|Home | About | Journals | Submit | Contact Us | Français|
Placenta growth factor (PlGF) plays an important role in pathological angiogenesis and is thought to be an independent biomarker in patients with coronary artery disease. However, little is known regarding the regulation of PlGF expression in heart tissue.
We determined expression changes in PlGF and its receptor, VEGFR1, in normal and abnormal biopsies from human cardiac allografts and in cardiomyocytes cultured under hypoxia or cyclical stretch conditions.
Human donor myocardium and biopsies from allografts without fibrin deposits expressed PlGF and VEGFR1 mRNA. Biopsies (n = 7) with myocardial fibrin, elevated serum cardiac troponin I titers (p < 0.03) and cellular infiltrates (p < 0.05), expressed 1.6-fold more PlGF mRNA than biopsies from allografts without fibrin (n=11; p < 0.05). PlGF protein was localized in cardiomyocytes, extracellular matrix and some microvessels in areas with fibrin deposition. VEGFR1 mRNA expression was not different between groups. Cultured neonatal rat cardiomyocytes constitutively expressed PlGF/VEGFR1 under normoxia. PlGF expression was increased 3.88 ± 0.62 fold after 12 hours (n = 6; p ≤ 0.05) and 3.64 ± 0.41 fold after 24 hours of hypoxia (n = 6; p ≤ 0.05). Shorter periods of hypoxia, conditioned media from hypoxic cells and cyclical stretch did not significantly alter PlGF or VEGFR1 expression.
Cardiomyocyte PIGF expression is upregulated by hypoxia in vitro and its expression increases significantly in allografts with myocardial damage. Collectively, these results provide important temporal and spatial evidence that endogenous PlGF could facilitate cardiac healing following myocardial hypoxia/ischemia.
A competent collateral circulation limits mortality/morbidity following myocardial infarction and members of the vascular endothelial growth factor (VEGF) family play an important role in the growth and maintenance of blood vessels. One VEGF family member that has received relatively little attention is placenta growth factor (PlGF). Although others have described the usefulness of plasma PlGF as an independent biomarker of favorable or unfavorable[5-7] outcomes in patients with acute coronary syndromes, little is known regarding its regulation in heart tissue.
The association between angiogenesis and increased PlGF expression during proliferative retinopathy, cutaneous wound healing[9, 10], bone fracture repair and tumor growth[9, 12] suggests PlGF can regulate vascularity during pathological conditions. PlGF binds VEGFR1 and neuropilin-1 (NP-1) receptors and amplifies VEGF-driven myocardial angiogenesis during ischemia in mice. Therefore, endogenous PlGF could play an important role in vascular growth/maintenance during myocardial ischemia.
Hypoxia is a potent stimulus for myocardial angiogenesis. However, hypoxia decreases PlGF expression in some cells [15, 16] but increases PlGF expression in others[17, 18]. No studies have directly assessed PlGF gene expression in ischemic human myocardium. Furthermore, cardiomyocyte stretch in vitro [19, 20] and in vivo [21, 22] increases VEGF expression and induces coronary angiogenesis independently of hypoxia. The effect of cyclical stretch on PlGF expression in cardiomyocytes is not also known. Clearly, there is a fundamental lack of information regarding cardiomyocyte PlGF expression. Our results are the first to show that 1) human myocardium expresses PlGF mRNA and protein, 2) its expression increases with cellular and biochemical indices of myocyte damage, and 3) that hypoxia, but not mechanical stretch, upregulates PlGF expression in isolated cardiomyocytes. Collectively, these results provide important temporal and spatial evidence that endogenous PlGF could facilitate cardiac healing following myocardial hypoxia/ischemia.
Right endomyocardial biopsy samples were obtained from donor hearts (n = 2) and cardiac allograft recipients (n = 16), embedded in medium, and snap frozen in liquid nitrogen. All biopsies after transplantation were for routine rejection surveillance and blood samples were obtained prior to biopsy to assess serum cardiac troponin-I titers as described[24, 25]. Additional paraffin embedded tissue was submitted for light microscopy assessment of allograft rejection according to the International Society for Heart and Lung Transplantation revised grades. Days between transplantation and biopsy were determined for each sample.
