Disturbed Fatty Acid Oxidation, Endoplasmic Reticulum Stress, and Apoptosis in Left Ventricle of Patients With Type 2 Diabetes

Chronic heart failure is a common complication in patients with type 2 diabetes mellitus (T2DM) (1,2). Being a complex metabolic disorder, T2DM is associated with accelerated atherosclerosis and coronary artery disease (CAD), leading to an increase in incidence of myocardial infarction, heart failure, and ischemic heart disease–associated death (1). Nevertheless, cardiac dysfunction in T2DM can occur even in the absence of CAD as a part of diabetic cardiomyopathy. Each 1% increase in HbA_1c was previously shown to result in 8% greater risk of heart failure, independent of other risk factors (3). Various mechanisms have been suggested as potential culprits of myocardial damage in diabetes, ranging from maladaptive inflammatory response and endocrine disorder to altered myocardial metabolism and utilization of energetic substrates (4).

T2DM also leads to altered metabolism of fat. This results in increased lipolysis, elevated levels of circulating fatty acids, and their enhanced uptake by peripheral tissues, including the heart (4). Consequently, diabetic myocardium “burns” more fatty acids than nondiabetic (5). However, it is not clear whether this is a physiological metabolic response to excess of fatty acid substrate and intracellular deficit of glucose (6), or it can be detrimental for the heart (7).

Data from patients with diabetes are very sparse and are mostly obtained by MRI and positron emission tomography (PET) measurements of cardiac contractile performance, lipid deposition, and energy substrate uptake (8). There are currently two studies that have directly investigated mitochondrial function in human diabetic atrial tissue, both showing that diabetes is associated with mitochondrial dysfunction (9,10). However, to date, there are no studies investigating these processes in human ventricular muscle, part of the myocardium most stricken by the diabetes complications. Also, atrial myocardium differs significantly from ventricular in mitochondrial respiratory capacity (11), activity of enzymes involved in aerobic metabolism (12), and response to noxious stimuli leading to chronic heart failure (13).

In the current study, we aimed to fill the gap between findings obtained in animal models of the disease, functional data recorded in imaging studies, and reports from human right atrium by directly assessing the mitochondrial capacity for oxidation of carbohydrates and fatty acids in the samples of left ventricle obtained from patients undergoing coronary artery bypass grafting (CABG) surgery. Additionally, we assessed some of the intracellular pathways that are linked to mitochondrial function and cellular energy balance and which can directly result in loss of functional myocytes, such as endoplasmic reticulum (ER) stress and apoptosis.

Thirty-seven hemodynamically stable CAD patients scheduled for elective CABG surgery at the University Hospital Split were included. Emergency patients, patients with left ventricular (LV) ejection fraction (LVEF) <50%, and patients with type 1 diabetes, concomitant valve replacement, and severe renal, hepatic, or pulmonary disease were excluded.

The included patients fell into one of the two group categories: nondiabetic (non-DM) group and diabetic (DM) group, based on clinical diagnosis of T2DM, the use of diabetes medication, and measurements of fasting plasma glucose (>7 mmol/L) or glycosylated hemoglobin (HbA_1c >6.5%). Pre, intra- and postsurgical procedures were performed according to the standard clinical routines of the Department of Cardiac Surgery, University Hospital Split. The study complies with the Declaration of Helsinki and was approved by the ethics committees of the University Hospital Split (2181-147-01) and University of Split School of Medicine (2181-198-03-04). All patients gave written informed consent to participate in the study before being enrolled. This was a single-center trial, registered as an observational study at http://www.clinicaltrials.gov under identification no. NCT03179137.

During the CABG procedure, performed without the use of cardiopulmonary bypass and cardioplegia (“off-pump”), one to two cylinder-shaped biopsies (∼15 × 1 mm dimensions) were taken from anteroseptal part of the LV as previously described (14). Tissue was cut into several pieces that were either transferred to the laboratory within 15 min in ice-cold storage solution (in mmol/L: 2.77 CaK₂EGTA, 7.23 K₂EGTA, 6.56 MgCl₂, 5.7 Na₂ATP, 15 phosphocreatine, 20 imidazole, 20 taurine, 0.5 dithiothreitol, and 50 K-methanesulfonate, pH 7.1, at 0°C) and used for mitochondrial respiration measurements or immediately snap-frozen in liquid nitrogen and stored at −80°C for later analyses (enzyme activity, expression analyses, histological staining).

All chemicals used for this study, unless otherwise noted, were purchased from Sigma-Aldrich.

