FAK (focal adhesion kinase) has been shown to mediate the hypertrophic growth of the left ventricle. Experimental results also suggest that FAK may contribute to the structural and functional deterioration of the chronically overloaded left ventricle. In the present study, we postulated that FAK expression and phosphorylation may be altered in the volume-overloaded heart in humans. FAK expression and phosphorylation at Tyr397 were detected by Western blotting and immunohistochemistry in samples from endomyocardial biopsies from patients with MR (mitral regurgitation; n=21) and donor subjects (n=4). Hearts from patients with MR had degenerated cardiac myocytes and areas of fibrosis. In this group, the myocardial collagen area was increased (18% in MR hearts compared with 3% in donor hearts respectively) and correlated negatively with left ventricular ejection fraction (r=−0.74; P>0.001). FAK expression and phosphorylation at Tyr397 (a marker of the enzyme activity) were increased in samples from MR hearts compared with those from donor hearts (3.1- and 4.9-fold respectively). In myocardial samples from donor hearts, anti-FAK staining was almost exclusively restricted to cardiac myocytes; however, in myocardial samples from MR hearts, staining with the anti-FAK antibody was found to occur in myocytes and the interstitium. There was a positive correlation between collagen and the interstitial areas stained with the anti-FAK antibody (r=0.76; P>0.001). Anti-FAK and anti-vimentin staining of the interstitial areas of samples from MR hearts were extensively superimposed, indicating that most of the interstitial FAK was located in fibroblasts. In conclusion, FAK expression and phosphorylation are increased and may contribute to the underlying structural and functional abnormalities in the volume-overloaded heart in humans.
Cardiac diseases often result in adverse remodelling of the left ventricle and progression to heart failure. Complex alterations of the architecture of the myocardium, including hypertrophy and loss of cardiac myocytes as well as interstitial fibrosis, lead to contractile and relaxing dysfunctions that underlie pump failure of the remodelled left ventricle [1,2]. These structural abnormalities are thought to result from the integrated responses of the various myocardial cell types, mainly myocytes and fibroblasts, to the continuous impact of mechanical stress and neurohumoral factors [3,4].
In recent years, evidence has accumulated suggesting that FAK (focal adhesion kinase) is a key mediator in mechanical stress and growth factor signalling in myocyte and non-myocyte myocardial cell populations [5–10]. FAK activation, elicited either by integrin engagement or growth factors, is dependent on its autophosphorylation at Tyr397 and co-operation with activated Src, which is responsible for the phosphorylation of additional tyrosine residues. This additional phosphorylation enables FAK to recruit molecules which modulate distinct signalling pathways involved in the regulation of multiple cell functions [11–13]. In cardiac myocytes, for instance, FAK activation has been demonstrated to be critical for the expression of the hypertrophic genetic programme in response to both mechanical stress and hypertrophic agonists [7,14–16]. The contribution of FAK signalling for fibroblast migration and differentiation into myofibroblasts is also relatively well established [9,10,17]. Furthermore, recent studies performed in cardiac-myocyte-FAK-conditional knockout mice [18,19] have confirmed the importance of FAK for the establishment of load-induced LV (left ventricular) hypertrophy and have indicated that activation of FAK signalling may play a role in preventing deterioration in the chronically overloaded heart. Results also exist, however, to suggest that chronic activation of FAK may contribute to the myocardial deterioration that culminates in heart failure. It has been shown  that, in hypertrophic left ventricles of aortic-banded rats, FAK levels are increased not only in compensated hypertrophy, but also in the transition from compensated hypertrophy to heart failure. Increases in FAK protein expression were seen in cardiac myocytes and in areas of active collagen deposition, suggesting a complex role of FAK signalling in the transition from hypertrophy to heart failure. However, FAK expression and activity in the chronically overloaded human heart has not been investigated.
In the present study, we examined FAK expression and distribution in samples from the left ventricle myocardium, obtained by intra-operative endomyocardial biopsy from patients with MR (mitral regurgitation). We found that FAK expression and phosphorylation at Tyr397, which is a hallmark of the active state of FAK, are increased distinctively in the interstitium of MR hearts compared with donor hearts.
