Acute kidney injury in the rat causes cardiac remodelling and increases angiotensin-converting enzyme 2 expression


  • L. Burchill, E. Velkoska and R. G. Dean contributed equally to this work.

Corresponding author L. M. Burrell: Department of Medicine, University of Melbourne, Austin Health, Heidelberg, Victoria 3081, Australia. Email:


Patients with kidney failure are at high risk of a cardiac death and frequently develop left ventricular hypertrophy (LVH). The mechanisms involved in the cardiac structural changes that occur in kidney failure are yet to be fully delineated. Angiotensin-converting enzyme (ACE) 2 is a newly described enzyme that is expressed in the heart and plays an important role in cardiac function. This study assessed whether ACE2 plays a role in the cardiac remodelling that occurs in experimental acute kidney injury (AKI). Sprague–Dawley rats had sham (control) or subtotal nephrectomy surgery (STNx). Control rats received vehicle (n= 10), and STNx rats received the ACE inhibitor (ACEi) ramipril, 1 mg kg−1 day−1 (n= 15) or vehicle (n= 13) orally for 10 days after surgery. Rats with AKI had polyuria (P < 0.001), proteinuria (P < 0.001) and hypertension (P < 0.001). Cardiac structural changes were present and characterized by LVH (P < 0.001), fibrosis (P < 0.001) and increased cardiac brain natriuretic peptide (BNP) mRNA (P < 0.01). These changes occurred in association with a significant increase in cardiac ACE2 gene expression (P < 0.01) and ACE2 activity (P < 0.05). Ramipril decreased blood pressure (P < 0.001), LVH (P < 0.001), fibrosis (P < 0.01) and BNP mRNA (P < 0.01). These changes occurred in association with inhibition of cardiac ACE (P < 0.05) and a reduction in cardiac ACE2 activity (P < 0.01). These data suggest that AKI, even at 10 days, promotes cardiac injury that is characterized by hypertrophy, fibrosis and increased cardiac ACE2. Angiotensin-converting enzyme 2, by promoting the production of the antifibrotic peptide angiotensin(1–7), may have a cardioprotective role in AKI, particularly since amelioration of adverse cardiac effects with ACE inhibition was associated with normalization of cardiac ACE2 activity.

Until recently, angiotensin-converting enzyme (ACE) was considered to be a key enzyme in the renin–angiotensin system (RAS), but this classical view of the RAS was challenged with the discovery of the enzyme ACE2 (Donoghue et al. 2000; Burrell et al. 2004). Angiotensin-converting enzyme 2 is expressed mainly in the heart and kidney, and its major role is to cleave angiotensin (Ang) II to the vasodilator and anti-fibrotic peptide, Ang(1–7) (Ferrario et al. 2002). There is increasing evidence that ACE2 may function to limit the vasoconstrictor action of Ang II both through Ang II inactivation and by counteracting the actions of Ang II through the formation of Ang(1–7) (Burrell et al. 2005).

The localization of ACE2 in the heart suggests an important role for cardiovascular function. Cardiac ACE2 increases after myocardial infarction (Burrell et al. 2005), and ACE2 knockout mice, which lack ACE2 protein (Crackower et al. 2002), have severely impaired cardiac contractility. It has also been shown that lentiviral delivery of ACE2 protects against hypertrophy and myocardial fibrosis induced by Ang II infusion in the rat (Huentelman et al. 2005). Taken together, the data suggest that ACE2 may play a role in the local control of cardiac structure and function in pathophysiological states (Grobe et al. 2006).

To date, the potential role of ACE2 in the cardiac complications of kidney failure has not been assessed, although patients with kidney failure frequently have left ventricular hypertrophy (LVH; Foley et al. 1995) and are at increased risk of a cardiac death (Foley et al. 1998; Liangos et al. 2006; Tonelli et al. 2006). Whilst hypertension is one cause of LVH in kidney failure, cardiac structural changes can occur independently of blood pressure (Amann et al. 1998b), and the pathogenesis of such changes remains under investigation. The rat model of acute kidney injury (AKI) induced by nephrectomy has been useful in this regard, since it is associated with cardiac remodelling, characterized by LVH and fibrosis, and impaired cardiac function (Amann et al. 1998a; Kennedy et al. 2003).