Myocardial biopsy samples were first dissected, permeabilized with saponin (50 µg/mL), and incubated in the reaction vessel filled with a respiratory buffer (in mmol/L: 2.77 CaK₂EGTA, 7.23 K₂EGTA, 1.38 MgCl₂, 3 K₂HPO₄, 20 imidazole, 20 taurine, 0.5 DTT, 90 K-methanesulfonate, 10 Na-methanesulfonate, and 0.2% BSA, pH 7.1, at 30°C), where mitochondrial respiration was evaluated with a Clark-type electrode (Oxygraph, Hansatech Instruments) (14,15). Fatty acid oxidation was assessed using palmitoyl-carnitine as substrate (40 µmol/L in the presence of 5 mmol/L malate) (Fig. 1A) and carbohydrate oxidation by providing mitochondria with pyruvate (10 mmol/L, with addition of 5 mmol/L malate) (Fig. 1B). The tissue oxygen consumption rate (in pmol O₂/s/mg wet tissue weight), a proxy of mitochondrial respiration, was sequentially recorded in the presence of substrates only, upon addition of a saturating amount of ADP (2.5 mmol/L), and trifluorocarbonylcyanide phenylhydrazone (FCCP) (1 µmol/L), an uncoupler of mitochondrial oxygen consumption from ATP synthesis. Ambient oxygen was maintained >210 µmol/L to avoid its diffusion limitation in the fibers (Fig. 1A and B).

Quantitative RT-PCR of total ventricular RNA was performed as previously described (16). mRNA levels for all target genes (PPARα [peroxisome proliferator–activated receptor α], FAT/CD36 [fatty acid translocase/cluster of differentiation 36], and CPT1 [carnitine palmitoyltransferase 1]) were normalized to Ywhaz and PolR2A levels. For detection of mature miRNA, cDNA was prepared in a reverse transcription reaction using miScript HiSpec Buffer from the miScript II RT Kit (QIAGEN). Real-time PCR was performed using miRNA-specific miScript Primer Assay (forward primer) and the miScript SYBR Green PCR Kit. PCR primers were obtained from Eurofins Genomics (Ebersberg, Germany) and are listed in Supplementary Table 1.

The activity of citrate synthase and pyruvate dehydrogenase (PDH) in homogenized LV tissue was assessed using commercial kits (CS0720, Sigma-Aldrich, and AAMT008-1KIT, Merck Millipore, respectively) as previously described (14). Activity of hydroxyacyl-CoA dehydrogenase (HADHA) in myocardial homogenate was assessed spectrophotometrically (at 340 nm [30°C]) by monitoring conversion of acetoacetyl-CoA to hydroxybutyryl-CoA.

For probing the blots of tissue protein, the following primary antibodies were used: anti-VLCAD (very-long-chain acyl-CoA dehydrogenase) (sc-376239; Santa Cruz Biotechnology), MitoProfile PDH WB Antibody Cocktail (ab110416; Abcam), total OXPHOS human antibody cocktail (against representative subunits of electron transfer chain complexes, ab110411; MitoSciences), anti-GRP78 (glucose-regulated protein 78) (3183; Cell Signaling Technology), anti-CHOP (sc-793; Santa Cruz Biotechnology), and anti–caspase-3 (sc-7148; Santa Cruz Biotechnology). Chemiluminescent substrate was Luminata Forte (Merck Millipore), and blots were imaged using the ChemiDoc imaging system (Bio-Rad). Actin served as a loading control.

Intramyocardial detection of neutral lipids was performed using oil red O (ORO) (Sigma-Aldrich) staining. Bright-field images of ×40 magnification were captured, and at least 10 frames per biopsy were used for analysis using ImageJ software. The extent of ORO staining is expressed as the pixel number over tissue area.

Intramyocardial ceramide levels were assessed by immunostaining of 10-μm cryosections of the ventricular biopsies. Primary anti-ceramide antibody was obtained from Sigma-Aldrich (Clone MID 15B4, product no. C8104, 1:300 dilution). Secondary antibody was goat anti-mouse Alexa Fluor 488 (ab150121; Abcam) used at 1:500 dilution. Fluorescence detection was done using ×40 objective using an AxioVision microscope (ZEISS).

TUNEL assay was performed in frozen myocardial biopsy tissue slices using a commercial kit (ApopTag Fluorescein in Situ Apoptosis Detection Kit; Merck). Apoptotic cells were quantified as percentage of FITC-stained nuclei in total number of cells. Image analysis was performed using Adobe Photoshop CS6 (Adobe Systems).

