The renin-angiotensin system (RAS) constitutes a key hormonal system in the physiological regulation of blood pressure through peripheral and central mechanisms. Indeed, dysregulation of the RAS is considered a major factor in the development of cardiovascular pathologies, and pharmacological blockade of this system by the inhibition of angiotensin-converting enzyme (ACE) or antagonism of the angiotensin type 1 receptor (AT1R) offers an effective therapeutic regimen. The RAS is now defined as a system composed of different angiotensin peptides with diverse biological actions mediated by distinct receptor subtypes. The classic RAS comprises the ACE-ANG II-AT1R axis that promotes vasoconstriction; water intake; sodium retention; and increased oxidative stress, fibrosis, cellular growth, and inflammation. In contrast, the nonclassical RAS composed primarily of the ANG II/ANG III-AT2R and the ACE2-ANG-(1–7)-AT7R pathways generally opposes the actions of a stimulated ANG II-AT1R axis. In lieu of the complex and multifunctional aspects of this system, as well as increased concerns on the reproducibility among laboratories, a critical assessment is provided on the current biochemical approaches to characterize and define the various components that ultimately reflect the status of the RAS.
the renin-angiotensin system (RAS) comprises a number of enzymatic pathways and bioactive components that underlie diverse functional actions (Fig. 1). Although originally characterized as a circulating endocrine system that targets both peripheral and central receptors, abundant evidence reveals a tissue-based RAS that influences local cellular actions and exhibits intracellular or subcellular components (2, 25, 54, 71, 74, 93, 134, 140, 151). The RAS is classically defined by the activity of angiotensin-converting enzyme (ACE) to form angiotensin II (ANG II) and the subsequent activation of the angiotensin type 1 receptor (AT1R) to mediate both peripheral and central mechanisms in the regulation of blood pressure. The ACE-ANG II-AT1R pathway is also associated with various pathologic responses including fibrosis, inflammation, metabolic dysregulation, heart failure, cancer, aging, and diabetic injury (25, 46, 71, 93, 134). Conversely, alternative RAS pathways may mitigate against the activation of the ACE-ANG II-AT1R, particularly in response to the pharmacological blockade of this axis (29, 43, 47, 119). Targeting ACE reduces ANG II expression but enhances the ANG-(1–7)-AT7/MasR axis that generally opposes the actions of the ANG II-AT1R (29, 43, 44, 103, 119, 120). AT1R antagonists may also increase the formation of ANG-(1–7) through ACE2, as well as shunt ANG II to the AT2R pathway that shares similar properties to the ANG-(1–7) system (12, 25, 29, 44, 47, 119). The RAS is composed of a complex array of components that can be functionally partitioned into distinct receptors, peptides, and peptidases that are downstream from the initial processing of the precursor angiotensinogen to form ANG I by renin (Fig. 1). The review assesses the current biochemical approaches to quantify these components that ultimately define the status of the RAS and to address the consistency of the expected values of the RAS, particularly in lieu of new methodologies applied to the quantification of endogenous angiotensin peptides in various biological compartments including the circulation, tissues, urine, and cells.
RAS Protein Components
A soluble glycosylated protein (molecular mass, 50–60 kDa), angiotensinogen is the sole precursor of angiotensin peptides and the only known substrate for renin (Fig. 1). Circulating levels of the protein are relatively abundant (10−5 g or 10 μg/ml; 200 nM), particularly compared with ANG II or ANG-(1–7); therefore, a very small quantity (<0.1 ml) of serum or plasma is required to quantify angiotensinogen (Table 1). A specific and sensitive enzyme-linked immunoassay (ELISA) for angiotensinogen developed by Kobori and colleagues (70) is commercially available and comprises antibodies directed against two distinct antigenic regions for the capture and detection of angiotensinogen. The angiotensinogen ELISA is quite sensitive (10−11 g or 10 pg; 1.6 amol), exhibits a broad detection range, and is available for multiple species including human, rat, and mouse. There are two angiotensinogen ELISAs that detect either total (ANG I or des-ANG I) or intact (ANG I) forms of angiotensinogen; therefore, the extent of the renin-processed protein can be determined (IBL-Japan, Gumma, Japan). Importantly, the angiotensinogen ELISAs are not compatible with tissue or cell extracts as opposed to plasma, serum, urine, and cell media, although partial purification of the tissue or cell extract may potentially remove the interfering substances. Tissue or cellular content of angiotensinogen can be quantified by immunoblot, provided a source of the protein is available to standardize this approach, or, alternatively, by addition of renin to quantify ANG I generation from intact angiotensinogen that essentially comprises a “reverse renin assay.”
Human urinary levels of angiotensinogen are lower than plasma and average 10–200 ng/mg creatinine (0.2 to 4 pmol/mg Cr) (Table 1). However, elevated excretion of angiotensinogen is considered a biomarker for an activated renal RAS in several models of renal injury that reflects the tubular release of angiotensinogen (5, 69, 72, 113). We applied the total angiotensinogen ELISA and immunoblot methods to assess angiotensinogen expression in the salt-sensitive mRen2.Lewis rat, a congenic model of an activated ACE-ANG II-AT1R axis, and observed a marked sex difference in urinary angiotensinogen levels (34) (Fig. 2). Male mRen2.Lewis exhibited 200-fold higher excretion of angiotensinogen than females (2 vs. 0.01 pmol/mg Cr) (Fig. 2). Estrogen depletion (ovariectomy) increased angiotensinogen excretion 20-fold, whereas a high-salt (HS) diet enhanced excretion 100-fold compared with intact mRen2 on a normal salt diet. Although both maneuvers increase blood pressure to a similar extent, excretion of angiotensinogen in the HS males remained sixfold higher than the HS females (Fig. 2). Despite the marked increases in excretion, renal mRNA and protein levels of angiotensinogen were not increased following ovariectomy or the HS diet, nor did we detect enhanced release of angiotensinogen from renal cortical slices in the HS mRen2.Lewis rat (34) (Fig. 2). The elevated urinary angiotensinogen in the mRen2.Lewis likely reflects a greater degree of filtration of the circulating protein rather than a renal source and is perhaps a sensitive biomarker of glomerular damage (34).
