The renin–angiotensin system (RAS) is widely recognized as one of the most important vasoactive hormonal systems in the physiological regulation of blood pressure and the development of hypertension. This recognition is derived from, and supported by, extensive molecular, cellular, genetic, and pharmacological studies on the circulating (tissue-to-tissue), paracrine (cell-to-cell), and intracrine (intracellular, mitochondrial, nuclear) RAS during last several decades. Now, it is widely accepted that circulating and local RAS may act independently or interactively, to regulate sympathetic activity, systemic and renal hemodynamics, body salt and fluid balance, and blood pressure homeostasis. However, there remains continuous debate with respect to the specific sources of intratubular and intracellular RAS in the kidney and other tissues, the relative contributions of the circulating RAS to intratubular and intracellular RAS, and the roles of intratubular compared with intracellular RAS to the normal control of blood pressure or the development of angiotensin II (ANG II)-dependent hypertension. Based on a lecture given at the recent XI International Symposium on Vasoactive Peptides held in Horizonte, Brazil, this article reviews recent studies using mouse models with global, kidney- or proximal tubule-specific overexpression (knockin) or deletion (knockout) of components of the RAS or its receptors. Although much knowledge has been gained from cell- and tissue-specific transgenic or knockout models, a unifying and integrative approach is now required to better understand how the circulating and local intratubular/intracellular RAS act independently, or with other vasoactive systems, to regulate blood pressure, cardiovascular and kidney function.
Since the seminal discovery of the rate-limiting enzyme renin by Tigerstedt and Bergman (in 1898)  and the landmark study of Goldblatt et al.  on the role of renin in the development of 2-kidney, 1-clip hypertension in 1934, the renin–angiotensin system (RAS) has since been the most extensively studied endocrine (tissue-to-tissue), paracrine (cell-to-cell), and intracrine (intracellular) hormonal system. A critical role for the RAS in the regulation of arterial blood pressure, cardiovascular and kidney function, and the development of hypertension is now firmly established from studies using genetically modified animals [3–9] and human clinical studies using the pharmacological inhibitors of the system to target this system in hypertension and other cardiovascular and kidney diseases [10–16]. The classic endocrine and paracrine paradigms as a powerful vasoconstrictor, a stimulator of the release of aldosterone, and a renal sodium-retaining hormone have led to one of the most successful drug discovery stories of the century, i.e. the development of the inhibitors of angiotensin-converting enzyme (ACE) and renin, and the blockers of the type 1 angiotensin II (ANG II) (ARBs) and aldosterone receptors. Indeed, ACE and renin inhibitors, and ARBs and aldosterone receptor antagonists are the mainstays for the treatment of hypertension, stroke, heart failure, diabetic nephropathy, and other kidney diseases [10–16]. However, recent studies have also shown that the classical RAS paradigm has evolved significantly following discoveries of several new members, enzymes, or receptors of the RAS and their new roles, including prorenin receptors (PRR) [17,18], ACE2 [19,20], and ANG (1-7)/Mas receptors [21–24]. Thus, the key members of the classical RAS, including renin, ACE, ANG II, and aldosterone, are no longer considered to be the only active effector molecules, but the classic renin/ACE/ANG II/ANG II type 1 (AT1) receptor axis still plays a predominant role in the regulation of arterial blood pressure, cardiovascular and kidney function, and the pathogenesis of hypertension [3–9]. The non-classical pathways, such as the prorenin/PRR/vacuolar H+-ATPase (V-ATPase) axis [18,25] and the intracrine (intracellular/mitochondria/nuclear) ANG II/AT1 and AT2 receptor axis [26–28] also appear to play an important role in the long-term transcriptional responses to the RAS stimulation. Conversely, the so-called protective arms of the RAS include the ACE2/ANG 1-7/Mas receptor axis, the aminopeptidase A (APA)/ANG III/AT2 receptor axis, and the ANG IV/AT4 receptor/IRAP axis serve counteracting roles of the renin/ACE/ANG II/AT1 receptor axis [19–24]. Based on the lecture at the XI International Symposium on Vasoactive Peptides held in Belo Horizonte of Brazil in 2017, this article aims to review the new roles of intratubular and/or intracellular RAS uncovered using genetically modified animals with either overexpression or deficiency of one key enzyme, ANG peptide, or receptor of the RAS in the kidney, and discuss their physiological relevance and perspectives.
Intratubular RAS in the kidney: current consensus and debates
Most of the investigators agree that the RAS (RAS) plays an indispensable role in the cardiovascular and renal regulation, normal blood pressure homeostasis, and the pathogenesis of hypertension [29–35]. There is also a general consensus that both circulating (endocrine) and local (paracrine) RAS act interactively to regulate vascular and sympathetic tones, renal pressure natriuresis response, and salt and water balance [29–35]. However, there are continuous debates with respect to: (i) the origins of the intratubular and/or intracellular RAS [30,36–39]; (ii) the relative contributions of the circulating compared with intrarenal RAS to the regulation of renal function [38–41]; (iii) the roles of intratubular RAS to the normal control of blood pressure and the development of ANG II-induced hypertension [29–31,42]; and (iv) the role of intracellular RAS [26–28,43–45]. Previously, it has been impossible to experimentally separate the roles of circulating compared with local intratubular RAS due to the lack of global, kidney-, tubule- or cell-specific genetically modified animal models. Furthermore, the findings that the renin derived from the kidney and angiotensinogen (AGT) derived from the liver are required for the activation of both circulating and intrarenal/intratubular RAS further complicate the respective roles of the circulating, intrarenal, and intratubular RAS in the regulation of blood pressure, cardiovascular and kidney function.