Fisher exact test was used for comparison of categorical parameters. For metrical parameters, normality of distribution was checked using the D’Agostino-Pearson test, and in case of normal distribution, unpaired Student t test was conducted for comparison. Otherwise, a nonparametric analysis was performed (Mann-Whitney test). Data in Table 1 are presented as means ± SD. Data in figures are presented as means ± SEM. Correlation analysis was performed using GraphPad Prism 6 software, with a two-sided P value <0.05 considered significant.

The data sets generated and/or analyzed during the current study are available from the corresponding author upon reasonable request. No applicable resources were generated or analyzed during the current study.

Main characteristics of the two patient groups are shown in Table 1. Compared with the non-DM group, patients with diabetes had slightly lower LVEF, as well as increased European System for Cardiac Operative Risk Evaluation (EuroSCORE) II and HbA_1c parameters. As expected, the two groups also differed in antidiabetes medication therapy. All patients successfully underwent the CABG surgery with no complications evidently related to the LV biopsy procedure.

When mitochondria were provided with palmitoyl-carnitine (in the presence of ADP), respiration was significantly lower in LV myocardium from DM patients compared with the non-DM group (Fig. 1C). On the other hand, ADP-supported mitochondrial oxidation of pyruvate was not different between the non-DM and DM groups (Fig. 1D). Administration of FCCP accelerated the respiration in both groups to the same extent, suggesting that maximal electron transfer chain capacity is comparable between them. The rate of mitochondrial respiration driven by palmitoyl-carnitine was negatively correlated with blood levels of HbA_1c (r² = 0.21). Such correlation was not present for pyruvate-fueled mitochondrial respiration.

Citrate synthase activity, a marker of mitochondrial content in the tissue, was not different between the two groups (Fig. 2A). Activity of HADHA (responsible for the second and third step of β-oxidation) was decreased in the DM group (Fig. 2B), revealing a reduced capacity for β-oxidation in diabetic hearts. Moreover, expression of VLCAD, an enzyme catalyzing the first step of β-oxidation, was significantly reduced in DM myocardium (Fig. 2C). There was no difference in mRNA expression of other key fatty acid metabolism factors: PPARα, FAT/CD36, and CPT1 (Fig. 2D). Also, levels of miR-33a, miR-33b, and miR-208a (miRNAs implicated in fatty acid metabolism [17,18]) were not different between the groups.

Activity and expression of the main subunits of PDH were unaltered in the DM group compared with the non-DM group (Fig. 3A and B). In addition, expression of the representative components of five mitochondrial respiratory complexes (OXPHOS I–V) was also not different between the non-DM and DM groups (Fig. 3C), thus supporting respirometry data, which suggest that there is no difference in total mitochondrial capacity for oxidative phosphorylation.

Myocardial staining using fat-soluble ORO dye revealed increased accumulation of intracellular lipid droplets in DM group myocardium (Fig. 4A and B). The staining extent was positively correlated with the ratio of pyruvate- to palmitoyl-driven mitochondrial respiration (Fig. 4C), indicating that decreased mitochondrial ability to oxidize palmitoyl relative to pyruvate is associated with increased accumulation of triglycerides (TGs) in cardiomyocytes. Also, anti-ceramide immunofluorescence staining revealed increased levels of ceramide in diabetic myocardium (Fig. 4D and E).

Diabetic myocardium exhibited significantly increased expression of GRP78 and CHOP, indicating activation of the ER stress response (Fig. 5A and B). This was associated with increased levels of cleaved caspase-3 and a tendency toward increased percentage of TUNEL-positive cells (P = 0.09) (Fig. 5C and D), pointing to increased apoptosis in diabetic myocardium.

In the current study, we found that the hearts of patients with type 2 diabetes, even in the absence of contractile failure, display 1) a decreased mitochondrial capacity for fatty acid–fueled respiration and unchanged mitochondrial oxidative capacity for carbohydrates, 2) a reduced expression/activity of β-oxidation enzymes, 3) an increased accumulation of intracellular TGs and ceramide, and 4) increased ER stress and apoptosis.

Due to insulin resistance in T2DM, there is a greater extent of lipolysis in adipose tissue, with increased fatty acid delivery to the myocardium (5). This is coupled with reduced insulin-stimulated entry of glucose into the cardiac myocytes (via GLUT4) (5), resulting in augmented myocardial reliance on fatty acid uptake and metabolism for production of ATP (19). Evidence for such cardiac metabolic impact of T2DM was mostly obtained in PET studies investigating substrate utilization at the level of whole heart (8) and were thus influenced by many variables, including plasma concentration of substrates.