Immunohistochemical assessment of the presence of myocardial fibrin, vascular antithrombin and cellular infiltrates were performed on frozen sections as described[24, 25, 27, 28]. For immunoperoxidase-based detection of PlGF in biopsies with (n = 7) and without (n = 7) fibrin deposition, 5um sections obtained from paraffin blocks were deparaffinized, hydrated, rinsed and antigen-retrieved using DAKO Target Retrieval solution (pH 6.0). Endogenous biotin was blocked using an avidin/biotin blocking system (DAKO, Carpinteria, CA). Endogenous peroxidase activity was blocked with 3% hydrogen peroxide. Rabbit anti-human PLGF antibody (ab9542, Abcam, Cambridge, MA) was applied for 60 minutes at room temperature, washed, and developed using DAKO’s EnVision+ Dual Link HRP kit. PlGF localization was evaluated by 2 investigators unaware of the biopsy fibrin characteristics.
Total RNA was isolated from frozen blocks, 500 ng was reverse transcribed as described, and cDNA was used for analysis of PlGF, VEGFR1, VEGFR2, and NP-1 mRNA expression by RT-PCR. We designed primers to amplify 318 bp of the 5′ region of human PlGF contained in all known splice variants (Table 1). Validity of the amplicon was confirmed by cloning into pCRII-TOPO vector and bi-directional sequencing of the insert. We also utilized primers to amplify 239 bp of human VEGFR1/flt-1 [15, 30], 204 bp of human VEGFR2/KDR [15, 31], and 210 bp of human NP-1  (Table 1). Appropriate PCR conditions were determined  to ensure linear amplification of cDNA target over a range of PCR cycles. Primers and competimers for 18S ribosomal RNA (Ambion Inc, Austin, TX) were used to determine relative differences in gene expression . Tissue samples were grouped according to the immunohistochemical presence or absence of fibrin and relative changes in mRNA expression and protein, cellular infiltrates, plasma Tn-I, and ISHLT rejection grades were compared between groups.
Ventricular cardiomyocytes were obtained from neonatal rats. The Institutional Animal Care and Use Committee approved all animal procedures in accordance with guidelines. Culture media (DMEM, 10% FCS, 1mM sodium pyruvate, 100 IU penicillin, 100 μg/ml streptomycin and 0.25 μg/ml amphotericin B) was changed every 24 hours. Approximately 90% of the cells began to beat spontaneously and cultures were used within 2-3 days of plating. Representative cell culture purity was assessed using monoclonal antibody to sarcomeric myosin (MF-20, University of Iowa Hybridoma Center) and rhodamine labeled goat anti-mouse IgG (1:100) with DAPI (1 μg/ml; Molecular Probes, Eugene, OR) counterstain. Cultures with >85% MF-20 positive or beating cells were used in the studies.
We used a stretch protocol known to significantly increase VEGF expression in isolated cardiomyocytes. Cardiomyocytes were plated on collagen I Bioflex plates and subjected to a 10% average surface elongation at 30 cycles/min for 1, 3, or 6 hours using a computerized Flexercell Strain Unit (Flexcell International, Hillsborough, NC). Cultures not subjected to cyclical stretch served as controls. At the end of each time period, total RNA was extracted and stored at −80°C until used.
Parallel hypoxic (< 2% O2) and control (21% O2) cardiomyocyte cultures were incubated for 1, 3, 6, 12, or 24 hours. In addition, conditioned media from cardiomyocytes incubated under hypoxic or normoxic conditions for 12 hours was collected, centrifuged, and added to fresh cardiomyocytes cultured under normoxia for 12 additional hours. Total RNA was isolated at the end of each time point and stored at −80°C.