The RAS in the kidney: roles of local intratubular expression and biosynthesis
Since 1980s, extensive molecular, biochemical, and immunohistochemical studies have established that all components of the RAS are localized in the rodent and human kidney (Figure 1) [29,30,46–48]. These include prorenin, renin, AGT, ANG I, ANG (1-7), ACE and ACE2, ANG II, ANG III, or ANG IV proteins. More recent studies have demonstrated that prorenin, renin, and PRR are also expressed in the cortical and/or medullary collecting duct (CD) [18,49–52]. Furthermore, prorenin, renin, AGT, and ANG II proteins have been reported in the urine [51,53–55]. These findings have been taken as evidence for the local intratubular RAS in the kidney. The concept of a local intratubular RAS in the kidney is also supported by well-recognized observations that the levels of ANG II peptides in the kidney [30,46,56–59], and the proximal tubular fluid compartment [56–58,60–62] are much higher than in the plasma. Navar et al. have been instrumental to demonstrate high levels of AGT, ANG I, and ANG II protein levels in the proximal tubular fluid [29,55,63,64]. Other studies from his group and others have shown that AGT mRNA and proteins were expressed in the proximal tubule or proximal tubule cells [29,55,63,64]. Since ANG I and ANG II levels in the proximal tubular fluid are estimated at the nanomolar range, whereas the circulating ANG II level is at the fetomolar range, these studies suggest that ANG II may be produced intracellularly and then secreted into the proximal tubular fluid [30,37]. Whether ANG I and ANG II levels can be as high as nanomolar ranges in the kidney, proximal tubules or proximal tubule and peritubular fluid compartments remain a subject of debate, given the difficulty in collecting adequate amounts of proximal tubular fluid samples and uncertainties about the specificity of antibodies especially in commercial RIAs. An alternative argument, however, may be that ANG I is generated in the proximal tubule fluid by renin and AGT that are filtered and taken up into the proximal tubule. ANG I generated in proximal tubule fluid or intracellularly may be converted into ANG II by the very high levels of ACE expressed on the apical membranes of the proximal tubule (see below).
The cellular and regional localization of major components of the renin-angiotensin system in the kidney
Intratubular RAS in the kidney: roles of circulating, liver-derived AGT
In addition to local onsite expression or synthesis of the RAS discussed above, an alternative, prevailing view is that circulating AGT, ANG I and ANG II may be filtered and taken up in the proximal tubule of the kidney [38,39,41,58,65–69]. As the only known substrate for the rate-limiting enzyme renin, AGT is the largest peptide of the RAS, an ~62 kDa protein similar to the molecular weight of albumin. Thus, like albumin, AGT was originally assumed to not be filtered by glomerulus, unless the glomerular filtration barrier is injured. Kobori et al.  determined whether circulating AGT is a source of urinary AGT by infusing exogenous human AGT in control or ANG II-induced hypertensive rats . No human AGT was detected in the urine of either sham control or ANG II-induced hypertensive rats in support of this view. Two possibilities may explain these findings; circulating human AGT may not have been filtered due to its poor glomerular permeability, and/or the filtered human AGT may have been metabolized in the renal tubules [37,55]. The data are consistent with their hypothesis that urinary AGT is primarily derived from AGT locally formed in the kidney tubules, rather than from the circulation in normal or ANG II-infused rats.
However, there is direct evidence that AGT may be filtered by the glomerulus and the filtered AGT is then taken up by proximal tubule cells . Pohl et al.  reported that immunoreactive or radiolabeled AGT found in the proximal tubule (S1) was due to its retrieval from the ultrafiltrate as well as the storage from endosomal and lysosomal compartments, whereas AGT mRNAs were detected mainly in the proximal straight tubule (S3), just as it has emerged as the major uptake pathway for filtered albumin. The endocytic receptor megalin appears to play an important role in the proximal tubular uptake of AGT . The study of Pohl et al.  thus supports the concept that AGT mRNA is locally expressed mainly in the proximal straight tubule (S3). The direct evidence supporting the hypothesis that the liver is the primary source of circulating and intrarenal AGT came from the elegant study of Matsusaka et al. . In this landmark study, these investigators generated mutant mice with hepatic-, kidney-specific, and double knockout of AGT to test their hypothesis . Surprisingly, both AGT and ANG II protein levels in the kidney were similar in both wild-type and mutant mice with kidney-specific AGT knockout . Conversely, AGT and ANG II protein levels in the kidney were almost abolished in mutant mice with liver-specific knockout of AGT . These striking findings were highly unexpected, since kidney AGT and ANG II levels should be much lower in kidney-specific AGT knockout than wild-type mice, if the kidney is the primary source of intrarenal AGT expression. Alternatively, kidney AGT and ANG II levels should not have been reduced significantly in the liver-specific AGT knockout mice, if the kidney is the primary source of renal AGT. These authors further showed that AGT proteins were selectively located in megalin-expressing proximal tubule cells, consistent with the study of Pohl et al. . Additional supporting data showed that genetic approach to induce podocyte injury, therefore damages the glomerular filtration barrier, increases urinary AGT protein excretion [38,71]. Taken together, the study of Matsusaka et al. provides the direct evidence that liver-derived AGT is the primary source of renal AGT and ANG II proteins in the kidney and that glomerular injury may explain the apparent activation of the renal RAS in CKD models .
So, how can we best reconcile the studies of Kobori et al. [37,55] and Matsusaka et al. [38,71], and two different hypotheses on whether AGT mRNAs and proteins in the kidney are primarily due to local onsite expression and biosynthesis or due to filtration of the circulating, liver-derived AGT? A careful analysis reveals that very different animal models and experimental conditions were used in those studies. While Kobori et al.  focussed on increased intratubular AGT mRNA expression and protein production during ANG II-induced hypertension , Matsusaka et al. [38,71] focussed on the glomerular filtration and proximal tubule uptake of liver-derived AGT using liver-specific AGT-KO mouse model with or without disruption of glomerular filtration barrier. It is likely that under physiological conditions and with an intact endocytic receptor megalin-dependent uptake mechanism, the liver-derived AGT is the primary source of AGT proteins in the kidney, filtered by the glomerulus and taken up by the proximal tubule in the kidney, along with ANG I, ANG II, proreinin, and renin (see later sections for details) (Figure 2). However, in ANG II-dependent hypertension and kidney diseases association with activation of the local RAS, AGT mRNA expression and AGT protein biosynthesis may also be up-regulated along the nephron in the kidney, independent of the circulating, liver-derived AGT . A better approach for further studies to test this hypothesis is to measure intratubular AGT and ANG II levels using ANG II-infused, liver- or kidney-specific AGT knockout mice [38,71].
