Trophic Effect of Human Pericardial Fluid on Adult Cardiac Myocytes
Differential Role of Fibroblast Growth Factor-2 and Factors Related to Ventricular Hypertrophy
Abstract Pericardial fluid (PF) may contain myocardial growth factors that exert paracrine actions on cardiac myocytes. The aims of this study were (1) to investigate the effects of human PF and serum, collected from patients undergoing cardiac surgery, on the growth of cultured adult rat cardiac myocytes and (2) to relate the growth activity of both fluids to the adaptive changes in overloaded human hearts. Both PF and serum increased the rate of protein synthesis, measured by [14C]phenylalanine incorporation in adult rat cardiomyocytes (PF, +71.9±8.2% [n=17]; serum, +14.9±6.5% [n=13]; both P<.01 versus control medium). The effects of both PF and serum on cardiomyocyte growth correlated positively with the respective left ventricular (LV) mass. However, the magnitude of change with PF was 3-fold greater than with serum (P<.01). These trophic effects of PF were mimicked by exogenous basic fibroblast growth factor (FGF2) and inhibited by anti-FGF2 antibodies and transforming growth factor-β (TGF-β), suggesting a relationship to FGF2. In addition, FGF2 concentration in PF was 20 times greater than in serum. On the other hand, the LV mass–dependent trophic effect, present in both fluids, was independent of FGF2 concentration or other factors, such as angiotensin II, atrial natriuretic factor, and TGF-β. These data suggest that FGF2 in human PF is a major determining factor in normal myocyte growth, whereas unidentified LV mass–dependent factor(s), present in both PF and serum, participates in the development of ventricular hypertrophy.
An increasing body of evidence supports the existence of paracrine and autocrine regulatory pathways in the heart.1,2 Both coronary vascular and endocardial endothelium release several diffusible agents that acutely modify cardiac myocyte function.3–5 Similarly, several factors (including polypeptide growth factors) may modulate the long-term adaptive quantitative and qualitative changes in cardiac myocyte gene expression and act on cardiac growth during the development of cardiac hypertrophy.6 Two major classes of growth factors, synthesized by both myocytes and nonmuscle cells in the heart, have clear paracrine/autocrine roles in the development and differentiation of cardiac myocytes: FGF1, FGF2, and TGF-β.6-8 FGF2 and TGF-β are found in cardiac myocytes and in cardiac interstitium9,10 after cardiac overload.7,8,11 Although the presence of these factors has been characterized in the heart using several approaches in animal models12,13 and in humans,14 their trophic role in human myocardium in vivo remains unclear. In addition to locally released growth factors, circulating factors responsible for initiating the molecular adaptive changes in cardiac hypertrophy have been proposed.15,16 In a recent study, Yu et al17 described the presence of circulating factors in the serum of patients with chronic renal failure that may contribute to the development of cardiac hypertrophy.
The present study was undertaken to investigate the impact of both locally released and circulating cardiac growth factors on cardiac growth and the development of cardiac hypertrophy in vitro. In order to evaluate their presence and biological activity, we studied the effect of PF and serum obtained from patients undergoing cardiac surgery on the phenotype and cell growth of isolated adult rat myocytes.
PF is an ultrafiltrate of plasma,18 but it also reflects the composition of cardiac interstitium, at least in cardiac ischemia,19 and the production and release of macromolecules (including growth factors) in normal and diseased myocardium.19,20 In fact, molecules with a molecular mass of ≤40 kD can diffuse through the epicardium into the pericardial space.21 It has been recently shown that isolated rat cardiac myocytes release FGF2 via sarcolemma transient disruption.22 However, whether human PF contains growth factors and peptides specifically synthesized within the myocardium and whether its characteristics reflect cardiac adaptation to chronic overload have not yet been established.
The results demonstrate that human PF has a potent trophic effect on adult cardiac myocytes, regardless of the underlying cardiac disease, which is probably related to its high FGF2 content. Furthermore, the trophic effects of both PF and serum correlated positively with ventricular mass, suggesting the presence of additional diffusible LV mass–dependent factor(s) in both fluids that contribute to the development of cardiac hypertrophy.
Materials and Methods
Seventeen patients (mean age, 54±4 years; range, 21 to 87 years; 9 male and 8 female) were selected from adult subjects undergoing cardiac surgery for either valvular, coronary artery, or congenital heart disease. Six patients had stable coronary artery disease, 4 had aortic stenosis, 3 had aortic regurgitation, 2 had mitral valve regurgitation, 1 had mitral stenosis, and 1 had tetralogy of Fallot with a ventricular septal defect. Patients with combined coronary artery and valvular heart diseases were excluded, as were those with unstable coronary syndromes requiring intravenous therapy and those with a prior history of myocardial infarction, because of the uncertain effects of acute ischemic syndromes on ventricular geometry and function.
All patients had a baseline two-dimensional Doppler echocardiogram 1 to 3 days before cardiac surgery. LV mass (in grams) was calculated using the cube function formula, according to the method of Devereux et al.23 All patients gave informed written consent before the study, which was approved by the Committee on Human Ethics of our institute.
