MK2 Deletion in Mice Prevents Diabetes-Induced Perturbations in Lipid Metabolism and Cardiac Dysfunction

Diabetes remains a worldwide health problem, which is associated with a high rate of mortality primarily as a consequence of cardiovascular diseases. Despite advances in our understanding of the pathophysiology of diabetes, its prevalence is estimated to rise in the next decade. Current drug therapies aim at normalizing glucose levels, thereby preventing complications (1). Although these therapeutic agents efficiently lower blood glucose levels, there is a progressive loss of their efficacy in maintaining glycemic control over time (2). Hence, the development of new strategies for the treatment of diabetes remains essential and constitutes an active field of research.

In addition to glucose, diabetes is associated with perturbations in lipid metabolism in various tissues, including the heart (3). As a result, targeting lipid metabolism has been proposed as an interesting alternative strategy to improve diabetes-induced cardiac dysfunction. Consistent with this notion, pharmacological agents, such as troglitazone (4) or peroxisome proliferator–activated receptor α (PPARα) agonists, which modulate fatty acid (FA) metabolism, have shown benefits in animal models of cardiac dysfunction (5,6). Progress in this area requires, however, a better understanding of the molecular mechanisms underlying these lipid alterations. In this regard, the current literature has involved mitogen-activated protein kinases (MAPKs), such as extracellular signal–related kinase 1/2, c-Jun N-terminal kinase, or p38MAPK in metabolic syndrome, obesity, and diabetes (7).

Among these MAPKs, p38MAPK was described as having a key role in the pathophysiology of diabetes (8). p38MAPK modulates both lipid and glucose metabolism (8), mediates insulin resistance (IR) (9), and contributes to diabetes-induced cardiac dysfunction (10). Inactivation or inhibition of p38MAPK improves insulin signaling in the liver (11) and alleviates the development of cardiac dysfunction in rodent models of diabetes (10). To the best of our knowledge, however, none of the p38MAPK inhibitors have been approved for medical use in human subjects because of numerous side effects, including hepatotoxicity or cardiotoxicity (12).

To circumvent p38MAPK inhibitor toxicity, an alternative strategy is to target one of the downstream targets of this kinase, such as PRAK/MK5, MSK1/2, MK3, or MK2. Interestingly, similar to p38MAPK, MK2 was recently reported to be activated in diabetic liver (11) and heart (13), and MK2 inhibition improves glucose homeostasis and insulin sensitivity in obese mice (14). Furthermore, mice lacking both MK2 and MK3 display differences in the expression of various genes involved in lipid and carbohydrate metabolism in skeletal muscle (15). In addition, cardiomyocytes from MK2/MK3 double-knockout mice showed both increased contractility and sarcoplasmic/endoplasmic reticulum Ca²⁺-ATPase 2a (SERCA2a) expression (15). However, whereas MK2 deletion in mice has been reported to be beneficial in models of cardiac ischemia-reperfusion and pressure overload–induced hypertrophy (16,17), to the best of our knowledge, the involvement of MK2 in diabetes-induced metabolic alterations and cardiac dysfunction has not yet been examined. Thus, the objective of this study was to investigate whether MK2 could impact on diabetes-induced complications using a murine pan-MK2 knockout model, in comparison with their control littermates, in a model of diabetes induced by repeated injections of a low dose of streptozotocin (STZ). Collectively, our results show that the absence of MK2 prevented diabetes-induced systemic and cardiac lipid metabolism perturbations as well as cardiac dysfunction.

A protocol timeline is shown in Supplementary Fig. 1.

Sources of chemicals, biological products, and ¹³C-labeled substrates as well as the albumin dialysis procedure (BSA fraction V; Intergen) have been described previously (18).

Animal experiments were approved by the Animal Research Ethics Committee of the Montreal Heart Institute in agreement with the guidelines of the Canadian Council on Animal Care. The generation of MK2-null (MK2^−/−) mice has been described previously (19), whereas genotyping is described in the Supplementary Data. MK2^−/− mice are viable and fertile and have no obvious physiological defects. Mice were housed in a specific pathogen-free facility under a 12-h light/dark cycle at constant temperature and were provided food and water ad libitum.

