Endothelin-1 (ET-1), a vasoconstrictor and mitogen, has recently been implicated in the pathogenesis of human glioblastoma, neuroblastoma, and meningioma. ET-1, formed by proteolysis of the propeptide big ET-1 by endothelin-converting enzyme-1 (ECE-1), mediates its cellular actions through ETA and ETB receptors. Because only immunoreactive ET-1 has been observed within human astrocytic tumor cells, the authors investigated the localization of the entire ET-1 system (ET-1 mRNA, ET-1, ECE-1, ETA and ETB receptors) in surgical samples of human diffuse astrocytomas WHO Grade II (n = 6).
ET-1 mRNA expression was elucidated by in situ reverse transcriptase polymerase chain reaction (RT-PCR) using synthetic primers. Polyclonal antibodies were used to localize ET-1, ECE-1, ETA and ETB receptors by immunocytochemistry.
All ET components were detected in the six tumor samples. Intense (3+) cytoplasmic ET-1 mRNA labeling was observed in more than 75% of cells in all 6 astrocytomas. Up to 75% of tumor cells displayed intense ET-1 and ECE-1 immunolabeling distributed throughout their cytoplasm. Immunoreactive ETA and ETB receptors, observed in 25% to 75% of astrocytic tumor cells, were of moderate intensity. In addition, all components of the ET system were seen within endothelial cells of tumor blood vessels.
The endothelins (ETs) are a family of 21 amino acid peptides with 3 distinct isoforms (ET-1, ET-2, and ET-3).1 Human ET-1 mRNA, transcribed from the EDN1 gene located on chromosome 6, is translated to its inactive precursor, pre-pro–ET-1, which is cleaved by endopeptidases to produce big ET-1. Big ET-1 is then converted into ET-1 by endothelin-converting enzyme (ECE-1). In mammals, the cellular actions of ET-1 are mediated by two G protein-coupled receptors, ETA and ETB.2 In addition to its vasoconstrictor effect, ET-1 is responsible for the control of sodium and water handling in the kidney, stimulation of pituitary hormone release, and, possibly, neurotransmission.2, 3 ET-1 may play an antiapoptotic survival role in endothelial4 and smooth muscle cells.5 ET-1 has been shown to induce mitogenesis by stimulating [3H]-thymidine uptake by cultured rat astrocytes.6 Further, ET-1 has been confirmed, by Western blot analysis, to induce the activation of extracellular matrix (ECM) and basement membrane-degrading proteinases,7 suggesting that ET-1 has a role in tumor invasion and metastasis.
In humans, increased expression of ET-1 mRNA and ET-1 have been demonstrated by in situ hybridization (ISH) and immunocytochemistry (ICC) respectively, in lung squamous cell carcinomas and adenocarcinomas,6, 7 cancerous islets of Langerhans cells8 as well as aldosterone-secreting adenomas.9 Immunoreactive ET-1 has also been detected in human ovarian,10 colon,11 and prostate12 cancer cell lines. Elevated plasma levels of ET-1, observed in patients with colon,13 prostate,12 and ovarian cancer,14 suggests ET-1 is a marker of disease progression. ETA and ETB receptor mRNAs have been localized by reverse-transcriptase polymerase chain reaction (RT-PCR) in human aldosterone-secreting adenomas.9
Astrocytes are the major neuroglia in the central nervous system (CNS).15 Studies have detected the presence of ET components within astrocytes in culture,16–19 but not in situ.20–23 High-performance liquid chromatography (HPLC) and radioimmunoassays have shown that cultured rat astrocytes produce ET-1,19, 21 with ECE-1 activity also being observed within these cells.23 In binding studies, [125I]-ET-1 has been observed within astrocytic cell bodies and processes in cultured explants of the rat cerebellum, brainstem, and spinal cord.24
An astrocytoma is a tumor mass occurring as a result of abnormal and unregulated astrocyte growth. These infiltrative brain tumors account for 25–30% of all human gliomas and do not metastasize through the lymphatic and vascular systems.25 They span a wide range of neoplasms with distinct clinical, histopathologic, and genetic features. Northern blot analysis demonstrated ET-1 mRNA labeling in cultured glioblastoma cells,26 whereas immunoreactive ET-1 was seen within tumor cells in astrocytomas27, 28 and glioblastomas.29–31 Using ISH and ICC, Egidy et al.,29 localized ECE-1 mRNA and immunoreactive ECE-1, respectively, within glioblastoma tumor cells. ETA and ETB receptor mRNAs were observed by ISH and RT-PCR within tumor cells of human glioblastomas,26, 29, 31 meningiomas,30 and neuroblastomas.31 Further, [125I]-ET-1 binding sites have been seen within human meningioma32 and glioblastoma tumor cells.29, 33, 34
The entire ET system (ET-1 mRNA, ET-1, ECE-1, ETA and ETB receptors) has not been previously mapped in situ in human gliomas. Thus, the aim of the current study was to determine the localization of all of the above ET components in diffuse astrocytomas by in situ RT-PCR and ICC and, thereby, possibly elucidate their role(s) in this human cancer.
