Cardiac fibroblast glycogen synthase kinase-3β regulates ventricular remodeling and dysfunction in ischemic heart

Abstract

Background: Myocardial infarction-induced remodeling includes chamber dilatation, contractile dysfunction, and fibrosis. Of these, fibrosis is the least understood. After myocardial infarction, activated cardiac fibroblasts deposit extracellular matrix. Current therapies to prevent fibrosis are inadequate, and new molecular targets are needed.

Methods and results: Herein we report that glycogen synthase kinase-3β (GSK-3β) is phosphorylated (inhibited) in fibrotic tissues from ischemic human and mouse heart. Using 2 fibroblast-specific GSK-3β knockout mouse models, we show that deletion of GSK-3β in cardiac fibroblasts leads to fibrogenesis, left ventricular dysfunction, and excessive scarring in the ischemic heart. Deletion of GSK-3β induces a profibrotic myofibroblast phenotype in isolated cardiac fibroblasts, in post-myocardial infarction hearts, and in mouse embryonic fibroblasts deleted for GSK-3β. Mechanistically, GSK-3β inhibits profibrotic transforming growth factor-β1/SMAD-3 signaling via interactions with SMAD-3. Moreover, deletion of GSK-3β resulted in the significant increase of SMAD-3 transcriptional activity. This pathway is central to the pathology because a small-molecule inhibitor of SMAD-3 largely prevented fibrosis and limited left ventricular remodeling.

Conclusions: These studies support targeting GSK-3β in myocardial fibrotic disorders and establish critical roles of cardiac fibroblasts in remodeling and ventricular dysfunction.

Keywords: fibroblasts; fibrosis; glycogen synthase kinase 3 beta; hypertrophy; myocardial infarction.

Figure 1. Activation of GSK-3β in ischemic heart and cardiac fibroblast-specific deletion of GSK-3β

A, WT mice were subjected to MI surgery at 2 months of age. Three wks post-MI, western blotting was performed on the heart lysates from border and remote zone. B&C, Quantification of GSK-3β phosphorylation in border vs remote zone. D, CFs were isolated from sham and MI operated hearts at 3 weeks post-MI and western blot analysis was performed. E, Quantification of P-GSK-3β in CFs from sham and MI hearts. F, Representative Immunoblot showing significantly increased phosphorylation of GSK-3β in the ischemic human heart vs control heart. G, Quantification of GSK-3β phosphorylation in failing vs control human heart. H, WT and Per-KO mice were subjected to MI surgery at 2 months of age. Two wks post-MI, cardiac fibroblasts were isolated from the area of injury (left ventricle) and Western blotting was performed. Representative Immunoblot demonstrates 65% deletion of GSK-3β in KOs. I, Quantification of GSK-3β expression in GSK-3β KO fibroblasts versus WT fibroblasts.

Figure 1. Activation of GSK-3β in ischemic heart and cardiac fibroblast-specific deletion of GSK-3β

A, WT mice were subjected to MI surgery at 2 months of age. Three wks post-MI, western blotting was performed on the heart lysates from border and remote zone. B&C, Quantification of GSK-3β phosphorylation in border vs remote zone. D, CFs were isolated from sham and MI operated hearts at 3 weeks post-MI and western blot analysis was performed. E, Quantification of P-GSK-3β in CFs from sham and MI hearts. F, Representative Immunoblot showing significantly increased phosphorylation of GSK-3β in the ischemic human heart vs control heart. G, Quantification of GSK-3β phosphorylation in failing vs control human heart. H, WT and Per-KO mice were subjected to MI surgery at 2 months of age. Two wks post-MI, cardiac fibroblasts were isolated from the area of injury (left ventricle) and Western blotting was performed. Representative Immunoblot demonstrates 65% deletion of GSK-3β in KOs. I, Quantification of GSK-3β expression in GSK-3β KO fibroblasts versus WT fibroblasts.