Five-week-old male control (MK2^+/+) and MK2^−/− mice were fed a standard diet (3.02 kcal/g; Rodent Diet 5001; LabDiet) and, after a 2-week adaptation period, received daily intraperitoneal STZ injections (40 mg/kg i.p.; Sigma-Aldrich) or vehicle (0.5 mol/L citrate buffer, pH 4.5). Mice were considered to be diabetic once their fasted blood glucose levels had reached or exceeded a concentration of 12.5 mmol/L on 5 consecutive days, which required between five and eight injections. Mice were observed for a total of 15 weeks. Body weight and food intake were measured weekly.

Intraperitoneal insulin tolerance tests (ipITTs; insulin 0.6 units/kg; Novolin ge) were performed at week 9 after a 5-h fast. For oral glucose tolerance tests (OGTTs), performed at week 10 after a 16-h fast, glucose (dextrose solution 2 g/kg; Hospira) was delivered into the stomach using a gavage needle (20-gauge × 1.5 inches, with 1.9-mm tip; Cadence Science). For both tests, blood was drawn from the saphenous vein, and the glucose level was measured prior to and 15, 30, 45, 60, 90, and 120 min after insulin injection or dextrose gavage. Insulin levels during the OGTT were examined at time 0, 15, 30, and 60 min. These tests were assessed, and results were analyzed according to the published standard operating procedure (20).

Fed mice were anesthetized with 2% isoflurane in 100% O₂ at a flow rate of 1 L/min, and body temperature was monitored and maintained at 37°C using a heating pad. Hemodynamic parameters were measured using a microtip pressure transducer catheter (1.4F; Millar Instruments) inserted into the left ventricle through the carotid artery as previously described (21). Data were analyzed using iox software version 2.5.1.6 (emka TECHNOLOGIES, Falls Church, VA).

Blood was collected at various time points. Plasma was obtained after centrifugation at 1,200g for 10 min at 4°C and stored at −80°C. Analyses performed after a 5-h fast included the following: 1) in blood (from saphenous vein), glycemia (Accu-Chek Aviva glucometer; Roche) and ketonemia (Blood β-Ketone test strips, Precision Xtra; Abbott) and glycated hemoglobin (HbA_1c) levels (Bayer A1C Now; Bayer) and 2) in plasma, insulin (Mouse Insulin ELISA; Albco). Other analyses performed in the fed state are as follows: 1) plasma free FAs (FFAs) (Fatty Acid Assay Kit; Biovision) and 2) plasma free carnitine and 17 acylcarnitines (ACs) using an electrospray ionization tandem mass spectrometry method (22) with some modifications (for details, see Supplementary Data).

Levels of transcripts (by quantitative RT-PCR), proteins (by immunoblotting), and triglycerides (TGs; assessed as described previously [23] from their FAs analyzed by gas chromatography–mass spectrometry) were performed in pulverized heart tissues, which were freeze clamped after assessment of in vivo cardiac function in fed mice (for details, see Supplementary Data).

Metabolic flux parameters were assessed in ex vivo working perfused mouse hearts, as previously described in detail (18). Briefly, hearts isolated from fed animals were perfused for 30 min under normoxia at fixed preload (15 mmHg) and afterload (50 mmHg) pressures with semi-recirculating modified Krebs-Henseleit buffer containing a mix of substrates and hormones (11 mmol/L glucose, 1.5 mmol/L lactate, 0.2 mmol/L pyruvate, and 0.7 mmol/L palmitate bound to 3% dialyzed albumin, 50 μmol/L carnitine, 0.8 nmol/L insulin, and 5 nmol/L epinephrine). These perfusion conditions were used for all mouse groups in order to compare the intrinsic capacity of the heart to metabolize a fixed concentration of glucose and palmitate, which we assessed by replacing, in any given perfusion, unlabeled palmitate or glucose with [U-¹³C₆]glucose or [U-¹³C₁₆]palmitate, at 50% and 25% initial molar percent enrichment, respectively. It is noteworthy that while substrate concentrations are within the physiological level, they do not, however, mimic those found in the diabetic state. Functional parameters were continuously monitored throughout the perfusion (iox2 Data Acquisition System; emka TECHNOLOGIES). At the end of the perfusion, hearts were freeze clamped with metal tongs chilled in liquid nitrogen and stored at −80°C for further analysis. Using gas chromatography–mass spectrometry, we determined the ¹³C-labeling pattern of 1) lactate and pyruvate in influent and effluent perfusates and 2) tissue citrate. The following flux parameters were calculated: 1) the efflux rate of labeled lactate and pyruvate arising from cytosolic glycolysis of exogenous [U-¹³C₆]glucose and 2) the relative contribution of exogenous [U-¹³C₆]glucose versus [U-¹³C₁₆]palmitate to mitochondrial acetyl CoA synthesis for citrate formation.