Despite valuable PET data on overall cardiac substrate utilization, it still remains unclear whether T2DM affects intrinsic function of human ventricular mitochondria. In the current study, cellular and mitochondrial influx of substrates was controlled by permeabilization of the tissue and providing it with fixed amounts of metabolites, thereby minimizing acute effects of plasma insulin and substrate availability on mitochondrial respiration. Also, with use of palmitoyl-carnitine, some of the rate-limiting steps in fatty acid utilization (sarcolemmal uptake by FAT/CD36 and mitochondrial translocation by CPT1) were bypassed. By virtue of using pyruvate, mitochondrial oxidation of carbohydrates independent of the insulin-mediated GLUT4 uptake was tested.

Mitochondrial respiration driven by palmitoyl-carnitine was decreased in diabetic myocardium, while the oxidation of pyruvate was unaffected. Patients with higher levels of HbA_1c exhibited greater impairment of mitochondrial palmitoyl-carnitine oxidation, pointing toward a relationship between severity of insulin resistance/chronic glycemia levels and mitochondrial dysfunction. Indeed, loss-of-insulin signaling in the heart by the selective cardiomyocyte deletion of insulin receptors in CIRKO mice induced significant mitochondrial dysfunction with reduced mitochondrial capacity for oxidation of substrates (20). Also, increased palmitate load was shown to induce mitochondrial and cellular damage, whereby rat cardiomyocytes incubated with high palmitate concentrations displayed diminished ability to oxidize fatty acids and intracellular steatosis (21). Considering that global insulin resistance causes elevated levels of circulating fatty acids (including palmitate), it is possible that their elevated load leads to cardiomyocytes’ impairment of fatty acid metabolization in mitochondria. This, combined with impairment of myocardial insulin signaling (which in itself was shown to induce mitochondrial dysfunction [20]), could further augment the intracellular accumulation of palmitate and start the cardiotoxic vicious cycle.

Alterations in mitochondrial fatty acid oxidation were previously reported in atrial tissue from patients with T2DM (9,10) and in animal models of the disease (20,22,23). Also, data from a recent PET study investigating LV substrate metabolism in patients with diabetes agree with our findings by showing that despite increased absolute rates of fatty acid oxidation, diabetic heart oxidizes relatively less of the extracted fatty acids, since the increase in fatty acid esterification was proportionately higher than the increase in oxidation (24). Therefore, findings from our and previous studies suggest that despite globally increased fatty acid usage by the diabetic hearts (due to increased fatty acid delivery and reduced glucose uptake), the mitochondrial machinery in diabetic myocardium actually has less capacity for processing fatty acids (the maximum capacity is not likely to be achieved in the in vivo setting but can result in reported lower percent of fatty acid oxidation relative to their extracted amount) (24). As a result, accumulation of fatty acids upstream from the bottleneck in their utilization pathway could lead to their excessive incorporation in various types of lipids and consequent lipotoxicity (25,26).

The finding of the unchanged expression of OXPHOS subunits in DM group, as well as the unchanged rate of maximal respiration driven by pyruvate (both ADP and FCCP stimulated), suggests that the observed reduction in mitochondrial oxidation of palmitoyl is not a result of defects in mitochondrial respiratory chain activity. Rather, these findings suggest the main disruption leading to the decreased capacity for fatty acid oxidation might be at the level of β-oxidation. Indeed, measurement of the expression and activity of VLCAD and HADHA, respectively, revealed that these β-oxidation steps are downregulated in diabetic myocardium. Along this line, proteomic analysis of insulin resistant myocardium from CIRKO mice revealed significantly altered expression of mitochondrial proteins with downregulation of β-oxidation enzymes (20).

Measurements of intracellular neutral lipids revealed a significantly increased amount of lipid droplets inside LV myocardium of patients with diabetes. This correlated with the degree of impairment of mitochondrial palmitoyl-carnitine oxidation, suggesting that the observed mitochondrial dysfunction might be responsible for the enhanced fat accumulation. These results are in agreement with earlier documentation of an increased lipid content in human diabetic myocardium based on MRI methodology (25) and on direct histological visualization (however, these were diabetic hearts in end-stage contractile failure) (27). Accumulation of neutral fat was previously shown to result in deterioration in cardiac function. For example, deficiency of adipose TG lipase (ATGL), an enzyme responsible for degradation of lipid droplets, induces massive cardiac TG deposition and fatal cardiomyopathy (28).