Typically regarded as the enzyme that initiates the RAS cascade, renin is an aspartyl-type protease (30–40 kDa) that cleaves angiotensinogen to form the inactive peptide ANG I (Fig. 1). The enzyme is synthesized, glycosylated, and released in both inactive (prorenin) and active forms by the juxtaglomerular cells of the renal cortex at a higher prorenin-to-renin ratio. A second source of renal renin is the collecting duct cells that secrete active renin into the tubular fluid (52, 71, 82, 92, 113). There is evidence for alternative gene products of intracellular renin in the kidney, brain, and adrenal that may imply a renin-dependent pathway for the intracellular expression of angiotensin peptides (65, 76, 106, 107, 142). Apart from its catalytic activity, prorenin may constitute an additional circulating hormone of the RAS through binding and activation of the prorenin receptor within various tissues (52, 71, 94). A high molecular mass complex (50–60 kDa) comprised of renin and the renin binding protein N-acetyl glucosamine epimerase is typically revealed by immunoblot analysis; the complex renders renin inactive and may influence renin secretion (64). Regarding renin measurement, the enzyme is released by stress, dehydration, low-salt diet, RAS blockade, and certain anesthetics (pentobarbital sodium); therefore, physiological assessment of activity should ideally be obtained by rapid decapitation or blood withdraw by an indwelling catheter in conscious animals. The classical method to detect renin activity is the direct generation of ANG I from angiotensinogen under assay conditions that protect ANG I from enzymatic degradation and does not attenuate renin activity. The assay requires a sensitive and quantitative method to detect the ANG I product at 10−12 g (pg) or 10−15 mol (fmol) typically by radioimmunoassay (RIA) or ELISA. To quantify the prorenin form, samples are enzymatically treated with trypsin to release the propeptide that occupies the active site and inhibits the enzyme; a trypsin inhibitor (i.e., soybean trypsin inhibitor) is then added before addition of the substrate. Prorenin activity constitutes the difference between the total or trypsin-activated renin and basal renin activity. Plasma contains high amounts of angiotensinogen; thus the assessment of plasma renin activity reflects both active renin and endogenous ANG I-angiotensinogen in the circulation. Elased et al. (39) initially used surface-enhanced laser deabsorption ionization (SELDI)-mass spectrometry (MS) to detect ANG I and demonstrate relative renin activity in very low volumes of plasma samples. Camenzind et al. (20) applied matrix-assisted laser deabsorption ionization-mass spectroscopy (MALDI-MS) to quantify ANG I generation from angiotensinogen with comparable renin activity values to that of RIA methods. Renin activity is typically expressed as nanograms or picomoles ANG I generated per milliliter of plasma per hour. Typical plasma renin activity values range from 0.5–10 ng·ml−1·h−1 (∼0.5–10 pmol·ml−1·h−1), whereas total renin (prorenin/renin) is 3- to 10-fold higher reflecting the greater abundance of prorenin. Addition of exogenous angiotensinogen yields maximal renin activity that is conventionally but inappropriately termed plasma renin concentration. An alternative method is a direct human renin assay using capture and detection antibodies with the latter antibody directed to the active or open form of renin. In this regard, addition of the nonpeptide inhibitor aliskerin induces a conformation change in the enzyme that facilitates detection of prorenin by this assay (139). Using this method, van den Huevel et al. (139) reported mean values of active and prorenin at 14 pg/ml (0.4 pM) and 66 pg/ml (1.9 pM), respectively, in human plasma.
Renin activity is also evident in the urine at ∼10% of the plasma levels of the enzyme (84, 139). In contrast to the circulation, there is little or no urinary prorenin, and this may reflect the predominant luminal secretion of renin from collecting duct cells (71, 82, 84, 92). Danser and colleagues (113, 139) suggest that urinary renin may be an alternative measure of an active RAS within the kidney. Quenched fluorescent substrates based on the ANG-(1–14) sequence are available to measure renin. Although this approach does not rely on the detection of ANG I or an exogenous source of angiotensinogen, the sensitivity and specificity of the peptide substrates are not comparable with angiotensinogen. A protease/peptidase cocktail of inhibitors must be added to prevent nonspecific cleavage of the quenched substrate, as well as addition of a renin inhibitor such as aliskerin to demonstrate specificity of the assay.
Peptidases are generally distinguished from proteases by their preference to hydrolyze peptides of 5 to 30 amino acids in length. In contrast to renin, the peptidases associated with the RAS do not exhibit exclusive specificity for angiotensins, although certain enzymes may exhibit preferential kinetics such as ACE2 for ANG II (112, 145). An extensive peptidase class is the metalloenzymes [ACE, ACE2, neprilysin (Nep), aminopeptidase A] that require divalent cations for activity; thus the assessment of activity in the circulation is performed in serum in the absence of strong chelating agents such as EDTA or phenanthroline.
The predominant pathway of the classical RAS is the hydrolysis of ANG I to ANG II by the metallopeptidase ACE, a dipeptidyl carboxypeptidase that cleaves two residues from the COOH-terminus of peptides. The peptidase is a membrane-bound and glycosylated protein (120–180 kDa); however, soluble forms of the enzyme are present in the circulation, cerebrospinal fluid, lymph, and urine (15). ACE plays a key role in the formation of ANG II, but the peptidase metabolizes other peptides including bradykinin, substance P, acetyl-Ser-Asp-Lys-Pro, and ANG-(1–7) (16, 145). ACE activity is typically measured by small peptide substrates such as hippuryl-His-Leu or furylacrylol-Phe-Ala-Gly-Gly in a buffer containing chloride (10–200 mM) and zinc (1–100 μM) for optimal peptidase activity (24, 122). Quenched fluorescent substrates Mca-Ser-Phe-Leu-Tyr-DNP or Mca-Tyr-Val-Ala-Arg-Ala-Phe-Lys-DNP have also been applied for ACE measurement; however, these are not specific substrates and appropriate assay conditions that include an inhibitor cocktail to prevent nonspecific hydrolysis by other peptidases must be considered, as well as separate samples sets for an ACE inhibitor (i.e., captopril, lisinopril) to confirm assay specificity (81). Finally, the degree of sample quenching of the fluorescent product should also be assessed for each source of activity.
Chymases comprise a family of serine peptidases that may form ANG II by hydrolysis of the Phe8-His9 bond of ANG I and pseudoprecursors [ANG-(1–12), ANG-(1–25)] (α-chymases) or metabolize ANG II at Tyr4-Ile5 to form ANG-(1–4) and ANG-(5–8) (β-chymases) (3, 27, 28, 89, 138). Humans express α-chymase, whereas rodents express primarily β-chymases, as well as other isoforms (mouse mast cell protease-4 and rat mast cell protease-5) that more closely resemble α-chymase in regard to the processing of ANG I to ANG II (27, 28, 126). The human and mouse enzymes may also play a role in the conversion of the endothelin precursor to the active peptide, as well as the activation of various inflammatory cytokines (28, 126). Chymases (35 kDa) are synthesized and stored in an inactive proform within mast cells and neutrophils that are released with other proteases (cathepsin G, tryptases, renin) upon degranulation following injury or inflammation (28). Although chymases are soluble enzymes, they associate with the cell membrane and may locate to the extracellular surface of tissues following release. Fluorescent-quenched substrates for chymase include succinyl-Ala-Ala-Pro-Phe-amido-4-methylcoumarin and succinyl-Leu-Leu-Val-Tyr-amido-4-methylcoumarin. Addition of the serine protease inhibitor chymostatin is typically used to demonstrate specificity; however, chymostatin inhibits other ANG II-generating enzymes (cathepsin G, elastase-2) and more selective chymase inhibitors should be considered (116–118, 123, 146). The extent that chymase or other peptidases participate in the formation of circulating or tissue ANG II through an ACE-independent pathway remains an area of significant debate (21).