Data are reported as the mean ± SE, and the statistical difference between groups was tested by two-way ANOVA, followed by a Bonferroni selected-comparison test to compare the respective source of variation (genotype and treatment) using GraphPad Prism software. When a significant interaction was observed, the following groups were compared using one-way ANOVA followed by Bonferroni selected-comparison test (MK2^+/+-vehicle vs. MK2^+/+-STZ, MK2^−/−-vehicle vs. MK2^−/−-STZ, and MK2^−/−-STZ vs. MK2^+/+-STZ). A P value <0.05 was considered to be statistically significant.

We first characterized the impact of our model of diabetes on systemic and metabolic parameters. Since MK2^+/+ and MK2^−/− vehicle-treated mice showed similar values for all anthropometric (Fig. 1) and glucose-related parameters (Fig. 2), these mice will be referred to as control mice, unless otherwise specified.

Induction of diabetes by multiple low-dose injections of STZ in MK2^+/+ mice (MK2^+/+-STZ) led to a progressive increase in cumulative caloric intake, which reached significance at week 12 (Fig. 1A). This protocol also provoked a significant decrease in cumulative weight gain (14%; Fig. 1B), body weight (11%; Supplementary Table 1), gastrocnemius (26%; Fig. 1C), and epididymal white adipose tissue (WAT; 3.6-fold; Fig. 1D) masses. In contrast, MK2^−/−-STZ mice displayed values similar to those of control mice for each of these parameters.

We next evaluated metabolic parameters in 5-h fasted mice. Compared with control mice, MK2^+/+-STZ mice showed a 50% decrease in plasma insulin levels (Fig. 2A), which was associated with an increase in blood glucose level (3.6-fold; Fig. 2B) and HbA_1c level (2.5-fold; Fig. 2C). MK2^−/−-STZ and MK2^+/+-STZ mice showed similar plasma insulin levels at 15 weeks (Fig. 2A). In contrast, whereas glucose levels (Fig. 2B) were initially similar in MK2^−/−-STZ and MK2^+/+-STZ mice, these values diverged over time, and at 15 weeks the glucose levels in MK2^−/−-STZ mice were lower (23%) than those in the MK2^+/+-STZ mice. However, these values remained twofold greater than those for controls. Furthermore, HbA_1c levels in the MK2^−/−-STZ mice were also lower than in the MK2^+/+-STZ mice at 15 weeks (Fig. 2C).

To gain further insight into the differences between MK2^+/+-STZ and MK2^−/−-STZ mice, we performed an OGTT after a 16-h fast. Both MK2^+/+-STZ and MK2^−/−-STZ mice demonstrated a similar rapid and sustained increase in glucose level, resulting in a 2.5-fold greater area under the curve (AUC) for glucose levels compared with control mice (Fig. 2D). This was consistent with the lower insulin levels observed in these animals (Fig. 2E). We next examined insulin sensitivity using an ipITT after a 5-h fast. As shown in Fig. 2F, MK2^−/−-STZ mice displayed a significantly lower AUC for glucose levels compared with MK2^+/+-STZ mice (−36%), suggesting an improvement in insulin sensitivity.

When these tests were assessed, we observed that initial glycemia was lower in STZ mice after 16 h of fasting than after 5 h of fasting (Supplementary Fig. 2A). The reduction in glucose levels observed in MK2^−/−-STZ mice compared with MK2^+/+-STZ mice after 5 h of fasting (−19.5%) was exacerbated after 16 h of fasting (−36%) (Supplementary Fig. 2A). This reduction was paralleled by decreased liver PEPCK expression (Supplementary Fig. 2B), suggesting reduced hepatic glucose production.

