Ski drives an acute increase in MMP-9 gene expression and release in primary cardiac myofibroblasts

Abstract

Many etiologies of heart disease are characterized by expansion and remodeling of the cardiac extracellular matrix (ECM or matrix) which results in cardiac fibrosis. Cardiac fibrosis is mediated in cardiac fibroblasts by TGF-β₁ /R-Smad2/3 signaling. Matrix component proteins are synthesized by activated resident cardiac fibroblasts known as myofibroblasts (MFB). These events are causal to heart failure with diastolic dysfunction and reduced cardiac filling. We have shown that exogenous Ski, a TGF-β₁ /Smad repressor, localizes in the cellular nucleus and deactivates cardiac myofibroblasts. This deactivation is associated with reduction of myofibroblast marker protein expression in vitro, including alpha smooth muscle actin (α-SMA) and extracellular domain-A (ED-A) fibronectin. We hypothesize that Ski also acutely regulates MMP expression in cardiac MFB. While acute Ski overexpression in cardiac MFB in vitro was not associated with any change in intracellular MMP-9 protein expression versus LacZ-treated controls,exogenous Ski caused elevated MMP-9 mRNA expression and increased MMP-9 protein secretion versus controls. Zymographic analysis revealed increased MMP-9-specific gelatinase activity in myofibroblasts overexpressing Ski versus controls. Moreover, Ski expression was attended by reduced paxillin and focal adhesion kinase phosphorylation (FAK - Tyr 397) versus controls. As myofibroblasts are hypersecretory and less motile relative to fibroblasts, Ski's reduction of paxillin and FAK expression may reflect the relative deactivation of myofibroblasts. Thus, in addition to its known antifibrotic effects, Ski overexpression elevates expression and extracellular secretion/release of MMP-9 and thus may facilitate internal cytoskeletal remodeling as well as extracellular ECM components. Further, as acute TGF-β₁ treatment of primary cardiac MFB is known to cause rapid translocation of Ski to the nucleus, our data support an autoregulatory role for Ski in mediating cardiac ECM accumulation.

Keywords: Cardiac fibroblast; MMP9; Ski; cardiac fibrosis; cell migration; fibroblast activation; myofibroblast.

Figure 1

Ski overexpression increases intracellular MMP2 expression in primary cardiac myofibroblasts. First passage (P1) cardiac MFB grown to 70% confluency, and infected with either Ad‐LacZ (LacZ‐100) or AdSki (Ski‐100) at MOI of 100 for 48 h. Untreated control cells were serum‐deprived, but not infected with either virus. (A) Representative immunoblots of total cell lysates probed for intracellular Ski, MMP‐2, β‐tubulin (control), and MMP‐9 with β‐tubulin (control). (B) Histogram showing data obtained by densitometry of blots as in A. The level of intracellular MMP‐2 was significantly increased in response to Ski overexpression, compared to untreated controls. Data are representative of the means ± SEM with n = 4 replicates. *P < 0.05 compared to uninfected controls. Untreated control lysates were serum‐starved, but not infected with either virus. (C) Histogram showing data obtained from Western blots of total cell lysates probed for intracellular MMP‐9. Data are normalized to β‐tubulin +/‐ SEM . The levels of intracellular MMP‐9 was not altered in response to Ski overexpression (AdSki‐100). Data are representative of the means ± SEM; n = 8; P = 0.7394.

Figure 2

Ectopic Ski expression increases MMP2 gelatinase activity in primary cardiac myofibroblasts. First passage (P1) primary cardiac MFB were grown to 70% confluency, and infected with either AdLacZ or AdSki at 50 and 100 MOI for 48 h. Untreated control lysates were serum‐deprived, but not infected with either virus. (A and C) Representative gelatin zymographs of total cell lysates for intracellular MMP‐2 in MFB treated with 50 MOI adenovirus (Fig. 2A) or 100 MOI adenovirus (Fig. 2C). Native human MMP‐2 protein (5 μg) was used as a positive control in Figure 2A. (B and D) Histograms showing data obtained in Figures 2A and C, respectively. The level of intracellular MMP‐2 activity was significantly increased in response to Ski overexpression, compared to Ad‐LacZ and untreated controls. Data are representative of the means ± SEM with n = 3. **P < 0.01, compared to uninfected and Ad‐LacZ infected controls.