MATERIALS AND METHODS
The study has been carried out in accordance with the Declaration of Helsinki (2000) of the World Medical Association, and has been approved by the Institutional Ethics Committee. All patients with MR gave informed consent. In the case of donor hearts, consent for research was granted by the donor family.
Protein A labelled with 125I was purchased from DuPont, rabbit polyclonal anti-FAK (SC-558) and mouse monoclonal anti-vimentin (SC-32322) antibodies were from Santa Cruz Biotechnology, the rabbit polyclonal anti-[phospho-FAK (Tyr397)] (44-625G) antibody was from Biosource International, the anti-CD45 (M0701) antibody was from Dako, streptavidin–peroxidase (RPN1231) and streptavidin–Cy2 (PA42001) were from Amersham Biosciences, and TRITC (tetramethylrhodamine β-isothiocyanate)–phalloidin (P1951), anti-(rabbit IgG)–biotin (B8895), anti-(mouse IgG)–biotin (B7264) and TRITC-conjugated (T5393) and FITC-conjugated (F0382) secondary antibodies were from Sigma.
A total of 21 patients with isolated MR underwent pre-operative clinical and echocardiographic examination. Patients had no evidence of coronary artery disease, hypertension or diabetes. Of these, 14 patients had degenerative MR (floppy and flail valve), and seven had rheumatic valve disease. The echocardiographic examination was performed with 2.5-3.5 MHz transducers. Dimensions of left ventricle [LVEDD (LV end-diastolic diameter) and LVESD (LV end-systolic diameter)] were assessed from two-dimensional guided M-mode tracings. LVMI (LV mass index) was estimated by LV mass calculated by standard echocardiographic methods normalized to body-surface area. LVEF (LV ejection fraction)  was calculated with the use of the Theicholz method. LAD (left atrial diameter) was assessed by two-dimensional measurements. PASP (pulmonary artery systolic pressure) was estimated through the velocity of the tricuspid regurgitant jet and estimated RAP (right atrial pressure) by using the formula: PASP=4×(peak tricsupid regurgitation velocity)2+mean RAP.
During open heart surgery, endomyocardial biopsy samples taken from the left ventricle free wall were immediately frozen in liquid nitrogen and stored at −80°C for later use in Western blotting or immunofluorescence. Endomyocardial biopsy specimens of the left ventricle free wall were also obtained from donor hearts (n=4) immediately before implantation and were used as controls for immunoblotting, immunohistochemistry and histology.
Tissue samples were homogenized in PBS containing freshly added protease and phosphatase inhibitors (1% Triton X-100, 10 mmol/l sodium pyrophosphate, 100 mmol/l sodium fluoride, 10 μg/ml aprotinin, 1 mmol/l PMSF and 0.25 mmol/l sodium orthovanadate). The samples were centrifuged for 20 min at 11000 g, and the soluble fraction was resuspended in 50 μl of Laemmli loading buffer [120 mmol/l Tris/HCl (pH 6.8), 2% (w/v) SDS, 20% (v/v) glycerol, 280 mmol/l 2-mercaptoethanol and 0.04 mg/ml Bromophenol Blue] before separation by SDS/PAGE [8% (w/v) polyacrylamide gels]. The same amount of total protein from individual samples was loaded on to the gels. Proteins were transferred from the gels on to a nitrocellulose membrane. Membranes were blocked for 2 h at room temperature (25°C) with 5% (w/v) non-fat dried milk in TBST buffer [10 mmol/l Tris/HCl (pH 8), 150 mmol/l NaCl and 0.05% Tween 20] and were then exposed to primary antibodies and 125I-labelled Protein A. The control for protein loading was obtained by resolving total myocardial extracts on SDS/PAGE gels followed by silver staining.
Serial 5-μm-thick sections of myocardium samples were placed on to poly(L-lysine)-precoated slides, and were stained with Masson's trichrome or Picro-Sirius Red.