Nep (∼95 kDa) and other endopeptidases (thimet oligopeptidase, 80 kDa; and prolyl oligopeptidase, 76 kDa) process ANG I directly to ANG-(1–7) without the requisite formation of ANG II (Fig. 1). These peptidases hydrolyze interior bonds of the peptide and prefer aromatic and/or hydrophobic residues, but prolyl oligopeptidase specifically cleaves the COOH-terminus of proline and hydrolyzes the Pro7-Phe8 bond of both ANG I and ANG II to form ANG-(1–7) (29, 112, 145). Circulating forms of soluble Nep are present in the serum, urine, and cerebrospinal fluid (111, 129). Following chronic ACE inhibition, circulating levels of ANG-(1–7) are markedly higher that reflect processing of ANG I by vascular Nep, as well as the reduced metabolism by ACE (Fig. 1) (29). Various quenched substrates are available to measure Nep and other endopeptidases (81). The quenched peptide MCA-Arg-Pro-Pro-Gly-Phe-Ser-Ala-Phe-Lys-DNP may be an appropriate dual substrate for ACE and Nep (M. C. Chappell, unpublished observations). Again, the assay conditions require inhibitors to prevent nonselective hydrolysis of the substrate, as well the addition of a Nep inhibitor (thiorphan, SCH39370) to demonstrate specificity. Miners et al. (86) immunoprecipitated Nep before the addition of the fluorescent substrate, and this maneuver may enhance the specificity of the assay. This approach should be applicable to other peptidases provided the appropriate antibodies are available to pull down the enzyme and do not inhibit the active site.
ACE2 is a membrane-bound monocarboxypeptidase (120 kDa) that converts ANG II directly to ANG-(1–7). Circulating levels of ACE2 are typically quite low particularly compared with ACE; however, serum ACE2 activity is elevated in diabetes, heart failure, and hypertension (41, 136, 148). ACE2 has a potentially significant role in the RAS pathway as a single catalytic step degrades ANG II but activates the ANG-(1–7) axis (Fig. 1) (29). ACE2 assays use the quenched fluorescent substrates Mca-Ala-Pro-Lys-DNP or Mca-Tyr-Tyr-Val-Ala-Asp-Pro-Lys-Dnp (111). These substrates are not specific for ACE2, and addition of other inhibitors to block ACE, Nep and prolyl oligopeptidase, are critical to demonstrate assay specificity, as well as additional samples that include a specific ACE2 inhibitor (MLN4760, DX600). These fluorescent assays can be quantified for enzyme concentration given a standardized source of each peptidase is available (R&D Systems, Minneapolis MN). Using this approach, Rice et al. (111) reported the concentration of ACE in human serum averaged 7 nM in over 500 subjects, whereas ACE2 content was 200-fold lower (33 pM) and detectable in a minor fraction (<10%) of this population. Circulating Nep content (290 pM) was also lower than ACE and evident in <30% of these patients (111).
The role of aminopeptidases in the RAS is generally considered a degradative pathway that initiates the metabolism of ANG II (Fig. 1). Recent studies suggest that the NH2-terminal processing of ANG II to ANG-(2–8) or ANG III by aminopeptidase A may yield a peptide selective for brain AT1 receptors and kidney AT2 receptors (18, 25, 68, 103, 109). Further NH2-terminal processing of ANG III to ANG-(3–8) or ANG IV yields another bioactive peptide that interacts with the insulin-regulated aminopeptidase (4). Fluorescent substrates are available for aminopeptidases A and N as well as general (amastatin, bestatin) and more selective inhibitors to distinguish these enzymes (18).
Endogenous peptide substrates including ANG I, ANG II and ANG-(1–12) can be directly used to quantify peptidase activities (79). One advantage is that the contribution of various peptidases for a given peptide can be directly compared to determine the predominant activity in a particular tissue or treatment. In contrast, peptidase activities derived by different synthetic substrates are not directly comparable unless standardized to enzyme content. Moreover, use of endogenous peptides may reveal novel peptidase activities involved in angiotensin processing (62, 144). Peptidase assays developed in the author's laboratory use 125I-radiolabled peptides coupled to HPLC-based separation and in-line γ-detection (129). Typically, microliter amounts of serum or microgram quantities of tissue are required with this approach. We find that circulating ACE and ACE2 activities were significantly increased in a model of diabetic hypertension; however, ACE activity remained far higher than ACE2 which may contribute to the greater content of ANG II in diabetic mRen2.Lewis rats (148). Kinetic analysis using both labeled and unlabeled ANG II confirmed a sevenfold increase in the Vmax value for serum ACE2 in the female diabetic rat (148).
Angiotensin processing in tissues, cells, and plasma was recently assessed by HPLC-MS. Hildesbrand et al. (62) used a HPLC-tandem quadropole system (HPLC-MS/MS) to reveal metabolism pathways from ANG I to its NH2-terminal metabolites ANG-(5–10) and ANG-(4–10), as well as ANG II and ANG-(1–7) in immobilized proteins from plasma. Suski et al. (133) find that ANG I was primarily converted to ANG-(1–7) in vascular smooth muscle cells (VSMC) by HPLC-MS/MS and support an earlier study (32) that thimet oligopeptidase directly processed ANG I to ANG-(1–7) in rat VSMCs. Finally, Grobe and colleagues (6, 53) applied in situ MALDI to characterize both renal and cardiac metabolism of exogenous ANG II. In this approach, ANG II was incubated on the tissue surface and the MS matrix subsequently layered over each tissue section. The matrix was subjected to MALDI-MS analysis, and the metabolism products were identified and localized to distinct areas of the kidney. ANG-(1–7) was the primary product from ANG II in the renal cortex, whereas ANG III was the major metabolite in the medulla (53). In the heart, ANG III and ANG-(1–7) were major products of ANG II metabolism, catalyzed by aminopeptidase A and ACE2, respectively (6). These data confirm earlier HPLC-based studies on the contribution of ACE2 to ANG-(1–7) formation in the mouse and human heart (49, 152). Caveats with this approach are that MALDI-MS cannot distinguish intracellular versus membrane or extracellular processing and requires relatively high substrate concentrations to detect peptidase activity that may not reflect endogenous processing pathways. However, given the rapid advancement of this technology, future systems will likely develop the required sensitivity and resolution to appropriately detect peptides in situ, as well as characterize the extent of enzymatic processing.