Total RNA was reverse transcribed as we have described. Two microliters of cDNA was combined with 50 pmol of primer in the presence of PCR beads (Amersham Pharmacia Biotech) and subjected to PCR amplification for 1 minute each at 95°C, 61°C, and 72°C. We designed primers that amplified 224bp of rat PlGF based on its published sequence (Table 1). The single PCR product was purified and sequence analysis confirmed 99% homology to the published sequence of rat PlGF. Primers were also used to amplify 600bp of rat VEGFR-1(Table 1). Multiplex PCR was performed using primers and competimers for 18S ribosomal RNA (Ambion Inc, Austin, TX) as we have described. PCR products were separated by gel electrophoresis and PlGF/18S and VEGFR1/18S ratios were compared between culture conditions. No amplicon signals were produced from RNA samples not subjected to reverse transcription or from samples without cDNA template added.
Significant increases in PlGF expression by end-point PCR were confirmed by real-time PCR. Gel electrophoresis and melting point analysis confirmed single products of the correct size. Primers for ribosomal protein L32 were used as an endogenous standard and relative changes in PlGF expression were calculated using:
where E = efficiency of the PCR reaction for a primer set, CT = threshold cycle for each product under each culture condition.
Immunohistochemistry was used to localize PlGF and VEGFR1 protein in mixed cultures of cardiomyocytes and non-cardiomyocytes. Cells were fixed with 10% buffered formalin for 30 minutes at 4°C. For PlGF, cells were incubated overnight at 4°C with a 10 μg/ml mixture (1:1) of affinity-purified goat polyclonal antibodies against the C-terminus and an internal sequence of mouse PlGF (Santa Cruz Biotechnology, Santa Cruz, CA) in PBS containing 0.15% Triton X-100. For VEGFR1 cells were reacted with 4 μg/ml goat polyclonal antibody against VEGFR1 (Santa Cruz Biotechnology). Cells were washed and incubated with 10 μg/ml fluorescein isothyiocyanate (FITC) conjugated donkey F(ab’)2 anti-goat IgG and counterstained with 1ug/ml DAPI. To confirm cardiomyocyte localization, some cultures were sequentially reacted with monoclonal antibody (MF-20) against sarcomeric myosin primary antibody and 10 μg/ml of rhodamine (RITC) conjugated goat F (ab’)2 anti-mouse IgG. Specificity of PlGF/VEGFR1 immunoreactivity was confirmed by incubating cells with primary antibody pre-absorbed overnight with 15-fold excess immunizing peptide.
All data are presented as mean ± SEM and all assays were performed in duplicate. Kruskal-Wallis analysis for multiple groups (cell culture data) or the Mann-Whitney U-test for two groups (biopsy data) was used to determine significant differences in immunohistochemical scores as well as fold changes in gene expression. Statistical differences in cardiac troponin I values were assessed by Wilcoxon’s rank sum test. Differences in the proportion of biopsies with ISHLT rejection grade ≥R1 or with positive plasma Tn-I titers between groups were assessed by Fisher’s exact test. All differences were considered statistically significant if P < 0.05.
As we have shown, fibrin deposition in cardiac allografts occurs early after transplantation and biopsies with fibrin were characterized by a significantly shorter period between transplantation and biopsy compared to biopsies without fibrin (Table 2). Biopsies with fibrin were associated with increased macrophage and neutrophil infiltrates; however, there was no significant difference in T-cell infiltrates or proportion of biopsies with ISHLT rejection grade ≥ 1R between groups (Table 2). Serum cardiac troponin-I titers were significantly greater in patients with fibrin in their biopsies (Table 2). Thus, biopsies with fibrin demonstrated immunohistochemical and biochemical indices of myocyte damage in the absence of cellular rejection.
Biopsies from donor hearts and transplanted hearts without myocardial fibrin expressed less PlGF mRNA compared to biopsies from transplanted hearts with fibrin (Figure 1). Fibrin deposition was associated with 1.6-fold increase in PlGF expression (no fibrin = 0.298 ± 0.05 relative units, positive fibrin = 0.472 ± 0.045 relative units; p < 0.05). By immunohistochemistry, biopsies with myocardial fibrin also demonstrated PlGF cytoplasmic immunoreactivity in cardiomyocytes and more intense myocyte plasma membrane and interstitial immunoreactivity. Microvessels in/near areas of fibrin deposition also demonstrated PlGF immunoreactivity (Figure 2A). In contrast, areas within these biopsies that were distant from sites of fibrin deposition did not demonstrate PlGF immunoreactivity. Similarly, biopsies without fibrin demonstrated little/no PlGF immunoreactivity (Figure 2B).
