Cardiac autophagy is a maladaptive response to hemodynamic stress

Abstract

Cardiac hypertrophy is a major predictor of heart failure and a prevalent disorder with high mortality. Little is known, however, regarding mechanisms governing the transition from stable cardiac hypertrophy to decompensated heart failure. Here, we tested the role of autophagy, a conserved pathway mediating bulk degradation of long-lived proteins and cellular organelles that can lead to cell death. To quantify autophagic activity, we engineered a line of "autophagy reporter" mice and confirmed that cardiomyocyte autophagy can be induced by short-term nutrient deprivation in vivo. Pressure overload induced by aortic banding induced heart failure and greatly increased cardiac autophagy. Load-induced autophagic activity peaked at 48 hours and remained significantly elevated for at least 3 weeks. In addition, autophagic activity was not spatially homogeneous but rather was seen at particularly high levels in basal septum. Heterozygous disruption of the gene coding for Beclin 1, a protein required for early autophagosome formation, decreased cardiomyocyte autophagy and diminished pathological remodeling induced by severe pressure stress. Conversely, Beclin 1 overexpression heightened autophagic activity and accentuated pathological remodeling. Taken together, these findings implicate autophagy in the pathogenesis of load-induced heart failure and suggest it may be a target for novel therapeutic intervention.

Figure 1. Cardiomyocyte autophagy triggered by short-term nutrient deprivation or pressure-overload hemodynamic stress.

(A) Representative electron micrographs from the septa of 48-hour sTAB LVs (WT C57BL/6 mice) demonstrate the presence of double-membrane autophagosomes (arrows) and autolysosomes containing cellular material. These features are more prominent in sTAB ventricles compared with those of sham-operated controls. Scale bar: 120 nm. (B) Representative immunoblot for LC3 showing an increase in LC3-II abundance following sTAB as early as 24 hours after operations and persisting for at least 2 weeks. (C) Quantification of LC3-II/LC3-I levels demonstrates significant autophagic activity induced by pressure overload. *P < 0.05 versus Sham. (D) Representative immunoblots for LC3 in liver and kidney demonstrating that LC3-II abundance does not change in these tissues following sTAB. (E) Under baseline conditions, GFP-LC3 Tg fusion protein is diffusely distributed throughout the cardiomyocyte cytoplasm in α-MHC–GFP–LC3 mice. Following short-term (48 hours) starvation, GFP-LC3 aggregates as autophagosome-localized GFP dots. Representative images from basal septum are shown. Scale bar: 35 μm. (F) Following sham operation, GFP-LC3 Tg fusion protein is diffusely distributed throughout the cardiomyocyte cytoplasm in α-MHC–GFP–LC3 mice. Following imposition of pressure stress by sTAB (48 hours), GFP-LC3 aggregates as autophagosome-localized GFP dots. Representative images from basal septum are shown. Scale bar: 35 μm. (G) Quantification of GFP aggregates per microscopic field (14,479 μm²) demonstrates significant autophagic activity induced by starvation. For each group, at least 4 mice were studied. ^‡P < 0.01 versus fed. (H) Quantification of GFP aggregates per microscopic field (14,479 μm²) demonstrates significant autophagic activity induced by pressure overload. For each group, at least 4 mice were studied. ^†P < 0.01 versus sham. FW, free wall of LVs.

Figure 2. Increased abundance of lysosomal markers in sTAB ventricle.

LAMP-1 (A) and cathepsin D (B), detected by immunohistochemistry, are increased in sTAB ventricle at multiple time points, indicative of increased lysosomal activity in pressure-stressed LVs. Scale bars: 40 μm. (C) Representative immunoblot of ventricular lysates from sham-operated and sTAB LVs probed for cathepsin D. Mean data from 3 independent experiments. *P < 0.05.

Figure 3. Time course of autophagic activity in sTAB ventricle and changes in Beclin 1 protein levels.

(A) Autophagic activity induced in the basal septum in response to sTAB was quantified by counting GFP-LC3 dots per microscopic field (14,479 μm²). Representative immunoblot for Beclin 1 showed an increase in Beclin 1 abundance following sTAB (B) but not in response to short-term nutrient deprivation (C). *P < 0.05 versus sham. Ctl, control.

Figure 4. Pressure overload induces apoptosis in proximal aorta but not in LVs.

(A) Rare TUNEL-positive figures (pink circles) are detected in ventricular myocardium from sham-operated or sTAB hearts (basal septum) whereas a significantly increased signal is detected in aorta. Scale bar: 10 μm. (B) Microscopic fields (10–15 per section), each containing approximately 800 myocytes, were evaluated by TUNEL staining (n = 30–45 fields from each of 3 mice at each time point) in C57BL/6 mice. P = NS for each time point compared with sham. (C) DNA laddering (arrows) indicative of apoptosis in proximal aortic tissue (Ao), but not LVs, in animals subjected to sTAB.

