| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Studies |
University Department of Medicine and Therapeutics, Western Infirmary, Glasgow, Scotland G11 6NT
Address all correspondence and requests for reprints to: Dr. Shinichiro Ueda, Second Department of Medicine, Yokohama City University School of Medicine, 3–9 Fuku-ura, Kanazawa-ku, Yokohama 236, Japan. E-mail: sueda{at}med.yokohama-cu.ac.jp
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Studies that have examined the vascular effects of systemic hyperinsulinemia have consistently demonstrated skeletal muscle vasodilatation (7, 8, 9, 10, 12, 13, 14), and recent reports have suggested that this effect is physiologically relevant (14) and is mediated by nitric oxide (15, 16) and Na+,K+-adenosine triphosphatase (17). In contrast, studies that have examined the direct effect of local hyperinsulinemia in man have either shown no effect (18, 19, 20) or weak vasodilatation (21, 22). Several studies reporting a lack of effect of insulin on resting vascular tone have nevertheless demonstrated an effect to attenuate the vasoconstrictor effects of sympathetic activation (23) and vasoactive peptides (20).
Methodological differences (with respect to insulin dose and technique of flow measurement) taken together with interindividual heterogeneity in vascular responses to insulin may account for some of these discrepancies. For example, circulating concentrations of insulin during systemic hyperinsulinemia have often been supraphysiological; in addition, the vasodilation observed under these circumstances can be attributed to centrally mediated sympathetic changes. Local intraarterial infusions of insulin would be expected to circumvent these problems, but all previous studies have been open in design, and most investigators have failed to take account of the metabolic effect of insulin to decrease concentrations of glucose locally in the infused limb. We hypothesized that supplementation of intraarterial insulin with concomitant local (intraarterial) D-glucose might be more physiological; the principal aim of these studies was, therefore, to evaluate in a double blind, placebo-controlled manner the direct vascular effect of insulin alone and in association with physiological concentrations of D-glucose.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Seventeen normotensive healthy males, aged 20–40 yr, were recruited. All subjects had normal routine physical examination and laboratory screening tests; normal glucose tolerance was confirmed by a 75-g oral glucose tolerance test. The protocol was approved by the West Glasgow Hospitals University National Health Service Trust ethics committee, and written informed consent was obtained from each subject after a full explanation of the study was given.
General procedures
All studies were of a double blind, placebo-controlled, cross-over design. Volunteers attended on 3 separate study days for protocol 1 and on 2 study days for protocol 2 the Clinical Investigation Research Unit, Western Infirmary (Glasgow, Scotland). Studies were performed after an overnight fast, and subjects abstained from cigarettes, alcohol, and caffeine-containing drinks for at least 12 h before the study. Forearm blood flow was measured by bilateral venous occlusion strain gauge plethysmography (24) in a temperature-controlled (25 ± 1 C) quiet room. In this technique, flow is derived from the rate of change in forearm circumference during intermittent occlusion of venous return (with continued arterial inflow) using mercury in silicone rubber strain gauges. Pediatric cuffs (Hokanson SC5, PMS Instruments, Maidenhead, UK) were placed around the wrists and inflated to 200 mm Hg during each measurement period (3 min). Collecting cuffs (Hokanson SC10) were placed around the upper arms, inflated to 40 mm Hg, and deflated every 15 s throughout each measurement period. The strain gauges were 2 cm shorter than the maximum arm circumference and were placed 5 cm distal to the olecranon. A 27-gauge intraarterial needle (Coopers Needlework, Birmingham, UK) was placed in the brachial artery of the nondominant arm under local anesthesia (1% lignocaine). We were permitted by the local ethics committee (see above) to use only this size of needle; thus, we repeated cannulation three times at weekly intervals, but we were not able to sample arterial blood. Venous catheters were positioned in an ipsilateral antecubital deep vein and in a contralateral superficial vein for blood sampling. Deep vein cannulation was confirmed when the tip of the catheter was not palpable, and oxygen saturation of samples was less than 50%. Baseline forearm blood flow measurements were obtained 30 min after cannulation. Blood pressure and heart rate were measured by a semiautomatic sphygmomanometer (Dinamap, Critikon, Tampa, FL) in the control arm after each blood flow measurement.
Preparation of insulin solution
Four milliliters of each volunteer’s own blood were mixed with 45 mL saline and 1 mL diluted soluble human insulin (Actrapid, Novo Nordisk, Copenhagen, Denmark; 50 mU/mL for 1 mU/min and 250 mU/mL for 5 mU/min). The insulin solution was prepared in the sterile unit of the hospital pharmacy using glass syringes and bottles and was diluted in the research unit immediately before each study. With this method, the recovery of insulin was approximately 95% (data not shown).
