Meet the First Author, see p 1218

Epidemiological and clinical studies show that the prevalence of asymptomatic left ventricular dysfunction and heart failure are very high among patients with diabetes mellitus.^1–3 Moreover, diabetes mellitus is linked to a progressive increase of death and hospitalization for heart failure.⁴ Long-term hyperglycemia may, even in the absence of other risk factors, impair myocardial function.⁵ Hyperglycemia increases reactive oxygen species (ROS) generation by activating different oxidative pathways.^6,7 ROS trigger transcriptional programs leading to myocardial inflammation, mitochondrial derangement, apoptosis, fibrosis, and contractile dysfunction.^8,9 Hence, a detailed understanding of the molecular networks regulating cardiac oxidative stress may furnish new targets to prevent heart failure in people with diabetes mellitus.

Despite intensive investigation, the link between hyperglycemia and altered redox state triggering cardiac accumulation of ROS remains poorly characterized. Recent evidence from our group supports the notion that the JunD (Jund proto-oncogene AP-1 [activator protein-1] transcription factor subunit) is a key molecule implicated in ROS-driven cardiovascular aging.¹⁰ AP-1 is a collection of dimeric complexes composed of several proteins belonging to the c-Fos (proto-oncogene c-Fos subunit), c-Jun (proto-oncogene c-Jun subunit), ATF (activating transcription factor), and JDP (Jun dimerization protein) families.¹¹ AP-1 modulates gene expression in response to a variety of stimuli, including bacterial and viral infections, stress, cytokines, and growth factors. Moreover, this molecular complex regulates transcriptional programs of cellular differentiation, proliferation, and apoptosis.¹¹ A critical role for the AP-1 member JunD was demonstrated not only in cell growth and survival but also in redox signaling by modulating genes either involved in antioxidant defense or ROS production.¹² Accordingly, JunD^−/− mice exhibited short life span, increased incidence of cancers and premature endothelial dysfunction.^10,13 Genetic disruption of JunD promotes pressure overload-induced apoptosis, hypertrophic growth, and angiogenesis in the heart.¹⁴ In the present study, we have investigated the role of JunD in hyperglycemia-induced and ROS-driven myocardial dysfunction.

Methods

Data Availability

The authors declare that all supporting data are available within the article and its Data Supplement. All other materials that support the findings of this study are available from the corresponding author upon reasonable request. Please see the Major Resources Table in the Data Supplement.

Generation of Cardiomyocyte-Specific JunD Transgenic Mice

Transgenic mouse model with cardiomyocyte-specific overexpression of JunD was generated as previously described.¹⁵ Briefly, the α-MHC (α -myosin heavy chain)–JunD transgene construct consists of the entire JunD cDNA cloned into the Sall-digested pMHC (α myosin heavy chain promoter) poly A vector. The construct was injected into the male pronucleus of fertilized single-cell embryos to produce cardiac-specific α-MHC-JunD^tg. The α-MHC-JunD^tg mice were generously provided by the Center for Molecular Cardiology, University of Zurich, Switzerland.

Induction of Diabetes Mellitus

Diabetes mellitus was induced in 4-month old male C57BL/6 wild-type (WT) and cardiac-specific α MHC JunD^tg mice by a single high dose of streptozotocin (180 mg/kg, via intraperitoneal injection), dissolved in sterile 0.025M citrate buffer (pH 4.5) and injected within 10 minutes. An equal volume of citrate buffer was administered in control animals. Animals were randomly divided into 4 experimental groups: (1) WT control, (2) JunD^tg control, (3) WT diabetes mellitus, (4) JunD^tg diabetes mellitus (Figure I in the Data Supplement). Hyperglycemia was defined as 3 random blood glucose levels >13.9 mmol/L following streptozotocin injection (Figure II in the Data Supplement). Mice were housed in temperature-controlled cages (20-22°C), fed ad libitum, and maintained on a 12:12-hour light/dark cycle. After 4-week follow-up, cardiac function was assessed and then mice were euthanized and organs, such as heart, kidney, liver, spleen, and aorta, were harvested for molecular analyses. All animal experiments were approved by the regional ethics committee for animal care. The experimenter was blinded with regard to experimental group during outcome assessment.

