Allopurinol reduces oxidative stress and activates Nrf2/p62 to attenuate diabetic cardiomyopathy in rats

Abstract Allopurinol (ALP) attenuates oxidative stress and diabetic cardiomyopathy (DCM), but the mechanism is unclear. Activation of nuclear factor erythroid 2‐related factor 2 (Nrf2) following the disassociation with its repressor Keap1 under oxidative stress can maintain inner redox homeostasis and attenuate DCM with concomitant attenuation of autophagy. We postulated that ALP treatment may activate Nrf2 to mitigate autophagy over‐activation and consequently attenuate DCM. Streptozotocin‐induced type 1 diabetic rats were untreated or treated with ALP (100 mg/kg/d) for 4 weeks and terminated after heart function measurements by echocardiography and pressure‐volume conductance system. Cardiomyocyte H9C2 cells infected with Nrf2 siRNA or not were incubated with high glucose (HG, 25 mmol/L) concomitantly with ALP treatment. Cell viability, lactate dehydrogenase, 15‐F2t‐Isoprostane and superoxide dismutase (SOD) were measured with colorimetric enzyme‐linked immunosorbent assays. ROS, apoptosis, was assessed by dihydroethidium staining and TUNEL, respectively. The Western blot and qRT‐PCR were used to assess protein and mRNA variations. Diabetic rats showed significant reductions in heart rate (HR), left ventricular eject fraction (LVEF), stroke work (SW) and cardiac output (CO), left ventricular end‐systolic volume (LVVs) as compared to non‐diabetic control and ALP improved or normalized HR, LVEF, SW, CO and LVVs in diabetic rats (all P < .05). Hearts of diabetic rats displayed excessive oxidative stress manifested as increased levels of 15‐F2t‐Isoprostane and superoxide anion production, increased apoptotic cell death and cardiomyocytes autophagy that were concomitant with reduced expressions of Nrf2, heme oxygenase‐1 (HO‐1) and Keap1. ALP reverted all the above‐mentioned diabetes‐induced biochemical changes except that it did not affect the levels of Keap1. In vitro, ALP increased Nrf2 and reduced the hyperglycaemia‐induced increases of H9C2 cardiomyocyte hypertrophy, oxidative stress, apoptosis and autophagy, and enhanced cellular viability. Nrf2 gene silence cancelled these protective effects of ALP in H9C2 cells. Activation of Nrf2 subsequent to the suppression of Keap1 and the mitigation of autophagy over‐activation may represent major mechanisms whereby ALP attenuates DCM.


