Conceived and designed the experiments: CL VS IR DM TC. Performed the experiments: CL VS MM AC IR DM. Analyzed the data: CL VS MM AC IR DM TC. Contributed reagents/materials/analysis tools: DM SS CP TC. Wrote the paper: VS TC.
Current address: The Rayne Institute, King's College London, The Centre of Excellence in Medical Engineering, London, United Kingdom
The authors have declared that no competing interests exist.
The Hepatocyte Growth Factor (HGF) is a pleiotropic cytokine involved in many physiological processes, including skeletal muscle, placenta and liver development. Little is known about its role and that of Met tyrosine kinase receptor in cardiac development.
In this study, we generated two transgenic mice with cardiac-specific, tetracycline-suppressible expression of either Hepatocyte Growth Factor (HGF) or the constitutively activated Tpr-Met kinase to explore: i) the effect of stimulation of the endogenous Met receptor by autocrine production of HGF and ii) the consequence of sustained activation of Met signalling in the heart. We first showed that Met is present in the neonatal cardiomyocytes and is responsive to exogenous HGF. Exogenous HGF starting from prenatal stage enhanced cardiac proliferation and reduced sarcomeric proteins and Connexin43 (Cx43) in newborn mice. As adults, these transgenics developed systolic contractile dysfunction. Conversely, prenatal Tpr-Met expression was lethal after birth. Inducing Tpr-Met expression during postnatal life caused early-onset heart failure, characterized by decreased Cx43, upregulation of fetal genes and hypertrophy.
Taken together, our data show that excessive activation of the HGF/Met system in development may result in cardiac damage and suggest that Met signalling may be implicated in the pathogenesis of cardiac disease.
The cellular events occurring during the early stages of life, including pre- and perinatal phases, may have strong impact on long-term health. Epidemiological and experimental evidences suggest that development of cardiovascular diseases in the adult is influenced by stressful events during late prenatal or early postnatal life
The Hepatocyte Growth Factor (HGF) is a mesenchyme-derived multifunctional molecule that elicits mitogenic and morphogenic activities in development, as well as in many patho-physiological processes
In the heart, HGF has been shown to exert anti-apoptotic/cardioprotective effects in rats subjected to myocardial infarction
In this study, we aimed to investigate this issue by activating the HGF/Met system specifically in the heart. To this purpose we generated two novel gain-of-function transgenic models with tetracycline-suppressible expression of either HGF or activated Met under control of the α-MHC promoter. In the mouse embryo, the α-MHC promoter is expressed throughout the myocardium starting from E8
All animal procedures were approved by the Ethical Commission of the University of Torino, Italy, and by the Italian Ministry of Health, both of which accepted the use of mice for this study (A/R 0045 and A/R 0041).
The mouse HGF cDNA was cloned into the pBI-EGFP plasmid which is responsive to tTA transactivator
The Tpr-Met-TRE-GFP responder mouse
Hearts were excised, rinsed in ice-cold Tyrode solution and prepared in RNA later (Ambion). Total RNA was extracted with TRIzol (Sigma). Qiagen RNAeasy kit (Qiagen GmbH, Hilden, Germany) was used to enhance purification. After quantification (NanoDrop® ND-1000, NanoDrop Technologies), reverse transcription was performed using DNA Polymerase/Superscript III Reverse Transcriptase (Invitrogen). For Real-time PCR, primers and Taqman probe specific for the transgene were designed using the File builder 3.1 program (Applied Biosystems, Foster city, CA, USA). Real-time PCR was performed on a 7300 Real-time PCR instrument (Applied Biosystem). Sample reactions were performed in triplicate and normalized to 18S mRNA expression. For semi-quantitative RT-PCR, control samples were prepared without adding the RT enzyme to the reaction. Tubulin was used as control. See