The present study used the subtotally nephrectomized (STNx) rat model of AKI to examine whether cardiac ACE2 is activated in kidney failure. Since blockade of the vasoconstrictor and hypertrophic actions of Ang II using ACE inhibitors confers beneficial renal and cardiac effects in patients with kidney disease (Eijkelkamp et al. 2007), we have also investigated the effects of ACE inhibition on cardiac remodelling after AKI.


Experimental protocol

Experimental procedures were performed in accordance with the National Health and Medical Research Council of Australia guidelines for animal experimentation. Rats were housed with a 12 h–12 h light–dark cycle, with ad libitum food containing 0.4–0.6% NaCl (Norco, Lismore, NSW, Australia) and water. Subtotal nephrectomy (n= 28) or sham surgery (n= 10) was performed in Sprague–Dawley rats (body weight of 200–250 g) as described previously (Wu et al. 1997; Cao et al. 2000, 2001). In brief, rats were anaesthetized by an i.p. injection of sodium pentobarbitone (60 mg (kg body weight)−1; Boehringer Ingelheim, Artarmon, NSW, Australia). The right kidney was freed from its capsule, the vascular bundle ligated with surgical ties and transected beyond the sutures. The non-perfused right kidney was removed, and haemostasis was confirmed visually. The left vascular pole and its contained vessels were identified and individually ligated, leaving all but one extrarenal branch of the left renal artery intact. Following STNx, animals were randomly allocated to a vehicle (n= 13) or the ACE inhibitor ramipril (n= 15) by daily oral gavage at a dose of 1 mg kg−1 for 10 days. All sham-operated rats received vehicle. On day 9, rats were housed in metabolic cages for 24 h, water intake and urine volume measured, and a urine sample collected. On day 10, rats were anaesthetized (i.p. sodium pentobarbitone, 60 mg (kg body weight)−1), and the carotid artery cannulated for measurement of mean arterial pressure (MAP) and heart rate (HR). Rats were then killed by a lethal dose of sodium pentobarbitone. Blood was collected, and the heart and remnant kidney were removed and weighed. The left ventricle (LV) was dissected, fixed in 4% paraformaldehyde and embedded in paraffin for histopathology. The remainder of the LV was snap frozen in isopentane and stored at −80°C for in vitro autoradiographic studies, ACE2 activity assays and RNA extraction.

Urinary protein

50 μl Urine samples were diluted 10-fold in lysis buffer (2 m Tris HCl, 5 m NaCl and Triton X-100), and urinary protein was determined using the bicinchoninic acid (BCA) method with a commercially available BCA protein assay kit (Pierce, Rockford, IL, USA).

In vitro autoradiography for cardiac ACE

Cardiac ACE was assessed by in vitro autoradiography on LV sections (20 μm) using the specific radioligand 125I-MK351A (Ki= 30 pmol l−1; Burrell et al. 2005). Quantification of ACE binding density in four LV sections from each animal (n= 5–6 per group) was performed using a microcomputer imaging device (Imaging Research, Linton, Cambridge, UK) which measures the relative optical density of the radioactive labelling. Results are expressed as a percentage of binding in sham-operated rats.

Determination of collagen fraction in the LV

Left ventricular paraffin sections 4 mm thick were deparaffinized, rehydrated, and then stained with 0.1% Sirius Red (Polysciences, Warrington, PA, USA) in saturated picric acid (Picrosirius Red) for 1 h, differentiated in 0.01% HCl for 30 s, and rapidly dehydrated. Interstitial collagen volume fraction was determined by measuring the area of stained tissue within a given field, excluding vessels, artefacts, minor scars or incomplete tissue fields; 15–20 fields were analysed per animal. To measure perivascular collagen, all arteries in the LV section were analysed, and the whole artery including the adventitia was selected for assessment. For both interstitial and perivascular collagen, the area stained was then calculated as a percentage of the total area within a given field (Candido et al. 2003; Burrell et al. 2005; Dean et al. 2005).