Protocol 1: effect of insulin on forearm blood flow
Each volunteer (n = 9) received vehicle (saline and subject’s own blood, as described above), 1 mU/min insulin, and 5 mU/min insulin intraarterially for 90 min on separate study days. The infusion rate was 1 mL/min throughout. There was no concurrent glucose infusion. Forearm blood flow was measured every 10 min, and blood samples were taken from both ipsilateral and contralateral catheters every 10 min for the measurement of glucose and every 30 min for potassium and insulin determinations.
Protocol 2: effect of D- and L-glucose on insulin-mediated vasodilation
In a preliminary study, intraarterial infusion of glucose at 75 µmol/min was shown to maintain euglycemia during concomitant infusion of insulin (5 mU/min) in samples obtained from a deep vein of the infused arm (data not shown). Thus, we investigated the stereospecificity of the effect of glucose on the vasodilating effect of insulin by coinfusing D- or L-glucose (inactive isomer, Sigma Chemical Co., St. Louis, MO) on separate study days at 75 µmol/min for 30 min with saline and subsequently for 90 min with 5 mU/min insulin in a double blind, cross-over design (n = 8). Blood samples were withdrawn from the ipsilateral deep vein catheter every 10 min for measurement of glucose and every 30 min for measurement of insulin and potassium. The infusion rate (D/L-glucose plus saline) was 2 mL/min throughout.
Analysis of forearm blood flow data
All forearm blood flow data were obtained via a Mac Lab II Chart recorder (AD Instruments, London, UK). The percent change in the ratio of forearm blood flow from baseline was calculated. Serial measurements of blood flow were summarized as the mean response to avoid multiple comparisons (25).
Analysis of samples
Plasma glucose concentrations were measured by the (stereospecific) glucose oxidase method using a Beckman II glucose analyzer (Beckman, Fullerton, CA). Serum insulin concentrations were measured by direct RIA using a commercially available kit (Incstar, Stillwater, MN). Coefficients of variation for these assays were less than 5%.
Statistical analysis
Comparisons between summary measures of response to treatment were made by one-way ANOVA (protocol 1) and paired t test (protocol 2). Appropriate confidence intervals are quoted for these comparisons. In protocol 2, glucose, potassium, and insulin concentrations were compared between D- and L-glucose conditions using the treatment-time interaction in the three-way ANOVA, followed by Bonferonni-corrected paired t tests if appropriate at individual time points. Grouped data are shown as the mean ± SD.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Serum insulin concentrations reached a plateau within the first 30
min and remained constant thereafter (Table 1
). No change in insulin concentration
was noted in the control arm. Glucose and potassium concentrations
decreased in a dose-dependent manner in the infused arm (Table 1).
|
; F = 5.75 and
P < 0.01, by one-way ANOVA; 97.5% confidence
interval, 7.9, 35.7). Infusion of the lower dose (1 mU/min) did not
result in a statistically significant difference in blood flow from
that after placebo treatment.
|
The mean ± SD change in forearm blood flow (Fig. 2) during infusion of insulin with
D-glucose (47.1 ± 21.3%) was significantly greater
than that during infusion of insulin and L-glucose
(6.1 ± 12.2%; P < 0.01; 95% confidence
interval, 23.1, 58.8). Infusion of both D- and
L-glucose alone before insulin administration tended to
reduce forearm blood flow, but there was no significant difference in
this response between the two treatments. Plasma D-glucose
concentrations in blood withdrawn from the ipsilateral deep vein
catheter were significantly higher during infusion of
D-glucose than during infusion of L-glucose
(F = 4.06; P < 0.05 for treatment-time
interaction; P < 0.001 for all individual time
points). There were no differences in serum potassium or insulin
concentrations between the two treatments (Fig. 3).
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
As outlined above, the available data on the vascular effects of insulin are conflicting; differences between methods used for the measurement of limb blood flow (plethysmography vs. thermodilution), the limb chosen for study (upper vs. lower), and the dose and time course of the insulin infusion may explain some of the discrepancies. However, a major difference differentiating positive and negative studies is the route of insulin administration. Although there are now many reports of insulin-mediated vasodilation during systemic hyperinsulinemia in euglycemic clamp studies (7, 8, 9, 10, 12, 13, 14), local administration of insulin is reported to produce either no change in blood flow (18, 19, 20) or weak vasodilation (21, 22).
