Endothelin Mediation of Insulin and Glucose-Induced Changes in Vascular Contractility
Abstract Although the prevalence of hypertension in diabetic patients is high and many factors participate, hyperinsulinemia cannot be discarded as a contributing factor. Insulin could act directly on smooth muscle altering intracellular calcium levels that mediate contraction and glucose transport or could induce the secretion of endothelin by the endothelial cells lining the vessels. The aim of the present report was to study the effect of different glucose and insulin concentrations on rat vascular smooth-muscle contractile characteristics and to determine whether insulin effects are mediated by endothelin. Femoral arteries obtained from Wistar rats were placed in an in vitro chamber and superfused with different glucose and/or insulin solutions. The contractile response to KCl 80 mmol/L, measured by the force generated, showed a significant decrease with high extracellular glucose concentrations (11 mmol/L). Insulin caused a dose-dependent increase in arterial contraction induced by KCl. This increase was significant when arteries were stimulated with 80 mmol/L KCl in the presence of 5.5 mmol/L glucose, but when 40 mmol/L KCl was used, an increase was observed with both 5.5 and 11 mmol/L glucose. The insulin-induced contraction was significantly reduced in the presence of hyperimmune anti-endothelin serum and in the presence of endothelin receptor ETA and ETB antagonists PD 151,242 and BQ-788, respectively. These results suggest that hyperinsulinemia and hyperglycemia may contribute to hypertension in diabetes and that responses to insulin are mediated partially by endothelin, thus explaining why non–insulin-dependent diabetes mellitus patients show an increase in arterial pressure before the onset of nephropathy.
The clinical association of diabetes and hypertension is the focus of increasing attention.1 2 3 In IDDM, elevated arterial blood pressure usually develops after nephropathy and could be its consequence.2 In contrast, in NIDDM, hypertension often precedes renal damage, and its cause remains unknown.2 There is controversy regarding whether insulin is the cause of hypertension because there are reports showing that high insulin levels such as those found in patients with insulinoma do not cause hypertension.4 5 6 However, studies performed on rats have shown that there is a correlation between high insulin levels and elevated blood pressure.5 The mechanism by which the hormone would elevate arterial pressure remains unknown. Elevated plasma endothelin levels in patients with diabetes mellitus7 and in lean NIDDM patients during euglycemic hyperinsulinemic clamp8 have been reported and could constitute one of the mechanisms involved in hypertension. Furthermore, a higher level of endothelin-1 gene expression9 and changes in its release in cultured endothelial cells induced by insulin10 11 also have been found. Another possibility is that cellular abnormalities induced by hyperinsulinemia induce changes in intracellular divalent ion levels, such as calcium and magnesium, that mediate contraction.1 2 12 13 Despite these controversies, there are few reports describing the effects of insulin on vascular smooth muscle contractility and the mediation of its effects by endothelin.
There is a poor knowledge of the effects of glucose on vascular contractile responses. Although in the short-term, high glucose levels could inhibit calcium-dependent mechanisms that increase glucose transport,14 15 thus diminishing tension development, only the long-term effects of hyperglycemia have been reported.2 Therefore, the aim of the present report was to study the effects of different glucose and insulin concentrations on the contractile characteristics of the vascular smooth muscle of the rat and to test the participation of endothelin as mediator of the insulin effects.
Sample Preparation and Tension Recording
Wistar rats of either sex weighing 250 to 350 g were used. Animals were killed by cervical dislocation, and femoral arteries were dissected immediately and placed in oxygenated normal Tyrode’s solution. Arterial segments about 5 mm long were cleaned from connective tissue, and two 250-μm-diameter S-shaped silver strings (Medwire Corp) were inserted into the lumen to measure tension developed transversely by rings of the vessel. One of the silver strings was fixed to the bottom of an in vitro chamber, and the other was attached to a tension transducer that was connected to a Grass polygraph. A basal passive tension of 500 mg was applied after determination in preliminary tests that this was the optimal resting tension in our experimental conditions. Contraction was induced by the addition of 40 and 80 mmol/L KCl. These doses were chosen after a dose-response curve was elaborated. The higher dose induced the maximal response, whereas the lower one provoked submaximal contractions.
