hour urine in metabolic cages and blood plasma by venipuncture from the lateral tail vein at 3 points in time: prior to DOCA-salt, and after 4 and 7 weeks of DOCA-salt. At the end of the 7-week study period, rats were anaesthetized by inhalation of isoflurane and exsanguinated: 1 mL blood was collected in 10mL of 4 mol/L guanidine thiocyanate (Sigma Aldrich) and used for quantification of angiotensin metabolites; remaining blood was supplemented with EDTA and centrifuged at 16000 x g to obtain plasma. Liver, brainstem, cerebellum, kidneys, heart, epididymal adipose tissue (AT), inguinal AT and brown AT were excised, weighed, and snap frozen in liquid nitrogen for gene expression and protein analysis.

Biochemical measurements
In plasma, AGT was measured by enzyme-kinetic assay as the maximum quantity of Ang I generated during incubation, at pH 7.4 and 37°C, with rat kidney renin in the presence of a mixture of ACE, angiotensinase, and serine protease inhibitors. 23,24 The lower limit of detection (LLOD) of this assay was 0.2 nmol/L. We have previously shown that the results Downloaded from http://portlandpress.com/clinsci/article-pdf/doi/10.1042/CS20201239/901276/cs-2020-1239.pdf by guest on 12 January 2021 produced by this assays correlate strongly with those obtained from AGT ELISA (Immuno-Biological Laboratories Co. Ltd.). 22 Active plasma renin concentration (APRC) was measured by enzyme kinetic assay, by quantifying Ang I generation in the presence of excess porcine AGT (LLOD 0.17 ng Ang I/mL per hour). 23 In the cases that measurements were at or below the LLOD, this limit was applied to allow for statistical analysis. Ang metabolites in plasma, kidney, brainstem, epididymal AT and heart tissue (left ventricle) were measured by LC-MS/MS analysis as described before. 25,26 Briefly, tissue samples were homogenized under liquid nitrogen and extracted with a guanidinium based extraction buffer. Stabilized whole blood and tissue extracts were spiked with stable isotope labeled internal standards for each individual target analyte (Sigma Aldrich) before being subjected to C18 based solid phase extraction and subsequent LC-MS/MS analysis. Table S1 specifies the lower limits of quantification (LLOQ) for each metabolite per tissue. NT-proBNP was measured with a rat ELISA NT-proBNP kit (LLOD 15.6 pg/mL; Aviva Systems Biology, San Diego, USA). Plasma and urinary sodium and urinary potassium (both 24-hour urine) were measured at the clinical chemistry laboratory of the Erasmus MC.

Quantitative polymerase chain reaction (qPCR)
Total RNA was isolated from snap-frozen liver, kidney, brainstem, epididymal AT, inguinal AT and brown AT using TRI Reagent (Sigma Aldrich) and reverse transcribed into cDNA using the QuantiTect Reverse Transcription Kit (Qiagen, Venlo, The Netherlands). cDNA was amplified in triplicate in 40 cycles (denaturation at 95C for 3 min; thermal cycling at 95C for 3 sec, annealing/extension at 60C for 20 sec) followed by a melt curve with a CFX384 (Bio-rad, Veenendaal, The Netherlands) using Kapa SYBR® Fast (Kapa Biosystems). The intron-spanning oligonucleotide primers were designed with NCBI Primer-BLAST ( Table S2).
The Ct method was used for relative quantification of mRNA expression levels, using the geometric mean of β 2 -microglobulin (B2M) and β-actin (ActB) for normalization.

