Department of Medicine, University of Melbourne, Austin Health, Melbourne, Victoria, Australia
Correspondence: Professor Louise M Burrell, Department of Medicine and Cardiology, University of Melbourne, Austin Health, Level 7, Lance Townsend Building, 145 Studley Road, Heidelberg, Vic. 3084, Australia. Email: firstname.lastname@example.org
The renin–angiotensin system plays a major role in the pathophysiology of cardiovascular disease (CVD). The enzyme angiotensin-converting enzyme (ACE) converts angiotensin (Ang) I into the vasoconstrictor AngII and was thought, until recently, to be the main effector of the system.
The enzyme ACE2, discovered in 2000, can counterbalance the effects of ACE through degradation of AngII and generation of Ang-(1–7). Angiotensin-converting enzyme 2 is abundantly expressed in the heart and localized to the endothelial cells of coronary vessels and smooth muscle cells. Its catalytically active ectodomain undergoes shedding, resulting in ACE2 in the circulation.
There are 10 studies to date that have measured circulating ACE2 activity in humans, including in healthy subjects and those with heart failure, Type 1 diabetes, implantable cardioverter/defibrillator, elderly subjects undergoing emergency orthopaedic surgery and kidney transplant patients. The results suggest that circulating ACE2 activity may be a marker of CVD, with low levels in healthy individuals and increased levels in those with cardiovascular risk factors or disease. Whether increased plasma ACE2 activity reflects increased synthesis from tissue ACE2 mRNA or increased shedding of tissue ACE2 remains to be determined.
Angiotensin-converting enzyme 2 is located on the X-chromosome and circulating ACE2 levels are higher in men than in women.
Large clinical studies in CVD are needed to more precisely clarify the role of ACE2 as a biomarker of CVD, determine the prognostic significance of circulating ACE2 activity and assess whether the measurement of ACE2 will improve CVD risk prediction.
Globally, cardiovascular disease (CVD) is the leading cause of death and disability, due to coronary artery disease (CAD), hypertension, heart failure and stroke. Despite a decline in mortality in developed countries in recent decades, CAD remains one of the largest causes of mortality worldwide. The discovery of biomarkers to improve CVD risk prediction has been a rapidly expanding area. The identification of new biomarkers may help increase our understanding of the pathogenesis of CVD and lead to improved risk prediction and development of novel therapeutic targets.
“Does ACE2 activity predict cardiovascular risk?”
The renin–angiotensin system (RAS) is a major hormonal system involved in the pathophysiology of CVD.[2, 3] Within the RAS, the enzyme angiotensin-converting enzyme (ACE) converts angiotensin (Ang) I into the vasoconstrictor AngII, which mediates its effects predominantly through angiotensin type 1 receptors. Angiotensin II increases blood pressure through vasoconstriction and salt and water retention, and contributes to cardiac remodelling, fibrosis, inflammation, thrombosis and plaque rupture.
In 2000, ACE2 was discovered by two independent groups, who cloned it from a human cardiac left ventricle cDNA library and a human lymphoma cDNA library. Angiotensin-converting enzyme 2 acts as a monocarboxypeptidase and removes a single C-terminal amino acid from AngI to generate Ang-(1–9). However, the preferred substrate of ACE2 is AngII, from which ACE2 cleaves the C-terminal amino acid to generate the peptide Ang-(1–7),[4, 6] a ligand for the Mas receptor, which is reported to have vasodilatory and antifibrotic actions.[8, 9] In this way, ACE2 limits the vasoconstrictor action of AngII through its degradation and counteracts the actions of AngII through the formation of Ang-(1–7). The substrate-binding pockets of ACE and ACE2 are significantly different, explaining why ACE inhibitors cannot bind and inhibit the activity of ACE2.
Angiotensin-converting enzyme 2 has been identified in many tissues[12-20] and is expressed in abundance in the heart,[4, 21, 22] blood vessels[23-25] and kidney.[5, 18] The aim of the present review is to discuss the measurement of ACE2 activity and the evidence regarding circulating ACE2 activity in the 10 studies published to date in healthy subjects, as well as in those with cardiovascular (CV) risk factors and disease.