Quenched fluorescent substrate assay of ACE2 activity

The LV membranes were prepared as previously described (Burrell et al. 2005) and 100 μg protein was incubated in duplicate with an ACE2-specific quenched fluorescent substrate (QFS; (7-methoxycoumarin-4-yl)-acetyl-Ala-Pro-Lys(2,4-dintirophenyl); Auspep, Parkville, Victoria, Australia; Vickers et al. 2002; Burrell et al. 2005). Assays were performed with 50 μm QFS in a final volume of 200 μl ACE2 assay buffer (100 mm Tris, 1 m NaCl, pH 6.5). The final concentration of DMSO (used to solubilize QFS) was 0.7%. Reactions proceeded at 37°C for 90 min with continuous monitoring of liberated fluorescence (excitatory wavelength, 320 nm; emission wavelength, 405 nm) using a FLUOstar Optima plate reader (BMG Labtechnologies, Offenburg, Germany). Since the QFS can be cleaved by prolyl endopeptidase, a specific inhibitor of this enzyme, Z-Pro-prolinal (1 μm), was included in all wells (Wilk & Orlowski, 1983). Cleavage of the QFS was attributed to ACE2 by the use of the specific inhibitor MLN-4760 (a generous gift of Dr Natalie Dales, Millenium Pharmaceuticals, Cambridge, MA, USA) in replicate wells at 100 nm final concentration (Dales et al. 2002). The rate of substrate cleavage was determined by comparison with a standard curve of the free fluorophore 4-amino-methoxycoumarin (MCA; Sigma) and expressed as nanomoles of substrate cleaved per millgram protein per hour.

Plasma ACE2 was measured using a modified version of the above method, in that EDTA was used to inhibit ACE2. The results obtained using the two methods were highly correlated. For plasma levels, ACE2 was measured in 20 μl of plasma in the presence of 100 mm EDTA and 1 mm Z-Pro-prolinal, and results expressed as nanomoles of substrate cleaved per millilitre of plasma per hour.

Quantitative real-time polymerase chain reaction (QRT-PCR)

Total RNA was isolated from fresh frozen LV (sham, n= 8 per group; STNx, n= 10 per group) using the RNeasy kit method (Qiagen, Doncaster, Vic, Australia). Cloned DNA was synthesized with a reverse transcriptase reaction using standard techniques (Superscript II kit, Life Technologies, Gaithersburg, MD, USA) as previously described (Candido et al. 2003; Tikellis et al. 2003; Burrell et al. 2005). Quantitative RT-PCR is a fully quantitative method for the determination of amounts of mRNA. Briefly, a gene-specific 5′-oligonucleotide corresponding to the rat ACE gene (5′-CACCGGCAAGGTCTGCTT), an ACE 3′-oligonucleotide primer (5′-CTTGGCATAGTTTCGTGAGGAA) and ACE probe (FAM5′-CAACAAGACTGCCACCTGCTGGTCC-TAMRA) were used. For ACE2, a gene 5′-oligonucleotide (5′-GCCAGGAGATGACCGGAAA), an ACE2 3′-oligonucleotide primer (5′-CTGAAGTCTCCATGTCCCAGATC) and ACE2 probe (FAM5′-TTGTCTGCCACCCCACA-TAMRA) were used. For the brain natriuretic peptide (BNP) gene, a 5′-oligonucleotide (5′-GCCGCTGGGAGGTCACT), a 3′-oligonucleotide primer (5′-AGCTTCTGCATCGTGGATTGT) and BNP probe (FAM5′-TCCTAGCCAGTCTCC-TAMRA) were used. All primers and probes were designed using the software program ‘Primer Express’ (PE Applied Biosystems, Foster City, CA, USA). Quantitative RT-PCR was performed using the multiplex method and 18S VIC was used as a control. The relative expression method was applied in this study, using the control as the calibrator. The calculations for fold induction are extrapolated as follows. The cycle number (Ct) for 18S was subtracted from the Ct of the gene of interest (in this case ACE, ACE2 or BNP), resulting in the difference in Ct, i.e. ΔCt. The average ΔCt was then calculated for the calibrator group, and this value was subtracted from every ΔCt value in all groups, resulting in a ΔΔCt value for all samples. The ΔΔCt value was then entered into the expression 2(exp –ΔΔCt), which resulted in a fold induction value. All groups were compared with the calibrator group, which was the sham group for these experiments and valued ‘1’.