Glomerular filtration, proximal tubular uptake, and urinary excretion of major components of the renin-angiotensin system in physiology and diseases
The specific roles of liver-derived or intratubular AGT in the regulation of basal blood pressure can be deduced from studies using mutant mouse models with liver-, kidney-, or tubule-specific overexpression or deletion of the AGT gene [38,72,73]. Ying et al.  used the kidney androgen protein (KAP) gene promoter to study the effects of knockin of the mouse AGT gene in the proximal tubule on blood pressure. KAP has been suggested to specifically express in the proximal tubule of the kidney, and its expression is increased by androgen stimulation [5,72,74]. Although AGT mRNA expression in the proximal tubule and urinary AGT proteins were significantly increased in mice with KAP-mediated AGT overexpression in the kidney, basal blood pressure remained normal under normal salt intake . Since basal blood pressure remained unchanged in mice with KAP-mediated AGT overexpression in the kidney, Ramkumar et al.  bred AGT-floxed mice with mice expressing the Pax8‐rtTA and LC‐1 transgenes to generate inducible nephron-wide AGT deletion in a follow-up study. Blood pressure was found to be significantly lower in the nephron-specific AGT-KO mice than in wild-type mice, whether they were fed normal, low- or high-salt diet . These conflicting results are inconsistent with the study of of Matsusaka et al. . In the latter study, the authors created three different strains of AGT-KO mice, the liver-specific, kidney-specific, and liver-/kidney-double KO, and determined their basal blood pressure. Systolic blood pressure fell from ~120 mmHg in wild-type mice to ~75 mmHg in liver-specific AGT-KO mice or ~80 mmHg in liver-/kidney-double KO mice, respectively, but it was not significantly decreased in kidney-specific AGT-KO mice (~115 mmHg) (Figure 3) . While the marked hypotensive response is highly expected in liver-specific and liver and kidney double AGT-KO mice, the well-maintained blood pressure in kidney-specific AGT-KO mice was surprising. The reasons underlying the differences of the findings and conclusions reached from these studies remain unknown, but they may be due to the differences of the approaches to generate AGT-floxed mice and express Cre in the tissue- or cell-specific manner [38,40,72]. However, the currently available evidence appears to support the concept that the liver is the primary source of systemic and intrarenal AGT and ANG II, and liver-derived AGT plays an important physiological role in basal blood pressure homeostasis . This may also explain the close link between blood pressure and hepatorenal syndrome.
Basal systolic blood pressure phenotype in conscious wild-type, liver AGT-KO, kidney AGT-KO, and dual liver and kidney AGT-KO mice
Intratubular RAS in the kidney: roles of intrarenal renin, prorenin, and PRRs
It is established beyond any doubt that prorenin and renin are synthesized and expressed in juxtaglomerular apparatus (JGA) and cells of afferent arterioles, and are secreted into the circulation in response to various stimuli [75–77]. It is this renin that initiates the biochemical cascade of the RAS activation in the kidney, circulation and other target tissues by cleaving AGT to form ANG I, which is then converted by ACE into the most potent peptide of the RAS, ANG II. Recently, it has been argued whether renin and prorenin in the kidney and urine may in fact come from the circulating sources or be synthesized locally . Roksner et al. used multiphoton microscopy to study whether fluorescently labeled renin and prorenin were filtered by the glomerulus in C57BL/6J mice before and after induction of doxorubicin-induced damage of the glomerular filtration barrier. These authors reported that nearly all fluorescently labeled renin and prorenin (37 kDa) and prorenin (42 kDa) were filtered by the glomerulus, and their filtration efficiency exceeded that of albumin (67 kDa). The glomerular filtration of renin and prorenin was increased by disruption of the glomerular filtration barrier with doxorubicin. Nearly all the filtered renin and prorenin were taken up along the nephron, primarily in the proximal tubule, and none appeared in the urine . Furthermore, these authors found that overexpression of prorenin in the liver of CYP1a1-Ren2 transgenic rats increased circulating levels of prorenin by 200-fold, but prorenin was not detected in urine of these transgenic animals, suggesting it was completely reabsorbed along the nephron. However, urinary prorenin, renin, AGT, and albumin all were significantly increased in the urine of patients with Dent’s disease or Lowe’s syndrome, which is associated with defective megalin-mediated uptake of low molecular weight of proteins in the proximal tubule . These findings provide the evidence that circulating renin and prorenin are clearly filtered by the glomerulus and almost completely taken up in the proximal tubule of the kidney primarily through a megalin-dependent mechanism.
The roles of circulating and intratubular renin and/or prorenin in the regulation of blood pressure have been extensively investigated and confirmed in genetically modified animal models with either global or tissue-specific overexpression or deletion of renin. The early study of Mullins et al.  first overexpressed the mouse renin gene, Ren-2, globally in a rat model, TGR(mRen-2)27, and reported that global overexpression of the renin gene induces fulminant hypertension. TGR(mRen-2)27 rats showed significantly elevated circulating and kidney ANG II levels and up-regulation of AT1 receptor expression, rather than down-regulation, in the kidney despite severe hypertension, which were blocked by the AT1 receptor antagonist losartan [3,78,79]. Thus, the TGR(mRen-2)27 rat represents a global renin overexpression, ANG II-dependent hypertensive rat model. To determine the role of intratubular renin, Lavoie et al.  overexpressed the human renin gene (hREN) with or without the human AGT gene, under the control of the kidney androgen-regulated promoter (KAP) . Dual overexpression of the human renin and AGT genes in the proximal tubule significantly increased blood pressure, providing a proof of concept evidence for intratubular AGT and renin in basal blood pressure regulation . Conversely, global deletion of the renin gene in mice led to significant phenotypes in the kidney and/or CDs , and depletion of circulating PRA and caused marked hypotension . These mice showed impaired renal morphology, and urinary concentration ability, as those observed in global AGT-KO or ACE-KO mice. In mice with Ren1c disruption (Ren1c−/−), which is the only renin gene in C57BL/6J mice and humans , nearly 80% of the pups died of dehydration after birth. In those surviving Ren1c−/- mice, basal blood pressure was ~30 mmHg lower with markedly diluted urine, significant hydronephrosis, and fibrosis responses in the kidney . Pentz et al.  have ablated renin-expressing cells from the JGAs of the afferent arterioles during development by placing diphtheria toxin A chain (DTA) under control of the Ren1d mouse renin promoter. They found that the kidneys were smaller with the development of hyperplastic or atrophic glomeruli and renal tubules. These studies suggest that systemic and JGA renin is not only important to maintain basal blood pressure, but is also required for the normal kidney and/or nephron development.