siRNA quantification
siRNA quantification was performed as described previously. 27 Antisense levels in siRNA standard curve dilutions and homogenized liver, kidney and cerebellum samples were quantified by stem loop reverse transcription followed by qPCR. The primer and probe sequences were GTCGTATCCAGTGCAGGGTCCGAGGTATTCGCACTGGATACGACGAGCTATCCT System (Bio-Rad). Membranes were blocked with 5% bovine serum albumin in Tris-buffered saline containing 0.1% Tween-20, followed by incubation overnight at 4C with a primary antibody directed against AGT (1:100; polyclonal antibody raised in rabbits against a synthetic peptide modeled after the C-terminal part of mouse AGT and purified by antigen affinity; Immuno-Biological Laboratories Co. Ltd., Japan; lot #1E-711). After washing, blots were incubated with an anti-rabbit horseradish peroxidase-conjugated secondary antibody (1:3000; Bio-Rad). Signals were detected by chemiluminescence (Clarity Western ECL substrate; Bio-Rad) and quantified using ImageQuant LAS 4000 (GE Healthcare, Diegem, Belgium). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH; 1:5000; GeneTex, Irvine, CA, USA) was used for normalization of protein levels. Using the method described above, AGT antibody target specificity was validated by demonstrating a single band at ~53 kD in tissues obtained from wild-type but not in tissues from Agt knockout mice 28 (tissues were a kind gift from Prof. M. Bader; Figure S1). Specificity was confirmed for rat tissues by demonstrating a single band at ~53 kD in liver homogenates from untreated control rats but not in liver homogenates from AGT siRNA-treated rats ( Figure S1).
For comparison, brainstem AGT in western blot samples was also measured by AGT ELISA (Immuno-Biological Laboratories Co. Ltd.; LLOD 0.32 ng/mL).

Histology
Transverse heart sections fixed in 4% paraformaldehyde were dehydrated and paraffinembedded. Deparaffinized sections (5μm) were stained with Gomori silver to visualize individual cardiomyocytes. Only left-ventricular, transversally cut cardiomyocytes showing a nucleus were analyzed with Qwin (Leica, Cambridge, UK) for surface area quantification.

Myograph studies
Responses of iliac arteries were measured in a Mulvany myograph as changes in isometric force. Iliac arteries were preconstricted with U46619 to construct concentration-response curves to the endothelium-dependent vasodilator acetylcholine, in the absence or presence of the NO synthase inhibitor L-NAME, the small conductance Ca 2 + -activated K + channel inhibitor apamin and the intermediate conductance Ca 2 + -activated K + channel inhibitor TRAM34.

Statistical analyses
Parametric data are expressed as mean values ± SEM and were analyzed using a one-way analysis of variance (ANOVA). Before statistical testing, data on Ang metabolites were logtransformed to conform to normality. Non-parametric data that did not follow a normal distribution after log transformation (APRC) are expressed as median (interquartile range) and were analyzed using a Kruskal-Wallis test. Data obtained at multiple points in time were analyzed using a repeated-measures ANOVA. Post-hoc correction according to Dunnett or were considered statistically significant. All analyses were performed using Prism v8.0 (Graphpad Software Inc., La Jolla, USA)

Spironolactone with or without valsartan reverses DOCA-salt hypertension
Upon induction of DOCA-salt, rats developed polydipsia and polyuria, accompanied by an increase in plasma sodium concentration and urinary sodium and potassium excretion (Table S3). Prior to DOCA-salt, mean arterial pressure (MAP) was 103±1 mm Hg. Over a 4week period, DOCA-salt increased MAP by 19±2 mm Hg and decreased heart rate by 52±3 BPM (P<0.0001 vs. baseline for both; Figure 1A-B). The rise was greater for systolic than for diastolic blood pressure (24±2 vs. 15±2 mm Hg; P<0.0001; data not shown). The hypertensive response to DOCA-salt was not attenuated when DOCA-salt was supplemented with vehicle, AGT siRNA or valsartan for the final 3 weeks (Figure 1A). In contrast, spironolactone with or without valsartan fully normalized MAP (P<0.01 and P<0.05 vs. vehicle, respectively; Figure 1A). When combined with valsartan, the antihypertensive effect of spironolactone occurred faster and tended to be stronger (P=0.12 vs. spironolactone). Reductions in MAP were matched by increases in heart rate ( Figure 1B).
Although AGT siRNA did not lower MAP, heart rate also rose in these rats (P<0.05 vs. vehicle; Figure 1B). None of the treatments, nor DOCA-salt on its own, affected locomotor activity or physiological weight gain ( Figure 1C; Table S3).

Spironolactone with or without valsartan reverses the suppressive effect of DOCA-salt on circulating RAS activity
Prior to DOCA-salt, circulating AGT levels were 698±29 nmol/L and active plasma renin concentrations (APRC) were 11.8±0.8 ng Ang I/mL per hour. Although DOCA-salt suppressed APRC by 98±0.5% (P<0.0001 vs. baseline), circulating AGT remained unaffected (Figure 2A-B), most likely because rat AGT levels approximate the Michaelis-Menten constant for renin.