Angiotensin-converting enzyme 2 immunohistochemistry

Immunohistochemical staining for ACE2 (polyclonal antibody, T17, from Santa Cruz Biotechnology, Santa Cruz, CA, USA; diluted 1:100) was performed in rat LV sections as described previously (Burrell et al. 2005; Zulli et al. 2006). Staining was quantified (n= 6–7 per group) using computerized image analysis (Imaging Research, Linton, Cambridge, UK). All sections used for quantification were fixed, processed, sectioned and immunolabelled at the same time and under the same conditions to limit variability. Images were imported into the AIS Imaging program using a colour video camera and a standard light microscope (magnification ×20). The detection level threshold for positively stained areas (brown for 3,3-Diaminobenzidine staining) was set so that the processed image accurately reflected the positively stained areas as visualized by light microscopy on the unprocessed digital image. Myocyte ACE2 staining was determined by measuring the area of stained tissue within a given field, excluding vessels, artefacts, minor scars or incomplete tissue fields. Fifteen to 20 fields were analysed per animal. For arteries, the whole cross-section of the artery including adventitia was selected. The percentage area of chromogen staining was determined by calculating the number of selected pixels (positively stained areas) in a given area and expressed as a percentage of the entire image (Dean et al. 2005).

Statistical analysis

Data are presented as means ±s.e.m. Significant differences were obtained when P < 0.05, and all P values were calculated using Student's unpaired t test. For QRT-PCR, values for sham-operated animals were arbitrarily standardized to ‘1’ by taking the average of the results of all hearts, and the data for sham treatment and for STNx rats with and without treatment were expressed relative to this value.


Acute kidney injury

The STNx rats gained less weight over the experimental period compared with sham-operated animals (P < 0.05; Table 1). Left kidney weight and the ratio of left kidney/body weight increased in STNx rats (P < 0.001; Table 1); these animals also drank more water and had polyuria (P < 0.001; Table 1). Urinary protein excretion increased in STNx rats (P < 0.001). Ramipril increased water intake (P < 0.05) but had no effect to reduce proteinuria (Table 1). Kidney injury led to increased whole heart (P < 0.001), LV (P < 0.001), right ventricle (P < 0.05) and atrial mass (P < 0.001; Table 1), and ramipril reduced whole heart (P < 0.001), LV (P < 0.001) and right ventricular mass (P < 0.05; Table 1).

Table 1.  End organ weights and physiological parameters
Vehicle (n= 10)Vehicle (n= 13)Ramipril (1 mg kg−1 day−1) (n= 15)
  1. *P < 0.05, **P < 0.01 and ***P < 0.001 versus sham vehicle; †P < 0.05, ††P < 0.01 and †††P < 0.001 versus vehicle-treated STNx. Abbreviations: BW, body weight; LV, left ventricle; RV, right ventricle; and SBP, systolic blood pressure.

Body weight (g)232 ± 5  211 ± 6*   217 ± 5  
Renal parameters
 Left kidney weight (g)0.818 ± 0.020 1.014 ± 0.030***0.992 ± 0.031
 Left kidney/BW (g (100 g)–1)0.353 ± 0.008 0.475 ± 0.014***0.461 ± 0.020
Cardiac parameters
 Heart weight (g)0.719 ± 0.051 0.872 ± 0.011***   0.704 ± 0.010†††
 LV weight (g)0.509 ± 0.014 0.651 ± 0.014***   0.496 ± 0.008†††
 RV weight (g)0.139 ± 0.0030.151 ± 0.002*  0.137 ± 0.001†
 Atrial weight (g)0.034 ± 0.001 0.041 ± 0.002***0.039 ± 0.003
 Heart weight/BW (g (100 g)–1)0.311 ± 0.009 0.422 ± 0.014***   0.326 ± 0.004†††
 LV weight/BW (g (100 g)–1)0.220 ± 0.006 0.314 ± 0.010***   0.229 ± 0.003†††
 RV weight/BW (g (100 g)–1)0.062 ± 0.0010.074 ± 0.002**  0.063 ± 0.001††
 Atrial weight/BW (g (100 g)–1)0.015 ± 0.0010.019 ± 0.001**0.018 ± 0.001
Physiological parameters
 Water intake (ml (24 h)–1)30 ± 3 54 ± 4***62 ± 5†
 Urine output (ml kg–1 (24 h)–1)37 ± 7 134 ± 18***147 ± 12 
 Urinary protein (mg (24 h)–1)5.4 ± 0.211.6 ± 3.0***11.3 ± 3.7 