The roles of CD renin and/or prorenin have been studied using CD-specific renin transgenic, knockout, or PRR-KO mice [49,50]. Ramkumar et al. [49,50] used the Cre/Lox recombination approach to generate mutant mice with CD-specific overexpression or deletion of renin to determine its role in the blood pressure regulation. Under basal conditions, CD-renin expression increased five-fold, whereas plasma renin was suppressed by more than 50%, but blood pressure was similar to wild-type mice . Blood pressure increased significantly in response to high salt intake . These authors also reported that there were no significant differences in basal blood pressure between wild-type and CD-specific renin-KO mice, but they did find attenuated blood pressure and epithelial Na+ channel (ENaC) responses to ANG II . Whether intratubular and/or CD-PRR plays a role in maintaining blood pressure homeostasis is still controversial [52,84–86]. PRR is also termed as Atp6ap2, ATPase H+-transporting lysosomal accessory protein 2 and a type 1 transmembrane protein and an accessory subunit of the V-ATPase. In the kidney, PRR is primarily expressed in the intercalated cells and/or principle cells of CD in the medulla , which was up-regulated during ANG II-induced hypertension [35,51,88]. Trepiccione et al.  generated inducible, conditional deletion of Atp6ap2 throughout the nephron segments, but primarily in the intercalated cells of CD in the medulla. These authors found that, apart from alterations in urine concentration and acid-base balance, nephron-specific Atp6ap2 depletion had no effects on intrarenal ANG II formation, sodium handling, and basal blood pressure regulation . Using a similar approach, Ramkumar et al. developed an inducible renal tubule-wide PRR KO using the Pax8/LC1 transgene recombination approach. They found that there were no differences in basal blood pressure between control and PRR-KO mice . Peng et al.  generated CD-specific PRR-KO mice primarily in the principal cells using the PRR-floxed and AQP2-Cre approach, and also failed to find any effect on blood pressure. By contrast, Prieto et al.  generated CD-specific PRR-KO mice using the PRR-floxed and Hox-b7-Cre mice. Interestingly, the latter CD-specific PRR-KO mice showed fewer nephrons or glomeruli, mild hypoplasia in the medulla, smaller kidney, and lower kidney weight etc. along with some abnormal morphological phenotypes pointing to impaired kidney development . Significant functional abnormalities observed in these mice include decreased basal systolic blood pressure and glomerular filtration rate, and elevated diuretic and natriuretic responses without altering the expression and activities of the ENaC in the CD . Overall, it remains difficult to reconcile the conclusions on the roles of CD-specific PPR in the regulation of blood pressure. The only consistent finding is that CD-specific PRR knockout attenuates ANG II-induced hypertension [52,85,86]. However, it is unlikely that CD-specific renin, prorenin, and PRR play a significant role in the normal regulation of blood pressure or kidney function.
Intratubular RAS in the kidney: roles of intratubular ACE
ACE a dipeptidyl carboxypeptidase I, kininase II or EC 184.108.40.206, plays a critical role in converting inactive ANG I into ANG II, as well as metabolizing bradykinin [89,90]. These dual roles of ACE have been well established in the regulation of cardiovascular and kidney function, and blood pressure homeostasis. Unlike AGT, renin, prorenin, ANG I and ANG II, which have been shown to be filtered by the glomerulus and taken up by the proximal tubule, it remains to be determined to what extent whether ACE in the kidney is derived from the systemic circulation or synthesized in the kidney. Previous studies using quantitative in vitro autoradiography have clearly demonstrated that ACE is expressed at the high levels and is found to localize in the apical membranes of the proximal tubule and endothelial cells of blood vessels in the kidney [91–93]. Since the radiolabeled ACE inhibitor Lisinopril selectively binds to, and inhibits the enzyme ACE, this technique has high specificity to localize intratubular ACE along the nephron segments in the rat and human kidneys [91–93]. Most, if not all of the radiolabeled ACE inhibitor is found in the proximal tubule in the kidney. High-resolution immunohistochemistry using the anti-human ACE antibody has been able to localize ACE in the microvilli and brush borders in the proximal tubule [94–96]. ACE immunostaining have also been visualized in the endoplasmic reticulum, mitochondria, and nucleus . In contrast, the glomerulus, distal tubules, and the inner medullary CDs express low levels of ACE [91,92,94–96]. These studies strongly suggest that intratubular ANG II may be generated at high levels in the proximal tubule, and much less is found in other nephron segments under physiological conditions.