Next, we assessed diabetes-induced cardiac alterations at the functional and molecular levels. Hemodynamic analyses using a Millar Mikro-Tip catheter revealed that MK2^+/+-STZ mice showed signs of cardiac dysfunction in vivo, especially diastolic dysfunction (Fig. 3) with preserved systolic function and hemodynamic parameters (Supplementary Table 1), as reflected by a significant increase in left ventricular minimum diastolic pressure (Fig. 3A), mean diastolic pressure (Fig. 3B), and end-diastolic pressure, although the latter did not reach significance (P = 0.1; Fig. 3C). While MK2^+/+-STZ mice had a heart weight-to-body weight ratio similar to that of controls (Supplementary Table 1), these mice showed an eightfold increase in the ratio of β-myosin heavy chain (MHC) to α-MHC mRNA (Fig. 3D) (24). Furthermore, there was a significant decrease in SERCA2a (−51%, Fig. 3E) as well as in phospholamban (PLB) phospho-Ser16 (−47%, Fig. 3E) immunoreactivity in MK2^+/+-STZ mice, which suggest decreased SERCA activity (25). There were, however, no changes in either transcript level or immunoreactivity for markers of endoplasmic reticulum stress such as ATF6, GRP78, or CHOP (Supplementary Fig. 3) (26). Finally, consistent with the previously reported increase in p38MAPK signaling in diabetic hearts (10), and in pathological cardiac remodeling (27), the phosphorylation of p38MAPK and MK2 was increased in MK2^+/+-STZ mice by 3.5-fold (Fig. 4A) and 3-fold (Fig. 4B), respectively, compared with control mice. Moreover, diabetes did not alter total p38 and MK2 immunoreactivity (Fig. 4) or the basal expression of proteins involved in insulin signaling, namely insulin receptor substrate 1 (IRS-1), AKT, and the protein-tyrosine phosphatase 1B, a negative regulator of insulin signaling (Supplementary Fig. 4). Moreover, in the absence of a prior insulin injection, IRS-1 phosphorylation (on serine 307) was undetectable, whereas AKT phosphorylation was similar among the four groups (Supplementary Fig. 4).

MK2^−/−-STZ mice were protected from all the aforementioned diabetes-induced cardiac alterations and showed values similar to their control counterparts for all functional parameters (Fig. 3A–C and Supplementary Table 1). In addition, the β-MHC–to–α-MHC ratio (Fig. 3D), SERCA2a expression, and PLB phosphorylation (Fig. 3E) in MK2^−/−-STZ mice were similar to those of their control counterparts. As expected, total and phosphorylated MK2 immunoreactivities were undetectable in control and diabetic MK2^−/− mice. As MK2 stabilizes p38α (28), total and phosphorylated p38 immunoreactivity was reduced (−87% and −72%, respectively) in MK2^−/− mice (Fig. 4).

Given our findings that the benefits of MK2 abrogation on diabetes-induced cardiac alterations 1) occurred despite sustained hyperglycemia and 2) were associated with preserved WAT mass, we hypothesized that the cardioprotective effect of knocking out MK2 could be associated with the normalization of diabetes-induced systemic and/or cardiac lipid alterations. As expected, at 15 weeks, MK2^+/+-STZ mice showed elevated circulating plasma FFA levels (44%; Fig. 5A) and β-hydroxybutyrate (β-HB) levels (384%; Fig. 5B) compared with their control counterparts. Remarkably, FFA (Fig. 5A) and β-HB (Fig. 5B) levels in MK2^−/−-STZ mice were similar to those in controls. We then examined plasma AC levels, which are monitored routinely when screening for defects in tissue FA β-oxidation (29) and have been found to be associated with IR and diabetes (30). The plasma levels of short-chain ACs and medium-chain ACs were similar among all four mouse groups (Supplementary Fig. 5). However, MK2^+/+-STZ mice displayed a significant increase in long-chain ACs (LCACs) (Fig. 5C), namely C14:2-carnitines (63%), C18-carnitines (238%), C18:1-carnitines (61%), and C18:2-carnitines (200%). In contrast, MK2^−/−-STZ mice maintained plasma LCAC levels that were similar to those in their control counterparts.

