Although the existence of a brain renin–angiotensin system (RAS) had been proposed five decades ago, we still struggle to understand how it functions. The main reason for this is the virtual lack of renin at brain tissue sites. Moreover, although renin’s substrate, angiotensinogen, appears to be synthesized locally in the brain, brain angiotensin (Ang) II disappeared after selective silencing of hepatic angiotensinogen. This implies that brain Ang generation depends on hepatic angiotensinogen after all. Rodrigues et al. (Clin Sci (Lond) (2021) 135:1353–1367) generated a transgenic mouse model overexpressing full-length rat angiotensinogen in astrocytes, and observed massively elevated brain Ang II levels, increased sympathetic nervous activity and vasopressin, and up-regulated erythropoiesis. Yet, blood pressure and kidney function remained unaltered, and surprisingly no other Ang metabolites occurred in the brain. Circulating renin was suppressed. This commentary critically discusses these findings, concluding that apparently in the brain, overexpressed angiotensinogen can be cleaved by an unidentified non-renin enzyme, yielding Ang II directly, which then binds to Ang receptors, allowing no metabolism by angiotensinases like ACE2 and aminopeptidase A. Future studies should now unravel the identity of this non-renin enzyme, and determine whether it also contributes to Ang II generation at brain tissue sites in wildtype animals. Such studies should also re-evaluate the concept that Ang-(1-7) and Ang III, generated by ACE2 and aminopeptidase A, respectively, have important functions in the brain.
The existence of a brain renin–angiotensin system (RAS) has been proposed five decades ago, after the seminal discovery of angiotensin (Ang) I-generating activity in brain tissue by Ganten et al. . Yet, whether such a system truly exists, and to what degree it functions independently of the circulating RAS, is still being debated. Here it is important to realize that circulating renin and angiotensinogen are far too large to pass the blood–brain barrier, and thus their presence in the brain requires either local synthesis or destruction of this barrier. This may be different for the much smaller Ang II, which itself is additionally capable of destructing the blood–brain barrier . Despite these uncertainties, brain RAS activation has been linked to (neurogenic) hypertension, potentially involving Ang II-mediated up-regulation of the sympathetic nervous system (SNS) and vasopressin release [3–5].
What is generally accepted after 50 years of research, is that brain renin levels are excessively low, and at most represent the small amount of renin present in blood in the brain . Obviously, this will be detected when measuring renin in homogenized, blood-containing brain tissue . This implies that there essentially is no renin in the brain outside the brain vascular compartment. Subsequently, it was proposed that the brain synthesizes prorenin, the inactive precursor of renin, which via binding to the so-called (pro)renin receptor, would be capable of generating Ang I locally . However, in reality, also prorenin could not be detected in brain tissue , and thus observing Ang I generation after infusing exceptionally high concentrations of prorenin into the brain is unlikely to be of physiological relevance. Finally, the concept came up that brain renin is actually an intracellular, non-secreted form of renin (icREN) that occurs exclusively in the brain . This renin isoform is derived from an alternative transcript of the renin gene, lacking the signal peptide and part of the prosegment. If true, this would imply that Ang generation in the brain occurs exclusively intracellularly, since truncated prorenin cannot leave the cell. This scenario, requiring the simultaneous presence of renin, angiotensinogen and ACE at the same time at the same intracellular site, was abandoned after observing that deletion of brain icREN actually increased blood pressure .
The occurrence of angiotensinogen in the brain, including its local synthesis by astrocytes, is less controversial [10,11]. In fact, it is this local presence of angiotensinogen, rather than that of renin, which has fueled the concept of independent Ang generation in the brain. Obviously, if true, this would require the simultaneous presence of some angiotensinogen-converting enzyme which then apparently is not renin, prorenin, or icREN. One condition where the brain RAS is proposed to be up-regulated is the deoxycorticosterone acetate (DOCA)-salt hypertension model [4,5,12]. DOCA-salt treatment will lower circulating RAS activity, and thus a simultaneous up-regulation of the brain RAS would be an indication of a fully independent RAS. Yet, disappointingly, a recent study that used small interfering RNA (siRNA) to selectively silence hepatic angiotensinogen showed that eliminating circulating angiotensinogen also made brain Ang II disappear in the DOCA-salt model . In other words, brain Ang II in this model still relied on hepatic angiotensinogen, and most likely represented Ang II taken up from blood. Nevertheless, even this study showed that brain angiotensinogen was locally produced, and remained unaltered after hepatic angiotensinogen deletion. This raises the question what the function of brain angiotensinogen truly is.