Figure 3

Ski overexpression upregulates MMP‐9 expression in primary cardiac myofibroblasts. First passage (P1) primary rat cardiac MFB were grown to 70% confluency, and infected with either Ad‐LacZ or Ad‐Ski at 50 and 150 MOI for 48 h. Untreated control MFB cultured cells were serum‐deprived, but not infected with either virus. Total cell lysates were assayed for Ski to confirm successful infection of cells. Conditioned media was assayed for MMP‐2 protein via immunoblotting. Data are normalized to the control expression levels ± SEM. (A) Representative immunoblots of concentrated conditioned media (20 μg/lane) for secreted MMP‐2, MMP‐9, and TIMP‐4 as well as cell lysates for HA‐tagged Ski. Specificity of bands for MMP‐2 and MMP‐9 were noted with the use of HT‐1080 cells treated with PMA and assayed from conditioned media using the same Western blot conditions as provided to our lab (Richard Schulz, University of Alberta, personal communication). (B) Histogram showing data obtained from Western blots, as represented in Figure 3A. The level of secreted MMP‐2 was unchanged in response to AdSki infection. Data are representative of means ± SEM, where n = 3. The level of secreted MMP‐9 was significantly increased in AdSki‐infected cells compared to Ad‐LacZ‐infected controls. Data are representative of the means ± SEM where n = 7; *P ≤ 0.05 compared to untreated and LacZ controls. Protein levels of secreted TIMP‐4 were significantly reduced with AdSki infection, compared to AdLacZ‐infected and uninfected controls. Data are representative of the means ± SEM where n = 7; **P ≤ 0.01 and ***P ≤ 0.001, when comparing to untreated and AdLacZ‐50 and AdLacZ‐150 controls, respectively. (C) qPCR assay for expression of MMP‐2 and MMP‐9 mRNA expression in AdSki treated samples versus controls. The histogram shows data normalized to Gapdh, ±SEM, where n = 3, ****P ≤ 0.001 using one‐way ANOVA, relative to untreated and AdLacZ controls.

Figure 4

Ectopic expression of Ski yields opposite effects on MMP‐2 and MMP‐9 gelatinase activity in cardiac myofibroblasts. First passage (P1) primary cardiac MFB were grown to 70% confluency, serum‐deprived and infected with either Ad‐LacZ or Ad‐Ski at 50 MOI (for the MMP‐9 assay) or a range of 10 to 150 MOI (for the MMP‐2 assay) for 48 h. Untreated controls were serum‐deprived, but not infected with either virus. Concentrated conditioned media (15 μg protein/lane) from all groups were assayed for MMP activity by gelatin zymography. Native human MMP‐2 protein (5 μg) served as a positive control in MMP‐2 studies. (A) Representative gelatin zymography of concentrated conditioned media for secreted MMP‐9. (B) Histogram showing data obtained in A. Data are normalized to the control expression ± SEM. The level of secreted MMP‐9 activity was significantly elevated in response to Ski overexpression at an AdSki MOI of 50. Data are representative of the mean ± SEM where n = 3; **P ≤ 0.01 compared to untreated control and AdLacZ‐50 control values. (C) Representative gelatin zymography of concentrated conditioned media (15 μg protein/lane) from all groups for MMP‐2 activity. The level of secreted MMP‐2 activity was significantly attenuated in response to Ski overexpression at MOI of 150, relative to LacZ‐infected controls. (D) Histogram showing data obtained in Figure 4C. Data are normalized to the control expression level ± SEM. MMP‐2 activity is significantly decreased in response to AdSki treatment compared to AdLacZ treated and uninfected controls. Data are representative of the means ± SEM where n = 3. *P ≤ 0.05 compared to untreated and AdLacZ‐150 controls.