Subjects were prepared for surgery in an identical fashion. All medical therapy was discontinued 24 hours before surgery, with the exception of β-blockers in patients with coronary artery disease. The anesthetic protocol was similar in all patients, and none received heparin before sampling. The chest was opened by median sternotomy. PF was collected using a catheter inserted into the pericardial cavity immediately after incision of the anterior surface of parietal pericardium. Only samples free of contamination with blood were used for the study. The mean PF volume collected was 20.5±3.5 mL. Cells in PF were mainly mesothelial cells (77%) and lymphocytes (22%), regardless of the underlying cardiac disease, with a mean concentration of 560±57 cells/mm.3 These were eliminated by centrifugation.
A sample of central venous blood was simultaneously drawn from the right atrial catheter, and platelet-poor plasma was prepared for measurement of growth factors.24 Blood was collected in citrate tubes to prevent clot formation. Both PF and plasma were immediately centrifuged (400g for 10 minutes at 4°C), and supernatants were retained for study.
In cardiac surgery, the right atrial appendage is removed to allow atrial cannulation. In each patient, myocardial specimens were obtained from the atrial appendage. Biopsies were weighed and frozen at −20°C, lysis buffer (TRAx buffer, T-Cell Sciences Inc) was added (3 mL/mg of tissue), and biopsy specimens were crushed and left at room temperature for 45 minutes. Then 1 μL of diluent buffer (TRAx buffer) was added per milligram tissue, and tubes were vortexed and centrifuged (1430g for 25 minutes at 4°C). Supernatants were then collected and frozen until analysis for growth factor content.
Cardiac Myocyte Preparation and Culture
Cardiac myocytes were obtained from 2-month-old male Wistar rats (weight, 200 g) as previously described.25 Briefly, after anesthesia by intraperitoneal injection of pentobarbital, hearts were excised and perfused retrogradely with a buffer containing collagenase (0.1%, Boehringer-Mannheim) and bovine serum albumin. At the end of the isolation procedure, cells were exposed to 1 mmol/L Ca2+ at 37°C for 10 minutes and then washed in a serum-free medium (BM 86 Wissler, Aqual, Biochrom) containing 100 IU/mL penicillin, 0.1 μg/mL streptomycin, 10−9 mol/L insulin, 10−9 mol/L transferrin, and 10−9 mol/L glutamine and triiodothyronine. The number of cells and percentage of viable rod-shaped myocytes were determined in a Malassez chamber. Viable rod-shaped myocytes constituted 85% of the total cells on day 1; this percentage fell slightly with time in all groups examined (67% and 53% on days 2 and 3, respectively), as previously described.25 Nonmyocyte cells were eliminated by sedimentation and subsequent wash and never exceeded 4% of the total cell number in culture.
The cell concentration in the suspension was adjusted to 100 000/mL, and cells were seeded either in T25 culture flasks or in six-well culture plates (Falcon), precoated with 20 μg/mL laminin (Collaborative Research), dissolved in sterile cold water, and incubated for 1 hour at 37°C. Excess substrate was removed, and the plates were washed with culture medium before seeding of cardiac myocytes in a volume of 3 mL (flasks) or 1 mL (wells) that contained 300 000 and 100 000 cells, respectively. Myocytes were kept in culture for 3 days, and culture medium was changed every 24 hours. Dead cells, which do not adhere to the culture substrate, were removed at each medium change. In all experimental conditions (ie, control, PF-treated, and serum-treated cells), both cell density (cells/field) and percentage of rod-shaped cardiomyocytes in culture decreased with time from 117±4 cells/field on day 1 to 77±3 cells/field and 54±0 cells/field on days 2 and 3, respectively, as previously described.25
Estimation of the Rate of Total Protein Synthesis
Protein synthesis in cultured cardiac myocytes in the presence or absence of PF or serum was estimated by the 24-hour incorporation of [14C]phenylalanine (specific activity, 472 mCi/mmol; Amersham Intl) into the total TCA-precipitable proteins, as previously reported.25 The [14C]phenylalanine was added to the culture medium at day 2 (1 μCi/mL per dish) for 24 hours.
At the end of the incubation period, the medium was removed, cell preparations were rinsed twice in cold HEPES buffer, and proteins were subsequently precipitated with 1 mL TCA (10%) at 4°C for 30 minutes. Cells were then scraped off and collected by centrifugation at 3000 rpm for 3 minutes at 4°C. The pellet was washed twice in TCA (10%) and then once in water and frozen for protein assay. After aliquots were dissolved in 1N NaOH, radioactivity was counted in a beta counter. The amount of protein was determined by the Bradford assay.26
Estimation of the rate of protein synthesis was derived from the ratio of counts per minute of incorporated [14C]phenylalanine per microgram protein. For each data point, experiments were performed independently in triplicate. Data are expressed as percent change in the rate of protein synthesis in the control medium (24-hour incorporation of [14C]phenylalanine) in the presence of appropriate dilutions of either PF or serum.
