PI3Kα-regulated gelsolin activity is a critical determinant of cardiac cytoskeletal remodeling and heart disease

Abstract

Biomechanical stress and cytoskeletal remodeling are key determinants of cellular homeostasis and tissue responses to mechanical stimuli and injury. Here we document the increased activity of gelsolin, an actin filament severing and capping protein, in failing human hearts. Deletion of gelsolin prevents biomechanical stress-induced adverse cytoskeletal remodeling and heart failure in mice. We show that phosphatidylinositol (3,4,5)-triphosphate (PIP3) lipid suppresses gelsolin actin-severing and capping activities. Accordingly, loss of PI3Kα, the key PIP3-producing enzyme in the heart, increases gelsolin-mediated actin-severing activities in the myocardium in vivo, resulting in dilated cardiomyopathy in response to pressure-overload. Mechanical stretching of adult PI3Kα-deficient cardiomyocytes disrupts the actin cytoskeleton, which is prevented by reconstituting cells with PIP3. The actin severing and capping activities of recombinant gelsolin are effectively suppressed by PIP3. Our data identify the role of gelsolin-driven cytoskeletal remodeling in heart failure in which PI3Kα/PIP3 act as negative regulators of gelsolin activity.

Fig. 1

Relationship between gelsolin and adverse cytoskeletal remodeling in DCM and in biomechanical stress-induced HF. a Schematic showing pathogenic and reverse remodeling process. Pathogenic remodeling in response to chronic injury leading to ventricular dilation, resulting in HF with reduced EF. LVAD placement results in reverse myocardial remodeling and improved cardiac function. b LVEF inversely correlates with myocardial gelsolin actin-depolymerizing activity in humans with DCM; n = 20 DCM hearts; age 52.3 ± 2.49 y; 16 male/4 female. c LV end-diastolic dimensions (LVEDD) from patients with DCM showing cardiac reverse remodeling in response to LVAD placement. d–g Representative images of F- and G-actin staining (d), F-actin to G-actin ratio (e), and actin-depolymerizing activity in human (f) and canine (g) hearts showing increased actin depolymerization in DCM samples compared with NFC. LVAD placement reduced actin-depolymerizing activity and recovered F to G-actin ratio. h Schematic showing the role of gelsolin in actin depolymerization. i Kaplan–Meier survival curve showing markedly increased mortality in response to pressure overload for 18 weeks in WT mice. Loss of gelsolin significantly decreased mortality in response to pressure overload. j Masson trichrome staining showing increased ventricular dilation in WT mice, which was attenuated in GSNKO mice. k Lung water content showing increased pulmonary edema in pressure-overloaded WT mice, which was attenuated in the GSNKO mice. l–o M-mode echocardiography images (l), quantification of LVEF (m), representative PV loop images (n), and dp/dt_max/EDV (load-independent index of systolic function; o) showing severe HF with EF in WT mice in response to pressure overload-induced biomechanical stress. Loss of gelsolin markedly preserved cardiac function. Data represent means ± s.e.m. ^*P < 0.05 compared with the respective control groups (NFC or Sham), ^#P < 0.05 compared with respective WT—9 Wk or 18 week TAC group as determined by unpaired two-tailed Student’s t test (c, g) and one-way ANOVA analysis (e, f, k, m, o). ^$P < 0.05 for Kaplan–Meir survival analysis (i) as determined by log-rank test. Biological replicates: n = 20 (b), n = 8 (c–e, n–o), n = 6 (g), n = 50 (i), n = 4 (j) and n = 12 (k–m). Scale bars show 25 µm (d), 1 mm (j), 2 mm (y-axis of l), and 200 ms (x-axis of l)

