Increased In Vivo Expression and Production of Endothelin-1 by Porcine Cardiomyocytes Subjected to Ischemia
Abstract Circulating levels of the endothelium-derived vasoconstrictor peptide endothelin-1 (ET-1) are increased in association with myocardial ischemia and infarction. The present study investigates whether ET-1 is synthesized and produced locally in the ischemic heart. Sixteen pigs were divided into three groups. In the first group, the left anterior descending coronary artery was occluded for 90 minutes, followed by 150 minutes of reperfusion (group A, n=8). Two additional groups were subjected to 90 minutes (group B, n=4) or 240 minutes (group C, n=4) of ischemia without reperfusion. Biopsies from the nonischemic and ischemic myocardium were rapidly obtained from the beating heart and subsequently examined by Northern blot, in situ hybridization, and immunohistochemistry. Arterial plasma ET-1 was measured before ischemia and at the end of the experiments. Northern blot revealed a twofold increase in ET-1 mRNA in the ischemic myocardium compared with the nonischemic myocardium. In situ hybridization showed a considerable increase in ET-1 mRNA over the ischemic cardiomyocytes. Substantial ET-1–like immunoreactivity (ET-1-ir) was detected in cardiomyocytes in the ischemic region. In contrast, little or no ET-1-ir or mRNA was detected in nonischemic cardiomyocytes. Both in the ischemic and nonischemic regions, little ET-1-ir was detected in vascular endothelial cells or vascular smooth muscle cells. There was no difference in the intensity and distribution of ET-1 mRNA expression or ET-1-ir among experimental groups A, B, and C. Arterial plasma ET-1 was increased only in group A, the reperfused group. In conclusion, these findings provide a definitive evidence for a de novo synthesis of ET-1 by cardiomyocytes subjected to ischemia in vivo.
Endothelin-1 (ET-1), the most potent vasoconstrictor peptide yet known, was first isolated from cultured porcine aortic endothelial cells in 1988.1 ET-1 is believed to exert its effect mainly as a paracrine/autocrine hormone,2 although it also circulates in plasma in small amounts. In addition to being a potent vasoconstrictor, ET-1 has been shown to promote proliferation of smooth muscle cells3 and fibroblasts4 and to induce hypertrophy of cardiomyocytes in vitro.5 A cytotoxic effect on myocytes subjected to simulated ischemia in vitro6 and a proischemic effect on the isolated rat heart7 have also been reported. ET-1 exerts its effect through stimulation of specific receptors, widely distributed in cardiac tissue.8
A substantial increase in circulating plasma ET-1 has been demonstrated in association with myocardial ischemia and infarction both in humans9 and in experimental studies.10 11 12 Whether ET-1 is synthesized de novo and produced locally in the ischemic heart is still to be elucidated. A few studies exist that might indicate a production of ET-1 in the ischemic heart.10 11 12 However, none of those studies investigated whether there was a de novo synthesis of ET-1. Furthermore, since ET-1 was determined only by radioimmunoassay in plasma and tissue extracts, the histological site of production was not addressed in any of those studies. Endothelial cells, vascular smooth muscle cells, and cardiomyocytes are potential sources of ET-1 production.
Accordingly, the aims of the present study were to investigate the level and sites of ET-1 synthesis and production in the ischemic heart. This was investigated by obtaining biopsies from the beating porcine heart subjected to ischemia alone or in combination with reperfusion. The biopsies were subsequently analyzed by Northern blot, in situ hybridization, and immunohistochemistry. Arterial plasma ET-1 was measured before the induction of myocardial ischemia and at the end of the experiments.
Materials and Methods
Sixteen domestic pigs of either sex (27 to 38 kg) were anesthetized with pentobarbital sodium, initially at 40 mg · kg body wt−1 IP, followed by a sustained intravenous infusion of 12 to 25 mg · kg−1 · h−1 according to the depth of anesthesia. The animals were ventilated by a volume-regulated respirator (model 101, Princeton Medical Instruments Inc) through a tracheostoma with a 50% O2/50% air mixture. A positive end-expiratory pressure of 5 cm H2O was added. Ventilation frequency and volume were adjusted to keep pH within physiological ranges. One polyethylene catheter was inserted into the right femoral vein for infusions, and another was inserted into the right femoral artery and advanced to the aortic arch for blood sampling and pressure recordings (Statham pressure transducer, model P23 Gb, Gould Instruments). Mean arterial pressures were within physiological ranges throughout the experiments. Body temperature was kept constant with wrappings and electric heating pads. Urine was drained continuously through a cystostoma.