In Situ Hybridization Technique
In situ hybridization with RNA probes specific for α- and β-MyHC RNA27 was adapted to the present experimental protocols to determine the distribution of α- and β-MyHC mRNA within myocytes cultured on laminin-coated slides in the presence or absence of either human PF or serum from both groups of patients, as previously described.28 The control condition was performed by the use of sense probes, which gave only background signal, as previously demonstrated.27,28
All slides derived from different sets of culture (control culture medium, PF, and serum) were fixed in paraformaldehyde, dehydrated, stored at −70°C until use, and then processed together for in situ hybridization, as previously described.27,28 Slides were examined by light- and dark-field illumination (Dialux, Leitz). Dark-field images, magnified 100-fold, were processed as previously reported.28 After computer image processing, hybridization signals were seen as black grains on a light field and were automatically quantified by the computer as dpd values and expressed in arbitrary units. A surface of 300 pixels allowed measurement within each myocyte. A minimum of 100 cells from 6 to 10 fields were randomly examined per slide, with the dpd values from each myocyte representing the mean of 3 to 5 dpd measurements randomly selected to scan the surface of each myocyte. Background was subtracted from the mean dpd within myocytes. Reproducibility of this method was established by the similar results obtained by two different operators on selected slices.
Experiments were performed on isolated cardiac myocytes exposed either to culture medium alone (control), supplemented as described above, or to the same medium containing 5% human PF or 5% human serum.
In a subgroup of experiments, the effect of PF and serum on cardiac myocyte growth was tested in the presence of (1) an antibody directed against human FGF2 (10 μg/mL anti-human FGF2 from goat, R&D Systems),22 (2) nonimmune goat IgG (10 μg/mL, R&D Systems), and (3) exogenous TGF-β (1 ng/mL, R&D Systems), added to either PF or serum. Finally, in a different set of experiments, we studied the effect of exogenous FGF2 (0.5 to 25 ng/mL, R&D Systems) and/or TGF-β (1 ng/mL), angiotensin II (10−10 to 10−5 mol/L), or ANF (10−9 to 10−6 mol/L) on isolated cardiac myocytes cultured in Wissler medium. Angiotensin II and ANF were purchased from Sigma.
The concentrations of Na+, K+, glucose, and total protein were measured in parallel in freshly sampled PF and serum (Kodak Ektakem 700, Johnson & Johnson). Osmolarity was determined by depression of the freezing point using an automatic osmometer (Roebling).
The supernatants of centrifuged PF and serum were stored at −70°C for use in the experimental protocols and for further laboratory analysis. The presence of growth factors (FGF2 and TGF-β) was measured using ELISA kits (R&D Systems).
FGF2 was measured by specific enzymatic immunoassay (Quantikine FGF basic, R&D Systems) in a solid-phase ELISA kit that measures human FGF2 in cell culture supernatants, serum, plasma, and other biological fluids. The accuracy of each series of assays was checked by adding a known amount of exogenous recombinant FGF2 (R&D Systems) to PF or serum.
The assay used for TGF-β detection (Quantikine, R&D Systems) uses a quantitative sandwich enzyme immunoassay technique and specifically detects active TGF-β1 concentrations in cells and serum, in biological fluids, and in culture supernatants using a receptor-based technique. Active TGF-β1 binds to TGF-β–soluble receptor coated onto a microtiter plate and is identified by an enzyme-linked polyclonal antibody, specific for the active form of TGF-β1. For ELISA quantification, the latent form present in serum and PF was converted into the active one by acidification and subsequent neutralization of PF and sera. In addition, since TGF-β is released in large amounts from degranulating platelets,29-31 only platelet-depleted serum was used for this assay.
Hormones (ANF, ET-1, and angiotensin II) were assessed by radioimmunological assay.32
Heparin-Sepharose Affinity Chromatography and Western Blot
Heparin-Sepharose affinity chromatography was used to enrich PF in FGF2, as described by Kardami and Fandrich.12 Briefly, 4 mL of either PF or serum samples, made up to 0.6 mol/L NaCl by adding solid NaCl, was mixed with 50 μL of settled heparin-Sepharose beads (Pharmacia), which were equilibrated in binding buffer (0.6 mol/L NaCl and 10 mmol/L Tris-HCl, pH 7). Suspensions containing beads were kept under slow agitation for 2 hours at room temperature. Heparin-Sepharose beads were allowed to settle, and the supernatant was carefully removed. After three washes with binding buffer (0.6 mol/L NaCl and 10 mmol/L Tris-HCl, pH 7), the settled beads were incubated for an additional 10 minutes with 1 mL of 1.1 mol/L NaCl and then washed twice with 0.1 mol/L NaCl and 10 mmol/L Tris-HCl, pH 6.8. Sedimented beads were finally incubated with 50 μL of electrophoresis gel sample buffer (0.1 mol/L Tris-HCl, pH 6.8, 2% SDS, 10% glycerol, and 5% [vol/vol] β-mercaptoethanol) and boiled for 5 minutes. The suspension was spinned for 1 minute, and the supernatant was used for electrophoresis and Western blotting.
A standard SDS-PAGE (0.75 mm thick, 12.5% polyacrylamide minigel) was used. Prestained protein molecular mass standards (200 to 7 kD, Bio-Rad), unstained chemiluminescent molecular mass standards (ECL Plus, Amersham), and human fibroblast lysates containing FGF2 (Affiniti) were included in the gel (10 μg/lane); 15 μL of either PF or serum extract was deposed in each lane.