We next ascertained whether these systemic changes in lipid metabolism would be paralleled by changes in their cardiac metabolism by conducting metabolic flux studies in ex vivo working hearts using ¹³C-labeled glucose and palmitate. It is noteworthy that functional parameters recorded during the perfusions demonstrated that alterations observed in vivo (Fig. 3 and Supplementary Table 1) were also present in ex vivo working heart preparations. Indeed, compared with control mice, MK2^+/+-STZ mice displayed diastolic dysfunction (−20% for minimum change in pressure over time (dP/dt; Fig. 6B), systolic dysfunction (−18% and −22%, respectively, for maximal pressure and maximum dP/dt; Fig. 6C and D), and hemodynamic alterations (−22% for stroke volume; Fig. 6E). MK2^−/−-STZ mice were, however, protected against all these diabetes-induced cardiac functional alterations.

Figure 7 and Supplementary Fig. 6A and B report metabolic flux parameters from these hearts. First, flux related to glucose metabolism did not differ significantly among groups. These fluxes include rates of lactate and pyruvate efflux formed from glycolysis of exogenous [U-¹³C₆]glucose (Supplementary Fig. 6A) as well as the relative contribution of glucose to mitochondrial acetyl-CoA formation through pyruvate decarboxylation (Supplementary Fig. 6B). Moreover, because myocardial oxygen consumption (MVO₂) was not significantly different among the four groups (in µmol/min, MK2^+/+-vehicle 1.50 ± 0.20, MK2^+/+-STZ 1.02 ± 0.23, MK2^−/−-vehicle 1.11 ± 0.11, and MK2^−/−-STZ 1.46 ± 0.30), this would also suggest a similar absolute flux of glucose to acetyl-CoA (calculated from the stoichiometric relationship between MVO₂ and citrate formation from glucose as determined from the measured flux ratio [31]). We did not assess the metabolism of other carbohydrates, such as exogenous lactate or pyruvate. The contribution of exogenous palmitate to acetyl-CoA via β-oxidation was, however, significantly increased (46%) in MK2^+/+-STZ mice compared with their control counterparts (Fig. 7A). In contrast, in MK2^−/−-STZ mice β-oxidation rates were similar to MK2^−/−-vehicle or MK2^+/+-vehicle mice.

The cardiac transcript levels for genes involved in carbohydrates and FA metabolism, which we assume to reflect protein levels, were consistent with the observed changes in cardiac metabolic flux parameters. Specifically, in MK2^+/+-STZ mice, although there was a significant decrease in the abundance of mRNA for GLUT1 (−26%), the other measured markers of glucose transport and glycolysis (GLUT4, PFK1, and PFK2) did not differ significantly from those of MK2^+/+-vehicle mice (Supplementary Fig. 6C and D). Moreover, MK2^+/+-STZ mice showed an increase in transcript levels for the FA transporters (CD36 25%, CPT1b 27%) and oxidation (long-chain acyl-CoA dehydrogenase 25%; Fig. 7B), as well as for the negative regulator of mitochondrial pyruvate dehydrogenase PDK4 (fourfold; Supplementary Fig. 6D). In contrast, for each of these markers, MK2^−/−-STZ and MK2^−/−-vehicle mice did not differ significantly.

We also investigated the mechanisms involved in mitochondrial FA oxidation, namely phosphorylation of AMPK and its downstream target acetyl-CoA carboxylase (ACC), transcript levels for PPARα, peroxisome proliferator–activated receptor α coactivator (PGC-1α), and uncoupling protein 3 (UCP3), as well as the level of mitochondrial electron transport chain (ETC) proteins. Although hearts from MK2^+/+-STZ mice displayed significantly increased levels of AMPK phosphorylation (330%; Fig. 7C) and UCP3 mRNA (210%, Fig. 7E), the abundance and phosphorylation of ACC (Fig. 7D), PPARα, and PGC-1α mRNA (Supplementary Fig. 7B), as well as immunoreactivity for components of the mitochondrial ETC (Supplementary Fig. 7A) remained unchanged. In contrast, MK2^−/−-STZ mice did not display any significant changes compared with their control counterpart in AMPK phosphorylation and transcript levels of UCP3. Moreover, as observed in MK2^+/+-STZ mice, ACC phosphorylation (Fig. 7D), PPARα or PGC-1α mRNA expression, and mitochondrial ETC protein levels (Supplementary Fig. 7A and B) were unchanged in MK2^−/−-STZ mice.