Human Cardiac Tissue

In collaboration with the Cardiothoracic Surgery Unit, the Heart and Vascular Theme of Karolinska University Hospital (NKS [Nya Karolinska Solna], Solna), left ventricular specimens were obtained from patients with and without diabetes mellitus (duration of disease >10 years), who had undergone elective open-heart surgery for aortic valve replacement (ASAP study [Advanced Study of Aortic Pathology]) or coronary bypass surgery (CABG PREFERS [Coronary Artery Bypass Graft surgery Preserved and Reduced Ejection Fraction Epidemiological Regional Study] Stockholm). Both groups were matched by age, gender, body mass index, and pharmacological treatments. Tissue samples were taken with needle biopsies from the lateral wall of left ventricle and immediately stored at −80°C for molecular analyses.¹⁶ The study protocol was approved by the Human Research Ethics Committee at Karolinska Institutet (license numbers 2018/255-32/1, 4-644/2018, and 2006/784-31/1) and performed in accordance with institutional guidelines. At the time of cardiac biopsy all participants were aware of the investigational nature of the study and signed a written consent.

Electron Spin Resonance Spectroscopy

Superoxide anion (O₂^–) was measured in mouse heart by electron spin resonance spectroscopy using the spin probe 1-hydroxy-3-methoxycarbonyl-2,2,2-tetramethyl-pyrrolidine. Additional information is provided in the Data Supplement.

Real-Time Polymerase Chain Reaction

Total RNA was extracted from mouse and human tissues with TRI reagent according to instructions provided by manufacturer (Sigma Aldrich St Louis). RNA conversion into cDNA was carried out with high capacity cDNA reverse transcription kit (Applied Biosystems Carlsbad) according to manufacturer´s protocol. Polymerase chain reaction (PCR) amplification was performed in an MX3000P PCR cycler (Stratagene) using the SYBR Green JumpStart kit (Sigma Aldrich, St Louis). Two microliters of cDNA, 10 pmol of each primer, 0.25 μL of internal reference dye, and 12.5 μL of JumpStart Taq ReadyMix (buffer, dNTP [deoxynucleotide], stabilizers, SYBR Green, Taq polymerase, and JumpStart Taq antibody) were mixed to a final volume of 25 μL. Real-time PCR amplification program was as follows: enzyme activation (2 minutes at 95°C), followed by amplification and real-time analysis (40 cycles at 95°C for 15 s and 60°C for 60 s). GAPDH (Glyceraldehyde 3-phosphate dehydrogenase) and ACTB (Actin beta) were used as endogenous controls. The differences in cycle threshold values between gene of interest and housekeeping gene were calculated either as 2^−ΔCt or as fold differences using the comparative 2^−ΔΔCt method and used in statistical analysis. Mouse and human primers used in this study are listed in Table I in the Data Supplement.

Coimmunoprecipitation

This method is reported in the Data Supplement.