Ceramide was found to be the critical cardiotoxin in a rodent model of lipotoxic cardiomyopathy, as reduction in myocardial ceramide resulted in improved cardiac function and survival (29). Moreover, obese diabetic rats (Zucker diabetic fatty rats) were previously shown to exhibit cardiac dysfunction associated with excessive myocardial storage of TG and ceramide (30). Despite the convincing results on ceramide accumulation obtained in animal models of diabetes and obesity, the content of ceramide was found to be unaltered in human atrial myocardium obtained from patients with T2DM (31). However, we also detected increased anti-ceramide staining in diabetic myocardium, with the most likely explanation for this discrepancy being the fact that we used LV myocardium, in contrast to the right atrial appendages used before (31).

ER stress is marked by an excess accumulation of misfolded proteins that can be triggered by intracellular buildup of saturated fatty acids (32). ER stress results in activation of cellular prosurvival response, called the “unfolded protein response,” aiming to restore the normal function of the ER by increasing expression of protein chaperones, such as GRP78, as well as by decreasing global protein synthesis and “unloading” the ER. However, if the unfolded protein response does not succeed in restoration of normal ER function, due to severe/chronic ER stress, the cell turns on its self-destruction program by inducing the proapoptotic transcription factor CHOP (33). In the current study, we detected an upregulation of GRP78 and induction of CHOP in diabetic myocardium, demonstrating a greater degree of ER stress in human diabetic hearts. These data are also in agreement with the results obtained in animal models of T2DM (34). Interestingly, induction of ER stress by tunicamycin can also induce profound mitochondrial remodeling, including decreased ability to process fatty acids (16). Whether ER stress is the initiator, or works concomitantly with other disturbances (e.g., fatty acid overload of cardiomyocytes) leading to the downward spiral of adverse mitochondrial remodeling in human diabetic myocardium, remains to be investigated.

Lastly, we demonstrate an augmented rate of apoptotic cell death in the DM group compared with the non-DM group. Our findings are in line with earlier studies on diabetes and cell death in human myocardium (35).

Since both patient groups suffered from CAD requiring surgical revascularization, we were not able to study factors influencing the diabetic myocardium in absence of potentially confounding CAD. Also, although there was no obvious difference in severity of CAD between DM and non-DM patients, we cannot assess possible differences in microvascular function between the two groups. The concentration of glucose and lipids in the blood was determined in a fasted state prior to the surgery. As we did not assess blood levels of insulin, free fatty acid, or glucose at the time of the biopsy procedure, we cannot assess their possible acute effects. However, by using permeabilized myocardial tissue and fixed concentrations of substrate for mitochondrial functional assay, we believe that acute effects of insulin and substrate availability on our results were minimized. Finally, due to the small amount of biopsy sample, we could not perform all of the measurements in tissues from all of the patients.

Our study shows that LV myocardium of patients with T2DM has significantly altered mitochondrial function, with decreased capacity to oxidize long-chain fatty acids through reduced activity and expression of β-oxidation enzymes. This is associated with excessive accumulation of fat in cardiac cells in form of intracellular lipid droplets and ceramide and is paralleled with ER stress and increased apoptosis. Therefore, based on our and previous data, it seems that insulin resistance, with abundant fatty acid loading of cardiomyocytes and ER stress, leads to decreased ability of mitochondria to handle large amounts of fatty acids being delivered to the cardiomyocyte. As a result, although total fatty acid usage in diabetic hearts is increased, it is actually insufficient to clear all intracellular fatty acids. This may lead to cardiac steatosis and lipotoxicity, extending to a loss of cardiac myocytes by apoptosis. Therefore, besides targeting insulin resistance as the primary aim of antidiabetes therapy, another target for treating diabetic myocardium might be further increasing its mitochondrial capacity for fatty acid oxidation.

Acknowledgments. The authors thank medical personnel of Department of Cardiac Surgery at University Hospital Split, especially Slavica Kotromanovic, for help with collection of patient data and biopsy samples, as well as Lucija Frankovic for technical assistance.

Funding. This work was supported by Croatian Science Foundation grants 6153 to M.L. and 3718 to D.B. and by the French grant ANR-10-LABX-33 to C.L. and M.G. as members of the Laboratory of Excellence LERMIT.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. M.L. and J.M. designed the study, researched data, and wrote the manuscript. M.L., M.G., C.B., M.C., D.B., D.F., I.G., C.L., and J.M. reviewed and edited the manuscript. M.G., C.B., M.C., D.B., D.F., and I.G. researched data. C.L. planned the experiments and researched data. M.L. and J.M. are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.