| INTRODUC TI ON
Diabetic cardiomyopathy (DCM), a common complication of diabetes mellitus, is the greatest mortality risk for patients with diabetes. 1 However, DCM is largely unrecognized in asymptomatic diabetic stage despite of the pathological progression. 2 Hyperglycaemia induced excessive production of reactive oxygen species (ROS), and the subsequent oxidative stress play critical roles in this pathology. 3,4 Cardiomyocyte hypertrophy, apoptosis and fibrosis have long been considered as three major features of DCM which eventually progress to heart failure. Antioxidant treatments as therapeutic strategies against DCM have been multi-dimensionally studied both clinically and experimentally. 5 However, convincing benefits of antioxidant therapy have not been seen in patients with diabetes because of limited understanding of the mechanism.
Nuclear factor erythroid 2-related factor 2 (Nrf2), a transcription factor that can be activated by oxidative stress, regulates antioxidant response elements (ARE) that control the basal and inducible expression of antioxidant genes in response to xenobiotics, antioxidants, heavy metals and UV light. 6,7 In the presence of oxidative stress, Nrf2 confers protective effects via translocation into the nuclear and subsequently activate ARE to induce antioxidant genes, such as NQO1, GCLC, heme oxygenase-1 (HO-1) and SQSTM1. 8,9 We recently showed that cardiac Nrf2 was reduced in streptozotocin (STZ)-induced diabetic rats and strategies that enhanced cardiac Nrf2 could activate myocardial HO-1 and attenuate cardiac hypertrophy and cardiac dysfunction in diabetic rats. 10 Also, recent studies suggest that targeting Nrf2 may represent a mechanism by which treatments that can attenuate DCM. 11,12 However, the potential impact of allopurinol (ALP) on cardiac Nrf2 expression in diabetes and its underlying mechanism of action have not been explored.
Nrf2 is activated following the disassociation with its repressor Kelch-like epichlorohydrin-associated protein 1 (Keap1), a cytosolic repressor protein of Nrf2. Keap1 binds and retains Nrf2 in the cytoplasm, preventing Nrf2's nuclear translocation. 13,14 Interestingly, among the downstream Nrf2-driven genes, p62, which is encoded by SQSTM1, 15 serves as a specific autophagy receptor for Keap1. 16 The autophagic cargo adaptor p62 conjugates with Keap1, embedding it together with aggregates of other ubiquitinated proteins for degradation, which liberates and activates Nrf2. 15 Autophagy is a dynamic process that is tightly regulated by cellular metabolic balance, and functional autophagy is essential for cell survival. Given the coexistence of elevated oxidative stress, increased protein and lipid oxidation, and impaired cellular energy balance in diabetes, functional autophagy may be of greater importance in maintaining cardiac cell integrity in diabetes. 17 Based on aforementioned findings, it is apparent that there exits an interconnected loop linking Nrf2, Keap1 and p62 in vivo, which functions collectively to maintain homeostasis during oxidative stress or other injury stresses.
However, whether or not the development of DCM involves this loop and how this loop varies in diabetes mellitus and in particular the diabetic myocardium is unclear.
Allopurinol, a competitive inhibitor of xanthine oxidase enzyme, has dose-dependent free radical scavenging property. 18 It attenuates the diabetes-induced increase of myocardial xanthine oxidase activity, 19,20 reduces ROS production and improves diabetes-induced cardiac dysfunction. 21 Our previous studies demonstrated that ALP could ameliorate diabetic myocardial ischaemia/reperfusion (MI/R) injury by lowering ROS production. 22  We suggested that the hyperglycaemia-induced oxidative stress and the over-activation of myocardial autophagy are key mechanisms that cause the imbalance of Nrf2-Keap1-p62 and the development of DCM. ALP treatment may activate Nrf2 to mitigate autophagy over-activation and consequently attenuate DCM. This hypothesis was tested both in type 1 diabetic rats and in H9C2 cardiomyocytes exposed to high glucose.  25 In brief, streptozotocin (STZ) was injected via tail vein at a dose of 65 mg/kg body weight (Sigma-Aldrich) in 0.1 mol/L citrate buffer (pH 4.5) or citrate buffer alone as control Keap1 and the mitigation of autophagy over-activation may represent major mechanisms whereby ALP attenuates DCM.

K E Y W O R D S
allopurinol, autophagy, diabetic cardiomyopathy, Nrf2, oxidative stress under anaesthesia. One week after STZ injection, rats exhibiting hyperglycaemia (blood glucose 16.7 mmol/L or higher) were considered type 1 diabetic and subjected to subsequent experiments. Consistent with our previous study which showed that STZ induced significant reduction in insulin and increase in plasma glucose and that antioxidant treatment did not have significant impact on plasma levels of insulin and glucose, 26 our preliminary experiment showed STZ significantly decreased plasma insulin level in diabetic rats compared with controls (data not shown) with concomitant increase in plasma glucose. Therefore, in the current study, hyperglycaemia-mediated cardiac complications and mechanism were our major targets; therefore, we did not specifically measure insulin in diabetic rats but confirmed the successful establishment of diabetic model by using plasma glucose as a major indicator.