Hearts were removed, rinsed in ice-cold Tyrode solution and fixed in 4% paraformaldehyde (PAF) in phosphate-buffered saline (PBS) for 4 hours at 4°C. After PBS washings, hearts were incubated in 30% sucrose in PBS overnight at 4°C to preserve GFP fluorescence. Stereomicroscopy was viewed by Leica MZ12 and imaged by Evolution VF colour cool camera and Image-Pro Plus software. Tissues for indirect immunofluorescence were embedded in OCT (Biooptica), quickly frozen in isopentane and stored at −20°C. Sections were 20 µm cut, post-fixed 5′ in ice-cold 4% PAF and washed with PBS. For Met staining, sections were incubated with SP260 primary antibody (Santa Cruz) overnight at 4° and subsequently with Alexa Fluor 546-conjugated goat anti-rabbit antibody (Molecular Probes) for 1h at room temperature. The double overlay pictures (Met/GFP) were viewed with a Leica DM6000 CS confocal microscope. Optical slices (1024 by 1024 pixels, frame resolution) were acquired at 10Hz and processed with LAS AF software (Leica Microsystems CMS GmbH). Quadruple overlay pictures (Laminin/Griffonia/DAPI/GFP) were obtained by staining with rabbit polyclonal antibody laminin (Sigma) followed by Alexa Fluor 647-conjugated goat anti-rabbit antibody (Molecular Probes), two hours incubation with rhodamine Griffonia Simplicifolia (Vector Laboratories) and 5 min with DAPI. Quadruple overlay pictures (Cx43/α-actinin/DAPI/GFP) were obtained by staining with rabbit polyclonal antibody Cx43 (Sigma) and mouse monoclonal antibody α-actinin (Sigma) and subsequently with Alexa Fluor 647-conjugated goat anti-rabbit antibody and Alexa Fluor 546-conjugated goat anti-mouse antibody (Molecular Probes). Confocal microscope imaging was performed with Leica TCS SP2 AOBS upright microscope. Optical slices (1024 by 1024 pixels, frame resolution) were acquired at 200 Hz, with a line average of 8, and processed with LAS AF software (Leica Microsystems CMS GmbH).
Ki67 positive nuclei were immunostained in 20 µm thick heart sections. Primary rabbit polyclonal Ki67 antibody (Novocastra) and secondary Alexa Fluor 546-conjugated goat anti-rabbit antibody (Molecular Probes) were used. Fluorescence imaging and processing were performed with Leica DM6000 CS confocal microscope and LAS AF software, respectively. 5 fields per area (Right Ventricle, Left Ventricle and Interventricular Septum) per mouse were analyzed. 3 mice per group were considered.
H9c2 cell line purchased from the American Type Culture Collection was grown as described
See Supplemental
Gap junction permeability assay was performed as described
Hearts were rinsed in PBS, dehydrated and embedded in paraffin. Sections (6–8 µm thick) were rehydrated, stained with hematoxylin-eosin or Masson's trichrome and analyzed with Leica DMRE microscope.
3 mice per group were analyzed. Transversal 20 µm thick cryo-sections of the middle region of the hearts were stained with rhodamine Griffonia Simplicifolia and rabbit polyclonal antibody laminin and subsequently with Alexa Fluor 488-conjugated goat anti-rabbit antibody (Molecular Probes) and DAPI. Fluorescence images were taken at 40× magnification with Leica DM6000 CS confocal microscope and LAS AF software was used for processing. 15 cross sectional areas of 6 fields per heart were measured. Small, medium and large-sized fibers were equally considered. Fiber CSA delimited by laminin staining was measured using ImageJ software. Density probability distribution curves were generated.
Images were obtained by classical microscopy analysis (Leica DMRE microscope) of ventricles and interventricular septum at 20×. Data from 10 samplings were averaged for each heart. 6 controls and 3 transgenics were analyzed. Signal intensity of staining was calculated as percentage of total tissue area using ImageJ.