Western Blotting

Mouse and human tissues were lysed in lysis buffer (150 mmol/L sodium chloride, 50 mmol/L Tris, 1 mmol/L sodium fluoride, 1 mmol/L dithiothreitol (DTT), 1 mmol/L EDTA, 10 µg/µL leupeptin, 10 µg/µL aprotinin, 0.1 mmol/L sodium vanadate, 1 mmol/L PMSF [phenylmethylsulfonyl fluoride], and 0.5% NP-40 [nonionic polyoxyethylene surfactant]). Equal volumes of 20 µg of protein samples were loaded to SDS-PAGE gel for electrophoresis followed by semidry transfer of proteins onto Immobilon-P polyvinylidene fluoride (PVDF) membranes (Millipore, Billerica). The membranes were then blocked with 5 % nonfat dry milk in PBS-Tween buffer (0.1% Tween 20; pH 7.5) for 1 hour and incubated with primary antibodies against JUND (sc-74 [clone 329]; Santa Cruz Biotechnology), NOX2 (NADPH [nicotinamide adenine dinucleotide phosphatase] oxidase subunit 2) (sc-74514 [clone G-1]; Santa Cruz Biotechnology), NOX4 (NADPH [nicotinamide adenine dinucleotide phosphatase] oxidase subunit 4) (sc-30141 [clone H-300]; Santa Cruz Biotechnology), ALDH2 (aldehyde dehydrogenase 2; sc-166362 [clone H-8]; Santa Cruz Biotechnology), SOD1 (superoxide dismutase 1; ab16831; Abcam), MEN1 (menin 1; A300-105A; Bethyl Laboratories), VCL (Vinculin) (ab129002 [EPR8185]; Abcam), and GAPDH (G9545; Sigma). Anti-mice or anti-rabbit secondary antibodies (sc-516102 [IgGκ BP] and sc-2357 [IgG], respectively; Santa Cruz Biotechnology) were used to detect the corresponding proteins of interest. Protein bands were detected by chemiluminescence system (Millipore, Billerica) and related signals were quantified by using Image J software (National Institutes of Health, Bethesda, MD).

Nuclear Factor Kappa B p65 binding activity

NF-κB (nuclear factor kappa B) p65 DNA binding activity was performed by using commercially available kit (Active Motif, Rixensart, Belgium). Additional information is provided in the Data Supplement.

Chromatin Immunoprecipitation Assay

Chromatin immunoprecipitation (ChIP) on mouse heart was performed as previously described.¹⁷ Briefly, mouse heart tissues were diced and fixed in 1 % of formaldehyde solution. The cross-linking of homogenate continued for 10 minutes at room temperature before quenching with 125 mmol/L of glycine for 10 minutes. After quenching, the homogenate was lysed in SDS lysis buffer and was sonicated to obtain chromatin fragments of 200 to 500 bp using a water bath sonicator (Diagenode: 30s On–30s Off for 15 minutes). Immunopurification of soluble chromatin was carried out by using antibodies against JunD (ab28837; Abcam), histone 3 lysine 4 monomethylation (H3K4me1; ab8895; Abcam), histone 3 lysine 4 trimethylation (H3K4me3; ab8580; Abcam), H3K9me3 (ab8898; Abcam), and an equal amount of rabbit IgG (171870; Abcam) antibody was used as negative control. Chromatin fraction bound to antibody was precipitated by using dynabeads coated with protein A (Invitrogen). Washing steps, reverse cross-linking, and purification of DNA conjugates were performed according to previously described protocol.¹⁷ Purified DNA sequences were detected by using real-time PCR system. Quantifications were performed by calculating percentage (%) input for each ChIP experiment, and results are reported as relative fold enrichment/ratio for target sequences compared between experimental and control groups. Primers used in ChIP quantitative PCR are listed in Table I in the Data Supplement.

Analysis of DNA Methylation

Genomic DNA isolated from mouse heart was sheared to 150 to 300 bp using Diagenode sonicator. Methylated DNA enrichment was performed using a Methylminer Kit (Invitrogen) as described.¹⁵ Briefly, methylated DNA fragments were captured via binding of MBD (methylcytosine binding domain) protein coupled to dynabeads by incubation for 1 hour at room temperature with rotation. Methylated DNA fragments were eluted with 2 M of NaCl buffer and precipitated with ethanol. Purified DNA fractions were measured using real-time quantitative PCR with MX3000P PCR cycler (Stratagene) and fluorescence-based SYBR green technology (Invitrogen). The primers used to detect DNA methylation levels at CpG (cytosine-phosphate-guanine) island in JunD gene are listed in Table I in the Data Supplement.