Figure 2. Cardiac fibroblast-specific deletion of GSK-3β leads to cardiac dysfunction and dilatative remodeling post-MI

Two-month-old WT and Per-KO mice underwent baseline transthoracic echocardiographic examination. Twenty-four hours later they were subjected to occlusion of the proximal left anterior descending coronary artery. Mice were then followed with serial echocardiography at the time points shown. A, Representative M-mode images from 6 weeks post-MI are shown. B, Left ventricular internal dimension at end-diastole (LVID;d). C, LVID at end-systole (LVID;s). D, left ventricular ejection fraction (LVEF). E, LV fractional shortening (LVFS). F, Increased hypertrophy in the Per-KO mice subjected to coronary artery ligation as shown by HW/BW ratio. G, Increased heart failure in the Per-KO mice. The ratio of lung weight to body weight (LW/BW, a measure of heart failure) was significantly increased in the KO mice. MI, myocardial infarction; BL, baseline; WT, Wild type; KO, Knockout.

Figure 2. Cardiac fibroblast-specific deletion of GSK-3β leads to cardiac dysfunction and dilatative remodeling post-MI

Two-month-old WT and Per-KO mice underwent baseline transthoracic echocardiographic examination. Twenty-four hours later they were subjected to occlusion of the proximal left anterior descending coronary artery. Mice were then followed with serial echocardiography at the time points shown. A, Representative M-mode images from 6 weeks post-MI are shown. B, Left ventricular internal dimension at end-diastole (LVID;d). C, LVID at end-systole (LVID;s). D, left ventricular ejection fraction (LVEF). E, LV fractional shortening (LVFS). F, Increased hypertrophy in the Per-KO mice subjected to coronary artery ligation as shown by HW/BW ratio. G, Increased heart failure in the Per-KO mice. The ratio of lung weight to body weight (LW/BW, a measure of heart failure) was significantly increased in the KO mice. MI, myocardial infarction; BL, baseline; WT, Wild type; KO, Knockout.

Figure 3. Cardiac fibroblast specific deletion of GSK-3β promotes post-MI scar expansion and fibrosis

Two-month-old WT and Pre-KO mice were subjected to MI surgery for 6 weeks, as described in Materials and Methods. A, Representative images of heart sections stained with Masson trichrome at 6 week’s post-MI vs sham surgery. B, scar circumference was measured and expressed as a percentage of total area of LV myocardium. C, Representative images of LV border zone showing increased fibrosis in the KO mouse heart. D, Bar graph showing fold changes in fibrosis in the border zone of WT and KO hearts 6 weeks post-MI. E, At 6 weeks post-MI, immunofluorescence staining was performed to visualize α-SMA positive cells (red), Sarcomeric-alpha-actinin (α-SA), a myocyte-specific marker (green) and DAPI (blue) was used to label nuclei. Representative merge images are shown. F, Quantification of α-SMA positive cells is expressed as a percentage of total cells counted.

Figure 3. Cardiac fibroblast specific deletion of GSK-3β promotes post-MI scar expansion and fibrosis

Two-month-old WT and Pre-KO mice were subjected to MI surgery for 6 weeks, as described in Materials and Methods. A, Representative images of heart sections stained with Masson trichrome at 6 week’s post-MI vs sham surgery. B, scar circumference was measured and expressed as a percentage of total area of LV myocardium. C, Representative images of LV border zone showing increased fibrosis in the KO mouse heart. D, Bar graph showing fold changes in fibrosis in the border zone of WT and KO hearts 6 weeks post-MI. E, At 6 weeks post-MI, immunofluorescence staining was performed to visualize α-SMA positive cells (red), Sarcomeric-alpha-actinin (α-SA), a myocyte-specific marker (green) and DAPI (blue) was used to label nuclei. Representative merge images are shown. F, Quantification of α-SMA positive cells is expressed as a percentage of total cells counted.