The vessels were superfused with normal Tyrode’s solution (NaCl 136.9 mmol/L, KCl 5.4 mmol/L, CaCl2 1 mmol/L, MgCl2 1.05 mmol/L, NaHCO3 11.9 mmol/L, NaH2PO4 0.33 mmol/L, and glucose 5.5 mmol/L) for 15 minutes to allow stabilization and then were made to contract by adding KCl (40 or 80 mmol/L); afterward they were washed so that the tension returned to the basal level, and the stimulus was repeated again. The mean force of these two contractions was used as the control value and was considered as the 100% contractile response. After this, the arteries were placed in one of the following situations in separate experiments repeating the stimulation with KCl to test the contractile response. (1) Arteries were bathed with Tyrode’s solution containing half the amount of glucose (2.75 mmol/L) or double the normal concentration (11 mmol/L). (2) Arteries were superfused with 150, 300, or 600 nmol/L of insulin (Eli Lilly). These concentrations were chosen after testing different doses in our experimental conditions. (3) Arteries were superfused with different glucose concentrations in the presence of 300 nmol/L of insulin. (4) Arteries in which the endothelium was removed by insertion of a metal wire into the lumen and gentle rolling of the vessel segment on a wetted filter paper were made to contract in the presence and absence of insulin (300 nmol/L). (5) Arteries were bathed with insulin (300 nmol/L) and with insulin in the presence of polyclonal anti-endothelin antibodies. (6) Arteries were bathed with insulin (300 nmol/L) in the presence and absence of mouse normal control serum without anti-endothelin antibodies. (7) Arteries were bathed with insulin (300 nmol/L) and with insulin in the presence of ETA-receptor antagonist PD 151,242 (10 μmol/L). (8) Arteries were superfused with insulin (300 nmol/L) and with insulin in the presence of the ETB-receptor antagonist BQ-788 (10 μmol/L) (antagonists were bought from Research Biochemicals International). At the end of each experiment, arteries were bathed again with normal Tyrode’s solution and made to contract with KCl to verify that there were no important effects secondary to the incubation time or to repeated exposures to KCl and to prove that glucose and insulin effects were reversible. This last contraction was reduced after adding the endothelin-receptor antagonists because they have long-lasting effects. Endothelial integrity or removal was verified by adding acetylcholine (1 μmol/L) to the contracted vessel, observing a decrease in the tension generated when the endothelium was intact and an increase in arteries without endothelium.
Hyperimmune Anti-Endothelin Serum
Human synthetic ET-1 (Sigma Chemical Co) was conjugated using glutaraldehyde11 with 100 μg hemocyanin from the crab Megatura crenulata (Boehringer Mannheim). Conjugated endothelin (15 to 20 μg) was emulsified in an equal volume of incomplete Freund’s adjuvant and injected intraperitoneally to immunize a Balb/c mouse. The inoculation was repeated after 3 and 5 weeks. After 7 weeks, blood was extracted by cardiac puncture from the ether-anesthetized mouse, and serum was separated by centrifugation at 1200 rpm for 10 minutes. Serum was stored at −80°C until used. A 1:2000 dilution in normal or modified (by the addition of insulin or KCl) Tyrode’s solution was used to perfuse the arteries.
The antibody’s specificity was demonstrated by enzyme-linked immunosorbent assay, which was performed as follows: 50 ng of ET-1 coupled to human albumin per well was adsorbed to a microtiter plate at room temperature overnight. Nonspecific binding sites were blocked with 0.01% of porcine gelatine for 2 hours at 37°C. After the plate was rinsed with PBS-0.1% Tween, 20 serial dilutions (1/500 to 1/64 000) of mouse anti-endothelin antiserum were incubated overnight at 4°C. The plate was then washed with PBS-Tween and a peroxidase conjugated polyclonal anti-mouse IgG (F[ab]-specific) was added. After 1 hour of incubation, the plate was rinsed with PBS-Tween, revealed with o-phenylenediamine, and read spectrophotometrically at 492 nm in a Microplate Autoreader EL311 (Bio-Tek Instruments). The titer of the anti-endothelin antiserum determined with this assay was 1/16 000. Negative controls were made using normal mouse serum and wells filled with uncoupled albumin in which no reaction was observed. Although it is possible that antibodies against hemocyanin (the protein to which ET-1 molecule was bound) were induced, this protein is not found in mammals, and therefore, this antibody would not interfere with our assay.
