Introduction

Diabetes mellitus results in severe cardiovascular complications, and heart disease and failure remain the major causes of death in patients with diabetes.^1–5 Given the increasing global tide of obesity and diabetes, the clinical burden of diabetes-induced cardiovascular disease is reaching epidemic proportions. Diabetic cardiomyopathy refers to ventricular dysfunction that occurs in the absence of any changes in blood pressure and coronary artery disease, with phenotypic features such as cardiomyocyte apoptosis, cardiac hypertrophy, myocardial fibrosis, and interstitial inflammation.^4,6,7 Several key mechanisms have been proposed and tested to explain diabetic myocardial dysfunction, some of which include increased oxidative stress, impaired calcium homeostasis, upregulation of the renin-angiotensin system (RAS), lipotoxicity, and mitochondrial dysfunction.^4,6

Editorial, see p 1270

In This Issue, see p 1265

Activation of the RAS plays a key role in the progression of diabetic complications and AT1 receptor blockers have reduced these complications.^8–11 Angiotensin-converting enzyme-2 (ACE2) is a carboxypeptidase that metabolizes angiotensin II (Ang II) to yield angiotensin 1 to 7 (Ang 1–7), essentially negatively regulating the RAS.^12–15 We hypothesized that loss of ACE2 accelerates diabetic injury in the cardiovascular system. We investigated cardiac and vascular structure and function, Ang II metabolism, signaling, and tissue reactive oxygen species generation in insulin-deficient diabetic Akita mice in response to genetic ablation of ACE2. In this study, we demonstrate that loss of ACE2 mediates development of systolic dysfunction and impaired vascular function in Akita mice. We identified augmented superoxide production, activation of the extracellular matrix (ECM)-degrading metalloproteinases, and differential signaling as the underlying mechanisms in this model of diabetic cardiovascular injury.

Methods

Experimental Animals and Protocol

C57BL/6J wild-type (WT) and diabetic heterozygous Akita (Ins2^WT/C96Y) and db/db mice were purchased from The Jackson Laboratory (Bar Harbor, ME). Male Akita mice were bred with female WT mice at the University of Alberta animal facility. Similarly, male heterozygous Akita mice were crossed with female ACE2^−/− mutant mice (ACE2KO) to obtain Akita/ACE2KO (Ins2^WT/C96Y/ACE2^−/y) double mutants (Detailed Methods are provided in the Online Supplement.). Throughout the period of study, animals were provided free access to water and standard 18% protein rodent chow (Harlan Teklad). A subgroup of Akita/ACE2KO mice also was treated with irbesartan, an AT1 receptor blocker (50 mg/kg⁻¹/d⁻¹) or Ang 1–7 (24 μg/kg/h; Bachem) for 1 month starting at 5 months of age. A subgroup of Akita mice was also treated with irbesartan (50 mg/kg⁻¹/d⁻¹) for 1 month starting at 5 months of age. The use of animals in this study conforms to the Care and Use of Laboratory Animals published by the United States National Institutes of Health (NIH Publication 85-23, revised 1996) and to the guidelines of the Canadian Council on Animal Care.

Echocardiography and Tissue Doppler Imaging

Transthoracic echocardiography was performed noninvasively to assess systolic and diastolic functions as described previously using a Vevo 770 high-resolution imaging system equipped with a 30-MHz transducer (RMV-707B; VisualSonics).^13,16

Tail-Cuff Systolic Blood Pressure Measurements

Systolic blood pressure of each mouse was measured by the tail-cuff method with an IITC blood pressure monitoring system (IITC Life Science) as previously described.¹³