Figure 4. GSK-3β regulates fibroblast to myofibroblast transformation in cardiac fibroblasts

A, Cardiac fibroblasts were isolated from 1–3 day old neonatal rat pups and were cultured up to three passages. Western blot analysis was performed to determine the expression of α-SMA and phosphorylation of GSK-3β. B, Line graph shows quantification of fold changes in α-SMA expression and GSK-3β phosphorylation. C, Neonatal cardiac fibroblasts were transfected with adenoviruses expressing constitutively active GSK-3β (Ad-GSK-3βS9) or GFP (Ad-GFP). After 24 h of transfection, viral medium was replaced with fresh serum free medium (SFM) and cells were treated with TGF-β1 (10ng/ml) for 48 h before harvesting. Western blot analysis was performed to determine the expression of α-SMA, endogenous and mutant GSK-3β, and HA-tag. D, Bar graphs show fold changes in α-SMA expression. E, MEF cells were cultured in complete DMEM medium (10%FBS and 1% antibiotics) to a confluency of ~75% and then were switched to fresh SFM medium for overnight starvation before harvesting. After starvation, cells were harvested and lysates were analyzed by immunoblotting with α-SMA and GSK-3α/β antibodies, as indicated. F, Bar graphs show fold changes in α-SMA expression in GSK-3β KO MEFs compared to WT cells. G, WT and GSK-3β KO MEFs were cultured on chamber slides and Immunofluorescence staining was performed. TGF, TGF-β1; GSKS9, GSK-3βS9

Figure 4. GSK-3β regulates fibroblast to myofibroblast transformation in cardiac fibroblasts

A, Cardiac fibroblasts were isolated from 1–3 day old neonatal rat pups and were cultured up to three passages. Western blot analysis was performed to determine the expression of α-SMA and phosphorylation of GSK-3β. B, Line graph shows quantification of fold changes in α-SMA expression and GSK-3β phosphorylation. C, Neonatal cardiac fibroblasts were transfected with adenoviruses expressing constitutively active GSK-3β (Ad-GSK-3βS9) or GFP (Ad-GFP). After 24 h of transfection, viral medium was replaced with fresh serum free medium (SFM) and cells were treated with TGF-β1 (10ng/ml) for 48 h before harvesting. Western blot analysis was performed to determine the expression of α-SMA, endogenous and mutant GSK-3β, and HA-tag. D, Bar graphs show fold changes in α-SMA expression. E, MEF cells were cultured in complete DMEM medium (10%FBS and 1% antibiotics) to a confluency of ~75% and then were switched to fresh SFM medium for overnight starvation before harvesting. After starvation, cells were harvested and lysates were analyzed by immunoblotting with α-SMA and GSK-3α/β antibodies, as indicated. F, Bar graphs show fold changes in α-SMA expression in GSK-3β KO MEFs compared to WT cells. G, WT and GSK-3β KO MEFs were cultured on chamber slides and Immunofluorescence staining was performed. TGF, TGF-β1; GSKS9, GSK-3βS9

Figure 5. GSK-3β regulates fibroblast to myofibroblast transformation in adult cardiac fibroblasts

A, adult CFs were isolated from GSK-3βfl/fl mice and Cre was driven by adenoviral transfection to delete GSK-3β. After 24 h of transfection, viral medium was replaced with fresh SFM and cells were further maintained for 48 h before harvesting. Western blot analysis was performed to determine the expression of α-SMA, Cre, and T-GSK-3α/β. B, Bar graphs show fold changes in α-SMA expression. C, CFs were isolated from WT mice and were transfected with control and GSK-3β targeted siRNA. After 24 h of transfection, medium was replaced with fresh SFM and cells were further maintained for 48 h before harvesting. D&E, Bar graphs show fold changes in T-GSK-3β and α-SMA expression.