Finally, MK2^+/+-STZ mice demonstrated a 38% increase in cardiac TG content in comparison with their vehicle controls and MK2^−/−-STZ mice (Fig. 8A). To further explore potential mechanisms underlying this TG accumulation, we assessed the expression of enzymes involved in TG synthesis (GPAT1) or hydrolysis (ATGL and HSL) by immunoblotting, as well as transcript levels of PPARδ/β, which has been inversely associated with TG content in diabetic hearts (32). Although GPAT1, ATGL, and HSL immunoreactivity were unchanged, we found significant increases in the phosphorylation of HSL on serine 565 (405%; Fig. 8B), a specific target of AMPK that inhibits HSL activity (33), as well as a reduction of PPARδ/β mRNA level (−58%; Supplementary Fig. 7B) in MK2^+/+-STZ mouse hearts, but not in MK2^−/−-STZ mouse hearts.

Collectively, all measured parameters relevant to FA metabolism convey a similar message, as follows: hearts from MK2^+/+-STZ mice show enhanced exogenous FA oxidation and esterification for storage, whereas their MK2^−/−-STZ counterparts showed no sign of altered FA metabolism, which is consistent with the lack of changes in systemic FAs in these mice.

Our results demonstrate that whole-body MK2 deletion in mice protects against diabetes-induced systemic and cardiac metabolic perturbations, as well as against cardiac dysfunction, despite sustained hyperglycemia. Specifically, the improvement in cardiac function was associated with normalization of diabetes-induced 1) cardiac molecular changes, including proteins involved in maintaining calcium homeostasis; 2) systemic lipid abnormalities; and 3) cardiac lipid abnormalities.

In this study, diabetes was induced by repeated administration of low doses of STZ to mice fed a standard chow diet. MK2^+/+-STZ mice showed a 50% decrease in plasma insulin levels in addition to IR, hyperglycemia, increased HbA_1c levels, and augmented ketonemia. These mice also displayed a loss of WAT and skeletal muscle mass, as well as loss of body weight despite an increase in caloric intake, which is indicative of a catabolic state. These phenotypic characteristics differ to some extent, particularly insulin secretion, from those induced by high doses of STZ, leading to almost total suppression of insulin secretion and type 1 diabetes (34), or by repeated low doses of STZ combined with a high-fat diet, which lead to IR and hyperinsulinemia, mimicking type 2 diabetes (35). Hence, our model seems to share characteristics of type 1 diabetes (catabolic state and increased ketonemia) and advanced type 2 diabetes (IR and 50% insulin level reduction).

Given that the main objective of this study was to test the impact of a whole-body MK2 deletion on diabetes-induced cardiac dysfunction, it was crucial to ascertain that our diabetes model develops cardiac abnormalities as previously described (36). Consistent with other model of diabetes, MK2^+/+-STZ mice displayed in vivo diastolic dysfunction, and changes in the expression of β- and α-MHC isoforms. Cardiac dysfunction was exacerbated ex vivo in isolated working heart preparations, where increased diastolic as well as systolic dysfunctions were observed. The latter finding suggests the presence of systemic homeostatic compensatory mechanisms (e.g., neuronal or hormonal influences), which are present in the intact animal but are absent in ex vivo preparations. Consistent with the observed diastolic dysfunction, MK2^+/+ diabetic mice displayed significant changes in proteins regulating cardiac excitation-contracting coupling, namely lower SERCA2a immunoreactivity and activity, as reflected by reduced PLB phosphorylation, suggesting disturbances in Ca²⁺ reuptake during diastole (37). Remarkably, MK2^−/− mice were protected against diabetes-induced cardiac dysfunction concomitant with normalization of SERCA2a expression and PLB phosphorylation. The latter finding is consistent with the recently proposed role of the p38-MK2/3 axis in regulating cardiomyocyte function through SERCA2a modulation (15). In this regard, results from our study demonstrate a specific role of MK2, which may act directly or through its capacity to modulate the abundance of p38α (28,38). Interestingly, the specific inhibition of p38αMAPK, but not p38βMAPK, was recently shown (39) to enhance diastolic Ca²⁺ uptake through increased PLB phosphorylation on the residue Ser16.