Biopsy samples were fixed by overnight immersion with 4% (w/v) paraformaldehyde in 100 mmol/l sodium phosphate buffer (pH 7.4) and processed to inclusion in Histotec (Merck). Sections (5 μm) were transferred to poly(L-lysine)-coated glass slides. Endogenous peroxidase activity was blocked by treatment with 0.03% H2O2 in 100 mmol/l PBS at room temperature for 30 min. Sections were pre-incubated in blocking buffer [5% (w/v) non-fat dried milk in 100 mmol/l PBS) for 45 min at 37°C, followed by overnight incubation with the primary anti-FAK or anti-vimentin antibodies (1:100 dilution) at 4°C. The sections were rinsed extensively in 50 mmol/l PBS and incubated with biotin-conjugated anti-(rabbit IgG) or anti-(mouse IgG) secondary antibody and streptavidin conjugated with peroxidase for 2 h at 25°C. After washing as above, sections were incubated in freshly prepared DAB (diaminobenzidine) solution containing H2O2 (0.8%) for 5 min. The specificity of the secondary antibodies was tested in a series of positive and negative control reactions. In the absence of primary antibodies, application of secondary antibodies (negative controls) failed to produce any significant staining.
For laser confocal microscopy, sections were preincubated in blocking buffer [3% (w/v) non-fat dried milk in 100 mmol/l PBS] containing 0.6% Triton X-100 for 45 min at room temperature, followed by overnight incubation with the primary anti-FAK, anti-(phospho-FAK), anti-vimentin or anti-CD45 antibody and then with biotin-conjugated goat anti-(rabbit IgG) or anti-(mouse IgG) secondary antibody and streptavidin–Cy2 (1:500 dilution in PBS) and TRITC–phalloidin at 4°C. Sections were also double-stained with anti-FAK and anti-vimentin antibodies. Anti-(mouse IgG)–TRITC (1:250 dilution) and anti-(rabbit IgG)–FITC antibodies (1:250 dilution) were used as secondary antibodies for anti-vimentin and anti-FAK primary antibodies respectively. Immunofluorescence was detected by laser confocal microscopy (Zeiss LSM510).
Quantification of fibrillar collagen and immunostained samples
Collagen density was estimated by digital image analysis from sections stained with Picro-Sirius Red by measuring the percentage of the total surface area pertaining to fibrillar collagen in each field (>10 fields of vision/sample chosen randomly), as identified by polarized light. The area occupied by anti-FAK in immunostained samples was also estimated. Furthermore, we compared FAK staining with the number of the interstitial cell nuclei ratio in the areas of dense fibrosis with those of normal interstitium in myocardial samples from patients with MR. All densities were determined at a magnification of ×250.
Samples were embedded in Epon, following a standard protocol [21a]. Ultra-thin sections were double-stained with uranyl acetate and lead citrate before examination with a transmission electron microscope (LEO 906).
Quantitative analysis was performed with values from standardized densitometric readings of staining with anti-FAK and anti-(phospho-FAK) antibodies in myocardial samples from MR and donor hearts. Statistical analysis was performed with the actual values from the densitometric readings. Differences between the mean values of the samples from MR and donor hearts were tested by Student's t test. Linear regression analysis and standard Student's t test were used to evaluate the relationships between the collagen area and LVEF or anti-FAK staining. P>0.05 indicated statistical significance. Results are expressed as means±S.E.M.
Pre-operative clinical and echocardiographic data from patients with MR are shown in Table 1. All patients were symptomatic and underwent surgical mitral valve repair (n=13) or valve replacement (n=8). No deaths occurred during the 6 months of follow-up after surgery. LVMI, LVEDD and LAD were increased in patients compared with the standard normal echocardiography reference. The average LVEF of patients with MR, determined at the time of admission, was 66±2.7%, but ten patients had an LVEF below 60%, which is considered a critical cut-off value for clinical outcome in patients with MR .
|Characteristic||Patients with MR (n=21)|
|Symptom duration (months)||18±2|
|Characteristic||Patients with MR (n=21)|
|Symptom duration (months)||18±2|
Histological analysis of endomyocardial biopsies from MR hearts (Figure 1A) revealed structural degenerative abnormalities (arrows). Interstitial fibrosis, indicated by Masson's trichrome staining, was seen in many biopsy samples from these patients (Figure 1B). Samples stained with Picro-Sirius Red had typical red birefringence of the collagen in the perivascular area as well as in the areas surrounding the cardiac myocytes, suggesting reactive and replacement fibrosis (Figure 1C). There was a negative correlation (r=−0.74, P>0.001) between the area stained with Picro-Sirius Red and LVEF in patients with MR (Figure 1D), suggesting the importance of fibrosis in cardiac functional deterioration in these patients. The areas of collagen staining were 18±4.3 and 3±3% in samples from MR and donor hearts respectively.