Figure 6. Deletion/inhibition of GSK-3β in fibroblasts leads to hyper-activation of TGF-β1-SMAD-3 signaling

A, WT and GSK-3β KO MEFs were serum-starved overnight before receiving TGFβ-1 for 1 h (10ng/ml). Western blot analysis was performed to analyze the phosphorylation of SMAD-3 at Ser204 and Ser423/425. B, Bar graphs show fold changes in phosphorylation of SMAD-3 at Ser423/425 in GSK-3βKO MEFs compared to WT cells in the presence or absence of TGF-β1. C, Bar graphs show fold changes in phosphorylation of SMAD-3 at Ser204. D, Neonatal cardiac fibroblasts were serum starved overnight before receiving GSK-3 inhibitor SB415286 (10μM) for 30 min and an additional 1 h of TGF-β1 stimulation. Western blot analysis was performed to analyze the phosphorylation of SMAD-3 at Ser204 and Ser423/425. E, Bar graphs show fold changes in phosphorylation of SMAD-3 at Ser423/425. F, Bar graphs show fold changes in phosphorylation of SMAD-3 at Ser204. G&H, Cells were transfected with SMAD-3 Cignal Lenti Reporter virus. Forty-eight h after transfection, cells were treated with 5ng/ml TGF-β1 for 24 h and the luciferase activity was quantified.

Figure 6. Deletion/inhibition of GSK-3β in fibroblasts leads to hyper-activation of TGF-β1-SMAD-3 signaling

A, WT and GSK-3β KO MEFs were serum-starved overnight before receiving TGFβ-1 for 1 h (10ng/ml). Western blot analysis was performed to analyze the phosphorylation of SMAD-3 at Ser204 and Ser423/425. B, Bar graphs show fold changes in phosphorylation of SMAD-3 at Ser423/425 in GSK-3βKO MEFs compared to WT cells in the presence or absence of TGF-β1. C, Bar graphs show fold changes in phosphorylation of SMAD-3 at Ser204. D, Neonatal cardiac fibroblasts were serum starved overnight before receiving GSK-3 inhibitor SB415286 (10μM) for 30 min and an additional 1 h of TGF-β1 stimulation. Western blot analysis was performed to analyze the phosphorylation of SMAD-3 at Ser204 and Ser423/425. E, Bar graphs show fold changes in phosphorylation of SMAD-3 at Ser423/425. F, Bar graphs show fold changes in phosphorylation of SMAD-3 at Ser204. G&H, Cells were transfected with SMAD-3 Cignal Lenti Reporter virus. Forty-eight h after transfection, cells were treated with 5ng/ml TGF-β1 for 24 h and the luciferase activity was quantified.

Figure 7. GSK-3β directly interacts with SMAD-3 but does not affect its nuclear accumulation after TGF-β1 stimulation

A, Nuclear and cytoplasmic fractionation was performed with WT and GSK-3β KO MEF cells treated with TGFβ1 (10ng/ml) for 1 h. The content of GSK-3β, SMAD-3, nuclear marker LaminA/C, and cytosolic marker GAPDH was determined. B, Bar graphs show fold changes in SMAD-3 and GSK-3β translocation to the nucleus after TGF-β1 stimulation. C, Neonatal cardiac fibroblasts were serum starved overnight before receiving GSK-3 inhibitor SB415286 (10μM) for 30 min and an additional 1 h of TGF-β1 stimulation. Nuclear and cytoplasmic fractionation was performed and content of GSK-3β, SMAD-3, nuclear marker LaminA/C, and cytosolic marker GAPDH was determined. D, Bar graphs show fold changes in SMAD-3 and GSK-3β translocation to nucleus after TGF-β1 stimulation. E, MEFs were lysed for endogenous co-immunoprecipitation assay using either IgG or the monoclonal antibodies as indicated, and followed by immunoblotting. F, G, physical interaction between SMAD-3 and GSK-3β was also examined in lysates from cardiac fibroblasts and human heart. H, Co-immunoprecipitation assay was performed with lysates from sham and MI operated hearts at 6 weeks post-MI, using either IgG or the monoclonal antibodies as indicated, and followed by immunoblotting.

