Direct Evidence that Myocardial Insulin Resistance following Myocardial Ischemia Contributes to Post-Ischemic Heart Failure

Abstract

A close link between heart failure (HF) and systemic insulin resistance has been well documented, whereas myocardial insulin resistance and its association with HF are inadequately investigated. This study aims to determine the role of myocardial insulin resistance in ischemic HF and its underlying mechanisms. Male Sprague-Dawley rats subjected to myocardial infarction (MI) developed progressive left ventricular dilation with dysfunction and HF at 4 wk post-MI. Of note, myocardial insulin sensitivity was decreased as early as 1 wk after MI, which was accompanied by increased production of myocardial TNF-α. Overexpression of TNF-α in heart mimicked impaired insulin signaling and cardiac dysfunction leading to HF observed after MI. Treatment of rats with a specific TNF-α inhibitor improved myocardial insulin signaling post-MI. Insulin treatment given immediately following MI suppressed myocardial TNF-α production and improved cardiac insulin sensitivity and opposed cardiac dysfunction/remodeling. Moreover, tamoxifen-induced cardiomyocyte-specific insulin receptor knockout mice exhibited aggravated post-ischemic ventricular remodeling and dysfunction compared with controls. In conclusion, MI induces myocardial insulin resistance (without systemic insulin resistance) mediated partly by ischemia-induced myocardial TNF-α overproduction and promotes the development of HF. Our findings underscore the direct and essential role of myocardial insulin signaling in protection against post-ischemic HF.

Figure 1. Cardiac function and dimensions, and myocardial insulin sensitivity in rats subjected to sham or myocardial infarction.

(A) Ejection fraction (EF) was progressively reduced in rats following MI over 8 wk when compared with sham rats. (B) Left ventricular end-systolic dimensions (LVESD) and (C) Left ventricular end-diastolic dimensions (LVEDD) were progressively elevated in rats with MI over 8 wk compared with sham. (D) Representative microPET/CT images of rats with sham or MI over 2 wk without or with insulin stimulation. (E) Quantification of maximum standardized glucose uptake values (SUV_max) obtained from multiple independent experiments as depicted in panel D. Data are mean ± SEM of 8 independent experiments. “Sham” means “non-MI + saline” at 30 min post-operation. **P < 0.01 vs. sham, ^#P < 0.05, ^##P < 0.01 vs. MI 1 wk, ^††P < 0.01 vs. vehicle in sham, ^§§P < 0.01 vs. insulin in sham, ^σσP < 0.01 vs. vehicle in MI 1d.

Figure 2. Insulin signaling and cardiac plasma membrane GLUT4 in sham or MI rats.

(A) Insulin stimulated phosphorylation of Akt was almost abolished 1 wk after MI. (B) Both basal and insulin stimulated phosphorylation of ERK1/2 was enhanced 1 wk after MI. In addition, 1 wk after MI, insulin did not further stimulate the elevated basal phosphorylation of ERK1/2. (C) Insulin treatment significantly increased phosphorylation of p38 MAPK 1 wk after MI. (D) Insulin-stimulated increase in cardiac plasma membrane GLUT4 was impaired 1d after MI and abolished 1 wk after MI compared with sham. The bar graphs below the respective immunoblots represent quantification of multiple independent experiments from 3 animals (mean ± SEM). “Sham” means “non-MI + saline” at 30 min post-operation. *P < 0.05, **P < 0.01 vs. vehicle in sham, ^##P < 0.01 vs. vehicle in MI 1d, ^σσP < 0.01 vs. insulin in sham.

Figure 3. Local adenoviral expression of TNF-α in the heart resulted in myocardial insulin resistance in non-MI hearts and led to contractile impairment of the heart.