“Tissue ACE2 undergoes shedding into the circulation”
Biochemistry and Regulation of ACE2
Ectodomain shedding of membrane-bound ACE2
Angiotensin-converting enzyme 2 is an 805 amino acid, Type 1, integral membrane protein consisting of a large N-terminal extracellular catalytically active ectodomain containing a single zinc-binding site with an HEXXH motif, a transmembrane region and a short C-terminal cytoplasmic tail.[4, 5, 26] Angiotensin-converting enzyme 2 shares approximately 42% homology with the N-terminal ectodomain of ACE and its active catalytic ectodomain is exposed to circulating vasoactive peptides in the circulation. The carboxyl end of ACE2 is homologous to collectrin a protein that has been shown to regulate renal amino acid uptake. More recently, Hashimoto et al. demonstrated that ACE2 is a key regulator in the uptake of neutral amino acids in the intestine and that genetic deletion of ACE2 in mice results in severe colitis following intestinal injury. Angiotensin-converting enzyme 2 undergoes cleavage or ‘shedding’ to release the catalytically active ectodomain into the extracellular milieu (Fig. 1). The shedding process involves the proteinase ADAM17 (a disintegrin and metalloproteinase), also known as tumour necrosis factor-α-converting enzyme (TACE),[30, 31] and can be regulated by various stimuli, including phorbol esters, calcium ionophores, growth factors and calmodulin.[30, 31] The shedding mechanism results in ACE2 that can be detected in the circulation. Western blot analysis in healthy subjects has shown plasma ACE2 protein to be of a smaller fragment size compared with full-length membrane-bound ACE2, thus likely to be the result of proteolytic cleavage.
The cleavage site for ADAM17-mediated ectodomain shedding of human ACE2 has been reported to occur between amino acid positions 716 and 741, which reside in the juxtamembrane region. The use of several mutant and chimeric ACE2 proteins has shown that the juxtamembrane stalk region, transmembrane and cytoplasmic domains were not required for constitutive ACE2 shedding. Another group used recombinant human ADAM17 and showed that it was able to cleave an ACE2 peptide mimetic that corresponded to the extracellular juxtamembrane region between the residues Arg708 and Ser709 of ACE2. Furthermore, deletion of the juxtamembrane coding region of ACE2 in a mutant cell line attenuates ADAM17-induced ACE2 shedding. Iwata et al. also showed the shedding of two distinct soluble forms of ACE2 with equivalent enzyme activities from cell lines overexpressing ACE2. The deglycosylated molecular masses of the two soluble forms were approximately 80 and 70 kDa. Furthermore, ADAM17 was responsible for shedding of the larger 80 kDa soluble form, but not the constitutive cleavage of the smaller soluble fragment, suggesting the involvement of one or more sheddases. This finding has been confirmed by Jia et al., who reported the shedding of two distinct ACE2 soluble forms in primary cultures of human airway epithelial cells. Further studies are needed to characterize the mechanisms involved in ACE2 shedding and to identify the involvement of other sheddases and the actual cleavage sites involved.
We have reported previously that ACE2 is present in both normal human vasculature and in diseased internal mammary and radial arteries taken from patients undergoing coronary artery bypass surgery. Others have shown that human carotid atherosclerotic lesions express ACE2 mRNA, with increased ACE2 tissue activity in ruptured atherosclerotic lesions compared with stable advanced atherosclerotic lesions. There are no studies that have simultaneously measured tissue and circulating ACE2 activity in humans, but it is possible that an increase in tissue ACE2 activity will result in increased circulating ACE2 activity. Certainly in the rat myocardial infarction (MI) model, we have shown that the increase in cardiac ACE2 after MI is associated with an increase in plasma ACE2 activity. We have also reported similar results in a rat model of acute renal injury, with increased cardiac ACE2 associated with increased circulating ACE2 activity.