The heart was exposed through a midsternal split and suspended in a pericardial cradle. A segment of the left anterior descending coronary artery (LAD) between the first and second major branches was carefully dissected free to allow attachment of a Mayfield clip (normally used to clip cerebral vascular aneurysms) for occlusion.
After the surgical preparation, a control period of 90 minutes with stable hemodynamics was allowed. Then, the LAD was occluded by a Mayfield clip. In one group of pigs (group A, n=8), cardiac biopsies were obtained after 150 minutes of reperfusion following 90 minutes of LAD occlusion. In two other groups, biopsies were obtained after 90 minutes (group B, n=4) and 240 minutes (group C, n=4) of LAD occlusion without reperfusion. In all animals, biopsies were taken at the end of the experiment from the beating heart both from a nonischemic area and the area subjected to ischemia in random order. To investigate whether there was an increased expression of ET-1 mRNA in the ischemic myocardium, Northern blot was performed (n=4 in each of the three experimental groups). In situ hybridization, which is a nonquantifiable technique, was used to localize the ET-1 mRNA to cell type, and immunohistochemistry was performed to localize mature ET-1. The nonischemic area was macroscopically distinguished from the dyskinetic miscolored ischemic area, and that biopsies were taken from within the two areas was later confirmed by histological examination. All biopsies were obtained by rapidly drilling in the myocardium with a sharp steel cylinder with a diameter of 1 cm. The biopsies were immediately snap-frozen in liquid nitrogen for the Northern blot analyses and fixed in 4% paraformaldehyde for in situ hybridization or in Bouin’s solution for immunohistochemistry. Northern blot, in situ hybridization, and immunohistochemical studies for all biopsies were carried out under similar conditions by one investigator (A.G.) in order to minimize technical error. Blood samples were collected from the aortic arch in the three experimental groups during control conditions and at 150 minutes of reperfusion following 90 minutes of ischemia (group A, n=8), at 90 minutes of ischemia (group B, n=4), and at 240 minutes of ischemia (group C, n=4). In three of the reperfused animals (group A), blood samples were also collected after 90 minutes of ischemia. To exclude an increase in plasma ET-1 due to anesthesia and the surgical procedure, plasma ET-1 was measured in two sham-operated animals serving as time controls. Plasma ET-1 averaged 4.8, 4.5, and 5.0 fmol · mL−1 at time points corresponding to the time of blood sampling in the experimental animals. These values indicate that plasma ET-1 was not influenced by anesthesia and the surgical procedure, which is in accordance with previous findings in our laboratory.13
Northern Blot Analysis
Total RNA was extracted from cardiac tissues by a modified method of Chomczynski and Sacchi.14 Briefly, fresh frozen cardiac tissue was homogenized in a solution of 4 mol/L guanidinium isothiocynate, 25 mmol/L sodium citrate, pH 7, 0.5% sarkosyl, and 0.1 mol/L β-mercaptoethanol. RNA was extracted by phenol-chloroform, precipitated in ethanol, and purified by pelleting through lithium chloride. RNA was glyoxal-denatured, electrophoresed on 1.1% agarose gel, and transferred to Hybond N nylon membrane (Amersham Int). A fragment of porcine ET-1 cDNA, described in detail elsewhere,1 derived from the 3′ noncoding region was used as the source for RNA probes. Blots were probed with 32P-labeled riboprobes for ET-1, β-actin, 18S, and 28S. Autoradiographs were quantified by densitometry, and the ET-1 mRNA levels were normalized for the loaded amounts of RNA by comparing 28S RNA levels as an internal control. The intensities of the autoradiographic bands of Northern blots from ischemic myocardial tissue were compared with the signal intensity of control bands corresponding to ET-1 mRNA derived from nonischemic myocardial tissue.