There was no significant association between PlGF mRNA expression and degree of macrophage, T-cell, or neutrophil infiltrates or ISHLT rejection grade in the biopsies. However, PlGF expression decreased with increased time since transplantation (R = −0.792; p < 0.001) and its expression was 1.6-fold greater (0.428 ± 0.06 relative units) in biopsies with reduced vascular antithrombin immunoreactivity compared to biopsies with unique capillary antithrombin (0.270 ± 0.04 relative units; p < 0.05). We have shown development of capillary antithrombin binding following fibrin deposition is associated with improved allograft outcome. There was no significant change in VEGFR1, VEGFR2 or NP-1 expression between groups of biopsies (Figure 3). Collectively, these results demonstrate spatial and temporal relationships between increased PlGF expression in biopsies with immunohistochemical and biochemical evidence of myocardial damage. The factor(s) regulating cardiomyocyte PlGF expression in vivo are not known but two potential mechanisms include hypoxic stress or physical stretch.
We assessed expression of PlGF and VEGFR1 in cardiomyocytes that had undergone cyclical stretch for 1, 3, or 6 hours. Although we have shown that this degree/duration of stretch increases VEGF expression in cardiomyocytes, it did not significantly alter PlGF or VEGFR1 mRNA expression (Figure 4).
Cardiomyocytes cultured in normoxia showed stable PlGF mRNA expression over 24 hours. Although 1, 3, or 6 hours of hypoxia did not significantly alter PlGF mRNA expression compared to time-matched controls; PlGF expression was significantly increased following 12 and 24 hours of hypoxia (Figure 5). By real-time RT-PCR, twelve hours of hypoxia produced a 3.88 ± 0.62 fold increase (n = 6; p<0.05) while 24 hours produced a 3.64 ± 0.41 fold increase (n = 6; p<0.05) in PlGF expression. Conditioned media from 12 hour hypoxic cultures did not alter PlGF expression (1.21 ± 0.40 of control; n = 5) compared to conditioned media from normoxic cultures. Hypoxia did not significantly alter VEGFR1 expression in cardiomyocytes (Figure 6).
Cultured cells demonstrated diffuse, cytoplasmic PlGF immunoreactivity that was consistent in both cardiomyocytes and noncardiomyocytes (Figure 7A-C). No immunoreactivity was detected when primary antibody, preabsorbed with excess PlGF antigen, was used (Figure 7D). Double immunofluorescence experiments also confirmed that cardiomyocytes demonstrated prominent VEGFR1 immunoreactivity (not shown).
PlGF plays a more prominent role in pathological rather than developmental cardiac angiogenesis[9, 36]. However, factors modulating endogenous PlGF gene expression in myocardial tissue are not known. Our results are the first to demonstrate 1) the presence and upregulation of PlGF expression in human myocardium with myocardial damage, 2) a time course for hypoxic upregulation of PlGF expression in cardiomyocytes, 3) a lack of effect of cyclical stretch on PlGF expression in cardiomyocytes, and 4) the presence of PlGF and receptor protein in isolated cardiomyocytes. Collectively, these results provide important temporal and spatial evidence of increased endogenous PlGF expression following myocardial ischemia.