Fig. 2

Loss of gelsolin attenuates pressure overload-induced cardiac remodeling and adverse cytoskeletal remodeling. a, b Histological analyses by PSR (a) staining showing increased myocardial fibrosis (b) in WT mice which were attenuated in GSNKO mice. c, d Taqman real-time PCR analyses showing increased mRNA expression of pro-collagen-I α1 (c) and pro-collagen-III α1 (d) in WT mice in response to pressure overload-induced biomechanical stress. Loss of gelsolin resulted in attenuation of pressure overload-induced mRNA expression of these extracellular matrix proteins. e–g Taqman real-time PCR analyses showing attenuation of pressure overload-induced increase in mRNA expression of cardiac disease markers including ANF (e), BNP (f), and β-MHC (g) in GSNKO mice compared with the WT mice at 9 and 18 weeks postsurgery. h–j Single cardiomyocyte contractility measurements (h) showing attenuation of decreased myofilament FS (i) and ±dL/dt (j) in cardiomyocytes isolated from GSNKO LVs compared with WT LVs in response to pressure overload for 9 weeks. k–m Representative images of F- and G-actin staining (k), F- to G-actin ratio (l), and actin-depolymerizing activity (m) showing increased actin depolymerization in WT hearts in response to pressure overload, whereas loss of gelsolin resulted in attenuation of actin-depolymerizing activity leading to increased F- to G-actin ratio. n Immunoprecipitation of gelsolin from WT—9 weeks TAC heart tissue homogenate resulting in marked attenuation of actin-depolymerizing activity. o–q Representative phase-contrast images (o) and quantification of viable cardiomyocytes (p) showing increased viability in GSNKO cardiomyocytes in response to 24-h cyclical stretch. Loss of gelsolin also resulted in greater increase F to G-actin ratio in response to 24-h cyclical stretch (q). In input, S supernatant from immunoprecipitate experiment. Data represent means ± s.e.m. ^*P < 0.05 compared with the respective sham group, ^#P < 0.05 compared with corresponding WT—9 Wk or 18 Wk TAC group as determined by unpaired two-tailed Student’s t test (n) and one-way ANOVA analysis (b–g, i, j, l, m, p, q). Biological replicates: n = 4 (a, b), n = 10 (c–g), n = 6 (h–j), n = 4 (k–m), and n = 3 (o–q). For in vitro experiments, each biological replicate was mean of technical replicates (o–q); only biological replicates are plotted and used for statistics. Scale bars show 25 µm (a, k) and 100 µm (o)

Fig. 3

PIP2 and PIP3 bind with and inhibit gelsolin. a Actin-depolymerization assay showing identical inhibition of gelsolin by equimolar PIP2 and PIP3 in an in vitro lysate-free assay. b Complete model of human gelsolin structure illustrating its 6 domains (G1-G6) and the C-terminal tail. c, d Molecular modeling illustrating potential sites of interaction of the N-terminus (c) and C-terminus (d) of gelsolin with PIP2. e, f Molecular modeling illustrating potential sites of interaction of the N-terminus (e) and C-terminus (f) of gelsolin with PIP3. Please also see Supplementary Movies 1–4. Data represent means ± s.e.m. ^$P < 0.05 compared with the 600 nM rpGelsolin group as determined by one-way ANOVA analysis (a). Biological replicates: n = 9 (a)

Fig. 4

PI3Kα-generated PIP3 is a major negative regulator of gelsolin activity. a Concentration-dependent effect of rpGSN in actin-depolymerization assay. b, c Inhibition of actin-depolymerizing activity of 600 nM rpGSN (b) and 60 nM rhGSN (c) by PIP3 in presence of VO-OHpic. d A schematic showing the interconversion pathway of PIP2 and PIP3 by p110α and PTEN. e PIP2-mediated inhibition of exogenous gelsolin actin-depolymerizing activity in GSNKO heart tissue homogenate is partially blocked by BYL-719 and VO-OHpic, suggesting that the conversion to PIP3 is necessary for PIP2-mediated inhibition of gelsolin activity in ex vivo actin-depolymerization assay. f, g Actin polymerization assay showing inhibition of actin-capping activity of rpGSN (f) and rhGSN (g) by PIP3 in presence of VO-OHpic. h PIP2-mediated inhibition of exogenous gelsolin actin-capping activity in GSNKO heart tissue homogenate is partially blocked by BYL-719 and VO-OHpic, suggesting that the conversion to PIP3 is necessary for the PIP2-mediated inhibition of gelsolin actin-capping activity in ex vivo actin polymerization assay. i A schematic showing the experimental plan for isolated cardiomyocyte stretching in the absence or presence of the PIP3 micelles. j–l Representative images of F- and G-actin staining (j), F- to G-actin ratio (k), and actin-depolymerizing activity (l) showing increased actin depolymerization in the p110α transgenic cardiomyocytes subjected to cyclical stretch for 24 h. Addition of PIP3 micelles in presence of PBP-10 and VO-OHpic attenuated the increased actin-depolymerizing activity in stretched p110α transgenic cardiomyocytes. m Representative gelsolin and PIP3 IF staining images showing increased PIP3 levels in PIP3 micelles-treated cells. Arrows indicate spatial colocalization of gelsolin and PIP3. Data represent means ± s.e.m. ^$P < 0.05 compared with respective controls, ^‡P < 0.05 compared with PIP2 + BYL-719 + VO-OHpic group, ^*P < 0.05 compared with nonstretch group, ^#P < 0.05 compared with WT-Ctrl—Stretch group, ^§P < 0.05 compared with respective Stretch group as determined by unpaired two-tailed Student’s t test (b, c, f, g) and one-way ANOVA analysis (e, h, k, l). Biological replicates: n = 6 (a–c, f–h, j–m) and n = 9 (e). For in vitro experiments, each biological replicate was mean of technical replicates (j–m); only biological replicates are plotted and used for statistics. Scale bars show 25 µm (j, m)