Assessment of Flow-Mediated Vasodilation

Six-month-old male WT, Akita, ACE2KO, and Akita/ACE2KO and irbesartan-treated Akita/ACE2KO mice were anesthetized with isoflurane (2% induction and 1% maintenance), and body temperature was maintained at 36.9°C to 37°C. A 40-MHz transducer (RMV-704; Visual Sonics) was used to visualize the left femoral artery. The left femoral artery was identified on the basis of its characteristic flow pattern.¹⁷ Doppler flow measurements from longitudinal sections of the femoral artery were obtained before and after 5 minutes of hind limb ischemia. Reproducible ischemia (seen as complete abrogation of Doppler flow wave) and reperfusion of the hind limb (reappearance of Doppler flow waveforms after 5 minutes of ischemia) were achieved with a traction suture as an arterial loop occluder that was positioned upstream of the site to be visualized, around the left common iliac artery, through a transfemoral access. The loop occluder consisted of a 7-0 nylon filament around the artery that was externalized and skin was closed with suture. Baseline readings were recorded after a 10-minute equilibration period followed by complete hind limb ischemia achieved by using the traction suture to occlude the common iliac artery. After 5 minutes of ischemia, the hind limb was reperfused by release of the occluder. Reactive hyperemia was measured by Doppler flow velocity of the left femoral artery at reperfusion and at 1, 2, 3, 4, and 5 minutes postreperfusion.

Biochemical Measurements

Random blood glucose was measured using Ascensia Contour glucometer (Bayer) as previously described.¹⁶ Identification and quantification of the major long-chain acyl CoA molecular species and myocardial ceramide levels (C18) were performed by high-performance liquid chromatography as previously described.¹⁶ Left ventricular (LV) myocardial ACE2 activity was measured using a specific fluorogenic substrate and calculated as the DX-600 suppressible activity as previously described.¹³

Histological Analyses, TUNEL Assay, and Immunofluorescence

Hearts were arrested in diastole with 1 mol/L KCl, fixed with 10% buffered formalin, and embedded in paraffin. Ten-micrometer-thick sections were stained with picro-sirius red and visualized using fluorescence microscopy as previously described.¹³ Myocyte cross-sectional area was measured using wheat germ agglutinin-oregon 488 (Invitrogen) staining in 5-micrometer-thick OCT-embedded cryosections. In situ DNA fragmentation was detected in 15-micrometer-thick cryosections using the commercially available terminal deoxynucleotidyltransferase-mediated dUTP nick-end labeling (TUNEL) assay according to manufacturer's instructions (Invitrogen). Nitrotyrosine immunofluorescence staining was performed in 5-micrometer-thick aorta cryosections using rabbit antinitrotyrosine (Millipore) primary antibody and TRITC-conjugated goat antirabbit (Abcam) secondary antibody.

Real-Time Reverse-Transcriptase Polymerase Chain Reaction and Western Blot Analysis

Myocardial mRNA expression levels were quantified by Taqman reverse-transcriptase polymerase chain reaction (Online Table I) as described previously.^13,16,18 Western blotting was performed to detect ACE2, protein kinase Cα (PKCα), phosphorylated and total levels of Akt (serine-473 and threonine-308), JAK2 (tyrosine-1007/1008), STAT3 (tyrosine-705), ERK-1/2 (threonine-177), and endothelial nitric oxide synthase (eNOS; serine-1177) using specific antibodies (Cell Signaling), pyruvate dehydrogenase-4 (Abgent Canada), and, for SERCA2a, phospho (serine 16)-phospholamban and total phospholamban (Santa Cruz), as previously described.^13,16,18 Blots were scanned and quantified using ImageQuant LAS 4000 (GE Healthcare, Biosciences).

Plasma and Myocardial Peptide Levels

Plasma and LV myocardial Ang II and Ang 1–7 levels were measured at the Hypertension Core Laboratory, Wake Forest University, Winston-Salem, North Carolina, as previously described.^13,14

Gelatin Zymography, Collagenase, and Gelatinase Activity

Gelatin zymography was performed as previously described.¹² Total collagenase and gelatinase activities were measured using fluorescent-based activity assays from EnzCheck (Molecular Probes) as described previously.^19,20 Samples were analyzed using a SpectraMax M5 microplate reader (Molecular Devices).

Superoxide Assay and Dihydroethidium Staining

The chemiluminescence lucigenin assay was used to measure NADPH oxidase activity as we have previously described.^13,14 The specific peptide inhibitor of NADPH oxidase, gp91phox ds tat (50 μmol/L), was used to confirm superoxide generation from NADPH oxidase.^13,14 Dihydroethidium (DHE) fluorescence studies were performed on 20-micrometer-thick frozen myocardial sections, which were washed with Hank's balanced salt solution incubated at 37°C for 30 minutes with DHE (20 μmol/L) in Hank's balanced salt solution, and then imaged using confocal microscopy.