Size and function of the left ventricle of the mice were evaluated by high-resolution echocardiography. M-mode examinations were performed using a dedicated small-animal high-resolution imaging unit (Vevo 770; VisualSonics, Toronto, Canada) and a 40-MHz high-frequency linear transducer (RMV 707B; VisualSonics, Toronto, Canada). Mice were kept anesthetized with tribromoethanol (Avertin, 350 mg/kg). Real-time imaging was performed with a frame rate of 100 Hz (temporal resolution of 10 msec). The following parameters were measured: systolic and diastolic thickness of the interventricular septum, end-systolic (LVESD) and end-diastolic diameter (LVEDD) of the left ventricle, systolic and diastolic thickness of the posterior wall of the left ventricle. Fractional shortening (FS) was then calculated
Data are expressed as the mean ± SD. Differences between groups were determined by independent T-tests (one or two-tailed T-tests have been used; details in each Figure Legend).
To examine the influence of HGF in normal prenatal and postnatal cardiac growth, we generated a bitransgenic α-MHC-driven tetracycline-suppressible system (
(A) Schematic representation of the two components (α-MHC-tTA and HGF-TRE-GFP transgenes) for bitransgenic conditional HGF expression. (B) Top: Experimental design: mice were never administered DOX to express HGF in the prenatal and postnatal period (HGF tg mice); pregnant mothers and suckling progeny received DOX to continuously repress HGF (HGF + DOX tg mice); pregnant mothers were not administered DOX to induce HGF
Neonatal heart of HGF tg showed specific GFP expression (
In a cohort of animals (prenatal HGF tg mice), DOX was administered to suckling progeny 1 day after birth (
Endogenous Met was localized all around the plasma membrane of cardiomyocytes in heart tissue isolated from littermate wild-type animals (control) and, at lower levels of expression, in HGF tg neonates (
(A) Immunofluorescence of Met receptor (red) and GFP (green) in neonatal (P7) heart samples of control (left panel) and HGF tg mice (middle panel). A negative control of secondary antibody was included (right panel). Bars: 50µm. (B) Western blot of Met (p140Met) protein in control and HGF tg mice at different ages post-birth (P2 n = 6 n = 7, P4 n = 8 n = 6, P7 n = 10 n = 11, P18 n = 9 n = 14). Representative blots are shown below densitometric quantification (normalized on GAPDH loading control, relative to P2 control). Controls vs HGF tg mice: *p<0.05 and †p<0.005 (two-tailed T-test). (C) Densitometric quantification (normalized on tubulin loading control) and representative Western blot of phospho-Erk1,2 (P Erk1,2), phospho-p38 MAPK (P p38) and phospho-Akt (P Akt) in HGF tg (n = 7) relative to control mice (n = 6) at two days post birth (P2). *p<0.05 (two-tailed T-test). (D) Western blot analysis of Met receptor and downstream signalling after treatment of H9c2 cardiomyoblast cell line with 10U/ml of HGF for different lengths of time. Densitometric quantification was normalized against tubulin and plotted as relative to time 0′ of treatment. Each condition was tested 3 times.
To examine whether the extra-dose of HGF was able to increase proliferation of cardiac cells, we analyzed Ki67 positive cells in tissue sections of 7 days-old neonatal hearts. We found a 3-fold increase (p<0.05) in HGF tg mice, compared to controls (
(A) Left panel: Quantification of Ki67 positive nuclei in tissue sections of 7 days-old neonatal hearts. Right Ventricle (RV), Left Ventricle (LV) and Interventricular Septum (IVS) were separately or totally analyzed and compared in controls vs HGF tg neonates (n = 3 animals per group). At least 5 fields per zone per sample were counted. *p<0.05 (one-tailed T-test). Right panel: Representative Ki67 staining in tissue sections of 7 days-old neonatal hearts of control (left) and HGF tg (right). Ki67: red-nuclear (white arrows). Bar: 100 µm. (B) AlamarBlue assay and (C) BrdU incorporation of H9c2 cell line not treated (nt) and treated with 10U/ml HGF for the indicated times. Experiments were done in 8 (B) and 2 (C) biological replicates for each sample group. †p<0.005 versus nt (two-tailed T-test). Right panels: representative IF. BrdU: green-nuclear; propidium iodide (PI): red-nuclear. Bar: 75 µm. (D,E) Densitometric quantification normalized to Erk2 loading control and representative Western blots of the indicated proteins. Results represent averaged values for immunoblot analyses performed on heart lysates in (D) P7 neonatal controls (n = 10) vs HGF tg (n = 11) and (E) P18 young adult controls (n = 9) vs HGF tg (n = 14). Myosin heavy chains (α and β-MHC), troponins (cTnT, cTnI and ssTnI) and Cx43 have been quantified. *p<0.05, †p<0.005 (two-tailed T-test).