Conventional Echocardiographic Measurements

After 4 weeks from diabetes mellitus induction, mice were anesthetized with 2% to 5% of isoflurane mixed with oxygen. Echocardiography was performed by high-resolution Micro-Ultrasound System (Vevo 2100; VisualSonics) equipped with a 22 to 55 MHz (MS550D) linear array transducer. Mice body temperature was maintained at ≈37°C using heating pad and heating lamp. The chest was shaved and prewarmed ultrasonic gel was applied throughout the measurements. M-mode images were obtained at the midpapillary level in the parasternal short-axis and long-axis views, and the average was reported. B-mode images were also recorded in the parasternal short-axis and long-axis views.

Two-Dimensional Speckle-Tracking–Based Strain Measures of Myocardial Deformation

Parasternal long-axis views were found to provide the most reproducible myocardial views for longitudinal strain analyses in mice, whereas parasternal short-axis views (at the midpapillary level) were obtained for circumferential and radial (short-axis) strain analyses. All images were acquired at a frame rate of >200 frames per second and at an average depth of 11 mm. Strain analyses were conducted by the same trained investigator on all mice from the four experimental groups using a speckle-tracking algorithm provided by VisualSonics (VevoStrain, VisualSonics).

Cell Culture Experiments

Human cardiomyocytes were purchased from Celprogen and grown in optimized growth media. Cells were detached by using Trypsin/EDTA for 2 minutes and reseeded in 3-cm cell culture dishes for the experiment. Neonatal rat ventricular myocytes (NRVM) were isolated from Sprague Dawley neonatal rat pups aged 0 to 3 days. Following euthanasia by cervical dislocation, the chest of the pups was opened with small scissors, and forceps were used to remove the heart. NRVM were then harvested by enzymatic dissociation and cultured in DMEM supplemented with 10% FBS and 1% penicillin/streptomycin and plated in 6-well plates at 37°C in a 5% CO₂ incubator. Before plating, viability and number of NRVM were assessed, as previously described.¹⁸ Cardiomyocytes and NRVM were exposed for 5 days either to normal glucose (5 mmol/L) or high glucose concentration (25 mmol/L). Mannitol was purchased from Sigma Aldrich (St Louis) and used at the final concentration of 25 mmol/L as osmotic control. Small interfering RNA transient cell transfection experiments were performed in 6-well plates at 24 hours after seeding (60% confluence). For small-interfering RNA–mediated Nox2 and Nox4 knockdown, 25 nmol/L of Nox2- or Nox4-specific silencer RNA (sc-35503 and sc-41586, respectively; Santa Cruz Biotechnology) or scrambled control silencer RNA (1374487, Microsynth) were transfected in human cardiomyocytes using Lipofectamine 3000 reagent (Thermo Scientific). Serum-free media were changed after 6 hours, and cells were further incubated in growth medium for a total of 48 hours. For in vitro miRNA (microRNA)-673 overexpression, 20 nmol/L mimic-673 or allStars negative control (mimic-NC [negative control], Qiagen) were transfected in NRVM.

Statistical Analysis

Data are presented as scatter plots with bars and mean±SEM for normally distributed data or box and whisker plots with data points and median±interquartile ranges for non-normally distributed data. The normality of continuous variables was assessed by Kolmogorov-Smirnov test. Unpaired 2-tailed Student t tests as well as the 1-way ANOVA corrected for multiple comparisons by Bonferroni post hoc test (all-pairwise comparisons) or by controlling the false discovery rate with the 2-stage linear step-up procedure of Benjamini, Krieger, and Yekutieli method were used to compare means of normally distributed continuous data. Statistical comparisons of non-normally distributed continuous data or data for which normality could not be assessed were made by using Mann-Whitney U test and Kruskal-Wallis followed by Dunn multiple comparisons test. Categorical variables were compared by using the χ² test. Spearman correlation test was used to assess the correlation between variables. The exact test used for each data set and the number of samples are mentioned in figure and table legends and P values are adjusted for multiple comparison when appropriate. We did not adjust for multiple testing across tests. Probability values <0.05 were considered statistically significant. All analyses were performed with GraphPad Prism (version 6.03, GraphPad Software), SPSS (version 25; IBM Statistics) and Statview (SAS Institute, Inc). Representative images for figures were selected manually to represent the mean value of each group.