Statistical Analysis

Comparison between two groups was made using a nonpaired Student t test (Figure 1A–C; Online Figure IA, E). Two-way ANOVA using diabetic state and ACE2 status as the two independent variables (factors) was performed to compare the data between the four experimental groups (WT, Akita, ACE2KO, and Akita/ACE2KO; Figure 1D–I, Figures 2–6, Online FigureIC, D; II–V, VIA). In experiments with multiple treatments, one-way ANOVA was followed by multiple comparison using the Student Neuman-Keuls test (Figures 7, 8, Online Figure IB, Online Figure VIB–VIII). Statistical analyses were performed using the SPSS Statistics 19 software. Averaged values are presented as means ± SEM. Statistical significance is recognized at P<0.05.

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

Results

Loss of ACE2 Increases Ang II Levels, Activates Pathological Gene Expression, and Leads to Systolic Dysfunction in Diabetic Hearts

Expression analysis and Western blotting showed a significant increase in ACE2 mRNA and protein levels (Figure 1A, B), resulting in increased ACE2 activity in Akita diabetic hearts (Figure 1C) without changes in ACE2 levels in the kidneys and a concordant increase in ACE2 levels in diabetic db/db hearts (Online Figure IA, B). We hypothesize that upregulation of ACE2 suppressed the activation of the RAS in a diabetic state and, as such, we generated Akita/ACE2KO double-mutant mice to probe the pathophysiological relevance of ACE2. Plasma and myocardial levels of Ang II were similar in WT, Akita, and ACE2KO mice, whereas Akita/ACE2KO mice showed significantly higher levels confirming activation of the RAS (Figure 1D, E). Interestingly, plasma and myocardial levels of Ang 1–7 were similar across all four experimental groups (Figure 1D, E). Loss of ACE2 conceivably could lead to differential impact on hyperglycemia and blood pressure. However, we found a markedly sustained and equivalent hyperglycemia in Akita and Akita/ACE2KO mice at 3 and 6 months of age and in 6-month old db/db mice (Online Figure IC–E). Similarly, systolic blood pressure was comparably elevated in Akita and Akita/ACE2KO mice with a mild reduction in body weight (Online Figure IIA–C). Ang II is a well-known mediator of pathological remodeling in the heart^13,21 and myocardial expression of α-skeletal actin, β-MHC, and BNP were increased in Akita/ACE2KO mice (Figure 1F–H), indicating pathological changes in the absence of overt hypertrophy based on morphometry and myocyte cross-sectional area (Figure 1I, Online Figure IID, E). These results illustrate that ACE2 is upregulated in diabetic hearts and loss of ACE2 results in increased Ang II levels and pathological myocardial gene expression without a differential effect on hyperglycemia or blood pressure.

The Akita mouse is a well-characterized nonobese model of type 1 diabetes and closely mimics the human diabetic condition.^16,22,23 WT, Akita, and ACE2KO mice at 6 months of age showed comparable and normal systolic function (Figure 2A, B and Online Table II). In contrast, LV fractional shortening (Figure 2C) and LV ejection fraction (Figure 2D) declined significantly in Akita/ACE2KO mice, indicating LV systolic dysfunction with mild LV dilation (Figure 2 and Online Table II). Tissue Doppler imaging is a novel technique to assess diastolic function and showed a reduction in E′ and E′/A′ ratio and elevation in E/E′ that, coupled with enlarged left atrial size, are indicative of elevated LV filling pressure and diastolic dysfunction in Akita and Akita/ACE2KO hearts (Figure 2 and Online Table II). The progressive nature of the cardiomyopathy in Akita/ACE2KO hearts resulted in a restrictive filling pattern with elevated E/A ratio and reduced isovolumetric relaxation time (Figure 2B and Online Table II). These data indicate that in a diabetic state, loss of ACE2 results in systolic dysfunction in a background of diastolic dysfunction.