Extrahepatic Agt gene expression is not affected by liver-targeted AGT siRNA
Agt gene expression in control rats that had not received DOCA-salt was lowest in inguinal adipose tissue (3±1% of liver Agt expression), followed by kidney (7±1%), brown adipose tissue (15±3%), epididymal adipose tissue (27±5%) and brainstem (55±8%; P<0.001 vs liver; Figure 3A). DOCA-salt increased Agt mRNA expression by 1.5-fold in liver and by 2-fold in brown adipose tissue (both P<0.05; Figure 3A). Treatment with liver-directed siRNA for 3 weeks resulted in siRNA levels of 87±17 g/g in the liver and 4±0.4 g/g in the kidney, whereas none could be detected in the brain. Accordingly, liver Agt mRNA was silenced by 98±1%, while expression remained intact in kidney, adipose and brainstem tissue of siRNAtreated rats (Figure 3A). In fact, in the brainstems of DOCA-salt-hypertensive rats given siRNA, Agt mRNA was doubled when compared to rats that had not received DOCA-salt (P<0.05) or only DOCA-salt (P<0.001; Figure 3A).

In contrast to kidney and adipose AGT, brain AGT is liver-independent
We quantified tissue AGT protein to determine whether locally transcribed Agt mRNA was also translated. In rats not given DOCA-salt, the lowest amount of AGT protein was detected in liver, followed by kidney (1.3-fold higher), brainstem (2.1-fold higher), brown adipose tissue (2.2-fold higher), epidydimal adipose tissue (3.2-fold higher) and inguinal adipose tissue (5-fold higher; Figure 3B). DOCA-salt on its own doubled AGT protein in brown adipose tissue (P<0.0001; Figure 3C). In DOCA-salt-treated rats given siRNA, hepatic Agt silencing eliminated AGT protein levels in liver and plasma concomitantly (Figure 3C, Figure   2A). Absence of AGT from the circulation also eliminated AGT protein in kidney and epididymal adipose tissue (P<0.0001 for both), whereas levels were reduced by threequarters in brown adipose tissue and halved in inguinal adipose tissue (P<0.0001 and P<0.01 vs. vehicle, respectively; Figure 3C). Depletion of AGT protein despite retention of mRNA Downloaded from http://portlandpress.com/clinsci/article-pdf/doi/10.1042/CS20201239/901276/cs-2020-1239.pdf by guest on 12 January 2021 expression indicates that AGT levels in these tissues are fully or partially dependent on circulating/liver AGT, consistent with previous observations. 6, 10 In contrast to kidney and adipose tissue, hepatic Agt silencing did not affect brainstem AGT protein ( Figure 3C). Equal brainstem AGT content was confirmed with an ELISA, which detected 7.1±1.1 ng AGT per g protein in DOCA-salt-treated rats given siRNA versus 9.1±1.6 ng AGT per g protein in rats that had not received DOCA-salt (n=7 for both, P=0.42).

DOCA-salt suppresses brain Ang II in parallel with circulating, kidney and adipose Ang II
In DOCA-salt-treated rats, suppression of circulating RAS activity was mirrored in the kidney, where renin expression was 88±6% lower (P<0.001; Figure S3), Ang I was 98±0.5% lower (P<0.001; Table S4) and Ang II was 99±0.3% lower than in rats not given DOCA-salt (P<0.05; Figure 4A). Except for Ang (3)(4)(5)(6)(7)(8), all Ang metabolites could be detected in the kidneys of rats not given DOCA-salt (Table S4). In contrast, only renal Ang I and II were detectable in the majority of DOCA-salt-treated rats given vehicle, and none of the Ang metabolites could be detected in any DOCA-salt-treated rat given AGT siRNA (Table S4). Kidney Ang I became detectable again in ~85% of the DOCA-salt-treated rats given valsartan, spironolactone or their combination, while renal Ang II was detected again in 25% of the rats given valsartan, and in ~85% of the rats given spironolactone with or without valsartan ( Figure 4A).
In the brainstem, only Ang II could be detected, and none of the other metabolites (Table S4). Brainstem Ang II was detected in 71% of the rats not given DOCA-salt, but only in 29% of the DOCA-salt-treated rats given vehicle, and in none of the DOCA-salt-treated rats given siRNA (P<0.05; Figure 4B). In contrast, treatment with valsartan, spironolactone or their combination reconstituted brainstem Ang II in 40-50% of the rats (Figure 4B). Ang II levels in epididymal adipose and heart tissue followed the exact same pattern of elimination and reconstitution as was observed for kidney and brainstem (Figure 4C-D).
DOCA-salt on its own halved the expression of Ang II type 1A receptors (P<0.05) in the brainstem, whereas the expression of Ang II type 1B and Ang II type 2 receptors remained comparable to that of rats that were not given DOCA-salt (Figure S3).