MATERIALS AND METHODS
Ethical permission for this study was granted by the Research Ethics Committee, Nelson R. Mandela School of Medicine, University of KwaZulu-Natal, South Africa.
Brain tumor tissue samples were collected with the cooperation of the attending neurosurgeon at Wentworth Hospital in Durban, South Africa, from 6 patients (2 males, 4 females; mean age 35 yrs), and fully informed consent from these patients and/or guardian(s) was obtained. The surgically excised samples (approximately 0.5 cm3) of diseased cerebral tissue were immediately immersed in 5% formal saline fixative (41% formaldehyde [Merck, Modderfontein, SA]/0.9% NaCl 1:8 v/v [Sabax, Port Elizabeth, SA]) for 24 hours at room temperature (RT), then dehydrated in absolute ethanol (Merck), cleared in xylene (Merck), and finally embedded in paraffin–wax.
Paraffin wax-embedded tissue sections (5 μm thick), mounted on 10% poly-L-lysine (Sigma, St. Louis) coated glass slides were stained with 0.1% Mayer's hematoxylin (Sigma) and 0.5% eosin (Sigma) for grading by an independent histopathologist. All six samples were identified as diffuse astrocytomas WHO Grade II based on the presence of nuclear atypia and the absence of mitosis, endothelial cell proliferation, and necrosis.35, 36 However, in three of the six samples mitotic cells were observed in some sections. The tumor cells were pleiomorphic with numerous elongated processes. In addition, there were numerous dilated and congested blood vessels as well as areas of microcystic degeneration (Figure 1A).
In situ RT-PCR
ET-1-specific primers for in situ RT-PCR, purchased from MWG-Biotech AG, Germany, were designed and used as previously described.37 ET-1 mRNA labeling has been demonstrated within amniotic epithelial cells in the human umbilical cord.38 Therefore, samples of human umbilical cord were obtained from cesarean section, fixed for 24 hours at RT in 5% formal saline and used as the positive method control for in situ RT-PCR.
Salamonsen et al.39 have demonstrated ET-1 immunoreactivity within endothelial cells of human umbilical veins and arteries and amniotic epithelial cells. Sections of human umbilical cord, described above, were used as positive method control tissue to demonstrate immunoreactive ET-1 (antibody diluted 1:600 in 0.01 M phosphate-buffered saline [PBS], pH 7.4/10% blocking reagent [Roche Diagnostics, Germany]) in the amniotic epithelium and endothelial cells of umbilical veins and arteries.
Immunoreactive ECE-1, ETA and ETB receptors have been previously localized in human renal tissue.40, 41 Therefore, wax-embedded tissue sections (5 μm) of normal human kidney, collected at autopsy, were used as positive method control tissue for immunocytochemical localization of ECE-1, ETA and ETB receptors. The primary, polyclonal, rabbit antihuman antibodies for ECE-1 (ECE-1 antiserum, 1:1000), and ETA (1:1000) and ETB receptors (1:1100) were all diluted in PBS/10% blocking reagent.