Gels were transferred onto nitrocellulose membrane (Hybond ECL, Amersham) by electrophoretic transfer (Transblot, Bio-Rad). After overnight blocking of nonspecific protein binding sites with PBS, 0.05% Tween 20, and 5% milk, the membrane was incubated with purified primary goat anti-FGF2 polyclonal antibody (diluted 1:250 in PBS, 0.05% Tween 20, and 5% milk; R&D Systems) for 1 hour at room temperature, washed three times with PBS and 0.05% Tween 20, and incubated with secondary anti-goat IgG conjugated with horseradish peroxidase (1:5000 in PBS, 0.05% Tween 20, and 5% milk for 45 minutes at room temperature; Amersham). Detection was performed by the chemiluminescence method (ECL Plus, Amersham). Autoradiograms from the Western Blot were obtained with films (Hyperfilm, ECL, Amersham) for enhanced chemiluminescence.
Student’s t test and one-way and two-way ANOVA were used when appropriate. Regression analysis between two variables was performed when appropriate. Data are expressed as mean±SEM. Results were considered to be statistically significant at P<.05.
In addition, multivariate factorial analyses (two-way ANOVA, followed by the Bonferroni test) according to age (young, <60; old, >60 years), sex, and the presence or absence of LV hypertrophy (LV mass, >250 g) were performed. Results were considered to be statistically significant at P<.05.
Human PF or serum (5% dilution) increased the rate of protein synthesis in adult cardiomyocytes after 3 days in culture (PF, +71.9±8.2% [n=17]; serum, +14.9±6.5% [n=13]; both P<.01 versus control medium). When the effect of PF was compared with that of serum from the same patient, the increase in the rate of protein synthesis was consistently greater with PF compared with serum (P<.01). In addition, the trophic activity of both fluids in vitro was positively related to the patient’s left ventricular mass (PF, r=.817 [P<.0001]; serum, r=.671 [P<.01]) (Fig 1⇓). Interestingly, this effect was not specifically related to LV hypertrophy, since PF and serum from one patient with tetralogy of Fallot, with exclusive RV hypertrophy, also produced a marked trophic effect on myocyte growth (+88% and +68%, respectively).
Table 1⇓ shows that the addition of PF and serum significantly increased the total protein content per cell (P<.01) but did not affect cell viability, as indicated by the similar cell density and percentage of rod-shaped myocytes under the different experimental conditions.
The cardiac β-MyHC is predominantly expressed in cardiac ventricular myocytes throughout embryonic and fetal development in the rat, whereas α-MyHC constitutes the main isoform in adulthood.33 Some pathological conditions, such as cardiac hypertrophy, are associated with reversal of this β-MyHC–to–α-MyHC ratio.27 When we analyzed the effect of PF and serum on MyHC mRNA level, we observed that control cells expressed predominantly α-MyHC, as expected,28 and that the addition of PF to the culture medium induced a rise of both α- and β-MyHC mRNAs in adult cardiac myocytes, whereas serum moderately increased β-MyHC only (Fig 2⇓), regardless of ventricular mass (data not shown). Since MyHC is one of the major proteins present in cardiomyocytes, the increase in α- and β-MyHC mRNAs strongly suggests that the increase in total protein synthesis is a pretranslational event.
To investigate whether the composition of PF and serum influenced the difference between the cell responses to PF and serum, we analyzed the ionic and protein concentration and the growth factor content in both fluids (Table 2⇓). As previously described in mammals,18 human PF is an ultrafiltrate of plasma, with higher [K+] and lower glucose, osmolarity, and protein content compared with serum. The concentration of peptide growth factors (FGF2 and TGF-β) in PF and in serum differed in PF, serum, and atrial biopsies: FGF2 concentration was high in atrial biopsies and PF and low in serum, whereas the highest levels of TGF-β were found in serum. It has been shown that TGF-β is in a latent form in biological fluids.29,30 It is noteworthy that in the present study, TGF-β was undetectable in serum and PF before in vitro activation by acidification; this indicates that the high concentration of TGF-β measured likely reflects the predominance of a latent form. The high concentration of FGF2 in myocardial biopsies suggests a cardiac origin, which is supported by the higher levels in PF than in serum. When a known amount of exogenous FGF2 was added to either PF or serum (166 and 664 pg/mL, respectively) to assess the sensitivity of the ELISA, the observed FGF2 values were 167±3 pg/mL in serum and 1314±246 pg/mL in PF versus theoretical values of 170±2 and 1247±276 pg/mL, respectively (n=6, P=.752 and 0.1). On the other hand, because the patient group was highly variable in terms of age, sex, and pathology, we verified that FGF2 was not dependent on these variables. The results of multivariate factorial analysis according to age, sex, and the presence or absence of LV hypertrophy are shown in Table 3⇓.