Cardiac myocyte degeneration, myocardial fibrosis and anti-FAK staining in samples from MR and donor hearts
In samples from MR hearts, cardiac myocytes and the interstitium surrounding the myocytes and blood vessels were consistently stained with the anti-FAK antibody (Figure 1E). Higher magnification of degenerated cardiac myocytes (Figure 1F) showed that the anti-FAK staining was characteristically intense with a gross texture, resembling aggregates that were prominent in the central part of the cells. In samples from donor hearts, anti-FAK staining was restricted to cardiac myocytes, with discrete staining of the interstitial areas (Figure 1G). A positive correlation was found between anti-FAK staining and the Picro-Sirius Red collagen staining in the samples from MR hearts (Figure 1H).
Further analysis of the FAK distribution in the samples of donor and MR hearts was performed with confocal microscopy. In the samples from donor hearts (Figures 2A and 2B), cardiac myocytes displayed the usual uniformity in size and compact arrangement with narrow interstitial spaces. There was consistent anti-FAK immunostaining of cardiac myocytes, reflecting the relatively high level of FAK expression in these cells. Anti-FAK antibody and phalloidin staining were not overlapping, indicating that FAK and sarcomeric actin were not co-localized in normal cardiac myocytes. Anti-FAK staining was distributed along the longitudinal axis of cardiac myocytes as spots organized regularly in the sarcoplasm (Figure 2B). No major anti-FAK immunostaining was seen in the interstitium of donor hearts.
Anti-FAK staining of samples from donor and MR hearts
In the myocardial samples from MR hearts, cardiac myocytes were typically variable in size and the spaces between the cardiac myocytes were clearly increased, reflecting interstitial fibrosis. These spaces also displayed staining with the anti-FAK and anti-(phospho-FAK) antibodies (Figures 2D, 2F and 2H). Cardiac myocytes from MR hearts had large spots stained strongly with phalloidin, which were mostly located in the perinuclear area of the cells (Figures 2C, 2E and 2G). Such spots stained with both anti-FAK and anti-(phospho-FAK) antibodies (Figures 2D, 2F and 2H). However, these spots were also co-stained with antibodies to vimentin or CD45 (as shown in Figure 3) and to distinct signalling and structural proteins [e.g. MEF2 (myocyte enhancer factor 2), NF-κB (nuclear factor κB) and myosin; results not shown]. The areas of cardiac myocytes from MR hearts outside of these aggregates displayed weaker and irregular anti-FAK and anti-(phospho-FAK) staining (Figures 2D, 2F and 2H). The weaker anti-FAK and anti-(phospho-FAK) staining of these areas of myocytes and in the interstitium was probably due to the fact that phalloidin fluoresced so brightly that it was necessary to scale the imaging gain to very low levels in order to obtain the image. The staining of cardiac myocytes and the interstitium with the anti-(phospho-FAK) antibody (Figure 2H) suggested an increased FAK activity in both structures.
Anti-vimentin and anti-CD45 staining of samples from MR and donor hearts
In samples from MR hearts, the interstitium surrounding the myocytes and blood vessels were consistently stained by the anti-vimentin antibody (Figure 3A), but the samples from donor hearts only had faint vimentin staining (Figure 3B). The preferential staining of the interstitium by the anti-vimentin antibody was confirmed by confocal microscopy (Figure 3C). Double-staining of endomyocardial samples from MR hearts with anti-FAK and anti-vimentin antibodies (Figure 3D) resulted in the superimposing of FAK and vimentin staining in interstitial areas, but not in the areas occupied by cardiac myocytes, indicating that most of the interstitial FAK is located in fibroblasts. This was strengthened by the demonstration that the anti-CD45 antibody (Figures 3E and 3F) stained discreet and disperse spots in the interstitium, indicating a limited number of leucocytes in the myocardium of patients with MR.