The roles of intratubular ACE in the regulation of renal tubular transport may be best recognized through its primary action on intratubular ANG II formation in the proximal tubule of the kidney, which stimulates proximal tubular Na+ reabsorption and contributes to maintenance of salt and fluid balance [30,61,92,93,97–100]. Before the invention of genetically modified animal models, the role of ACE in the regulation of proximal tubular Na+ transport and blood pressure was demonstrated primarily using in vivo micropuncture and ACE inhibitors [61,92,93,97,98]. The effects of ACE inhibitors appear to be relatively consistent in previous studies. With genetically modified mouse models becoming available, however, some confusing or unexpected roles of ACE on proximal tubular Na+ transport and blood pressure have been reported [42,99,101–104]. For example, Hashimoto et al.  used ACE2/2 mice, in which ACE expression was completely knocked out in the kidney, and found that the entire kidney and single nephron GFR were significantly reduced, and fractional proximal tubular reabsorption was increased significantly . This study strongly implies that intratubular ACE presumably does not play a significant role in stimulating proximal fluid Na+ transport . Thus, the results of studies using the ACE-KO mouse models reached a completely different conclusion on the role of intratubular ACE in the regulation proximal tubular Na+ reabsorption from those using ACE inhibitors or ANG II infusion in rats [61,92,93,97,98]. At present, it is difficult to reconcile these results. One plausible reason is that it is technically so difficult to perform in vivo micropuncture experiments in an anesthetized mouse, whose kidney, glomerular and proximal tubule size is only ~10% of those in an adult rat. It is nearly technically impossible to collect enough proximal tubule fluid from a single proximal tubule of a mouse for biochemical analysis of a single glomerular filtration rate and proximal tubule electrolyte transport. Proximal tubule fluid samples collected from many tubules and many mice would have to be pooled before any analyses, and mistakes may likely be introduced during the processes. The second possibility may be due to other compensatory mechanisms being activated in the absence of intratubular ACE and/or ANG II to maintain proximal tubular function. The third more plausible explanation would be that although intratubular or kidney ACE is knocked out, circulating ANG II is readily filtered by the glomerulus and taken up by the proximal tubule, and it maintains proximal tubular Na+ transport in the ACE mutant mouse models [58,65,68,69,105–107]. The only way to exclude this possibility is to measure intratubular ANG II levels in these ACE-KO mice.
The roles of circulating or intratubular ACE in the kidney in the blood pressure control and the development of hypertension may be distinguished using global and kidney- or tubule-specific ACE-KO mouse models [7,42,104,108]. The hypotension phenotype is surprisingly consistent and reproducible in mutant mouse models with global deletion of AGT [6,109], renin , ACE [7,108], or total AT1 receptors (AT1a and AT1b). These findings strongly support the scientific premise of the essential role of the renin/AGT/ACE/ANG II/AT1 receptor axis in the regulation of blood pressure. There was a consistent 20–30 mmHg fall in basal blood pressure in all these global mutant mouse models [6,7,69,108–110]. However, since ACE is widely expressed in different tissues or cells, the global ACE-KO mouse model may not be suitable to determine the relative contribution of extrarenal compared with kidney and/or intratubular ACE to blood pressure regulation. Bernstein et al. have been instrumental to generate a number of ACE mutant mouse models to determine the role of kidney or intratubular ACE in the blood pressure regulation and the development of hypertension [7,42,103,104]. Esther et al.  first generated ACE 2/2 mutant mice, in which circulating ACE levels are elevated but renal ACE are reduced, and found that ACE 2/2 mice had significantly lower blood pressure, urine concentrating defect, and increased proximal tubular reabsorption. A different strain of ACE-KO mice, ACE 3/3 mice, overexpressed ACE in the liver, but completely lost endothelial ACE in all tissues . Kidney ACE was reduced to just 14% of that of wild-type mice, thus representing kidney-specific ACE-KO. Yet, basal blood pressure, plasma ANG II, and kidney function were not different between ACE 3/3 and wild-type mice . Two additional ACE-mutant mouse models were developed by the same group of investigators, one expressing ACE only in the kidney, ACE9/9 , and the other having ACE-KO in the entire kidney, ACE10/10 . ACE 9/9 mice had lower plasma ANG II and hypotension , whereas ACE10/10 mice were able to maintain normal basal blood pressure (Figure 4) . These findings are consistent with those reported in kidney-specific AGT-KO mice, suggesting a key role of the circulating ACE and ANG II .
Basal systolic blood pressure phenotype in conscious wild-type, global ACE-KO , ACE2/2 with elevated plasma ACE , ACE3/3 with overexpression of ACE in the liver without kidney ACE , ACE9/9 with overexpression of ACE in the kidney , and ACE10/10 mice with kidney ACE knockout 
While it is well recognized that both systemic and tissue ACE play an undisputed role in generating circulating and intrarenal ANG II, stimulating proximal tubule Na+ reabsorption, and maintaining blood pressure homeostasis [61,92,93,97,98], it remains difficult to reconcile the conflicting results between studies using genetically modified ACE mouse models and those using systemic or intrarenal blockade of ACE with inhibitors [61,92,93,97,98]. Questions remain with respect to why mice with endothelial ACE-KO may have normal  or lower basal blood pressure [104,111], whereas mice with ACE-KO in the proximal tubule or the entire kidney may be normotensive [42,111]. Similarly, what are the mechanisms underlying normal circulating or kidney ANG II levels even though ACE has been deleted systemically in the endothelium or locally in the kidney and/or the proximal tubule [42,103,104,112]? Thus, further studies are necessary to resolve these issues.
Intratubular RAS in the kidney: roles of intratubular and intracellular ANG II
Local generation compared with uptake via endocytosis
ANG II levels in the kidney are consistently higher than in the circulation in normal rats or mice [30,46,56–59], and are further increased in response to long-term systemic infusion of ANG II, even though plasma renin activity and renin in the kidney JGA are suppressed [58,63,65,113]. This phenomenon supports the concept that ANG II is generated locally within the kidney. Indeed, all major components of the RAS from AGT, renin, to ACE are expressed in the kidney to allow local generation of ANG II [29,30,46–48]. Moreover, AGT mRNA and protein expression are significantly augmented in the proximal tubule or the kidney, rather than down-regulated, in ANG II-induced hypertension [30,63,64]. Augmentation of AGT expression provides further substrates for ANG II production in the kidney. In addition to the proximal tubule or the renal cortex, Prieto-Carrasquero et al.  showed increased renin and prorenin expression in the inner medullary CDs, therefore likely contributing to high intratubular ANG II levels in ANG II-dependent hypertension, such as 2K1C or ANG II-infused animal models [53,113,-115]. Overall, a positive feed-forward loop has been proposed to explain higher intratubular ANG II levels in the kidney during ANG II-dependent hypertension [29,30,54,63,93,113].