Taken together, the results from our study, in addition to those of others (15–17), demonstrate that suppressing MK2 activity has beneficial consequences for the heart under several pathological conditions. This study specifically substantiates its benefits against diabetes-induced metabolic alterations and cardiac dysfunction. However, similar to other studies, our mouse model is a whole-body MK2 knockout; thus, we were unable to discriminate between its cardiac and systemic effects. We observed additional effects of MK2 deletion in diabetic mice, beyond those relevant to cardiac calcium handling, which may directly or indirectly preserve cardiac function. Indeed, we showed that major diabetes-induced systemic and cardiac lipid abnormalities, which have been previously associated with cardiac dysfunction in both patients with diabetes (40) and several experimental diabetes models (41,42), were abrogated in MK2^−/−-STZ mice. At the systemic level, MK2^+/+-STZ mice displayed increased levels of plasma FFAs in combination with reduced WAT mass, suggesting an augmented WAT lipolysis (43). Furthermore, our finding of a selected increase in plasma LCACs, but not medium-chain ACs or short-chain ACs, suggests that FA cellular overload, namely increased FA cellular uptake and activation to acyl-CoA, exceeds the capacity of mitochondria to take up acyl-CoA for subsequent FA oxidation. The latter could be the consequence of a mismatch between expression or activity of the mitochondrial FA transporters CPT1, which converts LC-acyl-CoA to LCACs, and CPT2, which converts LCACs back to LC-acyl-CoA. While the increased levels of plasma LCACs reflect disturbances from multiple organs, our results in the heart support the proposed scenario. Indeed, in the MK2^+/+-STZ mice, cardiac transcript levels for the cellular long-chain FA (LCFA) transporter CD36 as well as CPT1 were increased, whereas those of CPT2 were unchanged. Along with this mismatch between LCFA uptake into cells and mitochondria, other data indicate that diabetic MK2^+/+-STZ mouse hearts are coping with the excessive LCFA uptake by shuttling proportionally more FAs to mitochondrial oxidation, while restricting TG hydrolysis. An augmented relative contribution of mitochondrial FA oxidation to acetyl-CoA generation for citrate formation is supported by metabolic flux measurements in ex vivo working hearts and by expression data for proteins involved in FA oxidation or its regulation. Moreover, while TG accumulation occurs without changes in the abundance of enzymes involved in TG synthesis (DGAT1) or hydrolysis (HSL and ATGL), there was, however, increased AMPK-dependent phosphorylation of HSL, which is known to inhibit its activity and is consistent with increased TG accumulation (33). These changes were also consistent with a decrease in PPARβ/δ expression, which was previously shown to parallel cardiac TG accumulation (44). The fact that the MVO₂ tended to be slightly decreased in diabetic MK2^+/+ hearts suggests, however, that globally mitochondrial oxidative metabolism, which encompasses absolute β-oxidation flux rate from FAs, both exogenously supplied and endogenously formed from TG, would also be decreased. All these metabolic changes, together with the observed cardiac dysfunction, suggest that the metabolic flexibility of diabetic MK2^+/+ hearts is impaired, while being preserved in MK2^−/− diabetic hearts. This conclusion needs, however, to be further substantiated by testing the metabolic and functional response of these diabetic hearts after a challenge such as high workload. Further studies are also required to determine the specific roles and mechanisms by which MK2 modulates diabetes-induced systemic versus cardiac lipid metabolism. Although PPARα, PGC-1α, and/or mitochondrial biogenesis are known to be factors determinant for lipid metabolism and have been reported to be part of diabetes-induced cardiac metabolic changes (45), our mRNA expression data do not support their involvement in our diabetes model, or as a MK2 target. Although additional data on protein expression and activity would be needed to ascertain this conclusion, there are several others factors that ought to be considered as potential, direct or indirect, players such as IR. Whole-body insulin sensitivity has been linked to perturbations in systemic lipid metabolism as well as ectopic lipid deposition, including the heart (46). Although we did not detect any sign of cardiac IR under the basal state, as reflected by undetectable IRS-1 phosphorylation levels (on serine 307) and unchanged AKT phosphorylation, additional studies should investigate the role of MK2 in insulin action/signaling in diabetes.