Fig. 5

p110α, but not p110β, interacts with gelsolin. a Co-immunoprecipitation and immunoblotting showing that p110α and gelsolin interact in murine and human hearts. b Co-immunoprecipitation and immunoblotting showing no interaction between p110β and gelsolin in murine and human hearts. c, d Representative IF staining images showing spatial colocalization between p110α and gelsolin in murine (c) and human (d) hearts. e Representative IF staining images for p110α and N-cadherin in murine hearts showing the localization of p110α at the intercalated discs. Surface plots were plotted using Fiji ImageJ (NIH) show a graphical representation of intensity profiles of the IF images. f–g Western blot analyses showing decreased myocardial levels of p110α in human (f) and canine (g) hearts with DCM compared with the NFC donor hearts. h–j Representative images of F- and G-actin staining (h), F- to G-actin ratio (i), and actin-depolymerizing activity derived from actin-depolymerization assay (j) showing a greater degree of actin depolymerization in p110α transgenic mice in response to pressure overload-induced biomechanical stress. Addition of PIP3, in the presence of VO-OHpic, inhibited the actin-depolymerizing activity (j). k–n Single cardiomyocyte contractility measurements (k) showing decreased myofilament FS (m) and ±dL/dt (l–n) in LV cardiomyocytes isolated from 2 weeks of pressure-overloaded p110α transgenic mice hearts. Data represent means ± s.e.m. ^*P < 0.05 compared with the respective sham/NFC groups, ^#P < 0.05 compared with WT-Ctrl—2 week TAC group as determined by unpaired two-tailed Student’s t test (f, g) and one-way ANOVA analysis (i, j, l–n). Biological replicates: n = 4 (a–j) and n = 6 (k–n). Scale bars show 25 µm (c, d, e, h)

Fig. 6

Loss of p110α leads to accelerated HF in response to pressure overload. a–e Representative M-mode echocardiography images of LV (a) and quantification of cardiac function showing severely decreased LVEF (b) and LVFS (c) along with increased LV end-diastolic dimension (LVEDD; d) and left atrium (LA) size (e) in p110α transgenic mice, αDN (PI3KαDN) and αCre (PI3Kα^flx/flx Cre), in response to pressure overload-induced biomechanical stress, compared with preserved cardiac function in the WT mice. f–h Taqman real-time PCR analyses showing a greater increase in mRNA expression of cardiac disease markers including ANF (f), BNP (g), and β-MHC (h) in p110α transgenic mice compared with the WT mice in response to pressure overload for 2 weeks. i–k Histological analyses by Masson trichrome staining (i) and WGA staining (j) showing a greater increase in ventricular dilation (i), cardiac fibrosis (j), and myocyte cross-sectional area (j, k) in p110α transgenic mice compared with the WT mice in response to pressure overload. l, m Representative IF staining images showing unchanged phosphorylation of FAK in WT LVs, in contrast to significantly increased phosphorylation of FAK in p110α transgenic LVs in response to pressure overload (Fig. 3d–f). Data represent means ± s.e.m. ^*P < 0.05 compared with the respective Sham groups, ^#P < 0.05 compared with WT-Ctrl—2 Wk TAC group as determined by one-way ANOVA analysis (b–h, k, m). Biological replicates: n = 12 (a–e), n = 10 (f–h) and n = 4 (i–m). Scale bars show 2 mm (y-axis of a), 200 ms (x-axis of a), 1 mm (i), and 25 µm (j, l)