| Experimental protocol
Rats were randomly divided into three groups (n = 6 per group): control (C), diabetes (D) and diabetes treated with ALP (DA). ALP (Sigma-Aldrich) were dissolved in drinking water for 4 weeks' duration of treatment starting 1 week after induction of diabetes. We used a dose of ALP at 100 mg/kg/d, which has been demonstrated to attenuate heart injury from ischaemia/reperfusion in our previous study. 22 The daily dose of ALP (100 mg/kg) was dissolved into 2/3 volume of the average daily amount of drinking water that was estimated based on preliminary study and previous studies to ensure that the total amount of ALP was taken by the rats before additional amount of water was supplied. Meanwhile, the plasma glucose, body weight, water intake and food consumption of rats were routinely monitored. At the end of the experiments, cardiac function was determined by the echocardiography (4 weeks after diabetes induction) and pressure-volume conductance catheter system (5 weeks after diabetes induction), respectively, before the rats was terminated. After the completion of experiments, rats were then deeply anaesthetized with sodium pentobarbital (65 mg/kg), and hearts were rapidly excised and frozen in liquid nitrogen for later analysis.

| Echocardiography
Rats from each group were anaesthetized by inhalation of 3% isoflurane in pure O 2 , which was continuously maintained till the end of the echocardiography. 27  Meanwhile, LV contractile function parameters FS (fractional shortening), EF (ejection fraction) and SV (stroke volume) were calculated using LVIDs, LVIDd, LVVd (LV end-diastolic volume) and LVVs (LV end-systolic volume) as described. 28 Three representative cardiac cycles were recorded and averaged for each measurement.

| Intracardiac heart function detection
The global cardiac functions were monitored by using a pressurevolume (PV) conductance catheter (AD Instruments) placed into the left ventricle through the right carotid artery and connected to a computer equipped with an advantage PV control box software (AD Instruments) as previously described. 25 The cardiac functional parameters were recorded, including heart rate (HR), left ventricu-

| Plasma and cardiac free 15-F2tisoprostane and SOD measurement
Free 15-F2t-IsoP, a specific marker of oxidative stress in vivo, was measured by using an enzyme immunoassay kit (Cayman Chemical) as described. 27 The value of free 15-F2t-IsoP was expressed as pg/ mL in plasma and as pg/mg protein in the heart tissue. Myocardial and plasma SOD activity was detected in cardiac tissue homogenates using commercially available kits (Cayman Chemical) as described previously. 29

| Detection of left ventricle tissue reactive oxygen species production by in situ dihydroethidium staining
Levels of in situ O − 2 production were detected by dihydroethidium (DHE; Sigma-Aldrich) staining as described. 27

| Apoptotic cell death detection using terminal deoxynucleotidyl transferase dUTP nick-end labelling
Terminal deoxynucleotidyl transferase dUTP nick-end labelling (TUNEL) reaction was performed using an in situ cell death detection kit (Roche Diagnostics GmbH) as previously described. 31 The sections were observed in the light microscope by an investigator who was initially blinded to treatment groups, and five randomly selected fields of each slide were analysed, and the apoptotic index was calculated as a percentage of apoptotic nuclei to total nuclei.

| Transfection and siRNA knock-down in H9C2 cells
Embryonic rat cardiac H9C2 cells from ATCC, at the passages

| Measurement of cardiomyocyte surface area
H9C2 cells were stained with phalloidin-tetramethylrhodamine conjugate (Santa Cruz Biotechnology) as previously described. 4 The cardiomyocyte picture was obtained with a fluorescence microscope (BX41 System microscope; Olympus), and the images were captured by a DP72 digital camera with 400 magnification. The cell surface area was determined with image analysis software ImageJ 1.48 (National Institutes of Health) and calculated as the mean of 100 cells of five randomly selected fields.

| Measurement of cellular ROS in cultured cardiomyocytes
Superoxide generation in cultured cardiomyocytes was estimated by DHE staining as previously described. 32 Briefly, cardiomyocytes were loaded with DHE at a concentration of 10 μmol/L for 30 minutes at 37°C. The DHE fluorescence of DHE-labelled positive nuclei was calculated in each of five randomly selected fields and was expressed as a percentage of the DHE-stained positive myocyte nuclei compared with control by a quantitative morphometric method.