In conjunction, the data presented in Tables 2⇑ and 3⇑ suggest that the greater FGF2 levels in PF may be responsible for its greater trophic effect. To validate this hypothesis, we verified that exogenous FGF2 increases the rate of protein synthesis in a dose-dependent manner (Fig 3⇓). Ultimately, the implication of FGF2 in the trophic effect of PF was assessed by the addition of specific anti-human FGF2 antibodies to the myocyte culture medium in the presence of PF. IgG anti-human FGF2 reduced the trophic effect of PF to levels similar to those observed with serum (Fig 4A⇓), in a specific manner (Fig 4B⇓), whereas an equivalent concentration of nonimmune goat IgG had no effect alone (data not shown) or on the PF-induced increase protein synthesis in cultured adult rat cardiac myocytes (Fig 4B⇓). Furthermore, PF-derived eluate from heparin-Sepharose beads reacted with specific anti-FGF2 polyclonal antibodies (Fig 4D⇓) and exhibited only one band at 18 kD, whereas no signal was detected for serum extract, indicating that FGF2 was below the detection threshold. Since human fibroblast lysate expressed a family of four peptides encoded by the FGF2 gene (Fig 4D⇓), we might hypothesize that human cardiac cells, including fibroblasts, are able to synthesize the four peptides but that the 18-kD isoform is preferentially released into PF. Fig 4C⇓ shows that FGF2 concentration was independent of LV mass, providing evidence of an additional LV mass–dependent growth factor (Fig 4A⇓), which is different from FGF2 and present in both fluids.
Active TGF-β is a known inhibitor of the trophic properties of many cytokines, including FGF2. To assess the potential role of TGF-β in vivo, exogenous active TGF-β (1 ng/mL)34 was added to the culture medium with PF or serum. This strongly inhibited the trophic effects of both fluids on isolated rat adult myocytes (Fig 5⇓). No effects were observed when 1 ng/mL active TGF-β was added alone to the culture medium (Fig 3⇑). Furthermore, 1 ng/mL TGF-β totally inhibited the trophic response of isolated adult cardiac myocytes to the highest dose of exogenous FGF2 (Fig 3⇑), consistent with previous observations.35,36
Thus, the prominent trophic effect of PF is possibly due to the higher concentration of FGF2. However, the LV mass–dependent effect observed in both PF and serum (Fig 1⇑) appears to be regulated via other mechanisms, since no correlation was found between LV mass and the levels of FGF2 (Fig 4C⇑) and TGF-β (data not shown) in serum and PF. To further characterize the LV mass–dependent effect shown in Fig 1⇑, we measured the PF and serum concentration of ANF, angiotensin II, and ET-1, all of which modulate hypertrophy in vitro. ANF was higher in PF (Table 2⇑), and its levels increased with LV mass (data not shown), as previously observed,37 whereas angiotensin II levels were higher in serum than in PF (Table 2⇑) but did not vary significantly, depending on LV mass (data not shown). Finally, ET-1 concentrations were similar in the two fluids and unrelated to LV mass.
To investigate the potential direct role of ANF and angiotensin II in the LV mass–dependent effects on protein synthesis (Fig 1⇑), the effects of exogenous angiotensin II (10−10 to 10−5 mol/L) and ANF (10−9 to 10−6 mol/L) were tested on isolated cardiac myocytes (Fig 6⇓). Only a moderate increase in the rate of protein synthesis was observed with angiotensin II at micromolar concentrations (Fig 6A⇓), whereas ANF significantly decreased the rate of protein synthesis in a dose-dependent manner (Fig 6B⇓).
The present study demonstrates that (1) human PF has a greater trophic effect than human serum on cultured rat adult cardiac myocytes, which is likely due to a high concentration of FGF2, and (2) both serum and PF from patients with cardiac hypertrophy have additional effects on cardiac myocytes that are related to the increase in LV mass and that suggest the presence of a circulating growth factor(s) involved in the process of hypertrophy.
One of the major findings of the present study is that PF has a hypertrophic effect on cardiac myocytes, as indicated by (1) the increase of MyHC mRNA level, (2) increased rate of protein synthesis, and (3) increase in total protein content. This trophic effect is not associated with a shift in myosin isoforms, since both α- and β-MyHC mRNAs increase. Furthermore, PF does not seem to enhance apoptosis, since the percentage of cells dying in culture was the same under all experimental conditions (Table 1⇑).
The greater trophic effect of PF on cardiac myocytes compared with that of serum is likely due to the respective composition and growth factor content (Table 2⇑). Furthermore, the high FGF2 concentration in PF (the present study and Reference 1919 ) raises the question of its origin. FGF2 is found in the heart in intracellular and extracellular locations,9,38 being both synthesized and released by most cardiac constitutive cells.14,22 Growth factors such as FGF1 and FGF2 (18 to 25 kD) are released through the epicardium from ex vivo beating rat hearts and are found in pericardial superfusate.20 Our data strongly indicate a decreasing gradient in FGF2 concentration from human myocardial tissue to serum, suggesting that the myocardium is the likely major source of FGF2 in PF (Table 2⇑). On the other hand, the greater concentration of FGF2 in PF is also associated with a lower osmolarity and protein concentration. This supports the evidence that FGF2 is not actively concentrated from serum but is most probably released into the pericardial space from the cardiac interstitium. PF is classically considered to be an ultrafiltrate of plasma,18 but it may also reflect the composition of cardiac interstitium in cardiac disease, as recently proposed by Fujita et al19 in patients with unstable angina. PF may contain molecules up to 40 kD, including growth factors such as FGF2, that can migrate from the myocardium through the epicardium into the pericardial space.19,20,21 The hypothesis that PF reflects peptide release within cardiac interstitium is also supported by the fact that ANF (molecular mass, 3 kD), which is produced and released by the heart, is more concentrated in PF than in serum (Table 2⇑).37 Therefore, PF is not a mere ultrafiltrate of plasma but a fluid substantially enriched in components of myocardial origin that may resemble the composition of human cardiac interstitium.