We then compared different regions of the interstitial space to determine the relationship between fibroblast number and the area stained with FAK. The regions were separated as those without major interstitial fibrosis (Figure 3G) and those with major interstitial fibrosis (Figure 3H). As shown in Figure 3(I), the ratio of anti-FAK staining to the number of fibroblasts was dramatically increased in areas of prominent fibrosis compared with areas without major interstitial fibrosis.
Ultrastructural examination of endomyocardial biopsy samples from MR hearts showed that some of the cardiac myocytes were severely degenerated and atrophic. In degenerated cells, there was extensive myofibril disorganization, a partial-to-complete loss of sarcomeres, sarcoplasmic reticulum and T-tubules as well as numerous mini-mitochondria (Figure 4A). Ultrastructural analysis also revealed fibroblasts in proximity to cardiac myocytes and surrounded by extensive collagen fibrils, as shown in Figure 4(B).
Ultrastructure of cardiac myocyte degeneration and fibrosis in MR hearts
To examine further the expression and phosphorylation of FAK, samples of endomyocardial biopsies taken from MR or donor hearts were homogenized, resolved by SDS/PAGE and immunoblotted with anti-FAK or anti-(phospho-FAK) antibodies. As shown in Figure 5(A), the amount of FAK detected by the anti-FAK antibody increased in the samples MR hearts compared with donor hearts. Equal amounts of total myocardial protein (50 μg) from individual samples were loaded on to the gels, as confirmed by silver staining. Quantitative analysis indicated that FAK expression was approx. 3.1-fold greater in samples from MR hearts compared with donor hearts (Figure 5B). Endomyocardial samples from MR hearts also had greater amounts of FAK phosphorylated at Tyr397 (approx. 4.9-fold). The ratio between the amount of phosphorylated and total FAK was increased by approx. 40% in samples from MR hearts compared with donor hearts.
FAK expression and phosphorylation at Tyr397 in the myocardium from donor and MR hearts
In the present study, we provide evidence that FAK protein expression and phosphorylation at Tyr397 are increased in the volume-overloaded myocardium of patients with chronic MR. It was demonstrated that increased FAK expression and phosphorylation were mainly related to myocardial fibroblasts and paralleled the increases in interstitial fibrosis of chronically overloaded myocardium. These results support the hypothesis that FAK may contribute to the fibrogenesis and structural abnormalities in response to chronic volume-overload in hearts from patients with MR.
MR is a prevalent valvular heart disease that evolves with progressive remodelling and eventually failure of the left ventricle [21,23,24]. Continuous impact of abnormal haemodynamic burden caused by the regurgitant flow and the later increases in the levels of systemic and local factors, such as AngII (angiotensin II), endothelin or cytokines, are thought to lead to the myocardial abnormalities that accompany cardiac remodelling and failure in patients with MR [25–27]. LV dysfunction in chronic MR has been shown to correlate with abnormalities in cardiac myocytes, such as reduced myofibril content and accumulation of lipids, and myocardial fibrosis [24,28]. Severe fibrosis has been reported in patients with MR and heart failure and is suspected to affect the overall LV myocardial remodelling . In the present study, we have shown abnormalities in cardiac myocytes and fibrosis in samples of endomyocardial biopsies taken from patients with moderate-to-severe MR. Considering that the average LVEF was within the normal range in the patients with MR included in the present study, the results shown indicate that insidious structural abnormalities in endomyocardial areas may be present in chronic volume-overload caused by MR, despite the preservation of LVEF. Although, to our knowledge, this is the first demonstration of the early structural changes of the myocardium in MR, previous studies have shown an early contractile impairment in hearts from patients with MR with normal LVEF . Volume-overload and the increased LV diameter/thickness ratio in MR, which tends to increase wall stress, have been suggested as the major factors responsible for the contractile dysfunction in chronic MR . Accordingly, these mechanical factors might possibly be the determinants of the damage in the endomyocardial layer of hearts from patients with MR as seen in the present study
Besides the already known degenerative abnormalities of cardiac myocytes, we also observed the presence of an amorphous material, probably constituted by protein aggregates, in the perinuclear areas of degenerated cardiac myocytes from the patients with MR. The assumption that these aggregates are composed mainly of proteins was indicated by the demonstration that they were stained strongly by phalloidin, as well as by antibodies directed to distinct signalling proteins. As the results in the present study are descriptive, further studies are necessary to determine the mechanisms leading to the formation, as well as the importance, of such amorphous material to cardiac myocyte degeneration in failing hearts. Noticeably, protein aggregates, similar to those seen in the present study, have been reported in cardiac myocytes of desmin- or αB-crystallin-related cardiomyopathies [32,33], where they have been suggested to contribute to cell death and the associated abnormalities in myocardial function.