Lyophilized ET-1 antigen (Sigma-Genosys, Haverhill, UK), reconstituted in sterile physiologic saline (0.9% NaCl, w/v, pH 7) was used to raise polyclonal anti-ET-1 antibodies in rabbit hosts. This antibody has been fully described by our group.42
Anti-ECE-1, anti-ETA, and anti-ETB receptor antibodies
Rabbit antihuman ECE-1 antiserum (473-17-A), previously fully characterized,40 was generously donated by Dr. Florence Pinet (INSERM U36, College de France, Paris, France). This antibody was raised against a synthetic peptide corresponding to amino acids 473–489 of the extracellular domain of human ECE–1, which recognizes both the monomeric and dimeric forms of ECE-1. Both the ETA and ETB receptor antibodies were kindly provided by Dr. Werner Muller-Esterl (Department of Pathobiochemistry, Johannes Gutenberg-University at Mainz, Mainz, Germany). Essentially, ETA receptor antibodies were raised against extracellular amino terminus (CDN25) and intracellular carboxyl terminus (CTS24) peptides. The resulting antisera were fully characterized and pooled (AS444) and demonstrated specific crossreactivity with the cognate protein.43 Similarly, ETB receptor antibodies were directed against ETB extracellular amino terminus (CGL26) and intracellular carboxyl terminus (CLK23) peptides. These were fully characterized, and a pooled sample (AS445) was shown to crossreact with the human ETB protein on Western blot.44
Localization of ET-1 mRNA by in situ RT-PCR
The highly sensitive in situ RT-PCR technique has the ability to detect low mRNA copy number. The method used in this study has been described previously.45 Briefly, after controlled fixation, human surgical astrocytoma tissue sections were treated with proteinase K (Roche, Basel, Switzerland) and DNase (Roche), followed by RT–PCR using specific forward and reverse primers for ET-1 mRNA, and digoxigenin (DIG)-labeled dUTP. PCR products that had incorporated the DIG label were then observed with an alkaline phosphatase conjugate and the chromogen NBT/BCIP (Roche).
Localization of ET-1, ECE-1, ETA and ETB Receptors by ICC
A modified avidin-biotin complex (ABC) method used for the immunocytochemical localization of ET-1, ECE-1, ETA and ETB receptors was described previously.42 Briefly, paraffin-embedded tissue sections (5 μm) were dewaxed in xylene and rehydrated through a graded series of alcohols. The tissues were then boiled in 0.1 M sodium citrate, pH 6 (Merck, Whitehouse Station, NJ) for antigen retrieval, allowed to cool to RT, and then incubated with 5% H2O2/methanol (Merck) to quench endogenous peroxidase activity. The primary antibodies, ET-1, ECE-1, ETA, and ETB, were diluted 1:1000, 1:1000, 1:800, and 1:1000, respectively, in PBS/10% blocking reagent, followed by incubations with an ABC system (LSAB® Plus; DakoCytomation USA, Carpenteria, CA). The bound complex was then rendered visible with 3,3′-diaminobenzidine (DAB, DakoCytomation). Slides were viewed with a Leica DMLB microscope (Leica Microsystems AG, Wetzlar, Germany) and images captured with a Leica DC100 digital camera (Leica), with AnalySIS Pro 2.11™ software (Soft Imaging Systems, Munster, Germany).
Semiquantitative Grading System for the Differentiation of Immunostaining
The grading system for ICC followed a two-tiered approach. For each field of view observed under the light microscope, the labeling of ET components was quantified according to 1) the intensity, and 2) extent of labeling. The symbol “-” was assigned when no cells were observed to be labeled, “1+” for low labeling, “2+” for moderate, and “3+” when the labeling was intense, The extent of labeling was assigned “1” when < 25% of cells per field of view were observed to be labeled, “2” if between 25–75% of cells labeled, and when > 75% of cells labeled, “3” was assigned. These semiquantitations were the mean results performed on multiple (at least five) fields of view for each component and were observer-dependant.