Mean and standard errors of at least six different arteries were calculated. When values were expressed as percentages, the percentage in each experiment was calculated, and the mean was then determined. Statistical significance of the differences was analyzed using paired Student’s t test and nonparametric Kruskal-Wallis analysis of variance. Differences were considered statistically significant when P<.05.
The effects of different glucose and insulin concentrations on the contractile activity of vascular smooth muscle were studied. No effects on passive basal tension were observed as the vessel was superfused with solutions containing different glucose or insulin concentrations. Nevertheless, when the arteries were induced to contract by the addition of KCl, tension development varied in the different experimental conditions.
Determination of KCl Concentrations
Several doses of KCl were tested to produce contraction of the vessel. KCl 20 mmol/L produced a small increase in the contraction force of 75±14 mg (n=10), which increased to 128±15 mg (n=9) with a KCl dose of 40 mmol/L. A maximal increase of 302±64 mg (n=12) was achieved with an 80 mmol/L dose.
Effects of Different Glucose Concentrations on Vascular Contractility
When arteries were perfused with different glucose concentrations and stimulated with KCl (80 mmol/L), their response in the presence of Tyrode’s solution with high glucose (11 mmol/L) was significantly reduced when compared with the response generated in Tyrode’s solution with the normal glucose concentration of 5.5 mmol/L (from 100% to 84.08±4.25%, n=6). However, no statistically significant changes were observed with low glucose concentrations (2.75 mmol/L) (Fig 1⇓). Control contraction at the end of this experiment was 103.01±4.7%, n=6.
Insulin Effects on Vascular Contractility
We found a dose-dependent increase in force generated per arterial contraction induced by KCl (40 and 80 mmol/L) in arteries bathed with normal Tyrode’s solution (glucose 5.5 mmol/L) in the presence of insulin. The maximal increase was obtained with an insulin dose of 300 nmol/L, reaching 156.9±7.62%9 and 120.69±4.08% (n=10) of the control value with 40 and 80 mmol/L KCl, respectively (Fig 2⇓). Control contraction at the end of this experiment was 98.31±1.07%, n=6. Insulin effect was not present in arteries without endothelium, and instead, a nonsignificant decrease of the response of 9.5±7.8% (n=5) was observed. Insulin also tended to modify force development in arteries bathed with other glucose concentrations. Although there were no significant changes with 2.75 mmol/L glucose with any of the KCl doses used, insulin provoked a significant force increase in arteries bathed with 11 mmol/L glucose only when the 40 mmol/L KCl was used (Table⇓).
Mediation of Insulin Effects by Endothelin
An insulin-induced modification of vasoconstrictor substances such as endothelin by endothelial cells was tested in our experiments using a hyperimmune anti-endothelin serum and antagonists of endothelin receptors ETA and ETB, PD 151,242 and BQ-788, respectively. The presence of the antibody decreased the response of arteries bathed in normal Tyrode’s solution without insulin and stimulated by KCl. This effect was small and nonsignificant with KCl 40 mmol/L, and although it remained small with 80 mmol/L KCl, it was statistically significant. Insulin increased the contractile response to KCl, showing a higher response with 40 mmol/L KCl than with 80 mmol/L KCl. Although a slight increase in the contractile response of arteries to 40 mmol/L KCl in the presence of insulin and anti-endothelin antibody was still observed, the response to the hormone was reduced by approximately 73% in the presence of the antibody. The insulin-induced increase in contractile response to KCl was no longer present with 80 mmol/L KCl, with which a nonsignificant decrease in force development was found (Fig 3⇓). Contraction without insulin or antibody at the end of the experiment was 95.92±5.22%, n=6. Control experiments in which normal mouse serum was used to perfuse the arteries while they contracted in the presence of 40 mmol/L KCl showed no changes in the tension developed by the arteries in the presence and absence of insulin.
The mediation by endothelin of the insulin effects also was tested using antagonists of the endothelin receptors ETA and ETB. Both antagonists significantly reduced insulin-induced increases in the contractile response of arteries; PD 151,242 caused a 100% decrease in the response, and BQ-788 reduced the response to even less tension than that shown in the control situation without insulin. Furthermore, a decrease of the tension developed by 40 mmol/L of KCl in the absence of insulin also was observed with the ETB receptor antagonist BQ-788 (Fig 3⇑). Control contraction at the end of the experiments was reduced in 15% and 45% in the experiments using the ETA and ETB receptor antagonists, respectively.