Rodrigues et al. have now taken a new step to address the role of angiotensinogen synthesized in the brain . They generated an elegant transgenic mouse model overexpressing full-length rat angiotensinogen in astrocytes (Figure 1). Remarkably, this approach resulted in a modest rise in brain angiotensinogen levels (by 30–35%) and yet a massive up-regulation (approx. ten-fold) of brain Ang II levels. The latter apparently was permanent, and did not result in an alteration in the expression of either endogenous mouse angiotensinogen or the Ang receptors in the brain. It also permanently up-regulated SNS activity and vasopressin release, evidenced by an enhanced response to hexamethonium and up-regulated copeptin levels in 13–20-week old mice (Figure 1). Moreover, renal renin expression was permanently suppressed, and this was accompanied by lower circulating Ang II levels. As a consequence, cardiovascular hemodynamics and renal function were unchanged. Lastly, brain Ang II up-regulation increased erythropoiesis, most likely via its effects on the SNS. These intriguing findings offer important insights, but still do not answer all questions.
Scheme summarizing the findings obtained by Rodrigues et al. in mice
The permanent up-regulation of brain Ang II, SNS activity, and vasopressin implies that apparently in this model, there are no counterregulatory mechanisms at the level of the brain to normalize the consequences of angiotensinogen overexpression (for instance, a down-regulation of mouse angiotensinogen, or of the unknown enzyme cleaving angiotensinogen). This is most unusual, and in full contrast with the proposal of Sigmund et al., who claim that up- or down-regulation of renin in the brain is an important determinant of brain RAS activity . In fact, the latter group has suggested that icREN is a master regulator of classical, extracellular renin, and that deletion of this icREN would strongly up-regulate the brain RAS by increasing local renin synthesis. If so, a 30–35% up-regulation of angiotensinogen might have been easily matched by a modest renin down-regulation. Outside the brain, such renin alterations are perfectly normal, and explain why Ang II levels remain remarkably similar over a wide range of angiotensinogen levels, e.g., in men and women (despite estrogen-related up-regulation of angiotensinogen in women) [15,16], and even in patients with heart failure in whom the circulating angiotensinogen levels are greatly diminished . Similarly, siRNA targeted at hepatic angiotensinogen is capable of lowering circulating angiotensinogen by >95%, yet without affecting circulating Ang II, simply because renin can be up-regulated several 100-fold [18,19].
If alterations at the level of Ang generation are not possible, there are alternative ways of dealing with excess Ang II levels. One is receptor down-regulation (to diminish the consequences of excess Ang II exposure), and a second is increased Ang II metabolism. Yet, Rodrigues et al. observed no changes in AT receptor expression, and were unable to observe other Ang metabolites than Ang II. Not demonstrating Ang I, Ang III or Ang-(1-7) at brain Ang II levels that are phenomenally increased strongly argues against the concept that these metabolites are ever formed in the brain. Normally, at least outside the brain, it is unavoidable that a massive Ang II up-regulation is accompanied by alterations in all of these Ang peptides. Based on this one might conclude that Ang III is not the major activator of Ang receptors in the brain, that Ang-(1-7) does not exert protective effects in the brain, and that in the brain apparently Ang II is generated directly from angiotensinogen, without Ang I as intermediate. Finally, to explain the entire lack of metabolites, it seems that Ang II is somehow captured directly, without being exposed to angiotensinases. This could be indicative of immediate binding of Ang II following its generation to Ang II type 1 (AT1) receptors, which are known to protect Ang II against metabolism [20,21].