ANG II and ANG III interact with two G protein-coupled receptors identified as the AT1 and AT2 subtype, whereas ANG-(1–7) and [Ala1]-ANG-(1–7) recognize the Mas and Mas-related G protein-coupled receptor member D (mRG-D) receptors, respectively (12). Immunoblot techniques can provide a relative quantification of protein expression, as well as tissue and cellular distribution; however, the commercially obtained AT1 and AT2 receptor antibodies are notorious for their nonspecific nature (59, 61). Receptor antibody specificity is complicated by 1) posttranslational processing events such as glycosylation or proteolysis that may yield unexpected molecular size bands for the receptor in different tissues or species, 2) the very low level of receptor expression particularly compared with other RAS proteins, and 3) the lack of appropriate controls for each antibody. The sole reliance on antibody-based immunoblot methods to assess receptor expression underlies the need to standardize the receptor antibodies to ensure some degree of reproducibility among laboratories. Issues with antibody specificity are certainly not limited to the RAS components and present a major concern regarding the unreliability of antibodies against their intended targets (13). Indeed, the National Institutes of Health has recently promoted a more rigorous approach to ensure reliability and overall reproducibility in preclinical experimental studies (http://www.nih.gov/science/reproducibility/). Uses of knockout transgenic mice, knockdown approaches in cells, or overexpression of the target protein are appropriate controls to test antibody specificity for immunoblot and immunocytochemistry methods, although these approaches are not directly comparable with other species or different tissues. Moreover, knockdown of the functional domain does not ensure that the remaining protein will not be recognized by the antibody. Blockade of the immunoreactive signal by the immunizing peptide or full-length protein is an additional approach to address specificity of the immunoblot or immunocytochemistry signal provided a pure source of the peptide/protein is available; however, antigenic sites exposed by SDS denaturation, formalin fixation, or antigen retrieval may not be accessible for proteins in solution.
The characterization of angiotensin receptors is further complicated by the fact that these proteins do not reside or exclusively localize to the plasma membrane. In lieu of earlier studies on intracellular peptide receptors, ANG II and ANG-(1–7) receptors were characterized on isolated nuclei from rat and sheep kidney (54). In rat nuclei, binding studies with the nonselective antagonist 125I-[sarcosine1, threonine8]-ANG II (sarthran) and competition with selective antagonists revealed exclusive expression of the AT1 receptor (104, 105) (Fig. 3, A–C). Fluorescence-activated cell sorting demonstrated essentially complete overlap of the fluorescent signature for the AT1 receptor and the nuclear marker nucleoporin; the immunoblot confirmed a single protein band for the renal nuclei with the same AT1R antibody (Fig. 3, D–E). Functional studies revealed that ANG II stimulated reactive oxygen species and was blocked by both an AT1R antagonist and a NAD(P)H inhibitor (Fig. 3F). A similar approach was taken for the AT7/Mas receptor in isolated nuclei of the sheep kidney (55–58). An alomone antibody (Alomone Labs, Tel Aviv, Israel) revealed Mas receptor expression along the tubular elements of the kidney that was abolished by antibody preabsorption with the immunogenic peptide; a single band (∼35 kDa) was evident on the full-length immunoblot of isolated nuclei (Fig. 4, A–F). Competition of sarthran binding in nuclei revealed complete additivity with the AT1R antagonist losartan and the AT7R antagonist [d-Ala7]-ANG-(1–7) (A779), but partial additivity with the AT2R antagonist PD123319 suggesting the AT7R and AT2R antagonists are not entirely selective for the sarthran ligand (Fig. 4G). However, only the [d-Ala7]-ANG-(1–7) antagonist and the nitric oxide synthase inhibitor NG-nitro-l-arginine methyl ester abolished the ANG-(1–7)-stimulated release of nitric oxide on renal nuclei as detected by diaminofluorescein fluorescence (Fig. 4H). In regard to the AT2R, Padia et al. (103) find that the receptor protein resides primarily in the cytosolic compartment and rapidly inserts into the apical membrane of the proximal tubules following dopaminergic stimulation to facilitate a natriuretic response. Interestingly, translocation of the AT2R in spontaneously hypertensive rats was blunted compared with normotensive Wistar-Kyoto rats (103). Biotinylation studies combined with an AT2R antibody was used to reveal localization of the receptor on the plasma membrane (103). It is not currently known whether the other angiotensin receptor subtypes exhibit a similar type of regulation. In this regard, a careful assessment of the subcellular fractions or intracellular organelle for angiotensin receptor expression is warranted. Moreover, these and other studies support the intracellular expression of G protein-coupled receptors within the kidney and other tissues, as well as emphasize the necessity for multiple approaches in the identification and characterization of angiotensin receptors (17, 35, 40, 74, 132, 151).
Interest in angiotensins other than ANG II and ANG III was stimulated by identification of endogenous ANG-(1–7) over 25 years ago (30), as well as subsequent evidence for an ANG-(1–7) receptor (121) and a converting enzyme (137). [Ala1]-ANG II and [Ala1]-ANG-(1–7) were recently identified in human plasma, likely arising from decarboxylation of the NH2-terminal aspartic acid residue (67, 75). Functionally, [Ala1]-ANG II may not distinguish AT1 or AT2 receptors (67, 149), whereas [Ala1]-ANG-(1–7) recognizes the Mas-related receptor mRG-D to induce vasorelaxation (12, 75). Both [d-Pro7]-ANG-(1–7) and the AT2 antagonist PD123319 blocked the actions of [Ala1]-ANG-(1–7), but not [d-Ala7]-ANG-(1–7) (75). Jankowski et al. (66) identified [Pro1, Glu2]-ANG II in human plasma and this peptide exhibits a greater affinity for the Mas receptor than ANG-(1–7); however, the pathway for the peptide's formation is unknown. The ANG II metabolite ANG-(3–8) or ANG IV has distinct functional properties through the interaction with the insulin-regulated aminopeptidase, although the peptide also stimulates the AT1R (4, 80, 150). ANG-(1–9), a potential product of ACE2 processing of ANG I, appears to functionally interact with the AT2 receptor (101). Finally, Nagata and colleagues (89, 90) identified the intermediate precursors ANG-(1–12) in rat and ANG-(1–25) in human urine. ANG-(1–12) is processed by ACE and chymase to ANG II, and Nep directly generates ANG-(1–7), whereas ANG-(1–25) is converted by a chymase-like enzyme to ANG II (3, 89, 143). The circulating and tissue pathways for the generation of ANG-(1–12) and ANG-(1–25) are not currently known but are likely renin independent.
The accurate assessment and identification of angiotensins are critical to ascertain the status of the RAS, particularly in pathological conditions or following therapeutic intervention; however, their evaluation is hampered by the extremely low level of peptide expression in the fentomole per gram or fentomole per milliliter range, interfering substances in samples that may be present at a far higher concentration, susceptibility to metabolism in sample collection and processing, the peptides' shared sequence, and the specificity of detection. All these factors may contribute to the marked variability in angiotensin values evident in the literature and raise concerns on the identity of ANG II or other angiotensins detected in tissues and the circulation, as well as the extent that altered peptide content truly reflects or contributes to a particular phenotype or treatment response.