FGF2 was proposed to be one of the signals triggering the qualitative and quantitative phenotypic changes of cardiac myocytes.39 In adult cardiac myocytes, it increases total protein synthesis without inducing a hypertrophic phenotype.36 Our data provide strong evidence that FGF2 mediates the trophic effect of PF, since (1) it is present in significant amount (Fig 4D⇑, Table 2⇑), (2) its effect is inhibited by antibodies specifically directed against human FGF2, and (3) its action is mimicked by exogenous FGF2. The high concentration of FGF2 in the PF of every patient studied, regardless of age, sex, underlying cardiac disease, or degree of ventricular hypertrophy, suggests that FGF2 has a role in controlling human myocyte protein synthesis and possibly participates in the regulation of cardiac mass under normal conditions.
When one compares the effects of endogenous and exogenous FGF2, it should be noted that the actual concentration of FGF2 in PF is 2 to 3 orders of magnitude lower than that required to produce a trophic effect of comparable amplitude in rat cardiac myocytes using exogenous FGF2. This raises an apparent discrepancy between the effects of endogenous molecules and purified exogenous peptides. Indeed, in vivo, the biological activity of factors with a paracrine action is strictly dependent on the local environment, and the final trophic response may be additive or synergistic compared with the intrinsic effect of each molecule.40,41
TGF-β has been reported to counteract the effect of other growth factors in different cell types, including cardiac myocytes11 and endothelial42 and smooth muscle cells.43 TGF-β and FGF2 have been found in cardiac myocytes and cardiac interstitium after increased mechanical stress, suggesting that they could mediate the response to mechanical load and be involved in myocardial remodeling secondary to hypertrophic stimuli.7,8,11 In addition, it has been reported that TGF-β opposes the action of FGF2 in the endothelium, smooth muscle cells, and cardiac myocytes.31,36,42 Recently, it has been documented that TGF-β is implicated in the negative control of skeletal muscle growth.44
It is shown in the present study that (1) TGF-β concentration is high in serum and low in PF, (2) exogenous active TGF-β alone has no intrinsic trophic action on adult cardiac myocytes but counteracts the trophic action of exogenous FGF2 (Fig 4⇑), and (3) the high concentration of endogenous TGF-β in serum does not blunt the LV mass–dependent effect of serum, whereas exogenous TGF-β has a strong inhibitory effect on myocyte growth (Fig 5⇑). The discrepancy between exogenous and endogenous TGF-β is only apparent, not actual, and reflects the fact that exogenous and endogenous TGF-β differ in their biological activities, the first being a pure active form, with the latter being a latent form requiring local activation.45
The other important finding of the present study is that both serum and PF have a LV mass–dependent trophic effect on cardiac myocytes (Fig 1⇑). This effect is not correlated with FGF2 and TGF-β concentrations and is unaltered by the addition of anti-human FGF2 antibodies to PF or by the action of the other known growth factors examined.
Factors such as ET-1 and angiotensin II have been suggested to act either directly or as cofactors/coeffectors in paracrine/autocrine signaling in cardiac hypertrophy in vitro.46-48 However, in the present study they seem unlikely to be directly responsible for the LV mass–dependent effect on cardiomyocyte growth. In fact, ET-1 levels did not vary, and angiotensin II levels only increased in one of the two fluids, which does not support any evidence of a LV mass–dependent factor present in both fluids. Furthermore, the addition of high doses of angiotensin II alone to the cardiac myocyte culture medium had only a moderate effect on total protein synthesis (Fig 6⇑), as previously shown49 and confirmed by our data.
ANF is a marker of cardiac hypertrophy, and its synthesis is enhanced in vivo by mechanical overload37,50 and in vitro by stretch.51 To our knowledge, its direct role on adult cardiac myocyte growth has not been investigated. It is shown here that ANF had no hypertrophic effect on myocytes per se. In fact, exogenous human recombinant ANF decreased the rate of protein synthesis in cardiac myocytes in a dose-dependent manner (Fig 6⇑). These results are consistent with previous findings in smooth muscle cells.52
Taken together, these data support the evidence of humoral substances in both PF and serum (different from ET-1, angiotensin II, and ANF) that are involved in the process of ventricular hypertrophy. These findings also confirm and extend the work of Yu et al,17 which showed an increase in the rate of protein synthesis in rat adult cardiac myocytes exposed to serum obtained from patients with chronic renal failure.