Thus, ACE2 undergoes shedding to release the catalytically active ectodomain into the circulation, which involves the proteinase ADAM17. Further studies are needed that measure both tissue and circulating ACE2 activity in humans to determine whether elevated circulating ACE2 activity is associated with increased tissue ACE2 synthesis from mRNA and/or increased ACE2 shedding from tissue.
Human Studies of Circulating ACE2
Measurement of ACE2
There are only 10 studies that have measured circulating ACE2 activity in plasma or serum from human subjects. These studies are summarized in Table 1 and include healthy subjects, subjects with heart failure, Type 1 diabetes, implantable cardioverter/defibrillator (ICD), elderly subjects undergoing emergency orthopaedic surgery and kidney transplant subjects.[32, 38-46] Not all the published studies include a control group and, to date, there are no studies of ACE2 in subjects with other CVD, such as Type 2 diabetes, essential hypertension, hyperlipidaemia, CAD or MI.
Table 1. Summary of studies of human circulating angiotensin-converting enzyme 2 activity
Median ACE2 activity (IQR) 15.9 (8.4–26.1), 9.2 (3.9–21.9) and 13.1 (5.7–21.9) pmol/min per mL for predialysis, dialysis and KT patients, respectively
All published reports on circulating ACE2 in humans have used an ACE2-specific quenched fluorescent substrate (QFS), based on the initial design of Vickers et al., who demonstrated specific hydrolysis of this substrate with recombinant human ACE2. The QFS-based assays are widely used for the detection of proteolytic activity and consist of a short synthetic peptide sequence that is specifically recognized and cleaved by the enzyme of interest and flanked by a quencher and fluorophore (Fig. 2). All studies,[32, 38-41, 43-46] apart from that of Soro-Paavonen et al., have used the ACE2 QFS (7-methoxycoumarin-4-yl)acetyl(Mca)-Ala-Pro-Lys-2,4-dinitrophenyl (Dnp). Angiotensin-converting enzyme 2 cleaves the Pro-Lys bond and fluorescence of the released Mca-Ala-Pro is measured (Fig. 2). This is a specific substrate for ACE2 that has also been widely used in animal studies.[36, 37, 48-51] In contrast, the substrate used by Soro-Paavonen et al. (o-aminobezoic acid-Ser-Pro-Tyr(NO2)-OH) has only been used in one other animal study and no validation studies have been reported. In all but three studies,[38, 44, 45] the assay sample is set up in parallel with an ACE2-specific inhibitor to determine non-specific activity and the relative fluorescence of the inhibited sample is then subtracted from the non-inhibited sample. The actual amount of ACE2 enzyme activity in the assay is then determined from a standard curve of relative fluorescent units plotted against known concentrations of either the fluorophore[32, 43] or recombinant ACE2 enzyme.[38-42] The inhibitors used in the ACE2 assays include MLN-4760, DX600[39-41] and EDTA.[42, 43, 46] Although MLN-4760 has been shown to be the most potent inhibitor of ACE2, it is not readily available commercially. Human ACE2 has been shown to be more sensitive to DX600 compared with rat or mouse ACE2, and Epelman et al. demonstrated a dose-dependent reduction in ACE2 activity with DX600. It has been shown that EDTA is as effective as MLN-4760 in rat tissue, and EDTA was initially used by Vickers et al. to quench the assay reactions.