Figure 5

Chronic overexpression of Ski yields and increase in MMP‐9 secretion over time. P1 cardiac MFB were grown to 70% confluency and were infected with either Ad‐LacZ (MOI of 50) or AdSki (MOI of 50) for various durations up to 96 h. Untreated controls were serum‐deprived, but not infected with either virus. Expression of secreted MMP‐9 was assessed at 24, 48, 72, and 96 h after infection with AdSki. Using mean ± SEM where n = 3, we observed an significant increase in MMP‐9 expression over time, as seen at72 versus 48 h (*P ≤ 0.05) and 96 versus 24 h samples (*P ≤ 0.05).

Figure 6

Paxilin expression and FAK phosphorylation are negatively regulated by Ski in cardiac myofibroblasts. First passage (P1) cardiac MFB were grown to reach 70% confluency and infected with either Ad‐LacZ or AdSki (MOI of 50, 100, or 150) for 48 h. Untreated controls were serum‐deprived, but not infected with either virus. Expression of exogenous HA‐tagged Ski was confirmed using anti‐HA antibodies and β‐tubulin was used as a loading control. (A) Representative immunoblots of total cell lysates probed for paxillin. (B) Histogram showing data obtained in A. Paxillin protein levels are significantly decreased in response to AdSki infection, as compared to AdLacZ‐infected and uninfected controls. Data shown are representative of the means ± SEM where n = 3; *P ≤ 0.05 for AdSki50 compared to untreated and AdLacZ‐50 controls; ***P ≤ 0.001 for AdSki150 compared to untreated and AdLacZ‐150 controls. C: Representative immunoblots of whole cell lysates probed for immunoreactivity to total FAK and phospho‐FAK (Tyr 397). n.s., non‐specific antibody binding. (D) Histogram showing data obtained in A. FAK phosphorylation at Tyr 397 is significantly decreased with Ad‐Ski infection at 150 MOI, compared to AdLacZ‐50 and uninfected controls. Data are representative of the mean ± SEM where n = 4; *P ≤ 0.05 for AdSki150 compared to controls. (E) Representative immunoblots of total cell lysates probed with vinculin antibody. (F) Histogram showing data obtained in E. No significant differences were observed among treatment groups. Data shown are representative of the means mean ± SEM where n = 8; P = 0.2619. (G) Representative immunoblots of total cell lysates for integrin‐β1 antibody. (H) Histogram showing data obtained in A. Data shown are representative of the mean ± SEM where n = 4; P < 0.6506.

Figure 7

A schematic depicting Ski function in acute cardiac MFB activation. Herein we summarize our main findings, indicating that overexpression of Ski alone is sufficient to specifically and MMP‐9 secretion in MFB. We suggest that upon acute activation of the cell, Ski moves from the cytosol to the nucleus to repress TGF‐β ₁ and, as our current data indicates, serves as a driver of MMP‐9 expression and release. We hypothesize that Ski exerts its anti‐fibrotic effects, in part, by causing enhanced release of MMP‐9 in acutely stimulated cardiac fibroblasts and in cardiac myofibroblasts. To effect this specific response in combination with decreased MFB marker proteins and adhesion proteins (FAK and paxillin) Ski may serve as a scaffold for number of cofactors including Smads (Tecalco‐Cruz et al. 2018). On the other hand, in chronic cardiac injury (infarcted hearts) we have found that the intracellular distribution of Ski is shifted to the cytosol (Cunnington et al. 2011). If Ski cannot reach the nucleus, its effectiveness as a repressor of TGF‐β ₁ is compromised. Thus, in the chronically damaged and fibrosed heart, compartmentalization of Ski results in chronic elevation of MFB marker proteins and activation of the MFB phenotype.