Haemodynamics and cardiac parameters

Acute kidney injury increased MAP and HR (P < 0.001; Fig. 1), caused LVH (P < 0.001; Fig. 2A) and increased cardiac BNP mRNA (P < 0.01; Fig. 2B). Rats also had interstitial (P < 0.001; Fig. 2C) and perivascular fibrosis (P < 0.001; Fig. 2D). Ramipril decreased blood pressure (P < 0.001; Fig. 1), improved the histological parameters of cardiac injury (P < 0.01; Fig. 2) and decreased cardiac BNP mRNA to normal values (P < 0.01; Fig. 2B).

Figure 1.

Graphs showing mean arterial pressure and heart rate in sham-operated and STNx (vehicle or ramipril) rats
Data are expressed as means +s.e.m.***P < 0.001 versus sham and ###P < 0.001 versus STNx (n= 10–15 per group).

Figure 2.

Graphs showing LV hypertrophy (A), LV BNP (B) and interstitial (C) and perivascular collagen (D) in sham-operated and STNx (vehicle or ramipril) rats
Data are expressed as means +s.e.m.**P < 0.01 and ***P < 0.001 versus sham and ##P < 0.01 and ###P < 0.001 versus STNx (n= 8–15 per group).

Cardiac expression of ACE gene and protein

There were no changes in relative levels of left ventricular ACE mRNA or ACE binding in AKI (Fig. 3A and B). However, ramipril did significantly decrease cardiac ACE in rats with AKI (P < 0.05; Fig. 3B).

Figure 3.

Graphs showing relative LV ACE mRNA levels (A) and LV ACE binding (B) in sham-operated and STNx (vehicle or ramipril) rats
Data are expressed as means +s.e.m.#P < 0.05 versus STNx (n= 6 per group).

Cardiac ACE2 mRNA and activity

Acute kidney injury increased cardiac ACE2 mRNA levels (P < 0.01; Fig. 4A), ACE2 activity (P < 0.05; Fig. 4B) and myocardial ACE2 protein measured by immunohistochemistry (Figs 4C and 5). There was no change in perivascular ACE2 protein (Fig. 4D). Ramipril significantly reduced myocardial ACE2 protein (P < 0.01; Figs 4C and 5), and there was a trend to a reduction in cardiac ACE2 activity (Fig. 4B). Angiotensin-converting enzyme 2 mRNA levels were unchanged (Fig. 4A).

Figure 4.

Graphs showing relative LV ACE2 mRNA (A), LV ACE2 activity (B) and myocardial (C) and perivascular ACE2 protein levels (D) in sham-operated and STNx (vehicle or ramipril) rats
Data are expressed as means +s.e.m.*P < 0.05 and **P < 0.01 versus sham and ##P < 0.01 versus STNx (n= 7–11 per group).

Figure 5.

Light microscopic images of ACE2 immunohistochemical labelling (brown staining) of cardiomyocytes in left ventricle of sham-operated (A), STNx (B) and ramipril-treated STNx rats (C)
Scale bar represents 10 microns.

The discrepancies between changes in ACE2 mRNA, activity and protein noted may reflect limitations in the methods. When quantifying ACE2 protein using immunohistochemical staining, we can quantify ACE2 expression separately in myocardium versus vessels. However, when ACE2 mRNA and activity are measured, ‘whole’ LV preparations (i.e myocardium and vessels) are used, which may lead to an underestimation of the true differences in mRNA expression.

Plasma ACE2 activity

Plasma ACE2 activity increased with STNx (P < 0.05; Fig. 6), and was reduced with ramipril (P < 0.01; Fig. 6).

Figure 6.

Graph showing plasma ACE2 activity in sham-operated and STNx (vehicle or ramipril) rats
Data are expressed as means +s.e.m.*P < 0.05 versus sham and ##P < 0.01 versus STNx (n= 10–15 per group).


The major findings of this study are that kidney injury of a short duration of 10 days results in cardiac remodelling, as evidenced by hypertrophy and by interstitial and perivascular fibrosis, and that these changes are associated with increased expression of the novel peptidase, ACE2. Indeed, marked increases in myocardial expression of ACE2 mRNA and protein occurred early after AKI, in an analogous fashion to the changes we observed in a rat model of myocardial injury (Burrell et al. 2005).