| Determination of cellular injury
Cell lactate dehydrogenase (LDH) content was measured with a LDH Cytotoxicity Assay Kit (Roche) as described. 33 Cell viability was measured using the CCK-8 kit (Solarbio). In brief, H9C2 cells were seeded at 5 × 10 3 cells per well in 96-well plates in triplicate. After the transfection and 48 hours treatments, 10 μL of CCK-8 solution was added to each well. After 2 hours of incubation, absorbance was measured at 450 nm.

| Isolation of myocardial cytosolic and nuclear fractions
Hearts from control and STZ-treated rats were cleared of blood by washing thoroughly in Tyrode buffer and aortic and atrial sections removed from the ventricles. Ventricular tissue was freezeclamped in liquid nitrogen and stored until fractionated to isolate cytosolic and nuclear fractions according to the manufacturer's protocol as described in the Nuclear and Cytoplasmic Extraction Kit (Thermo).
The reverse concentration is 500 ng of total RNA in 10 μL reaction system. The data of qPCR were analysed via mean relative content

| Western blot analysis
Protein samples from rat heart homogenate and H9C2 cells were resolved by 7.5%-12.5% SDS-PAGE and equal protein amount (50 µg) subsequently transferred to polyvinylidene nitrocellulose membranes and processed as described. 35 The primary antibodies against

| Statistical analysis
All values are presented as means ± standard error of the mean (SEM). One-way analysis of variance (ANOVA) was used for statistical analyses (GraphPad Software, Inc) of data obtained within the same group and between groups, respectively, followed by Tukey's test for multiple comparisons of group means. P values less than .05 were considered to indicate statistically significant differences.

| General characteristic variations at termination
As shown in Table 1

| Changes of left ventricular dimensions and function and the effects of ALP on diabetic heart dysfunction
Through echocardiography, left ventricular inner diameter and thickness of myocardium were measured and the heart function was detected (Table 2), synergizing with pressure-volume loop system (Table 3)

| ALP attenuated oxidative stress and apoptosis induced by hyperglycaemia
Both the plasma and cardiac 15-F2t-isoprostane (15-F2t-IsoP) concentration was significantly higher in diabetic rats than in controls, and ALP normalized the 15-F2t-IsoP level in diabetic rats ( Figure 1A,B). Similarly, the SOD activity in the plasma and heart tissue significantly increased in diabetic group compared with control group, and the ALP significantly reduced its levels comparable to controls ( Figure 1C,D). DHE staining of heart tissue sections showed that superoxide anion (ROS) production (number of DHE-labelled nuclei, Figure 1E,F) was significantly increased in D rats compared with

| The feed-forward loop linking Nrf2, Keap1 and P62 was broken in diabetes and ALP restored it
In the 5-week STZ-induced diabetic rats, both the cardiac cytosolic and nuclear Nrf2 significantly decreased compared with controls, but the nuclear Nrf2 did not decrease as obviously as cytosolic Nrf2 ( Figure 2A-C). In the meantime, protein expression of p62 ( Figure 2F) and Keap1 ( Figure 2D) went down obviously and the cardiac ratio of LC3 II/I ( Figure 2G), an indicator of the extent about autophagy, was significantly increased (all P < .05 D vs C). The interruption of the feed-forward loop in diabetic rats directly resulted in impaired antioxidant capacity manifested as significantly reduced HO-1 protein expression ( Figure 2E) and increased ROS production ( Figure 1E,F).
After 4-week treatment with ALP in diabetic rats, p62 significantly increased while the ratio of LC3 II/I was normalized to control level and was significantly higher than that in the untreated D groups. Of note, the nuclear Nrf2 expression was significantly higher in the diabetic treated group than that in the untreated D groups even though the Keap1 expression was obviously further decreased after ALP treatment (all P < .05 DA vs D). The mRNA levels of Nrf2 ( Figure 2H)