In summary, our data demonstrate that growth factors (mainly FGF2) with a prominent trophic effect on adult cardiac myocyte are normally present in human PF. We further demonstrate that PF and serum contain additional, stable, diffusible growth factors in patients with RV or LV hypertrophy. These factors, whose nature requires further investigation, are also present in the bloodstream and appear to be dependent on LV mass. They may act on cardiac myocyte growth during the development of ventricular hypertrophy and could also be implicated as humoral triggers in the signaling processes leading to ventricular hypertrophy.
Selected Abbreviations and Acronyms
|ANF||=||atrial natriuretic factor|
|dpd||=||dark pixel density|
|FGF1||=||acidic fibroblast growth factor|
|FGF2||=||basic fibroblast growth factor|
|LV, RV||=||left and right ventricle (ventricular)|
|MyHC||=||myosin heavy chain|
|TGF-β||=||transforming growth factor-β|
This study was supported by Institut National de la Santé et de la Recherche Médicale, CNRS, and Fondation de France; by DSPT 5 Ministère Français de l’Education Nationale, de l’Enseignement Superieur et de la Recherche; by DRC-Assistance Publique-Hôpitaux de Paris; and by Université de Paris 7 Denis Diderot. Dr Stefano Corda is a recipient of a postdoctoral fellowship of the European Commission (Human Capital and Mobility Program, ERB CH BGCT 930373). The authors wish to thank the staff of cardiac anesthesiologists and cardiac surgeons at Hôpital Lariboisière. The outstanding assistance of Drs Geneviève Maistre and Alain Carayon and the valuable advice of Dr Saul Winegrad are gratefully acknowledged. The authors are very grateful to Dr Bernard Prendergast for the revision of the manuscript.
- Received April 7, 1997.
- Accepted August 28, 1997.
- © 1997 American Heart Association, Inc.
Brutsaert DL, Andries LJ. The endocardial endothelium. Am J Physiol. 1992;263:H985-H1002.
Ramaciotti C, Sharkey A, McClellan G, Winegrad S. Endothelial cells regulate cardiac contractility. Proc Nat Acad Sci U S A. 1992;89:4033-4036.
Mebazaa A, Mayoux E, Maeda K, Martin LD, Lakatta EG, Robotham JR, Shah AM. Paracrine effects of endocardial endothelial cells on myocyte contraction mediated via endothelin. Am J Physiol. 1993;265:H1841-H1846.
Shah AM, Mebazaa A., Wetzel RC, Lakatta EG. A novel cardiac myofilament desensitizing factor released by endocardial and vascular endothelial cells. Circulation. 1994;89:2492-2497.
Schneider MD, Parker TG. Cardiac myocytes as targets for the action of peptide growth factors. Circulation. 1990;81:1443-1456.
Pasumarthi KBS, Kardami E, Cattini PA. High and low molecular weight fibroblast growth factor-2 increase proliferation of neonatal rat cardiac myocytes but have differential effects on binucleation and nuclear morphology: evidence for both paracrine and intracrine actions of fibroblast growth factor-2. Circ Res. 1996;78:126-136.
Takahashi N, Calderone A, Izzo NJ, Maki T, Marsh J, Colucci WS. Hypertrophic stimuli induce transforming growth factor-β1 expression in rat ventricular myocytes. J Clin Invest. 1994;94:1470-1476.
Speir E, Tanner V, Gonzalez AM, Farris J, Baird A, Casscells W. Acidic and basic fibroblast growth factors in adult rat heart myocytes: localization, regulation in culture, and effects on DNA synthesis. Circ Res. 1992;71:251-259.
Weiner HL, Swain JL. Acidic fibroblast growth factor mRNA is expressed by cardiac myocytes in culture and the protein is localized to the extracellular matrix. Proc Natl Acad Sci U S A. 1989;86:2683-2687.
Roberts AB, Roche NS, Winokur TS, Burmester JK, Sporn MB. Role of transforming growth factor-β in maintenance of function of cultured neonatal cardiac myocytes. J Clin Invest. 1992;90:2056-2062.
Kardami E, Fandrich R. Basic fibroblast growth factor in atria and ventricles of vertebrates hearts. J Cell Biol. 1989;109:1865-1875.
Liu L, Pasumarthi KBS, Padua R, Massaeli H, Fandrich R, Pierce GN, Cattini PA, Kardami E. Adult cardiomyocytes express functional high affinity receptors for basic fibroblast growth factor. Am J Physiol. 1995;37:H1927-H1938.
Casscells W, Speir E, Sasse J, Klagsbrun M, Allen P, Lee M, Calvo B, Chiba M, Haggroth L, Folkman J, Epstein SE. Isolation, characterization and localization of heparin-binding growth factors in the heart. J Clin Invest. 1990;85:433-441.
Hammond GL, Wieber E, Markert CL. Molecular signals initiating protein synthesis in organ hypertrophy. Proc Natl Acad Sci U S A.. 1979;76:2455-2459.
Long CS, Kariya K, Karns L, Simpson PC. Trophic factors for cardiac myocytes. J Hypertens. 1990;8:S219-S224.