The STNx rat model of acute kidney injury is an excellent model to assess the effect of kidney failure on the heart, and is characterized by hypertrophy of the remnant kidney, proteinuria, polydipsia and polyuria, as well as hypertension and cardiac remodelling (Tornig et al. 1996; Cao et al. 2001; Kennedy et al. 2003). In general, most studies assess the long-term effects of kidney failure (4–8 weeks duration). Few studies have assessed the acute effects of AKI on the heart, and our results indicate that significant cardiac remodelling occurs as early as 10 days after the kidney injury. Although we did not assess cardiac function directly, the changes in cardiac structure we observed were associated with increased cardiac BNP expression, an indirect marker of cardiac damage (Candido et al. 2003).

Components of the classical RAS, such as ACE and Ang II, are activated in the kidney after kidney injury (Cao et al. 2001), but to date the effects of AKI on novel components of the RAS, such as ACE2, have not been assessed in either the heart or the kidney. In this study, ACE2 was assessed by a range of techniques, including specific assessment of ACE2 catalytic activity (Burrell et al. 2005) as well as immunohistochemistry. We have demonstrated that ACE2 in the heart is increased at the gene and protein level and, using immunohistochemistry, have localized the increase in ACE2 protein to cardiomyocytes.

The precise role of ACE2 is under intense investigation, and a number of studies suggest that it may have a cardioprotective role (Crackower et al. 2002; Burrell et al. 2005; Huentelman et al. 2005; Yamamoto et al. 2006). In particular, an increase in cardiac ACE2 may limit the adverse effects of elevated Ang II, the main product of ACE activity, by increasing the levels of the vasodilator Ang(1–7) (Ferrario et al. 2005). Angiotensin(1–7), a major product of ACE2 activity, is found in normal myocardium (Averill et al. 2003) and is a potent anti-fibrotic peptide (Grobe et al. 2006, 2007) that can inhibit the remodelling effects of cardiac fibroblasts (Iwata et al. 2005). Although we were not able to measure cardiac Ang(1–7) in this study, we suggest that AKI significantly increased cardiac ACE2 in a compensatory and protective role to promote the production of the anti-fibrotic peptide Ang(1–7). Since ACE2 is also involved in the degradation of Ang II, which promotes cardiac fibroblast collagen production, increased ACE2 may also reduce Ang II and therefore decrease fibrosis (Brilla et al. 1994; Lee et al. 1995; Kawano et al. 2000).

In this study, we also confirmed the finding of other groups, albeit in rats with a longer duration of kidney failure (Tornig et al. 1996), that ACE inhibition with ramipril ameliorates cardiac hypertrophy and fibrosis. The benefits of ACE inhibition in improving cardiac function and structure, particularly after myocardial infarction, are well established, and are due to a reduction in cardiac ACE and Ang II levels (Duncan et al. 1996; Burrell et al. 2005). However, little is known about the potential mechanism for the cardiovascular benefits of ACE inhibitors in AKI. In this study, ramipril reduced cardiac ACE binding and improved cardiac structure in association with normalization of cardiac ACE2 gene and myocardial ACE2 protein expression. The lack of effect of ramipril on proteinuria is most probably due to the early time point studied. We and others (Cao et al. 2001) have shown that longer-term use of ACE inhibitors in this model will result in a reduction in proteinuria.

These data suggest that AKI promotes adverse cardiac remodelling with cardiac fibrosis and hypertrophy as early as 10 days after injury. The results also suggest that activation of cardiac ACE2 in AKI may serve as a counter-regulatory mechanism, to protect the heart against the adverse effects of an activated RAS, hypertension and renal impairment. Although ACE inhibitors are not first-line treatment in AKI, this study shows that ACE inhibition can promote favourable effects on the heart. It is not yet clear whether these improvements are a result of the reduction in blood pressure or result from direct modulation of the cardiac RAS. Further studies using ACE2 inhibitors, infusion of Ang(1–7) and non-RAS-blocking agents, such as calcium channel blockers, are needed to increase our understanding of the underlying mechanism of the novel changes observed in this study.



This work was supported by funding from Austin Hospital Medical Research Foundation and the National Health and Medical Research Council.