Pathological Signaling, Lipotoxicity, and Downregulation of SERCA2

Activation of PKC plays a fundamental role in mediating the pathological effects associated with a hyperglycemia diabetic state,^24–26 whereas insulin and Ang II activates ERK1/2, PI3K/Akt, and eNOS signaling cascades in the myocardium.^27–29 To elucidate the basis of the systolic dysfunction in Akita/ACE2KO hearts, we chose to study the activation of these various signaling pathways implicated in the development of cardiomyopathy. PKCα expression is significantly upregulated in Akita and Akita/ACE2KO myocardium, with a greater elevation in the latter group (Figure 3A). Phosphorylation of Akt at the serine-473 (Figure 3B) and threonine-308 (Figure 3C) residues diminished significantly in Akita hearts and, to a greater extent, in Akita/ACE2KO hearts compared to WT and ACE2KO hearts. Phosphorylation of eNOS at serine-1177 residue was drastically increased in ACE2KO hearts but was lost in the Akita/ACE2KO hearts (Figure 3D), whereas a comparable increase in phosphorylation of Janus-activated kinase-2 (JAK2) and signal transducer and activator of transcription-3 (STAT3; Figure 3E, F) and ERK1/2 (Online Figure IIIA) were observed in Akita and Akita/ACE2KO hearts. Thus, loss of ACE2 triggers PKCα expression and a greater loss of Akt and eNOS signaling in diabetic hearts.

• Download figure
• Download PowerPoint

Lack of insulin action could lead to fatty acid accumulation and lipotoxicity, which has been linked to diastolic dysfunction.^30,31 Myocardial long-chain fatty acid (palmitoyl CoA, oleoyl CoA, stearoyl CoA) and ceramide were elevated by approximately two-fold in the Akita hearts at 6 months and this was not increased further in Akita/ACE2KO hearts (Online Figure IIIB–D). Loss of insulin signaling also was associated with upregulated mRNA and protein levels of pyruvate dehydrogenase-4 in Akita hearts (Online Figure IIIE, F). Diastolic dysfunction is linked to suppressed activity of sarcoplasmic reticulum Ca²⁺-ATPase2a (SERCA2a) pump, which is responsible for the removal of approximately 90% of Ca²⁺ from the cytoplasm.^32,33 SERCA2a expression declined dramatically in both Akita and Akita/ACE2KO compared to WT and ACE2KO, whereas the expression of PLN and phospho-PLN was not altered in any of the experimental groups (Online Figure IVA, B). Analysis of apoptosis showed no significant upregulation of apoptosis in Akita hearts, which was not exacerbated by loss of ACE2 (Online Figure IVC–E). We conclude that the similar extent of lipotoxicity and downregulation of SERCA2a likely underlies the comparable diastolic dysfunction in the Akita and Akita/ACE2KO models.

Greater Activation of the NADPH Oxidase and Matrix Metalloproteinases in Akita/ACE2KO Hearts

Hyperglycemia and activation of the RAS are well-known stimulants of the NADPH oxidase system.^13,14,34 Diabetes significantly increased the NADPH oxidase subunit, NOX2, levels in the heart (Figure 4A). Phosphorylation of the p47^phox subunit, a key mediator of Ang II-induced NADPH oxidase activation,^14,35 was approximately two-fold higher in Akita myocardium and was further exaggerated in Akita/ACE2KO hearts (Figure 4B). In line with these findings, NADPH oxidase activity was significantly higher in Akita/ACE2KO hearts, resulting in enhanced superoxide generation, as assessed by lucigenin-enhanced chemiluminescence (Figure 4C) and DHE staining for superoxide levels (Figure 4D). NADPH oxidase activity and DHE fluorescence intensity were increased further in Akita/ACE2KO compared to Akita hearts. The specificity of NADPH oxidase activity was verified by using the specific NADPH oxidase inhibitor, gp91phox ds tat peptide.