In parallel, we analyzed the levels of sarcomeric proteins in newborn mice. We observed that the levels of cTnT (p<0.005) and cTnI (p<0.05) were significantly lower in hearts of HGF tg neonates, as compared to controls (
Adult HGF tg mice were analyzed by echocardiography and compared with controls. No difference was found between single transgenics (silent HGF and α-MHC-tTA) and wild-type mice (
(A) Representative images of Left Ventricle long-axis echocardiogram (2D and M-Mode) of control (upper panels) and HGF tg mice (lower panels). (B) Densitometric quantification normalized to Erk2 loading control and representative Western blots of heart ventricles from control vs HGF tg (upper graph) n = 9 mice per group and prenatal HGF tg (lower graph) n = 3 mice per group. In the latter, HGF expression was suppressed after birth. Re-expression of β-MHC and decreased Cx43 are evident in both bitransgenic mice compared to controls. *p<0.05 (two-tailed T-test). (C) Representative images of Lucifer yellow dye diffusion in HGF tg and control edge-cut hearts (upper panels) and zoom-in of the areas included in dashed boxes (lower panels). Bottom graph: quantification of pixel area showed that cell-to-cell spread of Lucifer yellow was significantly decreased in HGF tg mice vs controls (n = 3 mice per group). †p<0.005 (one-tailed T-test). Bars: 100µm.
wild-type | HGF tg | prenatal HGF tg | silent HGF | α-MHC-tTA | |
(n = 11) | (n = 20) | (n = 6) | (n = 9) | (n = 6) | |
0.49±0.08 | 0.34±0.09 |
0.33±0.06 |
0.46±0.07 | 0.46±0.13 | |
1.133±0.181 | 1.172±0.281 | 1.259±0.176 | 1.128±0.116 | 1.200±0.291 | |
3.772±0.402 | 4.052±0.484 | 3.970±0.473 | 3.861±0.371 | 3.732±1.140 | |
1.009±0.128 | 1.002±0.232 | 1.010±0.125 | 0.981±0.091 | 0.970±0.130 | |
1.654±0.203 | 1.602±0.414 | 1.614±0.227 | 1.648±0.140 | 1.639±0.455 | |
1.948±0.435 | 2.663±0.504 |
2.665±0.449 |
2.113±0.456 | 2.149±0.915 | |
1.476±0.180 | 1.374±0.305 | 1.332±0.092 | 1.502±0.065 | 1.397±0.189 | |
0.571±0.068 | 0.539±0.115 | 0.577±0.090 | 0.550±0.059 | 0.635±0.223 | |
0.193±0.050 | 0.182±0.036 | 0.226±0.031 | n.d. | n.d. | |
33.17±5.08 | 32.71±3.76 | 30.36±5.76 | n.d. | n.d. | |
0.006±0.001 | 0.006±0.001 | 0.007±0.001 | n.d. | n.d. |
Wild-type: littermate wild-type control; HGF tg: bitransgenic mice conceived in the absence of DOX; prenatal HGF tg: bitransgenic mice treated with DOX at birth and maintained in DOX thereafter; silent HGF and α-MHC-tTA: littermate single transgenics. FS, fractional shortening; IVSTd, interventricular septum thickness in end diastole; LVEDD, left ventricle end diastolic diameter; PWTd, posterior wall thickness in end diastole; IVSTs, interventricular septum thickness in end systole; LVESD, left ventricle end systolic diameter; PWTs, posterior wall thickness in end systole; h/r, heart rate; HW, heart weight; BW, body weight; HW/BW, heart weight/body weight ratio. n.d., not determined.