Figure 5. Heterozygous disruption of beclin 1 decreases cardiomyocyte autophagy.

(A) Autophagy induced by starvation (48 hours), manifested as punctate GFP-LC3 dots, is significantly diminished in beclin 1^+/– hearts. All images were taken from basal septum. Scale bar: 35 μm. (B) Quantification of GFP aggregates per microscopic field (14,479 μm²) in basal septum demonstrates significantly less autophagic activity in beclin 1^+/– hearts in response to starvation compared with WT. n = 3–5 microscopic fields in each of 3 mice. (C) Induction of autophagy by pressure overload (sTAB, 48 hours) is significantly diminished in beclin 1^+/– hearts compared with WT. All images are taken from basal septum. Scale bar: 35 μm. (D) Quantification of GFP aggregates per microscopic field (14,479 μm², basal septum) demonstrated significantly less autophagic activity in beclin 1^+/– hearts exposed to hemodynamic stress. n = 3–5 microscopic fields in each of 3 mice. *P < 0.05 versus α-MHC–GFP–LC3 mice.

Figure 6. Beclin 1 levels increase with pressure-overload stress but remain lower in beclin 1^+/– hearts relative to WT.

(A) Representative immunoblot demonstrating a time course of Beclin 1 levels following sTAB surgery in beclin 1^+/– hearts. (B) Direct comparisons of Beclin 1 protein levels in WT and beclin 1^+/– hearts following sTAB surgery.

Figure 7. Pathological remodeling in pressure-stressed ventricle is diminished when autophagy is inhibited by beclin 1 haploinsufficiency.

(A) Pressure overload–induced declines in systolic function, measured as %FS, are significantly decreased in beclin 1^+/– mice. Systolic performance was measured at 3 weeks after banding. n = 4 WT sham; n = 4 ± sham; n = 6 WT sTAB; n = 6 ± sTAB. (B) Four-chamber sections of hearts treated as listed and harvested at 3 weeks. Scale bar: 2 mm. (C) HW/BW is increased similarly in banded beclin 1^+/– mice compared with WT controls. n = 6 WT sham; n = 10 ± sham; n = 7 WT sTAB; n = 8 ± sTAB. ^‡P < 0.05 versus ± sTAB; *P < 0.05 versus ± sham.

Figure 8. Starvation-induced autophagy is increased in beclin 1 Tg mice.

(A) Starvation-induced (24 hours) autophagy, manifested as punctate GFP-LC3 dots, is significantly amplified in beclin 1 Tg hearts. All images were taken from basal septum, and nutrient deprivation was limited to 24 hours, rather than 48 hours, due to the amplified autophagic response in beclin 1 Tg hearts. Scale bar: 35 μm. (B) Quantification of GFP aggregates per microscopic field (14,479 μm², basal septum) demonstrated significant upregulation of autophagic activity in beclin 1 Tg hearts exposed to starvation. n = 3–5 microscopic fields in each of 5 mice. *P < 0.05 versus fed; ^‡P < 0.01 versus fed.

Figure 9. Increased abundance of lysosomal markers in TAB LVs and in beclin 1 Tgs.

Cathepsin D (A) and LAMP-1 (B), detected by immunohistochemistry, are increased in TAB ventricle and in beclin 1 Tg LV, indicative of increased lysosomal activity. Scale bars: 40 μm. (C) Representative immunoblots probed for cathepsin D in ventricular lysates as indicated. *P < 0.05 versus WT sham.

Figure 10. Stress-induced autophagy is increased in beclin 1 Tg mice.

(A) Four-chamber sections of hearts treated as listed and harvested at 3 weeks. Scale bar: 2 mm. (B) Trichrome staining for collagen (blue) demonstrates increased fibrotic change in banded beclin 1 Tg hearts. Scale bar: 60 μm. (C) Quantification of GFP aggregates per microscopic field (14,479 μm², basal septum) demonstrates significant upregulation of autophagic activity in beclin 1 Tg hearts exposed to standard TAB. n = 3–5 microscopic fields in each of 3 mice. *P < 0.05 versus sham. (D) Two-dimensional cardiomyocyte cross-sectional area measured as described. ^‡P < 0.05 versus WT TAB. (E) Banding-induced heart growth, measured as HW/BW, is amplified in beclin 1 Tg mice compared with WT controls. n = 6 WT sham; n = 6 Tg sham; n = 8 WT TAB; n = 6 Tg TAB. (F) In WT mice, moderate pressure stress induced by TAB did not alter systolic function. In beclin 1 Tg mice, however, %FS was diminished significantly at 3 weeks. Ventricular decompensation was observed in beclin 1 Tg mice exposed to TAB, as LVESD (G) and LVEDD (H) were both increased significantly at 3 weeks compared with control. n = 6 WT sham; n = 6 Tg sham; n = 5 WT TAB; n = 5 Tg TAB.