There is also solid evidence to suggest that higher levels of ANG II in the kidney may also be due to uptake of circulating and/or extracellular ANG II via AT1 (AT1a) receptor-mediated mechanisms [58,66,68,69]. As discussed previously, ANG II levels in the plasma and kidney are not different between wild-type mice and mice with the kidney- and/or proximal tubule-specific AGT-  or kidney ACE-KO [42,103,104,112]. This suggests that ANG II in the kidney is likely derived from the circulating ANG II via glomerular filtration and tubular uptake. Navar et al., Danser et al., and our groups have previously determined whether the kidney and/or the proximal tubule take up circulating ANG II via AT1 (AT1a) receptor-mediated mechanisms [58,65,68,69,105,107,116]. Zou et al. [65,107] used osmotic minipump to systemically infuse ANG II for 2 weeks, and measured ANG II levels. They demonstrated AT1 receptor-mediated accumulation of ANG II in the rat kidney. van Kats et al. [67,105] systemically infused [125I]-labeled ANG I or ANG II into anesthetized pigs, and showed that [125I]-labeled ANG II, but not [125I]-labeled ANG I, was accumulated in the kidney and adrenal glands, also via AT1 receptor-mediated mechanisms. In the kidney, high levels of internalized ANG II was primarily found in cortical endosomes [58,117], lysosomes , or the nuclei in the proximal tubule cells [105,106,118]. AT1 receptor-mediated accumulation of ANG II in the kidney or proximal tubule cells may be blocked by the AT1 receptor blocker losartan [106,119], siRNAs targetting AT1 receptors , or in AT1a-KO mice [68,69]. Thus, AT1 (AT1a) receptor-mediated uptake of ANG II has been confirmed and reproduced in cultured proximal tubule cells in vitro and in the proximal tubule of the kidney in vivo.
Novel role of intratubular ANG II in the blood pressure regulation
Although it is well recognized that the kidney plays a key role in the physiological regulation of blood pressure and the pathogenesis of hypertension, the precise contribution of intratubular ANG II is poorly understood [31,44,45,121–124]. In this regard, Coffman et al. have been instrumental in studying the role of the kidney in the blood pressure regulation using the kidney cross-transplantation approach and genetically modified global or proximal tubule-specific AT1a-KO mice [31,121,122]. Cross-transplantation of the kidneys between wild-type and global AT1a-KO mice confirmed the importance of AT1 receptors in the kidney in the development of ANG II-induced hypertension and cardiac hypertrophy . Subsequent studies used the Cre/LoxP approach to generate proximal tubule-specific AT1a-KO mice and showed that deletion of AT1a receptors selectively in the proximal tubule decreased basal blood pressure by >10 mmHg, even though systemic and/or vascular AT1a receptor expression and responses remained intact [123,124]. This probably reflects the full life-time compensatory state of extraproximal tubule AT1a receptors or other vasoactive systems in response to AT1 receptor deletion selectively in the proximal tubule. Instead, we used a short-term adenoviral transfer approach to rescue AT1a receptors in global AT1a-KO mice by overexpressing a full-length GFP-tagged AT1a receptor gene (AT1aR/GFP) selectively in the proximal tubule using a proximal tubule-specific sodium and glucose cotransporter 2 (sglt2) promoter . This approach likely mimics a short-term systemic AT1a-KO and proximal tubule-specific AT1a-knock in mouse model, which expressed AT1a receptors only in the proximal tubule of the kidney. In this model, blood pressure was elevated by 12–15 mmHg, 2 weeks after the rescue of AT1a receptor expression in the proximal tubule . These studies together provide evidence that the ACE/ANG II/AT1a receptor axis in the proximal tubule of the kidney contributes ~10–15 mmHg to basal blood pressure in mice under physiological conditions (Figure 5).
Basal systolic blood pressure phenotype in conscious wildtype, global AT1a-KO [9,31,70,113], proximal tubule-specific PT-AT1a-KO [126,127], proximal tubule-specific overexpression of an intracellular ANG II fusion protein, PT-iANG II-KI [44,45,130], and global AT1a-KO mice with proximal tubule-specific overexpression of AT1a receptors, AT1a-KO/PT-AT1a-KI 
Novel role of intratubular intracellular ANG II in the blood pressure regulation
Recently, the concept that there is a functional intracellular RAS in various tissues has gained traction [26–28,43,125]. It remains to be determined whether intracellular ANG II in the proximal tubule is formed within the cells or derived from AT1 receptor-mediated internalization [30,39,126,127], and whether actions of intracellular ANG II is detrimental or protective in the kidney and other tissues [28,118,128–130]. Indeed, we and others have shown that systemically administered [125I]-labeled ANG I or [125I]-ANG II accumulates inside the kidney cells [67,105,116,131,132], and that AGT, ACE, ANG II, and ANG II receptors are found in rat renal cortical endosomes [58,62], mouse proximal tubule mitochondria [126,128,133], and rat or sheep renal cortical nuclei [118,134–136]. With the substrate, enzymes, and ANG II receptors all found inside cells, ANG II, formed intracellularly or internalized, is expected to play an important role in the kidney.