RIA and ELISA are invaluable assays to characterize angiotensins, and they remain the most used biochemical methods to quantify endogenous peptides. Under optimal assay and extraction conditions, both methods should allow for the sensitive and relatively specific detection of ANG II and other angiotensins at fentomole (pg) quantities in plasma, tissue, urine, and cells. Since angiotensins share an identical NH2-terminus except for the Ala1-forms of ANG II and ANG-(1–7), antibodies for ANG II, ANG-(1–7), ANG-(1–9), ANG I, and ANG-(1–12) are directed against their unique COOH-terminus. The ANG II and ANG-(1–7) antibodies, for example, typically distinguish a single amino acid difference (Phe8) between ANG II and ANG-(1–7). Importantly, these antibodies are not selective against the NH2-terminal portion of the peptide (30). Thus an ANG II RIA distinguishes ANG-(1–7) and ANG I but will recognize ANG III, ANG IV, and ANG-(4–8) (37). Similarly, an ANG-(1–7) antibody recognizes the NH2-terminal metabolites ANG-(2–7) and ANG-(3–7), but not ANG-(4–7), ANG II, or ANG I (30). These assays can be used across species as the COOH-terminal portion is conserved for ANG-(1–7), ANG II, and ANG I; however, larger peptides such as ANG-(1–12) or ANG-(1–25) exhibit different COOH-terminal residues that are species specific and require unique antibodies for detection (89, 90). Apart from their nonselectivity for NH2-terminal metabolites of angiotensin peptides, RIA or ELISA may detect a significant degree of nonspecific or background immunoreactivity that reflects the sample source and extent of purification or extraction, as well as the assay antibodies (37). Therefore, chromatographic separation should be routinely coupled to RIA or ELISA to verify the identity of immunoreactive peptides, as well as account for nonspecific material in plasma or tissue samples. Peptide fractionation by HPLC provides this additional level of discrimination based on the resolution and high recovery of the sample peptides, rapid separation time, and general compatibility with RIA methods (22, 30, 37, 78, 99). In addition, the different characteristics for each antibody-based assay must be appreciated, particularly those obtained from commercial sources. The stated cross-reactivity of the Peninsula human ANG-(1–12) ELISA for ANG I is 0%; however, the Peninsula rat/mouse ANG-(1–12) ELISA cross-reacts 60% with ANG I (Peninsula Lab Int, San Carlos, CA). Thus HPLC fractionation would be essential to distinguish endogenous ANG I and ANG-(1–12) content in the rodent with the Peninsula ELISA.
We previously reported that renal ANG II content declined >80% in mice that lack membrane ACE, a transgenic model developed by Bernstein and colleagues (88) that expresses soluble but not the tissue-bound enzyme. ANG II content was quantified directly by a commercial ANG II RIA (ALPCO Diagonostics, Salem, NH) that recognizes ANG III, ANG IV, and ANG-(4–8), whereas ANG-(1–7) was detected by a custom RIA that also detects ANG-(2–7) and ANG-(3–7) (30). Renal tissue was extensively extracted with cold HCl-ethanol, proteins precipitated by heptafluorobutyric acid (HFBA) and peptides purified on C18 SepPak columns (30, 88). To confirm ANG II identity, pooled renal extracts were fractionated by HPLC using the HFBA system developed in the author's laboratory that achieves baseline separation of immunoreactive angiotensins, and the fractions were assayed by separate RIAs (30). ANG II was the primary immunoreactive peak in both the wild-type (WT) and transgenic ACE−/− mice, but the ANG II peak was markedly lower in the ACE-deficient mice (Fig. 5). Note that ANG III and ANG-(4–8) represent minor immunoreactive peaks that are essentially absent in the transgenic kidney. Renal ANG-(1–7) content was unchanged despite the marked reduction in ANG II and suggests that a non-ACE2 pathway contributes or maintains tissue ANG-(1–7) (Fig. 5). Overall, this approach confirmed that the two RIAs identified predominantly NH2-terminally intact ANG II and ANG-(1–7) in the mouse kidney, as well as demonstrate the greater ratio of ANG II to ANG-(1–7) in the WT kidney. However, these results do not ensure that ANG II or ANG-(1–7) RIAs primarily detect both peptides in all experimental situations and verification of the direct assay values is essential for each tissue source and species.
Campbell and colleagues (22–24, 78) developed an alternative approach with RIAs specific to the NH2-terminus of intact and smaller metabolites rather than the COOH-terminus of angiotensins. To distinguish peptides that differ in their COOH-terminal sequence, samples are extracted in guanidine thiocynate, acetylated, subjected to HPLC separation, and the fractions analyzed by the different NH2-terminally-specific RIAs. An advantage of the approach is that all peptides that share the identical NH2-terminus [i.e., ANG I, ANG II, and ANG-(1–7)] are assayed by one RIA antibody rather than multiple assays, and all reported values reflect HPLC fractionation. Disadvantages of the HPLC-RIA approach are the required equipment, expertise, and the lack of an automated method for analysis and that HPLC separation generates a significantly larger sample size for RIA or ELISA quantification. Moreover, the NH2-terminal RIAs are not commercially available and cannot be applied for direct peptide measurements. Finally, it is not clear whether HPLC-RIA can resolve [Ala1]-isoforms or other posttranslational modifications of ANG II and ANG-(1–7).
RIA and ELISA exhibit the requisite sensitivity and specificity to detect endogenous angiotensins; however, these methods are highly dependent on the characteristics for each antibody, as well as the optimal sample extraction and assay conditions. Angiotensin RIAs or ELISAs are typically directed against the unique COOH-terminus of the peptide, and they fail to distinguish the NH2-terminal metabolites unless coupled to HPLC. LC-MS/MS directly detects peptides based on their unique mass to charge spectra and exhibits the capability for quantification of endogenous peptides (83). Unlike the antibody-based RIA or ELISA, MS can potentially resolve posttranslational events such as peptide phosphorylation, amidation, acetylation, decarboxylation, or other peptide forms. Extracted samples are initially resolved by a chromatographic step [HPLC, ultrahigh-pressure LC (UPLC), or nano-LC], the eluent volatized, and the peptide or resultant fragment ions detected by MS [single or tandem, time of flight (TOF), ion trap, or their combination]. A key issue with MS is that a number of components are detected following the LC separation and may render very complex spectra with similar mass-to-charge ratios, particularly as the prior chromatographic step does not resolve all individual peptides, and endogenous peptides are typically 103-109 less abundant than other species in plasma and tissues (11, 83). To address this issue, Schulz et al. (125) developed a hybrid approach to selectively enrich ANG II from plasma extracts by absorption to an ANG II antibody before quantification by HPLC-MS/MS. These investigators quantified ANG II at ∼20 fmol/ml (20 pM) in human plasma that is comparable wth RIA or HPLC-RIA values, although other NH2-terminal metabolites of ANG II were not reported (125) (Table 1). Addition of isotopically labeled ANG II is key in the MS approach to account for extensive sample handling and extraction, as well as the extent of peptide ionization. Poglitsch and colleagues (60, 108) have recently developed a “fingerprint” HPLC-MS/MS approach to quantify 10 angiotensins in plasma with a comparable sensitivity to RIA; however, the addition of exogenous renin is apparently required for peptide measurements in human plasma. HPLC-MS/MS may potentially become the standard method to quantify angiotensins, although the complexity, expertise, and cost required for these systems are limiting factors for most laboratories at the current time. The fingerprint MS is available as a fee for service to quantify angiotensins in plasma and tissues (Attoquant, Vienna, Austria).