In Situ Hybridization
Two complementary cDNA probes from the 3′ noncoding region of the porcine ET-1 gene were subcloned into the polylinker site of the Bluescript vector. Antisense and sense riboprobes were generated by using T7 and T3 RNA polymerase, respectively. ET-1 probes (same as used for the Northern blot studies) were labeled with 35S-UTP by using a commercial kit (Ambion Inc). Cryostat sections of paraformaldehyde-fixed materials were rehydrated in PBS, permeabilized with proteinase K, and then fixed in 4% paraformaldehyde. This was followed by an acetylation step in triethanol amine and acetic anhydride to reduce background noise. Sections were dehydrated in increasing concentrations of ethanol, air-dried, and incubated with a hybridization mixture containing 1×106 cpm per section of radiolabeled probe for 16 hours at 42°C.15 Unbound RNA probes were removed by incubation in RNase A solution in 2× saline sodium citrate (SSC). Sections were then washed in decreasing concentrations of SSC (2× to 0.1× SSC), dehydrated in ethanol and ammonium acetate, air-dried, and processed for autoradiography. Negative control experiments involved the use of sense probe and incubation of sections with hybridization mixture in the absence of radiolabeled probe.
Multistep sections were dewaxed, rehydrated, and permeabilized in 0.3% Triton for 30 minutes. Sections were immunostained with antisera to the C-terminus of ET-1, C-terminus of big ET-1, and von Willebrand factor (factor VIII). The ET-1 and big ET-1 antisera were raised as previously described16 and have no cross-reactivity with big ET-1 and ET-1, respectively. Immunohistochemistry was performed by using a modification of the avidin-biotin-peroxidase complex method.17 Briefly, sections were incubated with hydrogen peroxide for 60 minutes to block endogenous peroxidase activity, followed by further incubation with normal goat serum to reduce nonspecific bindings of the antisera. Sections were then incubated with the relative antisera for 16 hours at 4°C, washed in PBS, and incubated with biotinylated goat anti-rabbit IgG for 45 minutes. After three washes in PBS, sections were incubated with the avidin-biotin-peroxidase complex (Vectastain Elite Kit, Vector Laboratories) for 45 minutes, and sites of immunoreaction were visualized by use of diaminobenzidine and hydrogen peroxide. Negative control experiments involved absorption of the first layer antisera with their respective antigens before incubation with sections or the use of normal serum instead of the first layer antiserum.
Extra sets of sections from each biopsy were stained with hematoxylin and eosin for the assessment of histological changes.
Determination of Plasma Endothelin
Plasma ET-1 was determined as described previously.13 Briefly, blood samples were collected directly into prechilled test tubes containing an inhibitor solution (2 mg EDTA and 1000 kallikrein inhibiting units of aprotinin per milliliter of blood) and then centrifuged at 1673g for 15 minutes at 4°C. Plasma was stored at −70°C and assayed within 2 weeks. ET-1 was measured by using an ET-1-21 specific [125I] radioimmunoassay system (RPA 555) from Amersham International. This assay system has no cross-reactivity with big ET-1, 144% cross-reactivity with ET-2, and 52% cross-reactivity with ET-3. The recovery of the extraction averaged 59%, and the values given in the study were not corrected for the loss during extraction.
Values are given as mean±SEM. Statistical evaluation of plasma ET-1 before (control) and after myocardial ischemia and of the densitometric Northern blot signals in the nonischemic compared with the ischemic myocardium were both performed by paired Student’s t test. A value of P<.05 was considered statistically significant.
Light-microscopic sections from the nonischemic control area revealed histological findings of a normal myocardium, such as centered nuclei and cross striation. In contrast, sections from the ischemic area showed variable degrees of tissue damage, such as loss of cross striation, decentralization of nuclei, cytoplasmic swelling, and early signs of irreversible cell damage, such as contraction band necrosis and coagulative necrosis.
Data of Northern blot analysis demonstrated the presence of ET-1 mRNA in nonischemic myocardial tissue from groups A, B, and C (Fig 1⇓). In myocardial tissue subjected to ischemia, there was an induction of ET-1 mRNA (Fig 1⇓). Densitometric analysis revealed an increase in ET-1 mRNA of 1.8±0.2-fold (P=.002) in the animals subjected to 90 minutes of ischemia followed by 150 minutes of reperfusion (group A). An increase in ET-1 mRNA was also observed in animals subjected to 90 minutes (group B) and 240 minutes (group C) of ischemia without reperfusion, and the increase was 2.2±0.3- and 2.4±0.4-fold, respectively (P<.05). The increase in the three groups was not significantly different.