“ACE2 activity is measured using an ACE2-specific QFS”
The published ACE2 assays have used modifications in the assay conditions that include differing temperatures and lengths of incubation time of the ACE2 QFS with the plasma and/or serum sample and different units for quantifying ACE2 activity. All studies report the assay units as ACE2 activity (either mol ACE2/min per L or mL), except for the studies of Lehmann et al. and Soler et al., which did not use an ACE2 inhibitor or a standard curve to convert the relative fluorescent units into ACE2 activity concentrations. We have reported that ACE2 circulates in human plasma, but its activity is masked by the presence of an endogenous inhibitor. The addition of neat plasma to recombinant soluble human ACE2 inhibited ACE2 activity in a dose-dependent manner, whereas removal of the endogenous inhibitor using anion-exchange chromatography allowed detection of plasma ACE2 activity using a QFS assay. Partial purification of the inhibitor suggested it was small, hydrophilic and cationic. The inhibitor was present in similar amounts in plasma samples from healthy volunteers because mean (±SEM) inhibition of recombinant ACE2 by 15 μL plasma was 37.7 ± 1.3% (n = 17). Hence, in healthy volunteers, the difference in ACE2 activity does not reflect a difference in ACE2 inhibitor levels. The studies of Chong et al. and Roberts et al. and data in abstract form in CAD subjects also reported removal of the endogenous inhibitor before the assay of plasma ACE2 activity. Roberts et al. determined whether the endogenous inhibitor was affected by reduced kidney function and compared ACE2 plasma activity in a subset of their patient groups (predialysis chronic kidney disease (CKD) n = 22; dialysis n = 17; transplant n = 22) with and without extraction of the endogenous inhibitor. They found that ACE2 plasma activity was substantially lower in plasma assayed without extraction of the endogenous inhibitor, but there was no difference in ACE2 plasma activity between the groups with or without removal of the endogenous inhibitor in this subset. Therefore, endogenous inhibitor levels appear not to be affected by reduced renal function, although further studies will need to be conducted to determine whether other CV risk factors or diseases affect endogenous inhibitor levels.
“Circulating ACE2 is higher in men than women”
“Plasma ACE2 activity is low in healthy individuals”
Only six of the 10 studies have provided assay reproducibility data.[32, 38-40, 45, 46] Of these, four studies had good intra- and interassay coefficients of variation, as follows: 3.7% and 8.8%, respectively, in the study of Rice et al.; 7.1% and 13.7%, respectively, in the study of Lew et al.; 4.7% and 6.2%, respectively, in the study of Soler et al.; and 5.6% and 11.8%, respectively, in the study of Roberts et al. Two studies[39, 40] reported a large interassay coefficients of variation (mean (±SD) 20.6 ± 5.5%). Thus, differences in assay conditions, measurement units and lack of reproducibility data make it difficult to compare circulating ACE2 activity across the subject groups.
Circulating ACE2 in healthy subjects
The first study to assess plasma ACE2 activity in humans was The Leeds Family study, which consisted of healthy subjects and family members totalling 534 individuals. Plasma ACE2 activity was detectable in 40 subjects who tended to be older, with greater abdominal adiposity and higher blood pressure, fasting glucose and lipid levels compared with those with undetectable ACE2 levels. In these subjects with CV risk factors, plasma ACE2 levels ranged from 3 to 460 pmol/L (overall mean 33.0 pmol/L; 95% confidence intervals (CI) 22.1–49.4 pmol/L). Furthermore, half the subjects with detectable ACE2 had at least one other family member with detectable ACE2, with up to 67% of the variability in circulating ACE2 levels being explained by hereditary factors. Using a similar QFS-based assay to Rice et al., we have reported that removal of the endogenous inhibitor allows ACE2 activity to be measured in plasma from healthy volunteers (n = 18); levels were low and lay within a tight range (mean (±SEM) 4.44 ± 0.56 pmol/min per mL; range 1.31–8.69 pmol/min per mL). Thus, circulating ACE2 levels are low in healthy subjects and increased in subjects with CV risk factors.
Gender and ACE2
The ACE2 gene is located on the X-chromosome and many of the genetic ACE2 studies have shown gender-specific associations (for a review, see Burrell et al.). This is of interest because the development of CVD is known to exhibit gender-specific characteristics. Of the 10 published studies, three have analysed circulating ACE2 activity according to gender.[42, 45, 46] These studies have shown increased ACE2 activity in men compared with women in healthy individuals and in subjects with Type 1 diabetes and renal disease.[45, 46] Future studies should analyse ACE2 activity according to gender and determine whether associations of ACE2 activity with CV risk or disease vary according to gender.