Effects on kidney function, endothelial function and cardiac hypertrophy
Glomerular filtration rate was 1.0±0.04 mL/min per 100 g body weight at baseline, and was not affected by DOCA-salt (P=0.34) or any other treatment (Table S3). Acetylcholine relaxed Downloaded from http://portlandpress.com/clinsci/article-pdf/doi/10.1042/CS20201239/901276/cs-2020-1239.pdf by guest on 12 January 2021 preconstricted iliac arteries of rats not given DOCA-salt by 87±5%, and by 79±5% after seven weeks of DOCA-salt (P=0.31; Figure S4). This indicates that endothelial function was also left intact by DOCA-salt. The development of cardiac hypertrophy in DOCA-salt-treated rats was indicated by greater heart weights (normalized to tibia length; P<0.05) and surface areas of individual cardiomyocytes (P<0.001) than those measured in rats that had not been given DOCA-salt (Figure 5A-B). Cardiac hypertrophy could be prevented by spironolactone combined with valsartan, which lowered heart weight, cardiomyocyte surface area and plasma levels of NT-proBNP -a marker for cardiac dysfunction (P<0.05 vs. vehicle for all; Figure 5A-C). Unexpectedly, AGT siRNA normalized the heart weights (P<0.01 vs. vehicle; Figure 5A), even though MAP had remained high throughout the treatment period ( Figure   1A). However, cardiomyocyte surface area and NT-proBNP were unaffected by AGT siRNA (Figure 5B-C).