Results

Hyperglycemia Downregulates JunD in the Heart

To investigate the contribution of JunD in the diabetic heart, we first assessed its expression in left ventricular specimens from diabetic and nondiabetic mice. Gene and protein expression of transcription factor JunD were significantly reduced only in the heart of diabetic WT mice as compared to controls (Figure 1A and 1B). Hyperglycemia-induced JunD downregulation was also demonstrated in vitro in cultured human cardiomyocytes exposed to normal or high glucose (Figure III in the Data Supplement).

• Download figure
• Download PowerPoint

Transgenic Mouse Model With Cardiac-Specific Overexpression of JunD

Cardiomyocyte-specific transgenic JunD mice (α MHC JunD^tg) were used to unmask the role of JunD in the diabetic heart. Genomic PCR and protein analysis confirmed selective JunD overexpression in the heart of α-MHC-JunD^tg as compared to WT mice (Figure IV in the Data Supplement). Accordingly, the expression of JunD did not significantly differ in aorta, liver, and kidney between the 2 groups (Figure IV in the Data Supplement).

Role of JunD in Myocardial Oxidative Stress

Downregulation of JunD in WT mice was associated with myocardial superoxide anion (O₂⁻) generation, as assessed by electron spin resonance spectroscopy (Figure 2A). By contrast, O₂⁻ levels were not elevated in α-MHC-JunD^tg diabetic mice (Figure 2A). Of note, no statistical difference was observed in O₂⁻ generation with cardiac-specific overexpression of JunD under control conditions (Figure 2A).

• Download figure
• Download PowerPoint

JunD Downregulation Impairs Expression of Redox Enzymes

Gene and protein expression of SOD1 and ALDH2 were decreased in the heart of WT diabetic mice (Figure 2B and 2C). Whereas, mRNA and protein levels of NADPH oxidase subunits NOX2 and NOX4 were upregulated as compared to WT controls (Figure 2D and 2E). Interestingly, such detrimental changes did not occur in the heart of α-MHC-JunD^tg mice with diabetes mellitus (Figure 2B through 2E). The specificity of Nox2 and Nox4 antibodies was validated using the small interfering RNA reverse transfection and Western blot (see Figure V in the Data Supplement). In agreement with JunD downregulation, ChIP assay showed that JunD binding to the promoter of antioxidant (Sod1 and Aldh2) and pro-oxidant (Nox2 and Nox4) genes was significantly reduced in WT mice with diabetes mellitus (Figure 3A through 3D).

• Download figure
• Download PowerPoint

JunD-Dependent and NF-κB–Dependent Inflammatory Pathways

Given the pivotal role of ROS in activating NF-κB–dependent pathways, we investigated inflammatory transcriptional programs in our setting. NF-κB activity was enhanced in the heart of WT diabetic mice as compared to controls (Figure 4A), whereas gene expression of the NF-κB inhibitory subunit IκB was reduced (Figure 4B). Accordingly, the expression of NF-κB–dependent inflammatory genes Mcp-1, Il-6, and Tnf-α was increased (Figure 4C). Mice with cardiac-specific overexpression of JunD were protected against such derangement of inflammatory pathways (Figure 4A through 4C).

• Download figure
• Download PowerPoint

Cardiac-Specific Overexpression of JunD Prevents Left Ventricular Dysfunction

We performed standard and speckle-tracking echocardiography to assess cardiac function. When compared to controls, WT diabetic mice showed an impairment of left ventricular (LV) systolic function, as reflected by a reduced fractional shortening and ejection fraction (Figure 5A). Myocardial strain measures revealed a clear impairment of LV performance in WT diabetic mice as compared to controls (Figure 5B). Interestingly, such abnormalities were not observed in α-MHC-JunD^tg diabetic mice (Figure 5A and 5B). Of note, α-MHC-JunD^tg control mice did not show any structural and functional abnormalities when compared to WT controls (Table II in the Data Supplement).