Figure 7. GSK-3β directly interacts with SMAD-3 but does not affect its nuclear accumulation after TGF-β1 stimulation

A, Nuclear and cytoplasmic fractionation was performed with WT and GSK-3β KO MEF cells treated with TGFβ1 (10ng/ml) for 1 h. The content of GSK-3β, SMAD-3, nuclear marker LaminA/C, and cytosolic marker GAPDH was determined. B, Bar graphs show fold changes in SMAD-3 and GSK-3β translocation to the nucleus after TGF-β1 stimulation. C, Neonatal cardiac fibroblasts were serum starved overnight before receiving GSK-3 inhibitor SB415286 (10μM) for 30 min and an additional 1 h of TGF-β1 stimulation. Nuclear and cytoplasmic fractionation was performed and content of GSK-3β, SMAD-3, nuclear marker LaminA/C, and cytosolic marker GAPDH was determined. D, Bar graphs show fold changes in SMAD-3 and GSK-3β translocation to nucleus after TGF-β1 stimulation. E, MEFs were lysed for endogenous co-immunoprecipitation assay using either IgG or the monoclonal antibodies as indicated, and followed by immunoblotting. F, G, physical interaction between SMAD-3 and GSK-3β was also examined in lysates from cardiac fibroblasts and human heart. H, Co-immunoprecipitation assay was performed with lysates from sham and MI operated hearts at 6 weeks post-MI, using either IgG or the monoclonal antibodies as indicated, and followed by immunoblotting.

Figure 8. Pharmacological inhibition of SMAD-3 attenuated the cardiac dysfunction, dilative remodeling and scar expansion in Per-KO hearts post-MI

WT and Per-KO mice underwent baseline transthoracic echocardiographic examination. Twenty-four hours later they were subjected to occlusion of the proximal left anterior descending coronary artery. Mice were then followed with serial echocardiography at the time points shown. Osmotic pumps were implanted 1 week post MI surgery. A, left ventricular ejection fraction (LVEF). B, LV fractional shortening (LVFS). C, Left ventricular internal dimension at end-diastole (LVID;d). D, LVID at end-systole (LVID;s). E, Representative images of heart sections stained with Masson trichrome 5 weeks post-MI. F, scar circumference was measured and expressed as a percentage of total area of LV myocardium. G, SIS3 rescues increased hypertrophy in the KO mice subjected to coronary artery ligation as shown by HW/BW ratio. H, SIS 3 rescued the failing heart phenotype the KO mice. The ratio of lung weight to body weight (LW/BW, a measure of heart failure) was significantly attenuated by SIS3 treatment in the KO mice. SIS, SMAD-3 inhibitor SIS3

Figure 8. Pharmacological inhibition of SMAD-3 attenuated the cardiac dysfunction, dilative remodeling and scar expansion in Per-KO hearts post-MI

WT and Per-KO mice underwent baseline transthoracic echocardiographic examination. Twenty-four hours later they were subjected to occlusion of the proximal left anterior descending coronary artery. Mice were then followed with serial echocardiography at the time points shown. Osmotic pumps were implanted 1 week post MI surgery. A, left ventricular ejection fraction (LVEF). B, LV fractional shortening (LVFS). C, Left ventricular internal dimension at end-diastole (LVID;d). D, LVID at end-systole (LVID;s). E, Representative images of heart sections stained with Masson trichrome 5 weeks post-MI. F, scar circumference was measured and expressed as a percentage of total area of LV myocardium. G, SIS3 rescues increased hypertrophy in the KO mice subjected to coronary artery ligation as shown by HW/BW ratio. H, SIS 3 rescued the failing heart phenotype the KO mice. The ratio of lung weight to body weight (LW/BW, a measure of heart failure) was significantly attenuated by SIS3 treatment in the KO mice. SIS, SMAD-3 inhibitor SIS3