• Download figure
• Download PowerPoint

Adverse remodeling of the ECM is a key pathogenic factor in systolic heart failure and has been linked to activation of the RAS and NADPH oxidase.^7,8,36 Fluorescence-based imaging of the ECM showed a marked degradation of the ECM network and loss of collagen content in Akita/ACE2KO hearts, whereas the WT, Akita, and ACE2KO hearts all showed an intact ECM (Figure 4E, F). Ang II–mediated oxidative stress is known to activate matrix metalloproteinases (MMPs), leading to degradation of ECM proteins. Gelatin zymography showed increased MMP9 level in Akita/ACE2KO hearts (Figure 5A, B). Importantly, loss of ACE2 increased pro-MMP2 and active MMP2 in Akita hearts (Figure 5A, C). MMP2 and MMP9 are major gelatinases, whereas MMP2 also is a potent collagenase in the heart.³⁷ Gelatinase and collagenase activities were increased only in the Akita/ACE2KO hearts, thereby providing further evidence for an ECM-degrading environment in Akita/ACE2KO hearts (Figure 5D). Expression analysis of mRNA levels showed increased MMP9, MMP12, MMP13, and MMP14 (MT1-MMP) levels (Figure 5E–H) and MMP8 level (Online Figure VA), whereas MMP2 level was unchanged (Figure 5I), suggesting posttranscriptional activation of MMP2 is a key factor in mediating increased MMP2 protein levels. Although TIMP1 expression was increased in Akita/ACE2KO hearts, TIMP2 and TIMP3 levels were unchanged and TIMP4 level was reduced (Online Figure VB–E).

• Download figure
• Download PowerPoint

Loss of ACE2 Impairs Flow-Mediated Vasodilation and Increases Vascular Oxidative Stress in Akita Model: Role of the AT1 Receptor

Endothelial dysfunction and vascular oxidative stress have been linked to the development of diabetic cardiomyopathy.^25,38–40 Given the systemic disturbance of the RAS, we assessed the in vivo endothelial function in the Akita diabetic model and in response to ACE2 deficiency. We measured changes in blood flow velocity before and after ischemia-reperfusion of the femoral artery using high-resolution vascular ultrasonography. Baseline flow velocity was not different among the four experimental groups. Transient ischemia was followed by reactive hyperemia seen as an increase in flow velocity immediately after reperfusion in WT, Akita, ACE2KO, and Akita/ACE2KO mice (Figure 6A, B). Femoral flow velocity progressively declined to baseline value within 5 minutes in all groups, except in the Akita/ACE2KO mice, in which the flow velocity remained elevated even after 5 minutes of reperfusion (Figure 6A, B). These data show that in vivo flow-mediated dilation (FMD) is impaired in the Akita/ACE2KO model, suggestive of endothelial dysfunction. Increased vascular oxidative stress has been linked to endothelial dysfunction, a key mediator of FMD. Consistent with impaired FMD, aortic NADPH oxidase activity (Figure 6C) and DHE staining for superoxide (Figure 6D) showed increased oxidative stress in Akita/ACE2KO aorta and greater formation of nitrotyrosine (Figure 6E). We showed that loss of ACE2 leads to increased Ang II levels, a key mediator of vascular oxidative stress. We examined the impact of AT1 receptor blockade on the FMD and vascular oxidative stress. Treatment with the AT1 receptor blocker, irbesartan, prevented the impairment in FMD in Akita/ACE2KO mice based on femoral artery flow profile (Figure 7A, B), whereas the elevated NADPH oxidase was normalized (Figure 7C). DHE and nitrotyrosine fluorescence staining showed a marked reduction in superoxide levels and nitrotyrosine formation in response to AT1 receptor blockade, respectively (Figure 7D, E). These results show that loss of ACE2 impairs FMD and is linked to activation of the vascular NADPH oxidase system and formation of nitrotyrosine.