Indirect studies have previously shown that AT1a receptor-mediated ANG II/AT1 receptor internalization was associated with ANG II-stimulated uptake of 22Na+ [137–139], inhibition of cAMP formation , NF-κB activation , and increased NHE3 expression in proximal tubule cells . However, these studies were unable to differentiate the actions of internalized ANG II compared with ANG II via the cell surface receptors. To overcome this technical problem, a number of studies on direct actions of intracellular ANG II have been reported in vascular smooth muscle cells [141,142], proximal tubule cells , freshly isolated renal cortical nuclei [118,129,135,136,144,145], or mitochondria [126,133]. In cultured VSMCs  or proximal tubule cells , we and Haller et al.  reported that direct microinjection of ANG II into the cells induced cytosolic and/or nuclear calcium mobilization, whereas AT1 receptor antagonists blocked this effect. In freshly isolated renal cortical nuclei, we showed that ANG II directly stimulated TGF-β, MCP-1, and NHE3 transcription responses , consistent with previous results in hepatic nuclei [144,145]. Similar nuclear transcriptional responses have been shown in cardiomyocytes or fibroblasts, in which intracellular ANG II regulated RNA synthesis and cell proliferation [129,146]. Furthermore, our lab, Cook et al., and Baker and Kumar  demonstrated that overexpression of an intracellular ANG II fusion protein, stimulated cell proliferation in A10 VSMCs via p38MAPK signaling  or CHO cells , or increased NHE3 expression in proximal tubule cells . Taken together, these studies suggest that intracellular ANG II contributes to cardiovascular and kidney diseases and hypertension. However, recent studies have suggested that intracellular ANG II may play a protective role in counteracting the detrimental effects of intracellular and/or nuclear ANG II [28,130,133,134]. Abadir et al.  have localized ANG II and AT2 receptors on mitochondrial inner membranes, and activation of AT2 receptors by ANG II led to mitochondrial nitric oxide production and modulate mitochondrial respiration in rat and human cells . In isolated sheep nuclei, Gwathmey et al. [134,135] reported that ANG II activated AT2 or ANG (1-7) receptors to increase NO production, whereas it acted on nuclear AT1 receptors to increase super oxide production . In the mitochondria, Wilson et al.  demonstrated that ANG (1-7) was formed to modulate NO production via ANG (1-7)/Mas receptors. Finally, Villar-Cheda et al.  recently reported that activation of nuclear receptors by intracellular ANG II induced a number of protective responses in neurones, such as increases in AT2 receptor and PGC-1α and IGF-1/SIRT1 expression. Whether these so-called protective effects of intracellular ANG II via activation of AT2 receptors or ANG (1-7) via activation of the Mas receptors are sufficient to counteract the detrimental effects of intracellular, mitochondrial, and nuclear ANG II via activation of AT1 receptors remains to be investigated. It is likely that these two systems may interact to play important ‘Ying’ and ‘Yang’ effects in the regulation of central, cardiovascular and kidney function, and blood pressure, but the ANG II/AT1 receptor axis likely plays a dominant role in the diseased states.
Although the abovementioned proof of concept studies demonstrate a biological role of intracellular ANG II in cultured cells and other in vitro settings such as isolated endosomes, mitochondria, or nuclei, the role of intracellular ANG II in the blood pressure regulation remains to be further determined. One of the most critical issues is the technical difficulty to separate its role from that of extracellular counterpart that activates cell surface ANG II receptors and initiates downstream signaling events. Again, some important proofs of concept experiments have been performed to prove a physiological role of intracellular RAS. Lavoie et al.  generated two different strains of renin transgenic mice, with one expressing a secreted form of prorenin, and the other expressing a non-secreted form of human intracellular active renin, icREN selectively in the brain, driven by the astrocytic glial fibrillary acidic protein (GFAP) promoter, GFAP-icREN. These authors further bred GFAP-icREN mice with transgenic mice expressing human AGT selectively in the brain using the same GFAP promoter to generate double transgenic mice expressing either the icREN or secreted renin. Interestingly, mean arterial pressure was significantly elevated by ~20 mmHg in GFAP-icREN mice. This response was blocked by ICV losartan administration . Since there was no difference in blood pressure response between transgenic mice expressing human AGT and icREN and transgenic mice expressing the secreted form of active renin in the brain, this study provides solid evidence for a physiological role of intracellular RAS in the brain . However the role of ANG II in astrocytes rather than neurones still remains to be determined. On the other hand, it has recently been debated whether there is actually a brain RAS, letting alone an intracellular system [152–154]. Danser et al. argued that renin in the brain may represent plasma renin that is taken up and trapped in the brain, whereas ANG II in the brain may represent ANG II taken up from blood rather than locally synthesized [152,153]. But Sigmund et al.  has argued that there is compelling evidence for the de novo production of all components of the RAS in the brain. Clearly, the unresolved issue in the brain just as in the kidney is how much ANG II accumulates compared with how much is synthesized locally.
Cook et al. used a different transgenic approach to globally overexpress an intracellular ANG II fusion protein, ECFP/ANG II, in mice with the mouse metallothionein promoter targetting all tissues . This fusion protein construct lacks a secretory signal, so it is not secreted outside the cells after its intracellular expression or synthesis , thus plasma ANG II was not altered in these mice . However, basal blood pressure was significantly elevated by 16 mmHg, associated with renal injury with thrombotic microangiopathy or microthrombosis . This model may be considered to be a global intracellular ANG II-knockin mouse model supporting the concept that intracellular ANG II has a physiological effect. We reasoned that ANG II taken up by the proximal tubule may escape the classical clathrin-coated pits/endosomes/lysosome degrading pathway, be transported to other organelles, and act as an intracellular peptide. To test this hypothesis, we overexpressed the non-secreted form of intracellular ANG II fusion protein, ECFP/ANG II [43,155], selectively in the proximal tubule of rats, C57BL/6J and global AT1a-KO mice using the proximal tubule-specific sglt2 promoter [44,45,149]. The effect of ECFP/ANG II overexpression in the proximal tubule on basal blood pressure was determined for 4 weeks. In both rats and C57BL/6J mice, ECFP/ANG II overexpression increased blood pressure by ~15–20 mmHg for 2 weeks after introduction of the overexpression. These responses was completely blocked by losartan or in AT1a-KO mice [44,45,149]. We further found that the ECFP/ANG II-induced hypertensive response was associated with increased p-ERK1/2 and NHE3 expression, augmentation of AGT expression, increased Na+ reabsorption in the proximal tubule, and a decrease in 24 h urinary sodium excretion [44,45,149]. Taken together, these proof of concept studies suggest that intracellular ANG II in the brain, cardiovascular tissues and the proximal tubule of the kidney likely play an important physiological role in the regulation of blood pressure independent of the circulating RAS and/or cell surface receptors.