Discussion
This study evaluated the concept that DOCA-salt selectively upregulates the brain RAS via locally synthesized AGT. Controversy has existed for decades regarding brain renin. Recently not only intracellular renin and secreted renin (i.e., the regular form of renin as it is released from the kidney), but also prorenin were ruled out as potential contributors to brain angiotensin generation. 18,19 This leaves increased brain AGT production, combined with its cleavage by an as of yet unidentified enzyme, as a potential mechanism for brain RAS upregulation. Making use of siRNA targeting liver Agt, our current data fully confirm the independent existence of AGT in the brain: decreasing circulating (i.e. liver-derived) AGT by 97% did not affect brain AGT levels. In fact, brain AGT levels (expressed per gram protein) were higher than liver AGT levels. However, DOCA-salt did not upregulate brain AGT.
Moreover, DOCA-salt did not increase brain Ang II. Instead, brain Ang II was lowered to values below the detection limit in 70% of the rats given only DOCA-salt, and in 100% of the rats given the combination of DOCA-salt and AGT siRNA. Yet suppression of brain Ang II by siRNA did not attenuate DOCA-salt hypertension (Figure 6). Surprisingly, upregulating circulating RAS activity in DOCA-salt rats using spironolactone, particularly in combination with valsartan, restored brain Ang II levels, despite the fact that these treatments fully reversed the hypertensive response to DOCA-salt. In other words, brain Ang II levels related neither to blood pressure, nor to brain AGT levels, but did parallel the levels of circulating Ang II. The latter was also true for kidney, cardiac and adipose tissue levels of Ang II, and suggests that at all these sites angiotensin generation depends on AGT of hepatic origin.
This raises the question what the function, if any, of locally synthesized AGT is in the brain. Previous studies demonstrated that deleting kidney Agt does not affect kidney angiotensin levels, both under normal and pathological conditions, while deleting hepatic AGT entirely abolishes kidney angiotensin generation. 6,30 Apparently, therefore, kidney angiotensin generation relies on hepatic AGT, and the function of kidney AGT is unknown.
The same now appears to be true for brain AGT, at least in the DOCA-salt model. Yet, brain AGT levels (expressed per gram tissue) were much higher than renal AGT levels. We confirmed that brain AGT truly is AGT by using both an immunoblot approach and a direct ELISA. Previous immunohistochemical studies have localized brain AGT intracellularly. 31 This is remarkable, as AGT is not typically retained but secreted, consistent with our observation that hepatic AGT levels are low. Taken together, a unifying concept might be that brain AGT cannot be converted to yield Ang I, either because there is neither renin, nor prorenin, nor any other enzyme capable of reacting with this substrate in the brain, or because AGT is located at an (intracellular) site where it cannot meet with any of these enzymes. Under either circumstance, the only possible source of Ang II in the brain is circulating Ang II. It may either diffuse into the brain at sites where the blood-brain barrier is disturbed 14,15 , or bind to AT 1 receptors. 32 In support of the latter, ARB treatment greatly diminished the brain/plasma Ang II concentration ratio in spontaneously hypertensive rats. 19 In the current study, DOCA-salt decreased circulating Ang II levels by >90%, and reduced brainstem Ang II levels in 5 out of 7 rats to undetectable levels, making the calculation of a tissue/plasma ratio impossible. Hence, a further reduction after valsartan could not be demonstrated.
Brainstem tissue was chosen for this purpose, since renin levels are highest here, 19 while the rostral ventrolateral medulla is an important cardiovascular control center that is fully sequestered from the circulation by the blood-brain barrier. 33 Additionally, the brainstem has often been used as an indicator of brain RAS activation in the DOCA-salt model. 34,35 Although the paraventricular nucleus has also been studied extensively and is similarly sequestered, it was less suited for our studies because it is known to respond to signals transmitted by binding of circulating Ang II to the subfornical organ 36 -an area that is not shielded by the blood brain barrier. Importantly, none of the other angiotensin metabolites could be demonstrated in brainstem tissue, in contrast with tissues like the heart and kidney, where Ang I is easily detectable. Indirectly, this argues against local Ang II formation from locally generated Ang I. It also argues against a role for angiotensin metabolites other than Ang II, like Ang-(1-7) and Ang III, in the brain.
Partial reversal of DOCA-salt hypertension has been demonstrated when administrating RAS blockers intracerebroventricularly instead of intravenously. [11][12][13] Yet in these studies, circulating and brainstem Ang II levels were not reported and there was severe hypertension, due to the combination of DOCA-salt with uninephrectomy. Such high blood pressures are likely to have disturbed the blood-brain barrier, thus allowing access of multiple RAS components from the circulation. Therefore, it cannot be excluded that the intracerebroventricularly administered RAS blockers were actually interfering with Ang I or Ang II taken up from the circulation. Secondly, several decades ago both AGT and AT 1 receptor antisense oligonucleotides were injected intracerebroventricularly in various hypertension rat models other than the DOCA-salt rat. [37][38][39][40][41] Although these approaches lowered blood pressure, intracardiac administration was equally effective, 40 while direct injection into the paraventricular nucleus (PVN) did not affect blood pressure. 39 Effects were short-lasting and could not be linked to consistent lowering of brain Ang II levels, while AGT suppression was not verified. Taken together, convincing evidence that suppression of brain AGT selectively reduces brain Ang II and thereby lowers blood pressure is still missing.
It has been suggested that brain angiotensin generation occurs in select nuclei only, e.g., the PVN or rostral ventrolateral medulla. If true, brain angiotensin levels would be severely diluted when highly localized angiotensin production sites are evaluated as part of a broader region, thereby reducing interpretability or utility of data. Lombard-Banek et al.
recently applied a novel micro-analytical capillary electrophoresis-coupled mass spectrometry approach to determine Ang II in the PVN. 42 Making use of 0.2 mg tissue, they reported levels of approximately 65 pmol Ang II/g. 43 At such levels, even if ''diluting'' our samples 10.000-fold with non-Ang II-containing brain tissue, we should still have detected brain Ang II given our detection limits (Table S1). In reality, Ang II levels of 65 pmol/g are many orders of magnitude above the levels measured in tissues with an abundance of the peptide, like the adrenal and kidney, and equivalent to the amount of AGT we observed in brain tissue (10 ng/g protein, which corresponds to 20 pmol AGT/g, given that 1 gram of tissue contains 100 g protein). Hence, before concluding that such high regional levels exist, ex-vivo AGT degradation needs to be excluded. Brainstem Ang II levels in the current Downloaded from http://portlandpress.com/clinsci/article-pdf/doi/10.1042/CS20201239/901276/cs-2020-1239.pdf by guest on 12 January 2021 study were detectable in 5 out of 7 normotensive SD rats, and this decreased to 0 out of 7 DOCA-salt-treated rats given siRNA. Had Ang II been independently upregulated in a regional manner after DOCA-salt, the opposite should have been observed. A final argument for brain RAS activation might be AT receptor upregulation independent of changes in Ang II levels. However, the opposite was observed for Ang II type 1A receptors, while no changes occurred in Ang II type 1B or type 2 receptors.
DOCA-salt suppressed the renal RAS by >95%. The RAS is essential to preserve renal function and glomerular filtration. 44 AGT siRNA on top of DOCA-salt lowered but not fully annihilated renal Ang II. This may explain why GFR remained intact. Alternatively, GFR in this model may be less dependent on renal RAS activity.
Spironolactone reversed the blood pressure rise observed after DOCA-salt, implying that this rise is mineralocorticoid receptor-mediated. It also restored RAS activity. Although valsartan (31 mg/kg/day) tended to reduce blood pressure even further on top of spironolactone, this effect was not significantly different from that of spironolactone alone.
Neither valsartan monotherapy nor AGT siRNA attenuated DOCA-salt hypertension, while applying the same or lower dosages robustly lowered blood pressure in spontaneously hypertensive rats. 22 Hence, as has been shown before 11,13 , systemic RAS blockade is of limited or no use in the DOCA-salt model. This comes to no surprise when considering that DOCA-salt suppressed circulating RAS activity by at least 95%. Nevertheless, AGT siRNA did fully reverse cardiac hypertrophy, to the same degree as spironolactone combined with valsartan. Since this occurred independently of a change in blood pressure or NT-proBNP, these data imply that it was due to a reduction of cardiac Ang II. Indeed, like in the brain, siRNA reduced cardiac Ang II to undetectable levels in 7 out of 7 DOCA-salt-treated rats.
Such observations were not made with valsartan, possibly because it did not fully suppress cardiac Ang II, unlike siRNA. Taken together, these findings confirm that cardiac angiotensin generation relies on hepatic AGT, and that liver-targeted AGT siRNA is therefore capable of exerting cardiac-specific effects in a blood pressure-independent manner.
Determining the functional significance of local, tissue-based AGT production may guide the development of tissue-targeted antihypertensive drugs. However, our data indicate that liver-derived AGT is the sole determinant of tissue angiotensin generation in blood, kidney, heart, adipose tissue and brain. We did not find selective brain RAS upregulation in DOCA-salt-hypertensive rats, and eliminating brain Ang II through hepatic Agt silencing did not lower blood pressure in this animal model for human primary hyperaldosteronism. To the best of our knowledge, brain-targeting of AGT siRNA (other than via intracerebroventricular injection) is not yet feasible. Yet this approach might help to identify a role of brain AGT, if any, in the various hypertension models that currently exist. Ideally, these studies will combine detailed measurements of both AGT and Ang II in the brain, and include a comparison with hepatic AGT siRNA.