In Situ Hybridization
In the myocardium not subjected to ischemia, there was little expression of ET-1 mRNA as displayed by the presence of scattered silver grains over cardiomyocytes (Fig 2A⇓) and endothelial cells. In contrast, a stronger hybridization signal was seen over the cardiomyocytes in areas subjected to ischemia (Fig 2B⇓ and 2C⇓). Only background noise of silver grains was seen over sections hybridized with sense probe (Fig 2D⇓). The stronger hybridization signal over ischemic cardiomyocytes compared with nonischemic myocytes was more appreciated when dark-field illumination of ET-1 mRNA was used (Fig 2E⇓ through 2H). No differences were detected between the three experimental groups (A, B, and C).
Immunohistochemical analyses of the biopsies revealed the presence of little or no ET-1–like immunoreactivity (ET-1-ir) in myocardial tissue not subjected to ischemia (Fig 3A⇓). Little ET-1-ir was found in cardiomyocytes and in vascular endothelial cells of both small and large extramural and intramural vessels. In some vessels, vascular smooth muscle cells showed diffuse staining for ET-1. There was, however, no obvious pattern with regard to which vessels showing ET-1-ir (eg, size or extramural or intramural localization). In the sections in which endocardium was present, endocardial endothelial cells also displayed the brown color of immunoreaction.
In contrast to the nonischemic myocardium, substantial ET-1-ir was observed in myocardial tissue subjected to ischemia (Fig 3B⇑ through 3D). The most apparent immunostaining was observed in cardiomyocytes, showing diffuse strong ET-1-ir. As with the normal cardiac tissue, little or no ET-1-ir was detected in the vascular endothelial cells (Fig 3D⇑), and some vessels showed diffuse immunostaining for ET-1 in vascular smooth muscle cells. No apparent differences were detected between the three groups of hearts subjected to 90 minutes of ischemia followed by 150 minutes of reperfusion (group A) or 90 minutes (group B) or 240 minutes (group C) of ischemia without reperfusion. Sections immunostained with an ET-1 antiserum preadsorbed with synthetic ET-1 peptide before the incubation with the section showed no staining; this finding was taken as a confirmation of the specificity of the antiserum (Fig 3E⇑). To verify that reagents in the avidin-biotin-peroxidase reaction did not preferentially stain ischemic muscle, sections from the ischemic region were stained in the absence of the ET-1 antiserum (Fig 3F⇑).
Immunostaining with antiserum to big ET-1 resulted in a pattern of immunostaining similar to that of mature ET-1, showing dense immunostaining in cardiomyocytes subjected to ischemia.
Arterial plasma ET-1 increased significantly from 3.5±0.9 fmol · mL−1 before the induction of ischemia to 4.4±0.7 fmol · mL−1 after 150 minutes of reperfusion (P<.001). In contrast, compared with the control condition (4.1±0.2 fmol · mL−1), no increase in plasma ET-1 was detected at the end of the 90-minute ischemic period (4.2±0.3 fmol · mL−1). In the third group of animals without reperfusion, plasma ET-1 averaged 5.4 fmol · mL−1 during control and 5.4 fmol · mL−1 after 240 minutes of ischemia.
The present study demonstrates an increase in the synthesis and production of ET-1 in ischemic cardiac tissue. Indeed, Northern blot analysis revealed that ET-1 mRNA in ischemic myocardial tissue is increased twofold over the nonischemic myocardial tissue. Also, in situ hybridization showed that cardiomyocytes are the main sites of ET-1 synthesis. Immunohistochemical analyses of the ischemic myocardial tissue revealed the presence of strong ET-1-ir in cardiomyocytes compared with a diffuse weak immunostaining in the nonischemic myocardial tissue. Furthermore, radioimmunoassay demonstrated an elevation of ET-1 plasma levels after 150 minutes of reperfusion following 90 minutes of ischemia. Our findings clearly demonstrate that the heart responds to an ischemic event by initiating ET-1 synthesis de novo.
An increase in plasma ET-1 has been demonstrated in association with myocardial infarction both in patients9 and in experimental studies.10 11 12 A few studies exist that might indicate the production of ET-1 by the ischemic heart.10 11 12 One study reported that reperfusion of the ischemic heart was necessary to increase venous plasma ET-1, indicating that the increase in plasma ET-1 was of cardiac origin.11 In another study, a venoarterial increase in plasma ET-1 was observed during reperfusion after myocardial ischemia.12 A third study speculated that an observed increase in ET-1 in tissue extracts from infarcted myocardium was caused either by increased production by vascular endothelial cells or by plasma ET-1 moving from the bloodstream into the cardiac tissue.10 Neither of those studies investigated whether the increase in ET-1 was due to de novo synthesis. Furthermore, since ET-1 was determined only by radioimmunoassay in plasma and tissue, the histological site of production was not addressed in any of those studies.