*p<0.005 versus wild-type (two-tailed T-test).
After echocardiographic analysis, bitransgenic animals and controls were sacrificed and heart/body weight ratios were measured (
We decided to extend observations to another gain-of-function model produced in our laboratory that, differently from the HGF tg mouse, allows activation of Met in the absence of the ligand and cannot be downregulated
To overcome the early lethality of Met hyperactivation and to evaluate effects of permanent Met activation in postnatal cardiomyocytes, Tpr-Met mice were conceived and delivered in the presence of DOX to suppress expression of Tpr-Met during
(A) Experimental design (left), RT-PCR (middle) and immunoblot (right) of Tpr-Met expression in postnatal Tpr-Met mice (P27) with Doxycycline (DOX) suppression until birth, compared to controls. (B) Control and postnatal Tpr-Met hearts were analyzed under stereomicroscopy for comparison (upper panel). Four-chamber cut hearts are also showed (lower panel). Bars: 5 mm. (C) Significantly increased heart weight (upper graph) and heart/body weight ratio (lower graph) indicate cardiac hypertrophy in postnatal Tpr-Met mice (n = 6 animals per group). †p<0.005 vs control (two-tailed T-test). (D) Mean cross-sectional area (CSA) of ventricular cardiomyocytes is significantly higher in postnatal Tpr-Met compared to controls (left panel). n = 300 cells from 3 biological replicates per group. †p<0.005 vs control (two-tailed T-test). Distribution curves (right panel) of counted CSA show a shift to the right side in postnatal Tpr-Met mice as respect to controls. (E) Representative transversal sections of left ventricles show increased size in postnatal Tpr-Met cardiomyocytes. Left panels: laminin (red-surface). Right panels: quadruple overlay with laminin (red- surface), Griffonia (blue-endothelial), DAPI (white-nuclear) and GFP (green-intracellular). Bars: 35µm.
(A) Semi-quantitative RT-PCR analysis of controls and postnatal Tpr-Met mice (P27) showed re-expression of ANF and β-MHC mRNA. Tubulin is used as loading control. (B) α to β isoform switch of MHC, increased phosphorylation of downstream Akt and Erk1,2 and strongly decreased Cx43, mild increase of ZO-1 and normal N-Cadherin and β-Catenin levels in postnatal Tpr-Met vs control hearts, analyzed by Western blot. Densitometric quantification normalized on Akt loading control and representative blots below graphs are shown. n = 10 mice for each group. *p<0.05 and †p<0.005 vs control (two-tailed T-test) (C) Left panels: representative confocal immunofluorescence images of left ventricle sections from postnatal Tpr-Met mice showed decreased staining of Cx43 (red), compared to controls (upper panels). Bottom panels: quadruple overlay with Cx43: red; α-actinin: blue; DAPI: white-nuclear; GFP: green-intracellular; Bars: 35µm. Quantification of Cx43 staining was performed with ImageJ. n = 6 controls and n = 3 postnatal Tpr-Met mice.