In sections of the human umbilical cord, cytoplasmic ET-1 mRNA as well as immunoreactive ET-1 labeling have been observed within amniotic squamous epithelial cells and intermittent cells in the Wharton jelly (Figure 1B).38, 39, 42, 45, 46
Immunolabeled renal ECE-1 was observed within capillary endothelial and interstitial cells of the glomerulus, as well as distal (dt), proximal convoluted (pt), and collecting tubular epithelial cells (Fig. 1C). Renal ETA and ETB immunostaining was seen within vascular endothelial cells in the glomerulus, as well as epithelial cells of distal and proximal convoluted and collecting tubules (Fig. 1D).
In situ RT-PCR Detection of ET-1 mRNA in Human Astrocytomas
In situ RT-PCR demonstrated intense (3+) ET-1 mRNA labeling throughout the cytoplasm of tumor astrocytes (Fig. 1E) and endothelial cells of stromal blood vessels (Fig. 1E inset) in all six samples of astrocytomas examined. Whereas > 75% of tumor astrocytes labeled positively for ET-1mRNA, all of the endothelial cells observed labeled positively. The specificity of ET-1 mRNA localization was confirmed by the absence of cytoplasmic labeling in astrocytic tumor cells following the omission of the RT and PCR primers (Fig. 1F). During in situ RT–PCR, Taq DNA polymerase attempts to repair nicked nuclear DNA (cleaved as a result of DNAse action), during which the DIG-labeled dUTP becomes incorporated into the repaired nuclear DNA fragments. This results in nonspecific DIG complexes in the nuclei of both experimental tissue and negative controls. These DIG nuclear complexes are immunodetected resulting in nonspecific nuclear labeling.
Immunolocalization of ET-1, ECE-1, ETA, and ETB Receptors in Human Astrocytomas
Immunohistochemical analyses of all 6 astrocytoma samples revealed intense (3+) ET-1 labeling throughout the tumor astrocytes, with up to 75% of cells observed to be labeled (Fig. 1G). ET-1 labeling was most intense in malignant astrocytes undergoing mitosis (red arrow, Fig. 1G). Intense (3+) ET-1 immunolabeling was also observed within all endothelial cells of tumor blood vessels (Fig. 1H). Immunoreactive ETA (Fig. 1I) and ETB receptors (Fig. 1J) were observed in 25–75% of malignant astrocytes. Moderate labeling for both receptors were detected throughout the tumor cells as well as in endothelial cells of tumor blood vessels in all six astrocytoma samples (not shown). Visually, ETA receptor immunolabeling appeared to be less intense than ETB labeling (subjective observation of at least 5 different fields of view). Intense ECE-1 immunolabeling was observed throughout the cells in up to 75% of astrocytes (Fig. 1K) in all 6 tumors. In negative control sections, ET-1, ECE-1, ETA and ETB labeling was completely abolished by the replacement of the primary antibodies with PBS (Fig. 1L).
In the mammalian CNS, vasoactive ET-1 may also modulate neuronal functions.22, 46, 47 Previous studies in normal human brain have mapped ET components to neurons but not within astrocytes.22, 46, 47 However, ET-1 mRNA, ET-1, ECE-1, ETA and ETB receptors have been observed predominantly in cultured astrocytes,19, 23, 24 and only immunoreactive ET-1 has previously been demonstrated in astrocytic tumors.27, 28 To our knowledge, this study is the first to describe the cellular distribution of the entire ET system in human diffuse astrocytomas WHO Grade II, indicating an upregulation of the endothelin system within astrocytes during tumorigenesis. ET-1 mRNA was observed by in situ RT-PCR within the cytoplasm of tumor astrocytes and tumor vascular endothelial cells, confirming local synthesis. Immunoreactive ET-1, ECE-1, ETA and ETB receptors were localized to numerous astrocytic tumor cells and endothelial cells of tumor blood vessels, suggesting that big ET-1, the substrate for ECE-1, is processed within these cells.