TA B L E 3
Haemodynamic variables and indices of cardiac systolic and diastolic function by a pressure-volume conductance catheter system and Keap1 ( Figure 2J) were increased in the hearts of diabetic rats, and ALP further increased Nrf2 expression but reduced Keap1 gene to a level comparable to the control. The cardiac mRNA levels of p62 did not significantly change in diabetic rats as compared to the control ( Figure 2I) but ALP treatment significantly increased cardiac p62 mRNA.

| ALP attenuated HG-induced increase of cardiomyocyte size, oxidative stress, apoptosis and cellular injury via Nrf2 signalling pathway
High glucose significantly increased cell surface area in cultured H9C2 cardiomyocytes ( Figure 3A,B) and elevated oxidative stress ( Figure 3D,E), concomitantly with increased apoptosis ( Figure 3G,H In vitro, HG caused significant reduction of nuclear of Nrf2, which was reversed by ALP ( Figure 4J). However, the protein expression of downstream targeted protein Keap1 and p62 decreased extraordinarily ( Figure 4D,E).
Besides, Bafilomycin treatment induced higher P62 protein level than NG group (Baf vs NG, P < .01), indicating that P62 was accumulated when autophagy was inhibited. Whereas P62 protein levels in the HG group was reduced relative to that in the NG group (HG vs NG, P < .05), indicating the excessive autophagy flux in cardiomyocytes under HG condition ( Figure 4K).