Although there is a high prevalence of hypertension in diabetic patients,1 2 13 few studies have assessed the effects of insulin and glucose on smooth-muscle vascular contractility as the present study has. This article also analyzes the participation of endothelin in mediating insulin vascular responses.
The contractile response of arteries was modified by the different extracellular glucose levels, showing a significant decrease with high glucose concentrations, a condition similar to that found in IDDM. This effect was probably secondary to the inhibition of mechanisms that increase glucose transport15 in vascular smooth-muscle cells through mobilization of calcium ions that mediate contraction. It has been postulated previously that hyperglycemia impairs contractile activity in muscle16 by diminishing phosphoinositide metabolism.17 18 Although the effects of hyperglycemia could be due to an alteration in endothelial cell metabolism, causing an elevated release of vasodilator substances such as nitric oxide, this possibility was not tested in the present work. The absence of an increase in the contractile response in hypoglycemic conditions in our experiments could be the consequence of a lack of the necessary energy needed to induce a contraction when glucose concentrations are low. The results observed in this experiment suggest that, in the short term, hyperglycemia does not contribute to hypertension in IDDM patients but instead decreases tension generation.
Long-term effects of hyperglycemia have been described previously. These include increased vascular rigidity due to toxic effects on the endothelial cells that result in a decreased endothelium-mediated vascular relaxation, muscle cell hyperplasia, vascular remodeling, and fibronectin and collagen IV overexpression.2 These long-term effects of hyperglycemia could contribute, together with nephropathy, to hypertension in IDDM.
We found an increase in force generated by arterial contraction in the presence of insulin, and the effect was greater as insulin concentrations rose. The insulin effect was absent in arteries in which the endothelium was removed, suggesting participation of the endothelial cells in the insulin response. There have been reports showing that high insulin levels such as those found in patients with insulinoma do not cause hypertension.4 5 6 However, elevated plasma endothelin levels have been found in patients with diabetes mellitus7 and during euglycemic hyperinsulinemic clamp in lean NIDDM men.8 Furthermore, in some reports in which no correlation was found between hyperinsulinemia and hypertension, hypertriglyceridemia was reported to enhance the stimulatory effect of insulin on endothelin release.6 Studies using other species have shown a correlation between high insulin levels and hypertension,5 and in bovine and porcine endothelial cells, insulin stimulates the production and secretion of endothelin.10 11 Insulin induces an increase in ET-1 gene expression.9 Hypertension in diabetes could be the consequence of changes in intracellular calcium levels induced by insulin.1 2 3 13 Insulin has been reported to influence different membrane components that increase or decrease intracellular calcium levels in diverse cell types and contribute to the changes found in vascular smooth-muscle cell tension development.1 2 13 17 18 19 The increase in the contractile response induced by insulin also could be the consequence of a modification of endothelial cell secretion, which was tested in our experiments using a polyclonal anti-endothelin antibody and antagonists of the endothelin receptors ETA and ETB. The use of the antibody slightly decreased the response to 40 and 80 mmol/L KCl of arteries bathed in normal Tyrode’s solution without insulin, and the ETB-receptor antagonist reduced the response to 40 mmol/L KCl. This last effect, together with the enormous decrease in the response to insulin to less than the control produced by the ETB-receptor antagonist, suggests that depolarization of endothelial cells by KCl causes endothelin secretion, which stimulates the ETB receptor enhancing the contractile response of vascular smooth muscle. The increase in the contractile response in the presence of insulin was reduced significantly when the anti-endothelin antibody or the antagonist of the endothelin receptors ETA and ETB was added. These results suggest that endothelin mediates the effects of insulin and that the hormone stimulates its liberation from intracellular stores in endothelial cells in the artery maintained in the in vitro chamber in our experimental conditions. Although our antibody was produced against the human ET-1 molecule, we cannot discard its interaction with endothelin-2 or -3 or big endothelin due to high molecular homologies. However, the specificity of the antibody against molecules of the endothelin family is suggested by its having the same kind of effects as the endothelin receptor antagonists. The smaller effect of the antibody compared with that of the receptor antagonists could be due to low affinity or avidity of the antibody. The differences found with the two antagonists suggest that the response is mediated predominantly by the ETB receptor. However, the fact that both antagonists blocked the insulin effect might be the consequence of the loss of specificity to one receptor subtype due to the concentration used.