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

AT1 Receptor Blockade Prevents Systolic Dysfunction in the Akita/ACE2KO Model

Given the rescue of the vascular changes by AT1 receptor blockade, we next tested the critical role of the Ang II/AT1R axis in mediating cardiac dysfunction in the Akita/ACE2KO murine model. Western blotting analysis showed significant increase of AT1R in Akita and Akita/ACE2KO hearts (Online Figure VIA). The increased AT1R density coupled with increased agonist stimulation likely perpetuate the Ang II–mediated adverse myocardial effects. Systolic dysfunction and the mild LV dilation in the Akita/ACE2KO mice showed a drastic normalization in response to treatment with irbesartan, whereas diastolic dysfunction persisted (Figure 8A–E, Online Table III), suggesting that the diastolic dysfunction is independent of the Ang II/AT1R axis. Irbesartan failed to prevent the diastolic dysfunction and elevated oxidative stress in Akita hearts (Online Figure VIB–I). In the Akita/ACE2KO mice, AT1R blockade significantly reduced the elevation in myocardial NADPH oxidase activity (Figure 8F) and superoxide generation (Figure 8G) while preventing the loss in phospho-Akt (Figure 8H) and restoring the elevation in phospho-eNOS levels (Figure 8I). The increased collagenase and gelatinase activity were also suppressed by irbesartan, resulting in normalization of the ECM architecture and collagen content in Akita/ACE2KO hearts (Online Figure VII). Although treatment with irbesartan and Ang 1–7 produced similar decrease in systolic blood pressure, Ang 1–7 failed to rescue the systolic dysfunction in Akita/ACE2KO mice (Online Figure VIII). These results show that we can uncouple systolic and diastolic dysfunction in a diabetic heart and pharmacological antagonism of the AT1 receptor reverses the systolic dysfunction and key pathophysiological processes in the Akita/ACE2KO hearts.

• Download figure
• Download PowerPoint

Discussion

Diabetic cardiomyopathy is characterized by early diastolic and vascular dysfunction, which progresses into systolic dysfunction, resulting in heart failure.^1–5 The RAS plays a central role in diabetic cardiomyopathy, and pharmacological inhibitors of the RAS are cornerstone to minimizing the cardiovascular complications. In an effort to determine the role of ACE2, we created ACE2-deficient Akita mice to determine the role of ACE2 in diabetic cardiomyopathy. The Akita murine model is a well-validated model of human diabetes^16,22 and also displays features compatible with type 2 diabetes, including insulin resistance.²³ Our study defines a critical role of ACE2 in suppressing the activation of the RAS in the heart such that loss of ACE2 results in increased Ang II/AT1 receptor signaling, adverse myocardial remodeling resulting in systolic dysfunction, vascular oxidative stress, and impaired flow-mediated dilation. In addition to its effects on the heart, Ang II affects virtually all vascular cells and is critical in endothelial dysfunction, which is a key determinant in the development and progression of diabetic complications.^4,25 Blockade of the AT1 receptor resulted in marked improvement in systolic dysfunction with reversal of the pathological effects in the vasculature and heart of Akita/ACE2KO mice. The failure of AT1 receptor blockade to reverse the diastolic dysfunction in the Akita model is consistent with clinical trials in patients with diastolic heart failure in which AT1 receptor blockers failed to improve clinical outcomes.^41,42

The upregulation of ACE2 in diabetic Akita and db/db hearts is likely a key compensatory mechanism responsible for inhibiting activation of the RAS. In the diabetic condition, Ang II responses in myocardium and vasculature are augmented because of upregulation of Ang II type 1 receptor, thereby increasing Ang II sensitivity.⁴³ Inhibition of the AT1 receptor reduces adverse outcomes from heart failure in patients with diabetes.⁹ ACE2, a homologue of ACE, is a carboxypeptidase that metabolizes Ang II to yield Ang 1–7 and lowers Ang II/Ang 1–7 ratio. The results from this study are consistent with other models of heart failure whereby loss of ACE2 exacerbates heart disease after myocardial infarction¹² and pressure overload.¹⁴ Loss of ACE2 also exacerbates diabetic nephropathy,⁴⁴ which is a known correlate of cardiovascular disease.⁴⁵ Although loss of ACE2 elevated Ang II levels as predicted, we did not observe a corresponding decrease in Ang 1–7 levels in our model. Moreover, Ang 1–7 supplementation failed to rescue the systolic dysfunction in the Akita/ACE2KO mice. These results suggest that alternative pathways for Ang 1–7 generation are activated in a diabetic state such as the potent and high-capacity neprilysin enzyme and/or prolyl carboxypeptidase.^46–49