| D ISCUSS I ON
In the current study, we have demonstrated that antioxidant ALP effectively maintains inner redox homeostasis and attenuates DCM via restoration of Nrf2/p62 signalling pathway through normalizing disordered autophagy and decreasing apoptosis in diabetic rats. We showed that reductions in cardiac nuclear translocation of Excessive oxidative stress induced by hyperglycaemia is a major mechanism of DCM and the subsequent cardiac dysfunction both in human and in animal models of diabetes. 29,36 In our study, diabetic rats showed asymptomatic cardiac dysfunction manifested as significant reductions in heart rate (HR), left ventricular eject fraction (LVEF), stroke work (SW) and cardiac output (CO), left ventricular F I G U R E 4 Nrf2/p62 signalling pathway played the key role in the ALP protection to high glucose-induced cardiomyopathy. A, B, Protein and mRNA expression of Nrf2 after siRNA treatment in the normal glucose. C, Protein expression of Nrf2. D, Protein expression of Keap1. E, Protein expression of p62. F, The ratio of LC3 II/LC3 I in the H9C2 cells. G-I, The mRNA expressions of Nrf2, p62 and Keap1, respectively, in H9C2 cells, calculated against the housekeeping gene GAPDH. J, Nuclear Nrf2 expression in high glucose cultured H9C2 with or without ALP. K, Treatment with Bafilomycin (Baf), a lysosomal inhibitor that can impair autophagosome-lysosome fusion and thus track autophagic flux, induced higher P62 protein level than NG group, whereas HG group displayed decreased P62 protein level than NG group. Data are mean ± SEM of two independent experiments each performed in triplicate, *P < .05. vs NG; # P < .05 vs HG; $ P < .05 vs HA; & P < .05 vs Nrf2 F I G U R E 5 Schematic of proposed signalling involved in ALP attenuates DCM and cardiac dysfunction via activating Nrf2 pathway. Hyperglycaemia-induced oxidative stress destructs Nrf2 signalling by inhibiting its nuclear translocation, which concomitantly with activated autophagy and subsequently reduced HO-1 and p62 expression, resulting in cardiomyocytes apoptosis and cardiac hypertrophy. ALP activates Nrf2 by increasing its nuclear translocation following with target on ARE elements (including p62 and HO-1) and concomitantly normalizes autophagy, which in turns reduces myocardial oxidative stress, attenuates cardiac hypertrophy and apoptosis and eventually improves cardiac function end-systolic volume (LVVs) as compared to non-diabetic control and by restoring Nrf2 and HO-1 expressions, 42 and enhancement of Nrf2 has recently been shown to be a mechanism by which treatments attenuate DCM. 43,44 We thus postulated that enhancement of Nrf2 may be a major mechanism whereby ALP attenuates DCM.
This mechanism may also explain why ALP could attenuated DCM despite it had no significant effect on plasma levels of glucose in diabetes.
There is general consensus that low levels of ROS act as intracellular signal transducers that activate autophagy, promoting cell survival, whereas high levels of ROS induce apoptosis by damaging cellular components. It is interesting to note that the redox system can potentially be a switch of autophagy and apoptosis and function as a crosstalk in between autophagy and apoptosis. 45 Coincidentally, our results showed increases in autophagy and apoptosis simultaneously occurred in the condition of high ROS both in vivo and in vitro.
ALP reduced apoptosis and attenuated autophagy in diabetic or high glucose condition. Findings of our current study provided a hint that Nrf2 pathway might regulate the crosstalk in between apoptosis and autophagy in the diabetic myocardium. It is worth noting that ALP treatment did not improve the Bcl2 levels in diabetic rats, which means that other pathways linking apoptosis and autophagy may also play a role in the protective effect of ALP. Transactivation of antioxidant genes may block apoptosis and serve as a feedback loop to reduce autophagy. In addition to oxidative stress, nutrient deprivation resulted from metabolic disorder of glucose also activates autophagy, which acts to restore metabolic homeostasis via the degradation of macromolecules to provide nutrients. 46 In the pres- This is largely because that the mechanism of antioxidant therapy is complicated. Of note, the cardioprotection of ALP was obviously observed in diabetic rats in this study, as well as in Szwejkowski's study. 51 In addition to the functional studies in vivo using both intra cardiac and echocardiographic approaches, we also detected the signalling changes that involve the aforementioned Nrf2 pathway in vivo and in vitro. After 4 weeks treatment with ALP in diabetic rats, the ratio of LC3 II/I reduced and P62 protein expression increased to levels comparable to that in the control group, while lower protein expression of Keap1 was accompanied with higher protein expression of Nrf2 in the nucleus as compared to untreated diabetic group. In vitro, Nrf2 gene silence cancelled the cardiac protective effects of ALP against high glucose-induced H9C2 injuries. Thus, it is plausible that ALP confers cardioprotection in diabetes by normalizing the disordered autophagy and to consequently restore the disrupted Nrf2-Keap1-p62 loop in diabetes.
Diabetic cardiomyopathy occurred in patients with both type 2 and type 1 diabetes, and the mechanisms and clinical features of diabetic cardiomyopathy may be partially distinct in type 1 vs type 2 diabetes. 52 With the development of diabetes, hyperglycaemia-induced oxidative stress and damage have been considered as a major mechanism of diabetic cardiomyopathy. 53 In the current study, we confirmed the development of diabetic cardiomyopathy STZ-induced type 1 diabetic rats by evaluating cardiac function, pathological changes and biological abnormality and demonstrated the effectiveness and mechanism of the antioxidant ALP on DCM.

| CON CLUS ION
The findings of the current study suggest that the impaired feed-forward loop linking Nrf2, p62 and autophagy is a critical mechanism for the out-of-balance redox homeostasis induced by hyperglycaemia in type 1 diabetes as illustrated ( Figure 5), which eventually result in the development of DCM. In the early stage of diabetic rats, diabetes exited some scathing heart functional changes that can be detected by the sensitive echocardiography and PV-loop system, including the lower HR, LVEF, stroke work and cardiac output. The excessive oxidative stress is the main pathogenesis for the broken of the feed-forward loop and subsequent recession of antioxidant capability. Thus, restoration of the disorganized autophagy and the subsequent repairing of Nrf2/p62 pathway may represent a major mechanism whereby ALP confers its cardioprotection against DCM. Nrf2 is the key director in this scenario.

CO N FLI C T O F I NTE R E S T
The authors confirm that there are no conflicts of interest. The data are available from the corresponding authors upon reasonable request.