Both endothelial cells1 and vascular smooth muscle cells18 have been shown to produce ET-1 in vitro. Each of these cell types could therefore potentially produce ET-1 in the ischemic heart in vivo. In the present study, however, there was no obvious increase in mature ET-1 production by ischemic vascular endothelial and smooth muscle cells as determined by immunohistochemistry. In sharp contrast, cardiomyocytes subjected to ischemia were the only subset of cells to exhibit an apparent increase in ET-1-ir. There was also a substantial increase in ET-1 mRNA expression over ischemic myocardial tissue as determined by in situ hybridization. The major increase in ET-1 mRNA expression was localized over ischemic cardiomyocytes, by far the cell type most frequently present in the histological sections. This was revealed by comparing adjacent sections stained with hematoxylin and eosin as well as with antisera to von Willebrand factor to display endothelial cells. Thus, although the present study cannot exclude an increased expression of ET-1 mRNA in ischemic vascular endothelial cells also, this would quantitatively be of minor importance. Although immunohistochemistry and in situ hybridization are powerful techniques in demonstrating sites of peptide cellular production and synthesis, respectively, they both are incapable of providing precise quantitative data. Therefore, we used Northern blot analysis as a means of quantifying the level of ET-1 synthesis. Indeed, Northern blot was in agreement with the immunohistochemistry and in situ hybridization data and showed a twofold increase in ET-1 in the ischemic myocardium. The synthesis and production of ET-1 by ischemic cardiomyocytes in vivo was an unexpected and interesting finding, which has not to our knowledge been previously reported. That cardiomyocytes are able to produce ET-1 is supported by two recent in vitro studies.19 20 However, it should be noted that both studies were performed with cultured neonatal cardiomyocytes, and the effect of ischemia was not investigated. One study showed an increase in secreted ET-1 by unstimulated cardiomyocytes after 3 days of incubation,19 whereas the other showed an increase in ET-1 secretion after stimulation with angiotensin II.20
The present study indicates an upregulation of ET-1 synthesis in ischemic cardiomyocytes initiated early in the ischemic period. This is supported by the findings that ET-1 mRNA and immunostaining were similar in hearts subjected to 240 minutes of ischemia compared with hearts subjected to 90 minutes of ischemia. Furthermore, although reperfusion was needed to increase plasma ET-1 in the present study, it did not influence the local myocardial ET-1 production. This is indicated by findings indicating no differences in ET-1-ir and its mRNA in hearts subjected to ischemia alone compared with those subjected to ischemia followed by reperfusion.
Synthesis and production of ET-1 by an ischemic heart may explain the observed increase in plasma ET-1 in association with myocardial ischemia and infarction. It might be speculated that ET-1 leaks along a concentration gradient from the site of increased production (ie, in the myocardium) into the capillaries and enters the circulation either from the ischemic border zones or from damaged myocardial tissue upon reperfusion. That the increase in plasma ET-1 in the present study originates from the myocardium is given further support by the fact that no increase in plasma ET-1 was seen in animals subjected to ischemia without reperfusion, which is in accordance with a previous study.11 However, it cannot be excluded that events taking place during reperfusion also stimulated a peripheral production of ET-1 in the present study.
Proteolytic leukocyte-derived enzymes have been shown to increase the conversion of big ET-1 to ET-1 in vitro,21 and such enzymes are likely to be increased in myocardial tissue subjected to ischemia. Thus, in addition to the ET-1 synthesized de novo, another possible source for the increased circulating and myocardial ET-1-ir seen in the present study is increased conversion of preexisting big ET-1 to mature ET-1. If increased conversion was the main mechanism, big ET-1 would be expected to decrease on behalf of the increase in mature ET-1. This was clearly not the case. Our immunohistochemical analysis using a specific antiserum to big ET-1 showed that big ET-1-ir was increased in the ischemic region compared with the nonischemic region. Furthermore, ET-1 mRNA levels were clearly upregulated in the ischemic tissue. These results indicate that the increase in mature ET-1 is mainly due to de novo synthesis.