In this article, first we demonstrate that during the early postnatal period of rapid growth, neonatal cardiomyocytes express the Met receptor
It has been shown that cardiomyocyte cell proliferation is accompanied by a decrease of cell-cell communication
The inducible character of our HGF mouse model demonstrates that prenatal development is the specific stage influenced by activation of Met signalling. In fact, at 3 months of age our HGF tg mice developed a contractile defect, even when HGF expression was suppressed after birth. This result is consistent with the finding that the endogenous Met receptor is physiologically downregulated in terminally differentiated cardiomyocytes, making the system insensitive to further HGF stimulation. The high susceptibility of prenatal age to Met stimulation is further confirmed by the fact that expressing Tpr-Met instead of HGF starting from prenatal age was lethal soon after birth. The Tpr-Met fusion protein lacks the extracellular, transmembrane and juxtamembrane domains of Met receptor and has gained the Tpr dimerization motif, which allows constitutive and ligand-independent activation of the kinase. The loss of juxtamembrane sequences necessary for the negative regulation of kinase activity and receptor degradation prolongs duration of Met signalling
Sustained activation of Tpr-Met in postnatal cardiomyocytes (1 to 4 weeks) leads to increased cross sectional area of cardiomyocytes, reactivation of fetal gene program, increased cardiac mass and, ultimately, to lethal congestive heart failure at P28. Thus, the constitutive activation of Tpr-Met gives to the cell a signal of growth, which, in terminally differentiated cardiomyocytes, results in switching on a hypertrophic program. Both Ras/RAF/MEK/ERK and Akt pathways, which are downstream to Met and Tpr-Met, are known to be involved in the growth promotion and protection of cardiomyocytes from apoptosis. However, their contribution in defining “physiological” versus “pathological” growth is still controversial. In the HGF model, the low level of Met receptor, which cannot be superinduced by HGF stimulation in the terminally differentiated cardiomyocyte, cannot shift the equilibrium to hypertrophic growth. Interestingly, in the postnatal Tpr-Met model we found only mild signs of interstitial fibrosis, albeit pathological hypertrophic growth is usually associated with scar tissue formation. This finding confirms the antifibrotic action of HGF/Met activation, which has been demonstrated in a variety of tissues, including the heart
Tpr-Met expression induced a dramatic decrease in Cx43 protein levels in postnatal cardiomyocytes. This reinforces the concept that Met receptor activation acts negatively on cell-cell communication, albeit the precise mechanism by which this suppression is mediated awaits elucidations. Evidence is in favour of the view that both formation and maintenance of gap junctions is critically dependent on the presence of correct mechanical stabilization
In conclusion, our mouse models support the idea that HGF/Met stimulation promotes cardiomyocyte growth. Although other studies have suggested that HGF may have a beneficial function in pathological conditions, such as ischemic injury, there are no experimental evidences in the current study to demonstrate that enhancement of HGF/Met signalling is favourable in a physiological setting. On the other hand, excessive HGF/Met signalling in prenatal period may raise adverse effects and might be linked to the pathogenesis of progressive cardiac disease.
Neonatal HGF tg hearts show no morphological defects. Haematoxylin-eosin staining of four-chamber cut sections of P7 control (left) and HGF tg (right) hearts. Bars: 2mm.
(0.88 MB TIF)
No signs of fibrosis nor hypertrophy were found in adult HGF tg mice. (A) Trichrome staining does not show fibrosis in either control or littermate HGF tg mice at 4 months of age. (B) Cross-sectional area of myocytes was not different between control and HGF tg mice at 4 months of age (green-surface: laminin; green-intracellular: GFP; blue-nuclear: DAPI; red-endothelial: Griffonia). Bars: 50 µm (A); 75µm (B).
(3.14 MB TIF)
Postnatal Tpr-Met mice at P27 display signs of congestive heart failure. (A) Tpr-Met mice exhibit dyspnea and lethargy. Extensive oedema and haemorrhage of Tpr-Met lungs shown by haematoxylin and eosin staining of lung tissue (B), stereomicroscopy inspection (C) and lung weight measurement (D), compared to littermate controls. n = 4 animals per group. * p<0.01 vs control (two-tailed T-test). Bars: 20mm (A); 100mm (B); 5mm (C).
(1.23 MB TIF)
Single immunofluorescence stainings of quadruple overlay shown in
(2.67 MB TIF)
Single immunofluorescence stainings of quadruple overlay shown in
(2.66 MB TIF)
Primers used throughout the study.
(0.31 MB PDF)
List of antibodies used in this study.
(0.59 MB PDF)
We gratefully acknowledge Sokol Kalaja for animal care, Marco Demaria for technical help in Real time-PCR and Guido Serini for Confocal Imaging with TCS SP2 AOBS. We also thank Stefano Gatti and Simona Gallo for technical assistance and Giovanni Losano for useful comments and critical reading of the manuscript.