• Download figure
• Download PowerPoint

Epigenetic Regulation of JunD Expression

To investigate whether JunD expression was modulated by epigenetic changes, first we assessed epigenetic remodeling on JunD promoter. DNA methylation level of CpG islands in the promoter region was analyzed using Methylminer kit and quantitative PCR (Figure 6A). Methylation of JunD promoter was significantly elevated in the heart of WT mice with diabetes mellitus as compared to controls (Figure 6B). Since it is well established that DNA methylation modulates gene expression by clustering with histone modifications, post-translational changes of histone H3 tail at JunD promoter were determined. Specific active and repressive histone marks were selected within ChIP-Seq data from human roadmap epigenomics and ENCODE (Encyclopedia of DNA Elements) projects in the genome browser of University of California Santa Cruz.¹⁹ Interestingly, the active marks H3K4me3 and H3K4me1 assessed by ChIP assay were significantly reduced in WT diabetic as compared to control hearts (Figure 6C). On the contrary, we found that repressive H3K9me3 was increased in the heart of WT diabetic mice as compared to controls (Figure 6C). Furthermore, post-translational repression by tumor suppressor MEN1, a critical modulator of JunD activity, also contributes to JunD downregulation. Indeed, MEN1 was significantly upregulated in the heart of WT diabetic mice as compared to controls (Figure 6D) and coimmunoprecipitation experiments showed binding of MEN1 to JunD (Figure 6E). Based on an unbiased mouse miRNome profiling in heart specimens of WT control and diabetic mice and an in silico prediction analysis, we found that miRNA-673 targeting Men1 was downregulated in hearts of WT diabetic mice (Figure 6F). Furthermore, in rat ventricular myocytes exposed to high glucose, we found increased MEN1 expression and JunD downregulation (Figure 6G and 6H). Overexpressing miRNA-673 in high glucose–treated myocytes was able to restore both MEN1 and JunD expression to control levels (Figure 6G and 6H).

• Download figure
• Download PowerPoint

JunD Expression in Myocardial Specimens of Patients With Diabetes Mellitus

To translate the findings observed in mice to the human clinical setting, we assessed JunD mRNA expression in left ventricle specimens obtained from patients with diabetes mellitus and matched control subjects. Patient’s clinical characteristics are reported in Table III in the Data Supplement. JunD mRNA expression was significantly reduced in patients with diabetes mellitus as compared to controls (Figure 7A). In accordance with our experimental findings, a significant correlation between hyperglycemia (assessed by HbA_1c [glycated haemoglobin]) and JunD downregulation was found (Figure 7B).

• Download figure
• Download PowerPoint

Discussion

The present study demonstrates for the first time a protective role of the AP-1 transcription factor JunD against deranged ROS homeostasis, inflammation, and myocardial dysfunction in the setting of experimental and human diabetes mellitus. Although JunD has recently emerged as a pivotal player in the setting of aging and metabolism, no previous work has investigated the effects of hyperglycemia, a hallmark of diabetes mellitus, on JunD signaling. Our findings indicate that high glucose levels strongly deregulate JunD expression thus fostering accumulation of free radicals and inflammatory transcriptional programs in the diabetic heart. Here, we show that JunD was downregulated in the myocardium of mice with type 1 diabetes mellitus as compared to nondiabetic controls, and this was associated with a downregulation of free radical scavengers, increased expression of ROS-generating NADPH oxidase and markedly enhanced myocardial O₂⁻ generation. These redox changes were paralleled by activation of NF-κB–dependent inflammatory pathways and LV dysfunction. Interestingly, in mice with cardiac-specific overexpression of JunD, these detrimental changes did not occur and LV function was preserved confirming the protective role of JunD in diabetes mellitus. Moreover, JunD expression was also significantly reduced in LV specimens of patients with diabetes mellitus as compared to age-matched controls, providing a translational perspective of our findings to the clinical context.