Yet, trials are being performed in hypertensive patients with the aminopeptidase A inhibitor firibastat which selectively blocks aminopeptidase A in the brain . This approach is based on the unproven concept that Ang III is the true agonist of brain AT1 receptors, and that aminopeptidase A generates Ang III from Ang II [23,24]. Moreover, ACE2 is generally believed to protect against excess brain Ang II levels, by rapidly converting it to Ang-(1-7) , while Ang-(1-7) itself has been proposed to exert beneficial neuroprotective effects in ischemic and hemorrhagic stroke . If none of this can be demonstrated at such massive brain Ang II up-regulation (i.e., the most ideal condition to allow the detection of these alternative metabolites), it seems unlikely that any of this truly plays a role in normal physiology. Finally, direct Ang II generation from angiotensinogen may help to identify the responsible enzyme. Clearly, this is not renin. A putative alternative player is tonin which indeed has been proposed to generate Ang II directly in the brain . However, this occurs at a 100-fold lower efficiency vs. renin-mediated Ang I formation, and requires a pH optimum of 5–5.5 . Hence, it needs to be proven that this is relevant under normal physiological (or even pathological) in-vivo conditions, and also occurs without substantial up-regulation of angiotensinogen. Indeed, in the present study angiotensins could not be detected in brain tissue without angiotensinogen overexpression, and previous studies even suggested that the small amounts of Ang II that could be detected in the brain are derived from blood, i.e. they do not represent local Ang II synthesis in the brain [7,13]. An important next step would therefore be to demonstrate that tonin-driven brain angiotensinogen cleavage truly has in-vivo relevance in either normal physiology or pathology. Here it is important to note that many investigators have claimed endogenous angiotensinogen in the brain to be present inside cells [29–31]. If true, it is not surprising that it did not result in Ang formation. A further option that Rodrigues et al. might evaluate is therefore that their rat angiotensinogen overexpression did result in angiotensinogen release from cells, as opposed to endogenous mouse angiotensinogen. Here they might also consider that the brain RAS, if existing, is generally believed to result in highly localized Ang II formation, e.g. in one or more specific brain nuclei . This is of course in full disagreement with the approach followed by Rodrigues et al., allowing angiotensinogen to be up-regulated in astrocytes all over the brain. Such non-specific up-regulation is unlikely to ever occur in vivo, and thus it would be intriguing to replace this type of up-regulation by angiotensinogen up-regulation in specific brain nuclei, and to link this to local Ang II up-regulation and tonin deletion in a future transgenic model.
Unexpectedly, brain RAS up-regulation in the study by Rodrigues et al. did not affect blood pressure, despite up-regulating the SNS and vasopressin . The author proposes that this is due to suppression of the peripheral, systemic RAS. In other words, this model results in a permanent resetting of the interactions among RAS, SNS and vasopressin release. Apparently, the renin release that normally occurs after sympathetic stimulation is prevented in this model, implying that the hypertensive consequences of sympathetic up-regulation (which would lower renin indirectly) overrule the direct consequences of sympathetic stimulation of the renin-producing cells in the kidney. To fully understand this, it would be crucial to measure catecholamines in both the kidney and circulation. These data also imply that this model of brain angiotensinogen overexpression is sensitive to blockers of the SNS. Intriguingly, sympathetic activation also up-regulated erythropoiesis. This is a topic that requires further study, as it has not been linked to brain Ang II before.
In summary, Rodrigues et al. have set the stage by firmly demonstrating that the only Ang metabolite in the brain is Ang II, and that its synthesis occurs directly from local angiotensinogen by an as of yet unidentified enzyme, provided that sufficient angiotensinogen is available . Currently, this concept could only be demonstrated in a transgenic model with supraphysiological brain angiotensinogen levels, and thus whether this is also relevant in normal physiology/pathology remains to be determined. Obviously, we now need to know the identity of the enzyme cleaving Ang II from angiotensinogen. Given the absence of other Ang metabolites, in particular Ang III and Ang-(1-7), even when greatly increasing Ang II generation, we need to reconsider how ACE2 activators and aminopeptidase A inhibitors act in the brain. Clearly, this must involve non-Ang peptides cleaved by these enzymes. Finally, it is still feasible that in the context of pathological conditions, characterized by blood-brain barrier damage (e.g., stroke and Alzheimer’s disease), blood-derived angiotensinogen is an additional determinant of the levels of angiotensinogen in the brain. Hence, future studies should evaluate the best approach to interfere with angiotensins in the brain under pathological conditions [6,31].
The authors declare that there are no competing interests associated with the manuscript.
This work was supported by the Consejo Nacional de Ciencia y Tecnología (CONACyT), Mexico [grant number 739513].
Figure was created with BioRender.com.