A substantial increase in ET-1 production in cardiomyocytes subjected to ischemia may have pathophysiological effects on the myocardium. ETA and ETB receptors have been demonstrated in the human myocardium,8 and both antibodies against ET-110 and the ET-1 receptor blocker BQ-12322 have been shown to reduce infarct size by almost 40% in experimental models. The infarct-reducing effect of the receptor blocker BQ-123 was shown to be independent of myocardial blood flow.22 This finding, together with our finding of a substantial local increase in ET-1 in ischemic cardiomyocytes, indicates that ET-1 may have deleterious effects locally on the ischemic myocardium that are not related to its potent vasoconstrictor properties. Further support for this hypothesis is given by a study showing that ET-1 elicits a cytotoxic effect on ischemic cultured myocytes.6 In another study, ET-1 has been shown to exert a proischemic effect on cardiomyocytes.7 A local myocardial ET-1 increase might also have other effects on the heart. ET-1 has been shown to promote proliferation of fibroblasts4 and hypertrophy of cardiomyocytes5 in vitro. ET-1 might therefore play a role in the repair process and myocardial adaptation associated with myocardial ischemia and infarction.
In conclusion, our data demonstrate for the first time that cardiomyocytes, in vivo, synthesize and produce elevated levels of ET-1 in response to ischemia. These findings are important prerequisites for the understanding of the pathophysiological role of ET-1 in myocardial ischemia/infarction and for potential future therapeutic approaches.
This study was supported by the Norwegian Council for Cardiovascular Diseases, the Anders Jahre’s Fund for the Promotion of Science, the Professor Carl Semb’s Medical Research Fund, the Norwegian Air Ambulance Fund, the Heart and Stroke Foundation of Québec, and the Medical Research Council of Canada. Dr Christensen was financed by the Research Forum, Ullevål Hospital, and Dr Giaid was supported by the Heart and Stroke Foundation of Canada. We want to thank Mette Ree Holte, Hilde Hyldmo, Dr Bruce Jamison, Bjørn Kristiansen, Sonia Lambert, Gerd Torgersen, and Annlaug Ødegaard for technical assistance. Prof Lars Aage Solberg is acknowledged for his advice concerning the interpretation of the histological sections.
- Received April 7, 1994.
- Accepted February 13, 1995.
- © 1995 American Heart Association, Inc.
Takuwa N, Takuwa Y, Yanagisawa M, Yamashita K, Masaki T. A novel vasoactive peptide endothelin stimulates mitogenesis through inositol lipid turnover in Swiss 3T3 fibroblasts. J Biol Chem. 1989;264:7856-7861.
Ito H, Hirata Y, Hiroe M, Tsujino M, Adachi S, Takamoto T, Nitta M, Taniguchi K, Marumo F. Endothelin-1 induces hypertrophy with enhanced expression of muscle-specific genes in cultured neonatal rat cardiomyocytes. Circ Res. 1991;69:209-215.
Grover GJ, Sleph PG, Fox M, Trippodo NC. Role of endothelin-1 and big endothelin-1 in modulating coronary vascular tone, contractile function and severity of ischemia in rat hearts. J Pharmacol Exp Ther. 1992;263:1074-1082.
Molenaar P, O’Reilly G, Sharkey A, Kuc RE, Harding DP, Plumpton C, Gresham GA, Davenport AP. Characterization and localization of endothelin receptor subtypes in the human atrioventricular conducting system and myocardium. Circ Res. 1993;72:526-538.
Watanabe T, Suzuki N, Shimamoto N, Fujino M, Imada A. Contribution of endogenous endothelin to the extension of myocardial infarct size in rats. Circ Res. 1991;69:370-377.
Velasco CE, Jackson EK, Morrow JA, Vitola JV, Ingami T, Forman MB. Intravenous adenosine suppresses cardiac release of endothelin after myocardial ischaemia and reperfusion. Cardiovasc Res. 1993;27:121-128.
Hsu SM, Raine L, Fanger H. Use of avidin-biotin-peroxidase complex (ABC) in immunoperoxidase techniques: a comparison between ABC and unlabelled antibody (PAP) procedures. J Histochem Cytochem. 1981;29:577-580.
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 cardiomyocytes. J Clin Invest. 1993;92:398-403.
Grover GJ, Dzwonczyk S, Parham CS. The endothelin-1 receptor antagonist BQ-123 reduces infarct size in a canine model of coronary occlusion and reperfusion. Cardiovasc Res. 1993;27:1613-1618.