IR alters fat deposition in WAT and liver (47). The improved insulin sensitivity in MK2^−/−-STZ mice may explain, at least in part, the observed increase of WAT mass and subsequently the decrease of plasma FFA levels compared with MK2^+/+-STZ mice, as well as the small, but significant, reduction in fasting hyperglycemia, which may result from normalization of the diabetes-induced increase in liver PEPCK expression in MK2^−/−-STZ mice (this study) and gluconeogenesis by MK2 inhibition (14). It is noteworthy that our results differ to some extent from those of de Boer et al. (48), who showed that MK2 deficiency reduces insulin sensitivity in high-fat diet–fed mice. Although in the latter study model mice are hyperinsulinemic, in contrast to our MK2^−/−-STZ mice, which are hypoinsulinemic, insulin level is unlikely to be the sole explanation. Indeed, expressing dominant-negative MK2 or administration of an allosteric MK2 inhibitor in ob/ob mice, a hyperinsulinemic mouse model, also improved liver and whole-body insulin sensitivity, respectively (14). Thereby, other factors are likely to be involved. In this regard, another mechanism that may underlie the beneficial effect of MK2 deletion on insulin sensitivity as well as systemic and cardiac lipid metabolism is an attenuated inflammatory response (47), although this remains to be further investigated. Activation of the p38MAPK/MK2 axis has been linked to a proinflammatory response (49), and MK2^−/− mice were shown to be more resistant to stress and endotoxic shock (19). A recent study provides valuable insight into the differential effect of p38MAPK versus MK2 inhibition in modulating proinflammatory signaling pathways. In both cases, there is inhibition of p38/MAPK signaling (50). However, inhibition of p38, without discrimination of the isoform, induces a sustained activation of c-Jun N-terminal kinase and nuclear factor-κB, whereas the inhibition of MK2 does not (50).

In summary, the current study demonstrates that the deletion of MK2 prevents diabetes-induced cardiac dysfunction, an effect that may involve alterations in the expression and activation of SERCA2a and PLB, proteins involved in maintaining cardiac calcium homeostasis. Our results also identify other effects of MK2 deletion on diabetes-induced lipid perturbations both at the systemic and cardiac level, which could also protect cardiac function. Future studies using mice with a cardiomyocyte-specific MK2 deletion are, however, needed to dissect 1) the molecular mechanisms by which MK2 modulate cardiac lipid metabolism and cardiac function and 2) the primary compartments in which these effects of MK2 occur, namely heart versus other tissues. Taken altogether, our findings suggest that MK2 inhibition may represent a new strategy to protect the myocardium from diabetes-induced alterations in systemic and cardiac FA metabolism, as well as cardiac dysfunction.

Acknowledgments. The authors thank Anik Forest, Research Centre, Montreal Heart Institute, for helpful comments for liquid chromatography–mass spectrometry methodology; Marie Eve Rivard, Isabelle Robillard-Frayne, and Nayla El Zyr, Research Centre, Montreal Heart Institute, the animal facility staff, for technical assistance; and France Thériault, Research Centre, Montreal Heart Institute, for editorial and secretarial assistance.

Funding. This work was supported by the Canadian Institutes of Health Research (grant 9575 to C.D.R.) and the Heart and Stroke Foundation of Canada (grant G-14-0006060 to B.G.A.).

The funders played no role in the study.

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

Author Contributions. M.R. contributed to the design of the project, performed the experiments, analyzed all of the data, and wrote, reviewed, and edited the manuscript. L.C. and B.G.A. contributed to the design of the project and wrote, reviewed, and edited the manuscript. D.L., V.H., C.M., J.T.L., M.-A.G., B.B., and C.D. performed the experiments. A.C.C. contributed to the design of the project. M.G. provided material essential for the study. C.D.R. contributed to the design of the project and wrote, reviewed, and edited the manuscript. C.D.R. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.