Additional methods to quantify angiotensin peptides in plasma and tissues include direct HPLC and capillary electrophoresis. Both methods are based on the direct determination of angiotensins by ultraviolet absorption (UV, 180–220 nm) following an initial chromatographic or electrophoretic separation. The concerns here are that neither approach achieves complete separation of all peptides (angiotensin and non-angiotensin), and UV detection does not discriminate among the peptide milieu in contrast to RIA, ELISA, or MS methods. Therefore, a UV peak may be comprised of multiple species, and alterations in peak area between experimental groups cannot be readily attributed to one peptide. Moreover, the reported sensitivity or limit of quantitation of 3 pmol for the HPLC analysis is not adequate to detect authentic levels of endogenous peptides (16). ANG II and ANG-(1–7) values derived from these methods are typically far higher in plasma (26, 33, 51, 135), kidney (33, 110, 114), vasculature (48), adipose tissue (33, 115), and heart (42) than by RIA, HPLC-RIA, or HPLC-MS methods (Table 1). Indeed, two recent studies quantified mouse renal ANG-(1–7) at >1,000,000 fmol/g tissue (110) and rat plasma ANG II at >300,000 pM by HPLC analysis alone (33). Rubio-Ruiz et al. (115) reported an ANG-(1–7) concentration of 6,000 pM in serum and >500,000 fmol/g in adipose tissue from male Wistar rats by capillary electrophoresis.
Sample handling and extraction.
Appropriate handling and subsequent extraction of plasma and tissue samples cannot be overly emphasized for the accurate measurement of endogenous angiotensins. Regardless of the detection method by RIA, HPLC-RIA, or HPLC-MS, improper sampling handling and extraction may yield artifactually high or low peptide values that do not reflect endogenous content. The sensitivity required to quantify endogenous angiotensins also increases the potential for artifacts or interfering substances whether using RIA or MS approaches. In general, animals should be maintained in a calm environment and anesthetic agents such as pentobarbital sodium that markedly activate the RAS avoided. Blood samples are collected directly into an inhibitor cocktail to block renin and other peptidases that process or metabolize angiotensin peptides; the obtained plasma should be subsequently stored at −80°C or immediately extracted. Tissue samples should be rinsed of blood and rapidly frozen in liquid nitrogen or an ethanol-dry ice bath or immediately extracted. Pepstatin, an aspartyl protease inhibitor, may not be sufficient to completely block renin activity, and a more specific renin inhibitor should be considered (73, 98). Moreover, the typical inhibitor cocktails for protein immunoblot analysis may be inadequate to attenuate angiotensin processing of samples in physiological buffers.
Various extraction methods that rapidly abolish peptide metabolism, deplete the sample of contaminating proteins and other substances that may interfere with the assay and enrich peptide content in a manageable volume include acid (HCl)-ethanol, pure methanol or acetonitrile, guanidine thiocynate, and immediate immersion in a boiling water bath with subsequent acidification and tissue homogenization (22, 24, 30, 37, 90, 130). The precipitated proteins are removed by high-speed/ultracentrifugation or molecular cut-off filters, and the supernatant is applied to solid-phase extraction columns (C18 or phenyl). Both angiotensin and nonangiotensin peptides are absorbed to these columns, whereas proteins and other substances are not retained; the peptides are subsequently eluted in an organic solvent (methanol or acetonitrile). Peptides standards (i.e., 125I-ANG II, ANG II, or isotopically labeled ANG II) should be routinely added to determine overall recovery in the extraction procedure for quantification methods, and particularly to account for matrix effects on MS ionization that may vary among samples or MS analysis runs. Recovery for plasma samples typically ranges from 70 to 90%, whereas tissues are generally lower in the 40–70% range. Appropriate blanks must be included that contain all components of the extraction method (apart from the actual sample) and assayed in parallel with samples.
The range of angiotensin values for plasma and tissues in Table 1 reflects studies using HPLC-RIA or HPLC-MS quantification, as well as consideration of the appropriate sample collection and extraction methods (22–24, 37, 85, 97–100, 127, 128). In lieu of these approaches, the NH2-terminally intact forms of ANG II, ANG-(1–7), and ANG I are the predominant peptides identified in plasma and various tissues. Quantitation of the cell and urinary data reflects direct RIA- or ELISA-based methods (9, 34, 38, 45, 50, 70, 77, 115, 131, 133, 139, 141, 148).
ANG II values in human, mouse, and rat plasma are quite low and average 5–50 fmol/ml (5–50 pM) (Table 1). Human plasma levels of the ANG II isoforms [Ala1]-ANG II and [Pro1, Glu2]-ANG II average 6 and 4 pM, respectively, as quantified by MALDI (66, 67). Circulating ANG-(1–7) values exhibit a similar range to ANG II, although the peptide is markedly increased following ACE inhibitor treatment, whereas endogenous levels of the ANG-(1–7) isoform [Ala1]-ANG-(1–7) have not been quantified (Table 1). ANG I values are somewhat higher than ANG II, provided renin activity is attenuated during the sample collection and processing, whereas ANG-(1–9) levels are very low (Table 1). Adequate plasma volumes to detect these peptides in this range require 3 to 5 ml of extracted plasma (4–10 ml blood) to assay.