Fig. 7

Loss of gelsolin attenuates the biomechanical stress-induced cytoskeletal remodeling and preserves the cardiac function in PI3KαDN mice. a–c Representative images of F- and G-actin staining (a), F- to G-actin ratio (b), and actin-depolymerizing activity (c) showing pressure overload-induced dysregulation of actin filaments (reduction of F/G-actin ratio) and increased actin-depolymerizing activity in the PI3KαDN LVs. Loss of gelsolin (in αDN/GSNKO mice) resulted in reduced actin-depolymerizing activity and preserved actin filament arrangement (preserved F/G-actin ratio) in response to pressure-overload. d, e Representative IF images (d) and their quantification (e) showing attenuated phosphorylation of FAK in αDN/GSNKO hearts in response to pressure overload. f–i Masson trichrome (f, g) and WGA staining (h) showing alleviation of ventricular dilation (f), cardiac fibrosis (g), and myocyte cross-sectional area (h, i) in αDN/GSNKO mice compared with the αDN mice in response to pressure overload. j–l Taqman real-time PCR analyses showing attenuation of increased mRNA expression of cardiac disease markers including ANF (j), BNP (k), and β-MHC (l) in αDN/GSNKO mice compared with the αDN mice in response to pressure overload-induced biomechanical stress. m–o Representative images of F- and G-actin staining (m), F- to G-actin ratio (n), and actin-depolymerizing activity (o) showing the attenuation of increased actin-depolymerizing activity in the cardiomyocytes isolated from αDN/GSNKO LVs compared with αDN LVs subjected to cyclic stretch for 24 h. p–s Single cardiomyocyte contractility measurements (p) showing attenuation of decreased myofilament FS (q) and ±dL/dt (r, s) in cardiomyocytes isolated from αDN/GSNKO LVs compared with αDN LVs in response to pressure overload. t–u Quantitative assessment of M-mode echocardiography of LV showing preserved cardiac function in αDN/GSNKO hearts in response to pressure overload-induced biomechanical stress. Data represent means ± s.e.m. ^*P < 0.05 compared to all the groups, ^#P < 0.05 compared with αDN—2 week TAC group as determined by one-way ANOVA analysis (b, c, e, i–l, n, o, q–u). Biological replicates: n = 4 (a, b, d–i), n = 6 (c, m–s), n = 10 (j–l) and n = 12 (t, u). For in vitro experiments, each biological replicate was mean of four technical replicates (m–o); only biological replicates are plotted and used for statistics. Scale bars show 25 µm (a, d, g, h, m) and 1 mm (f)

Fig. 8

Regulation of cytoskeleton density by PI3Kα. a Normal myocyte: active PI3Kα produces a pool of PIP3 that suppresses excessive activation of gelsolin (GSN) by Ca²⁺ during Ca²⁺ cycling leading to moderate gelsolin severing activity, normal cytoskeleton (F-actin) density, and good resilience to biomechanical stress. b Heart failure (e.g., dilated cardiomyopathy, DCM) or PI3Kα-deficient model under pressure overload: low-levels or absent PI3Kα activity leads to low levels of PIP3. Lack of PIP3 result in unhindered (high) gelsolin activation during Ca²⁺ cycling, excessive breakdown of cytoskeleton (F-actin), low-cytoskeleton density, and poor resistance to biomechanical stress leading to DCM. c Heart failure resilience due to GSN deficiency: in the absence of gelsolin (GSN) and actin-severing activity associated with it, myocytes are able to maintain a high density of cytoskeleton (F-actin) resulting in high resilience to biomechanical stress and linked heart failure. d Reverse remodeling (LVAD): in the presence of left-ventricular assist device, heart contraction and associated Ca²⁺ release are of much lower magnitude. Low levels of Ca²⁺ during Ca²⁺ cycling (release) result in less Ca²⁺-activation of gelsolin (inactive gelsolin) moderating gelsolin actin-severing activity that leads to improvement in cytoskeletal (F-actin) density, which in turn may drive reverse remodeling