Clinical perspectives
Background as to why the study was undertaken DOCA-salt hypertension -an animal model for human primary hyperaldosteronism -is thought to be mediated by independent generation of angiotensin II in the brain. If so, the development of brain-targeted antihypertensive drugs may help to treat this condition.

Brief summary of the results
In DOCA-salt-hypertensive rats, elimination of circulating angiotensinogen by liver-specific RNA silencing lowered brain angiotensin II to undetectable levels, even though brain angiotensinogen remained intact. Nonetheless, blood pressure remained high. In contrast, brain angiotensin II could be detected in normotensive control rats and after blood pressure normalization by spironolactone.

Potential significance to human health and disease
These results indicate that angiotensin II in the brain is derived from the circulation, and is not independently generated from locally synthesized angiotensinogen. These results argue against selective brain renin-angiotensin system activation during DOCA-salt hypertension or its human equivalent, primary hyperaldosteronism. Yet a role for brain angiotensinogen might still be identified if it becomes possible to target siRNA to the brain.

Data availability statement
All supporting data are included within the main article and its supplementary files.      Abbreviations: ND, not detectable; LLOQ, lower limit of quantification.  Liver-targeted angiotensinogen (AGT) siRNA eliminated circulating AGT in DOCA-salthypertensive rats and lowered brainstem Ang II to undetectable levels in all rats, even though brainstem AGT was intact. Yet blood pressure remained high. In contrast, brainstem Ang II was detectable in normotensive control rats and in DOCA-salt-treated rats after blood pressure normalization by spironolactone. This indicates that DOCA-salt hypertension is not mediated by brain RAS activation, and that Ang II in the brain is not generated from locally synthesized AGT, but instead must be derived from the circulation.