(A) Tumor necrosis factor (TNF)-α concentrations in myocardium 2, 4, and 7 d after infection. (B) Myocardial TNF-α was demonstrated by immunohistochemistry staining in the left ventricular 7 days post-adenovirus infection. Adenoviral expression of TNF-α in the heart abolished insulin-stimulated myocardial glucose uptake (C,D). Insulin-stimulated increase in Akt (E) and cardiac plasma membrane GLUT4 (F) was impaired in adTNF-α-treated rats. (G) P38 MAPK phosphorylation was increased in hearts infected with TNF-α with or without insulin stimulation. (H) Representative echocardiography images (major finding is I and J). Reduced Ejection fraction (I, EF) and elevated Left ventricular end-systolic dimensions (J, LVESD) were observed in the TNF-α overexpressed hearts at 1 wk after MI when compared with adGFP-treated rats. (K) Cardiac TNF-α levels in adTNF-α-treated or adGFP-treated rats at 1 wk after MI. Data are mean ± SEM of 8 independent experiments. **P < 0.01 vs. adGFP, ^†P < 0.05, ^††P < 0.01 vs. vehicle in adGFP, ^ττP < 0.01 vs. insulin in adGFP, ^#P < 0.05, ^##P < 0.01 vs. adGFP + MI.

Figure 4. Systemic etanercept treatment reduced myocardial TNF-α levels and improved myocardial insulin sensitivity and action at 1 wk post-MI.

(A) MI caused a significant increase in myocardial TNF-α expression that was somewhat blunted in etanercept treated rats. Etanercept treatment enhanced insulin stimulated glucose uptake 1 wk after MI when compared with vehicle treated rats (B,C). Insulin-stimulated increase in p-IRS-1 (D) p-Akt (E) and cardiac plasma membrane GLUT4 (F) was restored in etanercept-treated rats at 1 wk after MI. (G) P38 MAPK phosphorylation was decreased in etanercept-treated rats at 1 wk after MI. (H) Representative echocardiography images (major finding is I, J). (I) Ejection fraction (EF). (J) Left ventricular end-systolic dimensions (LVESD). Data are mean ± SEM of 8 independent experiments. **P < 0.01 vs. Non-MI + saline (Sham), ^#P < 0.05, ^##P < 0.01 vs. MI + saline, ^†P < 0.05, ^††P < 0.01 vs. vehicle in MI + Et 1wk, ^ξξP < 0.01 vs. vehicle in MI + saline 1wk, ^ττP < 0.01 vs. insulin in MI + saline 1wk, ^ΔΔP < 0.01 vs. Sham.

Figure 5. Early insulin treatment reduced cardiac TNF-α expression and improved myocardial insulin sensitivity and cardiac function following MI.

(B) TNF-α expression in myocardium. Early insulin treatment significantly improved acute insulin-stimulated myocardial glucose uptake 1 wk after MI compared with vehicle treated rats (A,C). Insulin-stimulated increase in Akt phosphorylation (D) was restored in early insulin-treated rats at 1 wk after MI. (G) Representative echocardiography images (major finding is E, F). Increased ejection fraction (E, EF) and reduced Left ventricular end-systolic dimensions (F, LVESD) were observed in early insulin-treated rats at 4 wk after MI when compared with saline-treated or late insulin-treated animals. Values presented are mean ± SEM (n = 8 per group). **P < 0.01 vs. Non-MI + saline, ^#P < 0.05, ^##P < 0.01 vs. MI + saline, ^††P < 0.01 vs. insulin in MI + saline 1wk, ^ΔΔP < 0.01 vs. Sham, ^ξP < 0.01 vs. MI + saline, ^τP < 0.05 vs. MI + insulin.

Figure 6. Aggravated LV dilation and dysfunction in tamoxifen-induced cardiomyocyte-specific insulin receptor knockout (TCIRKO) mice at 4 wk after MI.

(A) Representative immunoblots demonstrating absence of insulin receptor (INSR) in cardiac muscle of TCIRKO mice. Insulin-stimulated myocardial Akt phosphorylation (B) was almost entirely blocked in TCIRKO mice. No detectable differences in heart weight/body weight (C, HW/BW) were observed between controls and TCIRKO. (D) Representative echocardiography images (major finding is E–H). TCIRKO mice developed decreased ejection fraction (E, EF) and fractional shortening (F, FS) and elevated left ventricular end-systolic volume (G, LVESV) and end-diastolic volume (H, LVEDV) compared with littermate controls at 4 wk after MI. Insulin administration improved EF and FS and reduced LVESV compared with controls at 4 wk after MI. These benefits were absent in TCIRKO mice. Data are mean ± SEM of 8 independent experiments. ^ΔΔP < 0.01 vs. vehicle in control, **P < 0.01 vs. sham in control, ^#P < 0.05, ^##P < 0.01 vs. MI in control, ^τP < 0.05, ^ττP < 0.01 vs. MI + Ins in control.