The transcription factor JunD is a pivotal modulator of oxidative stress levels and modulates different genes involved in proliferation, growth, and survival.^20–22JunD^−/− immortalized cells showed higher ROS generation.¹² We recently demonstrated that deletion of JunD is associated with premature endothelial dysfunction and vascular aging via ROS production. Accordingly, JunD^−/− mice display reduced life span.^13,23,24 Given the important role of ROS in the pathogenesis of diabetic cardiomyopathy^5,25 we were prompted to investigate whether JunD modulates oxidative stress and inflammation in the diabetic heart.

In the present study, we observed that increased superoxide generation in the diabetic myocardium was coupled with an imbalance between ROS-scavenging and -generating enzymes. Indeed, expression of SOD1 and ALDH2 were significantly reduced in WT diabetic mice suggesting that JunD is involved in their transcription. ALDH2 has been recently reported to protect against diabetes mellitus–induced cardiac dysfunction.²⁶ In addition, ALDH2 also protects against cardiac arrhythmias by reducing ischemia/reperfusion injury.^27–29 Here, we extend these findings by showing that JunD downregulation impairs ALDH2 expression in the diabetic heart and its reduced expression may contribute to the abnormal myocardial redox state. Our findings are in line with the established role of AP-1 in activating several genes implicated in the detoxification defense system.³⁰ Indeed, antioxidant response elements have been found in the promoter regions of ROS-scavenging enzymes.³¹ We had previously reported a reduced vascular expression of SOD1 and ALDH2 contributing to oxidative stress and endothelial dysfunction in young JunD^−/− mice.¹⁰ However, we found that NOX2 and NOX4, key subunits of ROS-generating NADPH oxidase, are upregulated in the heart of WT diabetic mice as compared to controls. Interestingly, JNK1/c-Jun pathway leads to enhanced AP-1 activity and transcription of Nox2/4 in the absence of JunD, indicating that JunD plays an important role in suppressing AP-1 transcriptional activation of both NADPH oxidases.³² Consistently with the pivotal role of ROS in activating NF-κB-dependent inflammatory pathways,³³ gene expression of NF-κB inhibitory subunit was reduced, while NF-κB binding activity was increased and Mcp-1, Il-6, and Tnf-α genes were enhanced in the myocardium of diabetic mice.

By contrast, α MHC JunD^tg mice were protected against oxidative stress, myocardial inflammation, and LV dysfunction as assessed by conventional and 2-dimensional speckle-tracking echocardiography. These latter findings strongly suggest that modulation of JunD expression plays a key role in ROS-driven inflammation and cardiac dysfunction. Altogether, these findings suggest that downregulation of JunD with subsequent oxidative stress and inflammation represents a key mechanism of myocardial damage in the diabetic heart, thus opening perspectives for the development of new therapeutic approaches to prevent diabetes mellitus–induced LV dysfunction.

Previous work found ventricular dilation and decreased contractility in 8-week old cardiomyocyte-specific JunD transgenic mice as compared to age-matched WT animals.¹⁵ In our study, echocardiography in 20-week old α-MHC-JunD^tg control mice showed no significant alterations in left ventricular mass, heart contractility, as well as ventricular dilation as compared to WT. Although we do not have a thorough explanation for this discrepancy, it is possible that the modulation of JunD expression may promote either maladaptive or protective responses in the heart, depending on age and different experimental conditions. In line with our results, adenoviral overexpression of wild-type JunD blunts phenylephrine-mediated cardiomyocyte hypertrophy.³⁴ Moreover, another study showed that JunD protects against apoptosis and hypertrophy in the myocardium in response to pressure overload.¹⁴ We observed an upregulation of JunD mRNA expression in the kidney and liver of diabetic mice as compared to control mice, whereas no changes occurred in spleen and aorta (data not shown). Two recent studies have demonstrated that JunD is activated by fatty acid signaling and is critically implicated in the pathogenesis of liver and myocardial steatosis.^35,36 Indeed, mice lacking JunD were protected against obesity cardiomyopathy by repressing the activation of PPARγ (Peroxisome proliferator-activated receptor gamma) and PPARγ-related genes involved in myocardial triglyceride uptake and ceramide biosynthesis.³⁶ Therefore, available evidence indicates that JunD may act as a molecular switch of cellular stress in the heart, and different stimuli (ie, fatty acids, hyperglycemia) may turn JunD signaling into protective or maladaptive effects.