Coronary artery disease
We have reported previously that, in a group of high-risk elderly patients undergoing emergency orthopaedic surgery (n = 187), preoperative plasma ACE2 activity was similar between patients with an in-hospital CV event (n = 20; MI, heart failure, atrial fibrillation, major arrhythmia, cardiac arrest) and those with no CV event. However, postoperative plasma ACE2 activity was significantly increased in those with an in-hospital CV event. Although plasma ACE2 activity was not a significant predictor of CV events in multivariate analysis, in receiver operating characteristic (ROC) analysis, levels above the cut-off of 20.3 pmol/min per mL (area under the curve (AUC) = 0.68; P < 0.10) gave modest sensitivity and specificity for the prediction of in-hospital CV events. Further studies in large surgical patient cohorts are needed to confirm these findings. We have also reported (in abstract form) that in patients with suspected CAD, plasma ACE2 activity was significantly increased in those with angiographically defined CAD compared with those with normal coronary arteries, suggesting that ACE2 may be a useful biomarker of cardiac disease.
In a cross-sectional cohort study in subjects suspected of heart failure (n = 221), serum ACE2 activity strongly correlated with a clinical diagnosis of heart failure, left ventricular ejection fraction (LVEF) and increasing B-type natriuretic (BNP) peptide levels. Higher serum ACE2 activity also reflected the severity of heart failure according to the New York Heart Association function classification. Serum ACE2 activity was increased in heart failure of both ischaemic and non-ischaemic origin, as well as in those with heart failure and preserved LVEF. In a separate cohort of 113 patients with stable chronic systolic heart failure, those with a lower LVEF had higher ACE2 activity, and levels were correlated with higher plasma N-terminal (NT)-proBNP. Plasma ACE2 activity predicted the combined clinical end-point (occurred in 29%) of all-cause mortality, heart transplantation and heart failure hospitalization independent of LVEF and NT-proBNP levels. Similar results have been reported for patients with Chagas disease, a leading cause of heart disease in Central and South America. Plasma ACE2 activity was significantly higher in those with Chagas disease and heart failure compared with healthy controls and was predictive of cardiac death and heart transplantation.
As mentioned earlier, in the MI rat model of heart failure, increased cardiac ACE2 is reflected by increased plasma ACE2 activity. The results of increased circulating ACE2 in human heart failure are also consistent with our own study that reported ACE2 immunoreactivity is increased in explanted ischaemic failing human heart tissue. Others have shown that the ACE2 gene is upregulated in human idiopathic and ischaemic cardiomyopathy, and in myocardial biopsies from subjects with heart failure.
Increased plasma BNP and NT-proBNP levels are powerful independent predictors of morbidity and mortality in heart failure. The current ACE2 activity studies in heart failure subjects suggest a significant positive correlation between circulating ACE2 activity and BNP levels.[39-41] Epelman et al. reported that, in 113 heart failure subjects, ACE2 activity levels above the ROC-derived value of 28.3 ng/mL in combination with NT-proBNP levels above the median value of 1240 pg/mL (AUC = 0.78; P < 0.0001) predicted greater adverse events of all-cause mortality, heart transplantation and heart failure hospitalization compared with ACE2 activity (AUC = 0.66; P < 0.001) and NT-proBNP levels (AUC = 0.71; P < 0.0001) alone. In Chagas disease, the combination of ACE2 activity and BNP levels improved the predictive value of adverse cardiac events. Thus, the assessment of ACE2 activity in heart failure patients together with BNP levels may improve risk prediction in heart failure.