Concluding remarks and perspective
In summary, the available evidence supports the concept of the circulating (endocrine), local tissue (paracrine), and intracellular RAS system in the kidney. All major components of the RAS, including AGT [5,38–40,55,71,72], renin and prorenin [17,18,25,35,41,49–54,157], ACE [7,8,42,90–96,98,99,104,111,112] and ACE2 [19,20,34,48,], ANG II and other ANG metabolites [37,46,47,55–61], AT1, AT2 and the ANG (1-7)/Mas receptors [79,92,93,123,124,131–136,156], have been identified, especially in the proximal tubule. Their respective sources and roles in the cardiovascular and kidney, and blood pressure regulation have been extensively investigated using molecular, cellular, genetic, and pharmacological approaches during last three decades. The strongest evidence supporting the important roles of an intratubular RAS is that under physiological conditions, AGT, prorenin, renin, ACE, ANG II and AT1 and AT2 receptors are localized in the proximal tubule and/or the CD, and the ANG II levels in the proximal tubule are much higher than can be explained by the circulating ANG II level. Moreover, the expression and the levels of intratubular and urinary AGT, prorenin, ACE, and ANG II proteins are significantly increased in the kidney in response to ANG II infusion in spite of suppression of the circulating RAS. By contrast, the strongest evidence supporting the important roles of the circulating RAS is that also under physiological conditions, the liver-derived AGT, and systemic prorenin and active renin, ANG I and ANG II are all filtered by the glomeruli and taken up by the proximal tubule. The maintenance of normal plasma ANG II levels and basal arterial blood pressure in the kidney- or proximal tubule-specific knockout of AGT or ACE further supports the roles of the circulating RAS. However, it may be too simplistic to try and distinguish between the actions of the circulating (endocrine) compared with the local tissue (paracrine) or intracellular (intracrine) RAS, as their functions are intertwined and we ‘can’t see the forest for the trees’.
In reality, it still remains very difficult to separate the sources, contributions, and physiological roles of the intratubular/intracellular RAS from that of the circulating RAS due to the significant overlaps between two systems. A unifying or integrative hypothesis may be required to further improve our understanding of the important roles of the circulating and intratubular RAS in the cardiovascular, kidney, and blood pressure regulation. It is possible that the intratubular RAS may not be active under physiological conditions, since circulating AGT, prorenin and renin, and ANG I may easily reach the proximal tubule through glomerular filtration, and ACE on the apical membranes rapidly converts ANG I into ANG II in the proximal tubule lumen. ANG II may also be filtered into the proximal tubule, where ANG II stimulates proximal tubular Na+ reabsorption and contributes to the maintenance of salt and water balance and normal blood pressure homeostasis. In ANG II-dependent hypertension, increased circulating ANG II may suppress the expression and production of hepatic AGT, prorenin and renin, and ACE, but it significantly increases the synthesis and secretion of AGT and ACE and the uptake of ANG II in the proximal tubule and the expression of prorenin, renin and PRR in the CD. Augmentation of intratubular AGT, prorenin and renin, ACE and PRR expression, and increased uptake of extracellular ANG II into the mitochondria, endoplasmic reticulum, and nucleus in the kidney may represent important feed-forward mechanisms of ANG II-dependent hypertension.
This work was supported in part by the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) [grant number 2RO1DK067299-10A]; the NIDDK/National Institute of General Medical Sciences [grant number 2R01DK102429-03A]; the National Heart, Lung, and Blood Institute [grant number 1R56HL130988-01 (to J.L.Z.)]; visiting Associate Professor of Medicine from and supported by the Guangxi University of Technology School of Medicine, Liuzhou, Guangxi, China (to D.Z.); the visiting Professors of Medicine at Guangxi Medical University, Nanning, Guangxi, China (to X.Z. and J.Z.); and the National Natural Science Foundation of China [grant number #81360290].
The authors declare that there are no competing interests associated with the manuscript.
X.C.L., D.Z., and J.L.Z. contributed to overall conception and design of the manuscript, acquisition of data, analysis and interpretation of data, drafting the article or revising it for important intellectual content, final approval of the version to be published, and agreed to be accountable for all aspects of the work. X.Z. and J.Z. contributed to acquisition of data, analysis and interpretation of data, drafting and revising the article, and agreed to be accountable for all aspects of the work. All authors proofread and approved the manuscript.
angiotensin II type 1 receptor blocker
- ANG II
ANG II type 1
ATPase H+ transporting accessory protein 2
chinese hamster ovary
chronic kidney disease
the Cre recombinase
the Cre recombinase/locus of X-over P1 recombination
cytochrome P450, family 1, member A1-the murin renin gene 2
- ECFP/ANG II
an intracellular angiotensin II fusion protein tagged by enhanced cyan fluorescent protein
epithelial Na+ channel
extracellular signal–regulated kinases 1/2
glial fibrillary acidic protein
insulin-like growth factor-1/NAD-dependent deacetylase sirtuin-1
insulin regulated aminopeptidase
genetic gene deletion
the LC-1 transgene
nuclear factor kappa-light-chain-enhancer of activated B cells
Na+/H+ exchanger 3
National Institute of Diabetes and Digestive and Kidney Diseases
p38 mitogen-activated protein kinase
the Pax8 transgene promoter to drive inducible nephron-wide expression or deletion of a gene
peroxisome proliferator-activated receptor gamma coactivator 1-alpha
plasma renin activity
proximal convoluted tubule
proximal straight tubule
transforming growth factor beta
an angiotensin II-dependent transgenic hypertensive rat model overexpression the murine renin gene 2
sodium and glucose cotransporter 2
vascular smooth muscle
These authors contributed equally to this work.