In vascular tissue, ETA receptors predominate in smooth muscle cells of the tunica media, whereas ETB receptors are present in endothelial cells of the tunica intima.2 It has been shown that the radioligand PD 156707, which opposes the vasoconstrictor responses of ET-1,48 is selective for ETA receptors located in the vasculature of human gliomas and meningiomas. From the current study, we surmise that in human astrocytomas, ETA receptors in smooth muscle cells mediate the constriction of tumor blood vessels, whereas ETB receptors located in endothelial cells (Fig. 2) regulate this vasoconstriction by inducing the release of endothelium-derived relaxing factors (EDRFs). This regulation of tumor vascular resistance could be crucial in enhancing blood flow and thereby augmenting tumor proliferation.
Because astrocytes express G protein-linked receptors, which function through ligand-gated ion channels, it is possible that glia play a role in regulating neurotransmission through a variety of intercellular signal cascades. The ability of neurotransmitters, released at neuronal synapses, to enter astroglia,49 and regulate potassium channel conductance50 suggests that gap junctions may influence astroglial function51 as well as the transmission of apoptotic signals.52 In rodent astrocyte cultures, high ET-1 levels demonstrated a reduction in gap junction permeability, resulting in inhibition of intercellular communication, thus altering normal astrocytic function.53 Therefore, because ET-1 stimulates astrocyte proliferation,18, 21, 54 we speculate that reduced gap junction permeability may alter normal astrocytic function and is implicated in neoplastic transformation (Fig. 2). Further, ET-1 has been found to be mitogenic in various neoplasms11 as well as in astrocytes,54 and its presence in astrocytomas could well implicate ET-1 in tumor promotion by means of ETA and ETB receptor signaling in an autocrine and/or paracrine mode. The detection of ET-1 as a survival factor in endothelial16 and smooth muscle cells17 further supports an antiapoptotic role for ET-1 in the proliferation of tumor astrocytes.
A shortcoming in this study is that elucidation of the endothelin system is limited to diffuse astrocytomas WHO Grade II. It would be interesting to correlate the expression and distribution of the endothelin system in diffuse astrocytoma, anaplastic astrocytoma (WHO Grade III), and glioblastoma (WHO Grade IV). For the latter, it will be valuable to study the endothelin system components in both primary and secondary glioblastomas because these two forms differ in epidemiology, clinical presentation, as well as many molecular markers.35
The presence of ET-1 mRNA and immunoreactive ET-1, ECE-1, ETA receptors, and ETB receptors in human astrocytoma cells and tumor blood vessels and the paucity of these components in white matter of normal brain implies that most human astrocytomas possibly acquire ET-1 for survival. Astrocytoma proliferation may be facilitated by the degradation of the ECM and basement membranes by ET-1 and by vasodilatory action of ETB receptors. This study clearly suggests a possible function for the endothelin system in astrocytoma progression, and the development of specific antagonists may have therapeutic relevance in the treatment of patients with gliomas.
The authors acknowledge the following South African institutions for financial support: Medical Research Council (MRC), University of Natal Research Fund (URF), Cancer Association of South Africa (CANSA), and National Research Foundation (NRF).
In addition, they thank the Department of Neurosurgery, Wentworth Hospital, Durban, South Africa, for access to astrocytoma tissue samples; Dr Ashwin Bramdev (Lancet Laboratories, Durban) for grading the astrocytoma sections; Dr Florence Pinet (INSERM U36, College de France, Paris, France) for donating the ECE-1 polyclonal antiserum; and Dr Werner Muller-Esterl (Department of Pathobiochemistry, Johannes Gutenberg-University at Mainz, Mainz, Germany) for providing the ETA and ETB receptor antibodies.