In regard to the MS analysis of circulating angiotensins, the peptide profile reported by Poglitsch and colleagues (60, 108) is difficult to interpret as addition of exogenous renin to human plasma samples is apparently necessary to quantify circulating angiotensin peptides. This approach does not reflect the actual angiotensin content in the circulation as renin addition fails to account for the contribution of membrane-bound peptidases (i.e., Nep, ACE2, ACE, APA) nor the vascular release of angiotensin peptides (31, 63, 87). Wysocki et al. (147) recently evaluated plasma angiotensins from male WT mice by the fingerprint MS method without the addition of renin. Surprisingly, plasma levels of ANG III greatly exceeded ANG II (1,130 vs. 7 pM), whereas similar levels of ANG I and ANG-(2–10) (1,000 and 700 pM) were reported (147). Little detail was provided on the extraction method or inhibitors employed, and the high plasma content of ANG III and ANG-(2–10) may reflect inadequate blockade of aminopeptidase activity particularly given the use of the anesthetic pentobarbital (147). Indeed, their additional analysis of plasma peptides from female WT mice anesthetized with ketamine revealed far lower plasma ANG III, although the levels of ANG III were twofold higher than ANG II (160 vs. 74 pM) (147). Olkowicz et al. (102) quantified plasma angiotensin by MS in male WT mice and in apolipoprotein E−/− mice that exhibit a stimulated RAS. Plasma levels of ANG II, ANG III, and ANG-(1–7) were 36, 32, and 63 pM, respectively; however, ANG IV (158 pM) was the predominant peptide in WT mice (102). ANG II and its COOH-terminal intact metabolites were elevated 3- to 10-fold in the apolipoprotein E−/− mice, but ANG IV remained the major peptide (550 pM); plasma levels of ANG-(1–7) were unchanged (102). Although the ANG II and ANG-(1–7) values were comparable with RIA-based studies, the higher ANG III and ANG IV content may again reflect inadequate inhibition of aminopeptidases by addition of a standard protease inhibitor cocktail in contrast to extraction methods specifically designed to inhibit peptide processing (102). Finally, Ali et al. (7) quantified plasma ANG II and ANG-(1–7) content at 20,000 and 40,000 pM, respectively, in WT mice using UPLC-quardruple (Q)TOF. The presence of other ANG II or ANG-(1–7) metabolites were not reported, but the extreme ANG II and ANG-(1–7) values by this method likely reflect inadequate inhibition of plasma peptidases during blood collection and extraction (7).
Tissue expression of angiotensins is widespread; however, there are marked differences in the tissue content of these peptides (Table 1). The adrenal typically expresses the highest content of ANG II from 300 to 2,000 fmol/g tissue with far lower levels of ANG III and ANG-(1–7) (Table 1). ANG II content in other peripheral tissues including the kidney, lung, and liver average of 100 to 700 fmol/g (Table 1). Within the rat kidney, ANG II, ANG-(1–7), and ANG I levels were higher in the medulla than the cortex (direct RIAs) (91, 104); Pendergrass et al. (104) confirmed the predominant ANG II identify in both cortex and medulla by HPLC-RIA. In comparison, significant levels of ANG II (∼300 fmol/g tissue) were quantified in the ovary and uterus by Senanayake and colleagues (37). Apart from these tissues, central and peripheral tissues including the heart and adipose express lower and somewhat more variable ANG II levels of 2–150 fmol/g tissue (Table 1).
In regards to MS analysis, few studies have quantified ANG II and ANG-(1–7) in tissue. Ali et al. (7, 8) reported ANG II and ANG-(1–7) levels of 200,000 and 350,000 fmol/g in mouse kidney and 25,000 to 60,000 fmol/g tissue in rat kidney cortex by UPLC-QTOF (7, 8). They also quantified ANG II and ANG-(1–7) content in mouse adipose tissue at 150,000 and 350,000 fmol/g tissue, respectively (7). It is difficult to reconcile that the kidney and adipose exhibit the same synthetic capacity to generate essentially identical levels of ANG II or ANG-(1–7). Moreover, extraction of tissue with a very high protease content such as the kidney in a Tris lysis buffer at neutral pH is likely insufficient to abolish processing and may contribute to the high peptide content reported for kidney (7, 8). The stated limit of quantitation for the UPLC-QTOF for ANG II (500 fmol) and ANG-(1–7) (800 fmol) also does not appear adequate to quantify the authentic peptide content in plasma and tissues (7). In contrast to the Ali studies, Wysocki et al. (147) quantified far lower levels of ANG II and ANG-(1–7) (293 and 95 fmol/g tissue, respectively) in the mouse kidney by fingerprint MS; these values are clearly more consistent with the literature, although the ANG III content (185 fmol/mg tissue) was relatively high at 60% of the total ANG II (Fig. 5, Table 1). Finally, Burger et al. (19) found a higher ANG II-to-ANG III ratio (700 vs. 130 fmol/g tissue) in mouse kidney using guanidine extraction and MS. In this case, guanidine may attenuate aminopeptidase and other peptidase activities more effectively to prevent the artifactual generation of peptides during tissue extraction.
The extracellular or interstial content of ANG II in heart and kidney was assessed by direct RIA and HPLC-RIA. These studies use an implanted dialysis probe permeable to low molecular substances that sample the local release or formation of angiotensin peptides in the interstial space. Dell'Italia and colleagues (14, 36) reported interstial ANG II levels in the canine heart at 0.6 to 5 pmol/ml (0.6–5 nM) by HPLC-RIA. Intravenous infusion of ANG I alone or combined with an ACE inhibitor did not impact the interstial ANG II levels. Schuijt et al. (124) found low to undetectable levels of ANG II (<30 fmol/ml or 30 pM) in interstial fluid of the ovine heart also quantified by HPLC-RIA. In the rat kidney, Kemp et al. (68) report comparable ANG II and ANG III interstial content (∼80 fmol/ml or 80 pM) using the HFBA-HPLC system to separate the two peptides before RIA detection. This study finds a higher ANG III-to-ANG II ratio than that in the circulation or tissues, which may reflect the significant level of ectocellular APA activity in the renal tubules and suggests significant angiotensin processing in the interstial space. Finally, Nishiyama et al. (95, 96) reported higher renal interstial levels of ANG II and ANG I (1–3 nM) by direct RIAs. The interstial ANG II levels were not responsive to peripheral ACE inhibition or volume expansion, further suggesting a local mechanism of peptide expression distinct from the circulation, albeit the identity of ANG II or ANG I immunoreactivity was not established in these studies (95, 96).
The urinary components of the RAS are of interest regarding their potential application as biomarkers of renal injury or RAS activation (5, 72, 113). Although excretion of the RAS proteins including angiotensinogen, renin, and ACE2 are more routinely assessed, the urinary levels of ANG II and ANG-(1–7) were quantified by direct RIA or ELISA (Table 1). Urine contains multiple peptidase activities, and the samples should be collected under conditions that limit ex vivo metabolism (acidification, addition of inhibitors, or collection on dry ice) (50). Chan and colleagues (131) reported higher ANG II excretion in the diabetic Akita mice compared with WT (∼25 vs. 2 pmol/mg Cr), whereas urinary ANG-(1–7) levels were 50% lower in the diabetic mice (3 vs. ∼7 pmol/mg Cr), suggesting an imbalance of ANG II to ANG-(1–7) in the diabetic kidney; these peptides were quantified by Bachem ELISAs (Torrance, CA). Gilliam-Davis et al. (50) quantified ANG II (0.4 pmol/mg Cr) and ANG-(1–7) (0.2 pmol/mg Cr) in adult male Fisher 344 rats by direct RIAs. Both ANG II and ANG-(1–7) excretion increased twofold in aged Fisher rats, suggesting activation of the renal RAS in this model of aging. Chronic treatment with an AT1R antagonist lowered the excretion of both peptides despite a marked increase in circulating angiotensins and further distinguishes compartmentalization of the renal and circulating RAS (50). Urinary levels of both ANG II and ANG-(1–7) were elevated in female and male mRen2.Lewis on the HS diet; however, ovariectomy of the hypertensive females did not increase peptide excretion despite a 20-fold increase in urinary angiotensinogen (Fig. 2) (34). Finally, ANG-(1–7) excretion was significantly lower in a cohort of hypertensive patients compared with the normotensive group (0.05 vs. 0.11 pmol/mg Cr) using a custom RIA to quantify ANG-(1–7); ANG-(1–7) identity in human urine was verified by HPLC-RIA (45).