Altered signaling may be an important driver of the phenotypic changes observed in the Akita and Akita/ACE2KO hearts. Because PKC has been shown to inhibit myofibrillar ATPase and sarcoplasmic reticular Ca²⁺ pump^26,50 activities, it is possible that upregulation of PKCα in the diabetic heart may have exacerbated the cardiomyopathy in the Akita/ACE2KO model. Akita hearts also show suppressed insulin-dependent signaling and evidence of lipotoxicity coupled with increased expression of pyruvate dehydrogenase kinase-4 and downregulation of SERCA2a. Akt and eNOS are positive modulators of myocardial contractility,^29,51 and the greater loss of phospho-Akt and phospho-eNOS in the Akita/ACE2KO hearts may contribute to the impaired systolic performance. Adverse remodeling of the ECM is a key pathogenic factor in heart failure and has been linked to activation of the RAS and NADPH oxidase.^7,8,36 Lack of myocardial fibrosis in Akita/ACEKO mice was accompanied by upregulation of MMP8, MMP9, MMP12, MMP13, and MMP14 expression, MMP2 and MMP9 protein levels (and their activities), which are the key ECM-degrading enzymes,^37,52 leading to further adverse remodeling of the ECM, LV dilation, and systolic dysfunction.^37,52

Microvascular complications associated with diabetes also can adversely affect myocardial remodeling, and diabetic cardiomyopathy has been identified as a microvascular complication.^53,54 FMD is elicited by changes in blood flow, causing shear stress–induced release of nitric oxide by the endothelium.^55–58 Cellular oxidative stress and endothelial dysfunction are pivotal in the pathogenesis of cardiovascular diseases and are closely linked to circulating and tissue levels of Ang II.⁵⁹ Ang II can increase tyrosine-657 phosphorylation of eNOS, leading to decrease nitric oxide production and endothelial dysfunction.⁶⁰ Ang II–mediated reactive oxygen species production has the potential to further impair nitric oxide bioavailability by the consumption of nitric oxide and formation of peroxynitrite, leading to impaired endothelial relaxation and dysfunction.^39,58,61 Nitrotyrosine level, a marker of peroxynitrite formation, was elevated in Akita/ACE2KO aorta, which was reversed by treatment with irbesartan. AT1 receptor blockade clearly improved vascular oxidative stress, restored FMD, and prevented the development of systolic dysfunction in the Akita/ACE2KO model. Ang II activation of the AT1 receptor results in increased NADPH oxidase activity and enhanced reactive oxygen species generation.³⁴ The pathophysiological effects attributable to elevated Ang II levels in an ACE2-null environment are likely to be augmented by upregulation of the Ang II type 1 receptor in the Akita/ACE2KO hearts.

In summary, our study demonstrates loss of ACE2 exacerbates an underlying diabetic cardiomyopathy dysfunction resulting in a phenotype showing both diastolic and systolic dysfunction. We have identified increased oxidative stress, excessive ECM degradation, and vascular dysfunction brought about by upregulation of RAS in the absence of ACE2 as the underlying mechanism for the observed systolic dysfunction. Importantly, we have uncoupled diastolic and systolic dysfunction in our experimental models and have provided distinct mechanisms for each phenotype. Increased ACE2 expression and/or activity can represent a viable approach to minimize diabetes-induced cardiovascular complications.

Sources of Funding

This study was supported by operating grants from the Canadian Institute for Health Research ( GYO grant 86602 ) and the National Institutes of Health ( MBG grant HL110170 ). S.B. and J.C.Z. were supported by Alberta Innovates–Health Solutions (AI-HS) postdoctoral fellowships, R.B. is supported by a Motyl Graduate Scholarship , W.W. is partially supported by Mazankowski Alberta Heart Institute Graduate Scholarship , and J.L. is supported by an AI-HS Summer Studentship . G.Y.O. is a Clinician Investigator of the AI-HS and the Distinguished Clinician Scientist of the Heart and Stroke Foundation of Canada and Canadian Institutes of Health Research. Z.K. is a New Investigator of the Heart and Stroke Foundation of Canada and Scholar of the AI-HS.

Disclosures

None.

Footnotes