Another major finding of the present study was the increased expression of FAK and its phosphorylation in hearts from patients with MR. Morphologically, the increased expression of FAK in these samples appears to be predominantly associated with increased tissue fibrosis, rather than with cardiac myocytes. This contrasts with the distribution of FAK in samples from donor hearts, which is predominantly seen in cardiac myocytes. The demonstration of an extensive superimposing of FAK with vimentin staining and the scarce number of CD45-labelled cells (a marker of inflammatory cells) in our samples suggest that increased interstitial FAK staining is related to fibroblasts. In addition, we have shown that the interstitial cells were stained by the anti-(phospho-FAK) antibody, suggesting an increased activity of FAK in fibroblasts. Importantly, the fibroblasts in the areas of dense fibrosis were stained markedly with the anti-FAK antibody compared with those located in the interstitial areas without fibrosis. This might imply a specific role for FAK in the genesis of fibrosis in chronically overloaded myocardium. Accordingly, FAK has been demonstrated to influence several aspects of fibroblast biology. For instance, FAK is a critical regulator of fibroblast differentiation and survival in response to a number of humoral factors or mechanical stress [10,34,35]. Moreover, previous observations indicate that FAK plays a pivotal role in the expression, secretion and activation of MMP-2 (matrix metalloproteinase-2) [36,37], a major determinant of the extracellular matrix remodelling process in overloaded myocardium [38,39].
The contribution of cardiac myocytes to the increased expression of myocardial FAK in the endomyocardial samples from patients with MR could not be confirmed in the present study, as there was no apparent change in FAK staining in cardiac myocytes from MR compared with those from donor hearts. However, we did observe that cardiac myocytes from MR hearts were stained by the anti-(phospho-FAK) antibody, suggesting an increased activity of FAK. Moreover, we also noticed that the distribution of FAK in cardiac myocytes from patients with MR lacked the typical regular distribution characteristic of normal cardiac myocytes. This might reflect the advanced degenerative stage of the cardiac myocytes in the samples from patients with MR examined in the present study. It will be important to determine whether this atypical localization affects the basal state or the activation of FAK in cardiac myocytes. Consistent with a major role for FAK in the responses of cardiac myocytes to chronic overload, recent studies in myocyte-restricted FAK-knockout mice  have demonstrated that the lack of FAK is associated with premature cardiac myocyte degeneration, interstitial fibrosis and functional deterioration in the chronically overloaded left ventricle, indicating that FAK signalling plays a critical role in the adaptive hypertrophy to increased workload. However, it remains unknown whether its persistent activation contributes to maladaptive abnormalities in chronically overloaded hearts.
In conclusion, in the present study, we have investigated the expression, tyrosine phosphorylation and distribution of FAK in the volume-overloaded left ventricle from patients with MR. An important and novel finding of the present study is the co-existence of increased FAK expression and phosphorylation, and the increased myocardial interstitial fibrosis in chronically overloaded hearts in humans. As FAK is a critical signalling molecule for the control of a variety of basic cellular processes, either in cardiac myocytes or fibroblasts, these results may provide new insights into the mechanisms underlying the structural alterations of the chronically overloaded myocardium and, consequently, in the pathogenesis of heart failure.
focal adhesion kinase
left atrial diameter
LV end-diastolic diameter
LV ejection fraction
LV end-systolic diameter
LV mass index
pulmonary artery systolic pressure
right atrial pressure
This study was sponsored by grants from FAPESP (Fundação de Amparo à Pesquisa do Estado de São Paulo; Proc. 01/11698-1) and CNPq (Conselho Nacional de Desenvolvimento Científico e Tecnológico; Proc. 521098/97-1).