In an extensive analysis of angiotensin regulation in rat VSMCs using direct RIAs (Peninsula), Lavrentyev et al. (77) reported that ANG II content increased threefold (∼50 to 150 fmol/mg protein) in response to high glucose conditions. ANG II content also increased in the cell media with high glucose (100 to190 fmol/ml). In contrast, basal ANG-(1–7) (260 fmol/mg protein) was higher than ANG II, and ANG-(1–7) content was reduced fivefold under hyperglycemic conditions that was also associated with a reduction in ACE2 (77). In a human mesangial cell line, Boim and colleagues (141) reported comparable ANG II levels (150 fmol/mg protein), as well as secreted ANG II in the media (30 fmol/mg protein after 24 h) using a Peninsula ELISA. Basal ANG II levels were also similar (180 fmol/mg protein) in immortalized mouse podocytes, and peptide content increased with mechanical strain (38). Singh and colleagues (132) quantified ANG II content at ∼200 fmol/mg protein in cultured neonatal cardiomyocytes and fibroblasts; the cellular ANG II levels were also markedly elevated by high glucose conditions and exhibited a greater nuclear localization for ANG II. ANG II content was determined by a direct Peninsula ELISA, and the relative peptide expression in control and high glucose-exposed cardiomyocytes was confirmed by antibody capture and HPLC-MS (132). Finally, we assessed the effect of advanced glycation end products on angiotensin expression in renal NRK-52E cells (9). Basal cellular content was quantified by direct RIAs for ANG I (85 fmol/mg protein; Peninsula), ANG II (150 fmol/mg protein; ALPCO), and ANG-(1–7) (400 fmol/mg protein; custom RIA) (9). Advanced glycation end product exposure significantly reduced ANG-(1–7) and increased metabolism of the peptide that was associated with myofibroblast transition but did not impact ANG II or ANG I levels (9). As these cells express an intracellular RAS, nuclear content of ANG II and ANG-(1–7) was quantified at ∼60 fmol/mg protein consistent with nuclear immunostaining for both peptides (10).
In review of the cell studies, angiotensin content in various cell types is ∼200 fmol/mg protein (Table 1). The high angiotensin levels compared with tissues likely reflect the lower cellular protein to normalize peptide content, as well as the greater density of angiotensin-expressing cells maintained in culture; however, direct comparisons of angiotensins quantified by RIA (or ELISA) to HPLC-RIA and LC-MS are lacking, and the true cellular content of ANG II or ANG-(1–7) in these cell models is not known. A second issue with cell studies is the presence of circulating RAS components in fetal bovine serum (10). Cells can be maintained in low or serum-free conditions for short periods, but internalized angiotensinogen or renin may persist and contribute to the intracellular expression of angiotensin peptides, particularly if nonrenin proteases that can process the precursor (cathepsins, tonin) are also expressed. Furthermore, bovine angiotensinogen contains an Ile5 substitution for Val5 in the NH2-terminal portion of the protein, and RIAs will not distinguish these angiotensin isoforms unless samples are initially fractionated by HPLC or characterized by LC-MS (130). ANG II also undergoes internalization through both AT1R-dependent and -independent pathways that may further contribute to the cellular content of the peptide (40, 92, 140, 150). Indeed, the interaction between the circulating RAS components and the cellular or tissue system is likely a dynamic process in vivo that should be examined to a greater extent in cell studies. Finally, the subcellular compartmentalization of angiotensin peptides and other components is becoming increasingly evident (1, 2, 25, 35, 40, 54, 74, 140, 151). Apart from establishing the functional role of these intracellular systems, a greater emphasis should be placed on the quantification of angiotensins in cellular compartments, and particularly the extent that altered peptide content in tissue reflects discrete changes in the subcellular systems such as the nucleus and mitochondria.
The current review assessed the biochemical approaches to quantify the multiple components of the RAS, as well as attempt a consensus on expected values of endogenous angiotensin peptides in the circulation, tissues, and cells. In regard to the peptidase and receptor components of the RAS, an optimal approach should reflect some measure of the activity of these proteins, although it is acknowledged that the receptor expression is an order of magnitude lower than that of other RAS proteins and presents a more difficult challenge for receptor characterization This same issue holds for true for the quantification of angiotensin peptides given their very low abundance; optimal approaches should reflect a highly sensitive and specific method coupled with relatively high throughput to evaluate large sample sets. The LC-MS/MS approach appears to be the ideal choice to quantify angiotensin peptides; however, the required expertise and overall expense of these systems limits their widespread availability to a limited number of laboratories at the current time. Moreover, the application of MS to the determination of angiotensins does not necessarily guarantee an accurate profile or an absolute value in plasma and tissues. Thus RIA or ELISA methods will continue to be the predominant approach to quantify angiotensin peptides and assess alterations in peptide expression. The limitations of these assays, as well as the methods applied for sampling handling, extraction, and approaches for validation, must be considered for the accurate assessment of the peptide hormones of the RAS, as well as to establish that an altered phenotypic or treatment response truly reflects changes in peptide expression. Importantly, validation of these assays for either individual samples or a sample pool by HPLC-RIA should be a requirement for the analysis of endogenous angiotensin peptides in various compartments, and particularly in studies that assess the cellular expression of angiotensins.
This study was supported in part by National Institute of Health Grants HL-56973, HL-51952, HD-084227, HD-047584, and HD-017644 and American Heart Association Grants AHA-151521 and AHA-355741. An unrestricted grant from the Farley-Hudson Foundation (Jacksonville, NC), Groskert Heart Fund, and the Wake Forest Venture Fund is also acknowledged.
No conflicts of interest, financial or otherwise, are declared by the author(s).
M.C.C. drafted, edited, revised, and approved final version of manuscript.
I gratefully acknowledge K. Bridget Brosnihan and Debra I. Diz of the Hypertension and Vascular Research Center for their critical reading and insightful comments pertaining to this review.
- Copyright © 2016 the American Physiological Society