A recent study examined serum ACE2 activity in patients with decreased LVEF (≤ 35%) who had received an ICD and/or cardiac resynchronization therapy device (CRT) for primary prevention of sudden cardiac death. The study consisted of 57 patients, mostly men (93%), with ischaemic (n = 49) and non-ischaemic cardiomyopathy (n = 8) who had received ICD and/or CRT and were prospectively followed for subsequent arrhythmic episodes and appropriate ICD interventions over a mean follow-up time of 365 ± 90 days. Baseline serum ACE2 activity was significantly increased in patients who developed ventricular arrhythmias and had appropriate ICD intervention (n = 16) compared with subjects without arrhythmia (n = 41). Angiotensin-converting enzyme 2 was a significant univariate predictor of appropriate ICD intervention (P = 0.015). Interestingly, the proportion of subjects with a previous history of percutaneous coronary intervention was significantly higher in the ICD intervention group (63%) compared with the no ICD group (27%). Because increased ACE2 activity has been reported in coronary disease,[42, 56] the increased ACE2 levels in the ICD group may reflect existing coronary disease rather than ventricular arrhythmias. Serum samples for ACE2 measurement were based on a single sample at the time of the ICD surgery or a day before ICD implantation and not the ACE2 level at the time of the arrhythmic event. Larger studies are required to assess the prognostic value of plasma ACE2 levels in this group of patients, preferably with repeated sample collection to assess whether ACE2 levels increase with each arrhythmic episode and are independent of baseline coronary disease.
“Circulating ACE2 is increased in people with CVD”
Type 1 diabetes
In the largest clinical study to date of circulating ACE2 activity, there was no difference in serum ACE2 activity in men and women with Type 1 diabetes (n = 859) compared with healthy control subjects (n = 204). The study did suggest that serum ACE2 activity was sex dependent, with higher levels in men compared with women in both the control and Type 1 diabetes groups. Serum ACE2 was positively correlated with systolic blood pressure (SBP) and associated with coronary heart disease in both men and women, although no details were provided as to how the diagnosis of coronary heart disease was made and the numbers studied were small (35 men, 22 women). However, the results are consistent with our previously reported data in a non-diabetic population, namely that plasma ACE2 activity is increased in patients with angiographically proven CAD.
Cardiovascular disease is a major comorbidity associated with CKD. Our group and others have reported higher levels of ACE2 in men compared with women with renal disease.[45, 46] We have shown that plasma ACE2 activity is lower in patients undergoing haemodialysis than other forms of CKD. In dialysis patients, plasma ACE2 activity is lower in women than in men and was most strongly associated with SBP. In male dialysis patients, plasma ACE2 activity was most strongly associated with BNP levels. In Type 1 diabetes subjects, increased serum ACE2 activity was associated with the diabetic microvascular complication of microalbuminuria but not macroalbuminuria and only in men. In a longitudinal pilot study of serum ACE2 activity in 113 kidney transplant patients, ACE2 activity was detectable in kidney transplant patients and was increased in those with ischaemic heart disease compared with kidney transplant patients without ischaemic heart disease. In addition, ACE2 was increased in male compared with female subjects and was positively correlated with serum creatinine.
“The evidence linking ACE2 with CVD is increasing”
The evidence linking ACE2 with CVD is increasing, but data are limited to 10 studies, and differences in ACE2 assay conditions and lack of reproducibility data must be taken into account in the interpretation of the published data. The results from the 10 cross-sectional human studies do suggest that circulating ACE2 activity may be a marker of CVD, with low levels in healthy individuals and increased levels in those with CV risk factors or disease. Whether the increase in plasma ACE2 activity is a reflection of increased synthesis from tissue ACE2 mRNA or increased shedding of ACE2 from tissue remains to be determined. Data from three published studies suggest that circulating ACE2 levels are higher in men than in women. This issue has not been addressed in the study design of the majority of published reports, with only two studies[42, 46] performing analysis separately for men and women. Further studies examining the gender differences in circulating ACE2 levels are needed. Increased circulating ACE2 is associated with coronary heart disease, adverse CV outcomes in heart failure and with postoperative cardiac events in elderly orthopaedic surgery patients. In the future, large-scale, carefully conducted clinical studies are needed to more precisely clarify the role of ACE2 as a biomarker of CVD, to determine the prognostic significance of circulating ACE2 activity and to assess whether measurement of ACE2 will improve CVD risk prediction.
The authors’ work reported herein was supported by a grant from the National Health and Medical Research Council of Australia (ID1048285).