We next investigated the mechanisms underlying hyperglycemia-induced JunD downregulation and we uncovered a complex regulation of this transcription factor JunD by epigenetic changes. Growing evidence indicates that epigenetic signals, which include DNA methylation, histone post-translational modifications, and noncoding RNAs represent a key biological layer orchestrating gene expression and determining cell phenotype.^37,38 Indeed, although genetics play a pivotal role in diabetes mellitus–related cardiovascular diseases, their nongenetic regulation by environmentally induced epigenetic changes is gaining increasing attention.³⁹ Our quantitative analysis of JunD promoter methylation showed a significant hypermethylation of CpG dinucleotides in the heart of diabetic mice as compared to nondiabetic animals. Changes in DNA methylation were accompanied by an array of chromatin modifications including activating (H3K4me3/H3K4me1) and repressive (H3K9me3) marks. Together with chromatin remodeling, we also report an interesting function of the tumor suppressor MEN1, a pivotal repressor of JunD activity in our setting.^40–42 We found that MEN1 was negatively regulated by miR-673, and this axis was involved in the suppression of JunD transcriptional activity. Indeed, miR-673 overexpression prevented MEN1 upregulation thus restoring JunD expression in glucose-treated cardiomyocytes. Interestingly, MEN1 is also involved in chromatin remodeling as it has shown to recruit the histone methyltransferase Suv39h1 (Suppressor of variegation 3-9 homolog 1), eventually promoting H3K9me3 enrichment at the promoter of target genes. MEN1 is also a component of the MLL (Mixed lineage leukemia)/Set1 (SET domain containing 1) histone methyltransferase complex which leads to H3K4me1/H3K4me3 marks, and, most notably, shares the same binding pocket for JunD and MLL.^41,43,44 Taken together, these findings suggest that MEN1 activation in diabetes mellitus might also regulate key chromatin marks (H3K4me1, H3K4me3, H3K9me3) implicated in JunD downregulation.

Of clinical relevance, we also found that JunD expression was downregulated in heart tissues obtained from diabetic patients as compared with nondiabetic controls and was negatively correlated with HbA_1c. Although in our human cohort we cannot rule out the effect of insulin resistance and other risk factors (eg, hypertension, dyslipidemia, obesity), our data strongly indicate that downregulation of JunD is driven by hyperglycemia and may contribute to ROS accumulation and subsequent cardiac damage providing a translational perspective to our findings. Moreover, in the present study, we demonstrate that hyperglycemia-induced JunD inactivation is relevant both in type 1 (streptozotocin-treated mice) and type 2 diabetes mellitus (patients) and participates to myocardial oxidative stress, inflammation, and LV dysfunction.

The use of only male mice may represent a limitation of this study. However, it enables the reduction of variability thereby limiting the number of animals needed to detect significant differences among experimental groups.

In conclusion, the present work demonstrates a protective role of JunD in the heart, a mechanism that is defective in the presence of diabetes mellitus (Figure 8). A complex epigenetic machinery was found to fine-tune this process. These results contribute to uncover the role of cardiac epigenome in the etiological pathway linking hyperglycemia and diabetic cardiomyopathy and may set the stage for mechanism-based therapeutic approaches targeting JunD in the diabetic heart.

• Download figure
• Download PowerPoint

Supplemental Materials

Detailed Methods

Online Figures I–V

Online Tables I–III

Footnotes