Fujita M, Ikemoto M, Kishishita M, Otani H, Nohara R, Tanaka T, Tamaki S, Yamazato A, Sasayama S. Elevated basic fibroblast growth factor in pericardial fluid of patients with unstable angina. Circulation. 1996;94:610-613.
Clarke MSF, Caldwell RW, Chiao H, Miyake K, McNeil P. Contraction-induced cell wounding and release of fibroblast growth factor in heart. Circ Res. 1995;76:927-934.
Page E, Upshaw-Earley J, Goings G. Permeability of rat atrial endocardium, epicardium, and myocardium to large molecules: stretch-dependent effects. Circ Res. 1992;71:159-173.
Kaye D, Pimental D, Prasay S, Maki T, Berger HJ, McNeil PL, Smith TW, Kelly RA. Role of transiently altered sarcolemmal membrane permeability and basic fibroblast growth factor release in the hypertrophic response of adult rat ventricular myocytes to increased mechanical activity in vitro. J Clin Invest. 1996;97:281-291.
Dubus I, Samuel J-L, Marotte F, Delcayre C, Rappaport L. β-Adrenergic agonists stimulate the synthesis of noncontractile but not contractile proteins in cultured myocytes isolated from adult rat heart. Circ Res. 1990;66:867-874.
Schiaffino S, Samuel JL, Sassoon D, Lompré AM, Garner I, Marotte F, Buckingham M, Rappaport L, Schwartz K. Nonsynchronous accumulation of α-skeletal actin and β-myosin heavy chain mRNAs during early stages of pressure-overload-induced cardiac hypertrophy demonstrated by in situ hybridization. Circ Res. 1989;64:937-948.
Dubus I, Mercadier A, Lucas O, Contard F, Nallet O, Oliviero P, Rappaport L, Samuel JL. α-, β-MHC mRNA quantification in adult cardiomyocytes by in situ hybridization: effect of thyroid hormone. Am J Physiol. 1993;34:C62-C71.
Wakefield LM, Smith DM, Flanders KC, Sporn MB. Latent transforming growth factor-β from human platelets. J Biol Chem. 1988;263:7646-7654.
Roberts A, Sporn M. The transforming growth factors-βs. In: Sporn M, Roberts A, eds. Peptide Growth Factors and Their Receptors. New York, NY: Springer-Verlag; 1991:419-472.
Nadal-Ginard B, Mahdavi V. Molecular basis of cardiac performance. J Clin Invest. 1989;84:1693-1700.
Parker T, Packer S, Schneider M. Peptide growth factors can provoke ‘fetal’ contractile protein gene expression in rat cardiac myocyte. J Clin Invest. 1990;85:507-514.
Samuel JL, Dubus F, Farhadian F, Marotte F, Oliviero P, Mercadier A, Contard F, Barrieux A, Rappaport L. Multifactorial regulation of cardiac gene expression: an in vivo and in vitro analysis. Ann N Y Acad Sci. 1995;752:371-386.
Schneider MD, Parker TG. Molecular mechanism of cardiac growth and hypertrophy: myocardial growth factors and protooncogenes in development and disease. In: Roberts R, ed. Molecular Basis of Cardiology. New York, NY: Blackwell Scientific Publications; 1993:113-134.
Frelin C, Guedin D. Why are circulating concentrations of endothelin so low? Cardiovasc Res. 1994;28:1613-1622.
Bogoyevitch MA, Glennon PE, Andersson MB, Clerk A, Lazou A, Marshall C, Parker P, Sugden PH. Endothelin-1 and fibroblast growth factor stimulate the mitogen activated protein kinase signalling cascade in cardiac myocytes. J Biol Chem. 1994;269: 1110-1119.
Lyons RM, Jorma KO, Moses HL. Proteolytic activation of latent transforming growth factor-beta from fibroblast-conditioned medium. J Cell Biol. 1989;106:1659-1665.
Ito H, Hirata Y, Adachi S, Tanaka M, Tsujino M, Koike A, Nogami A, Marumo F, Hiroe M. Endothelin-1 is an autocrine/paracrine factor in the mechanism of angiotensin II-induced hypertrophy in cultured rat myocytes. J Clin Invest. 1993;92:398-403.
Fishel RS, Thourani V, Eisenberg S, Shai SY, Corson MA, Nabel EG, Bernstein KE, Berk BC. Fibroblast growth factor stimulates angiotensin converting enzyme expression in vascular smooth muscle cells: possible mediator of the response to vascular injury. J Clin Invest. 1995;95:377-387.
Booz G, Baker K. Role of type 1 and type 2 angiotensin receptor in angiotensin II-induced cardiomyocyte hypertrophy. Hypertension. 1996;28:635-640.
Mercadier JJ, Samuel JL, Michel JB, Zongazo M, De la Bastie A, Lompré AM, Wisnewsky C, Rappaport L, Levy B, Schwartz K. Atrial natriuretic factor gene expression in rat ventricle during experimental hypertension. Am J Physiol. 1989;257:H979-H987.
Itoh H, Pratt R, Dzau V. Atrial natriuretic peptides inhibits hypertrophy of vascular smooth muscle cells. J Clin invest.. 1990;86:1690-1697.