There is a paucity of data on the expression and function of PlGF in human myocardium. Others have shown that plasma PlGF titers are significantly increased in patients with coronary occlusion and developing collaterals and patients with greater plasma PlGF titers showed greater improvement in left ventricular ejection fraction. However, elevated plasma PlGF titers in patients with acute coronary syndromes have also been associated with greater morbidity/mortality[5-7] suggesting that PlGF is a biomarker of adverse outcomes. Unfortunately, the cellular source of PlGF was not determined in these studies. Our findings show that myocardium with fibrin deposition, cellular infiltrates and elevated troponin I titers exhibited 1.6-fold increase in PlGF mRNA expression while longer-term allografts and those with favorable vascular antithrombin binding had reduced PlGF expression. Furthermore, PlGF protein localization was confined to cardiomyocytes, interstitium and blood vessels in and around areas of fibrin deposition. Additional studies are needed to determine whether plasma PlGF protein titers are useful as a noninvasive biomarker in cardiac transplant patients, similar to acute coronary artery syndrome patients[4-6].
These findings in human allografts, which are similar to those we have shown with VEGF expression[25, 29], do not provide a mechanism for the increased expression. Accordingly, we determined the independent effects of cyclical stretch and hypoxia on cardiomyocyte PlGF expression in vitro. We chose to concentrate our studies on isolated cardiomyocytes since they are the major source of VEGF in mouse and human hearts and non-myocyte cells cannot compensate for deficiencies in growth factor production.
Cardiomyocyte stretch in vitro  and in vivo [21, 22] produces a significant increase in VEGF expression that contributes to increased myocardial capillarity[21, 22]. However, PlGF mRNA expression does not change under similar stretch conditions. These results are consistent with findings that PlGF plays a more relevant role in pathological angiogenesis rather than physiological vascular growth associated with organ growth[9, 36].
Our results show that chronic hypoxia significantly increased cardiomyocyte PlGF expression while shorter periods did not alter its expression. Others have shown increased PlGF expression in hypoxic cardiomyocytes and immortalized fibroblasts after 24 hours but its temporal effect on PlGF expression was not known. Furthermore, these results are strikingly different from those in human trophoblast, microvascular endothelial cells, and human hepatoma cells (hepG2) where hypoxia reduces or does not alter PlGF expression. The mechanism for the delay in hypoxia-induced PlGF expression in cardiomyocytes is not known; however, these results correspond to clinical findings where a single brief episode of provoked cardiac ischemia in humans did not significantly increase plasma PlGF titers. Clearly, the effect of hypoxia on PlGF expression is cell type specific and temporally regulated and additional studies are needed to understand PlGF gene regulation in cardiomyocytes.
Conditioned media from hypoxic cardiomyocytes did not induce PlGF expression suggesting that hypoxia directly upregulates PlGF expression. The degree of hypoxia used in the present study (<2% O2) produces ~10-fold increase in HIF-1 binding activity and HIF-1alpha, in the absence of hypoxia, has been shown to upregulate PlGF expression in cardiomyocytes. PlGF promoter reporter studies in mouse fibroblasts have shown that hypoxic induction of PlGF is regulated, in part, by metal response element-binding transcription factor 1. Nuclear factor – kappa B elements may also function to increase the hypoxic responses of PlGF in some cell types. Thus, it remains to be determined whether hypoxic induction of PlGF expression in rat cardiomyocytes is wholly regulated by HIF-1alpha activity. Our data also confirm that isolated cardiomyocytes express VEGFR1 mRNA and protein as others have shown in situ. Although the physiological effect of PlGF-VEGFR1 interactions in cardiomyocytes is not known, it has been shown to protect other cell types from stress-induced apoptosis[9, 44].
In summary, our results show that cardiomyocytes express PlGF as well as its receptor and that PlGF expression is increased by hypoxia in vitro and by ischemia in vivo. Coupled with functional data from relevant animal models[9, 45], we hypothesize that increased cardiomyocyte PlGF expression may function to amplify the angiogenic response and/or provide a novel protection mechanism against cell death following myocardial hypoxia.
We thank Joanna Schwartz, Sarah Ronnebaum, Joshua Sandquist, Lisa Stalheim, Ben Colton, Ngoc Dang, Paul Tran and Lee Ann Baldridge for excellent technical assistance. Supported by R15-HL72802-01 (R Torry), RO1-HD36830 (D Torry) and RO1-HL075446 (R Tomanek).
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.