Insulin effect on arteries stimulated with 80 mmol/L KCl was statistically significant only when glucose concentration was 5.5 mmol/L, whereas when stimulated with 40 mmol/L KCl, the effect was significant with both 5.5 and 11 mmol/L glucose. The absence of an insulin effect with high glucose concentrations when the artery is stimulated with 80 mmol/L KCl could be due to compensation of increased force induced by insulin with the effect of high glucose levels that diminish the contractile response. This compensation is not observed with the lower KCl dose. Therefore, hypertension in patients with NIDDM who have increased levels of glucose and insulin could be a consequence of hyperinsulinemia and hyperglycemia when stimulation of the vessel is not maximal. Although the higher prevalence of hypertension in diabetes could be the result of the above-mentioned intracellular effects of insulin and glucose on smooth muscle contractility, other possible causes such as nervous and renal alterations1 3 19 have been reported.
In conclusion, our results suggest that high levels of both glucose (11 mmol/L) and insulin (300 nmol/L) contribute to hypertension in diabetes, particularly in NIDDM patients, and that the effects of insulin are mediated partly by endothelin.
Selected Abbreviations and Acronyms
|ETA||=||endothelin receptor subtype A|
|ETB||=||endothelin receptor subtype B|
|IDDM||=||insulin-dependent diabetes mellitus|
|NIDDM||=||non–insulin-dependent diabetes mellitus|
- Received December 30, 1996.
- Revision received January 16, 1997.
- Accepted March 14, 1997.
Levy J, Zemel MB, Sowers JR. Role of cellular calcium metabolism in abnormal glucose metabolism and diabetic hypertension. Am J Med. 1989;87(suppl 6A):7S-16S.
Epstein M, Sowers JR. Diabetes mellitus and hypertension. Hypertension.. 1992;19:403-418.
Modan M, Halkin H, Almog S, Lusky A, Eshkol A, Shefi M, Shitrit A, Fuchs Z. Hyperinsulinemia, a link between hypertension, obesity and glucose intolerance. J Clin Invest.. 1985;75:809-817.
Hall JE, Brands MW, Hildebrandt DA, Mizelle HL. Obesity-associated hypertension: hyperinsulinemia and renal mechanisms. Hypertension.. 1992;19:I-45-I-55.
Piatti PM, Monti LD, Conti M, Baruffaldi L, Galli L, Phan CV, Guazzini B, Pontiroli AE, Pozza G. Hypertriglyceridemia and hyperinsulinemia are potent inducers of endothelin-1 release in humans. Diabetes.. 1996;45:316-321.
Ferri C, Carlomagno A, Coassin S, Baldoncini R, Cassone Faldetta MR, Laurenti O, Properzi G, Santucci A, De Mattia G. Circulating endothelin-1 levels increase during euglycemic hyperinsulinemic clamp in lean NIDDM men. Diabetes Care.. 1995;18:226-233.
Oliver FJ, de la Rubia G, Feener EP, Lee ME, Loeken MR, Shiba T, Quertermous T, King GL. Stimulation of endothelin-1 gene expression by insulin in endothelial cells. J Biol Chem.. 1991;266:23251-23256.
Hu E-M, Levin ER, Pedram A. Insulin stimulates production and secretion of endothelin from bovine endothelial cells. Diabetes.. 1993;42:351-358.
Resnick L. Hypertension and abnormal glucose homeostasis. Am J Med. 1989;87(suppl 6A):17S-22S.
Suzuki E, Kono T. Evidence that insulin causes translocation of glucose transport activity to the plasma membrane from an intracellular storage site. Proc Natl Acad Sci U S A.. 1980;77:2542-2545.
Clausen T. The role of calcium in the activation of the glucose transport system. Cell Calcium.. 1980;1:311-325.
Avrameas S, Ternynek T, Guesdon JL. Coupling of enzymes to antibodies and antigens. Scand J Immunol. 1978;8(suppl 7):7-23.
Reaven GM, Hoffman BB. Hypertension as a disease of carbohydrate and lipoprotein metabolism. Am J Med. 1989;87(suppl 6A):2S-6S.