When insulin alone was administered for 90 min in the present study (protocol 1), a relatively small (20%) change in forearm blood flow was observed. Although this response was statistically significant and potentially physiologically relevant, forearm blood flow varies considerably over time as a result of nonexperimental stimuli, and it is conceivable that it would not have been detected had a placebo comparison not been incorporated, if a control limb had not been studied (29), or if the main end point of the study had been more complex. Although the hypothesis that insulin is a direct local vasodilator was supported by the data, the observed changes in blood flow, even at the highest dose of insulin, were considerably smaller (20% increase) than those obtained from studies using systemic insulin administration (50–100% increase) (26). We hypothesize that the reason for this discrepancy is that euglycemic conditions are maintained during systemic hyperinsulinemic clamp studies, whereas a nonphysiological dissociation (local hyperinsulinemia and mild local hypoglycemia) occurs in local studies. Under physiological circumstances, insulin is secreted in response to a considerable postprandial rise in blood glucose; thus, hyperinsulinemia in any vascular bed is likely to be accompanied by an increase in the supply of substrate (glucose). If the vasodilating effect of insulin was dependent on stimulation of tissue glucose uptake, this effect might be masked during isolated local hyperinsulinemia. Alternatively, local hypoglycemia might itself affect the response in some way. Having considered these lines of reasoning, we designed the second experiment to assess the effects of glucose on insulin-mediated vasodilation.
Local supplementation of glucose to maintain blood glucose levels during local hyperinsulinemia was associated with a vasodilator response similar to that observed in studies employing systemic hyperinsulinemia and euglycemia (47%). This is not due simply to an osmotic effect of glucose, as glucose infusion per se did not increase forearm blood flow; indeed, infusions of both D- and L-glucose in the absence of insulin tended to reduce blood flow.
The present results highlight both the importance of local vs. systemic insulin administration and the role of tissue glucose uptake; they appear to explain previous discrepancies in the literature regarding the presence or absence of a physiologically relevant vasodilator effect of insulin. However, they contrast in some respects with data from one previous, carefully performed, open study (22). Tack et al. studied local insulin-mediated vasodilation for a longer time course (180 min) in the forearm vascular bed of healthy male volunteers using bilateral plethysmography, achieving insulin concentrations in the ipsilateral venous effluent similar to those reported herein. The vasodilator response, corrected for increased flow in the contralateral arm, was approximately 12% by 90 min and approximately 40% by 180 min, i.e. at 90 min the effect observed was slightly smaller than (but compatible with) that observed in the present study. These findings highlight an additional factor distinguishing positive and negative studies; most local insulin infusion studies examined a shorter time course than systemic studies.
Like ourselves, Tack et al. were intrigued by the
observation that local glucose concentrations in the ipsilateral venous
effluent fell during local insulin infusion (by 2 mmol/L) and attempted
to assess the importance of tissue glucose uptake on insulin-mediated
blood flow by coinfusing insulin and glucose in a subgroup of
individuals. With a dose of glucose over 50 times higher than our own,
resulting in glucose concentrations of 15 mmol/L (3 times normal) in
the ipsilateral venous effluent, no augmentation of flow was observed.
To account for the apparent conflict with our own results, we can only
speculate that the dose response for augmentation of insulin-mediated
vasodilation by glucose may not be linear; indeed, there may be a
paradoxical effect at very high concentrations. In view of the lack of
augmentation of insulin vasodilation by glucose in their study, Tack
et al. argue that local hyperinsulinemia alone is a
sufficient model of physiological insulin-mediated vasodilation,
stating that venous hypoglycemia occurs both in these studies and
during euglycemic clamp studies (which are generally conducted
according to arterial or arterialized samples). Our own simultaneous
arterialized and venous glucose data during systemic euglycemic
hyperinsulinemia, collected in the course of a study examining the
methodology of arterialization (30), indicate that differences are much
smaller (
0.6 mmol/L) than those observed by Tack et al.
between arterial and venous samples during local hyperinsulinemia.
Hence, we suggest that coinfusion of a low dose of glucose is a more
physiological model than local hyperinsulinemia alone; in addition, our
model allows experiments to be conducted over a shorter time course.
The latter point is particularly desirable in studies using
plethysmography, as measurement error becomes higher at the end of
prolonged studies (25).
Therefore, the results of the present studies provide new insights into mechanisms underlying insulin-mediated vasodilation. They are consistent with in vitro data demonstrating that insulin-mediated vasorelaxation is augmented in the presence of D-glucose (31, 32). Although tissue glucose uptake was not measured (for ethical reasons outlined in Materials and Methods), the stereospecificity of the response supports the concept that glucose uptake is an early and essential step for insulin-induced vasorelaxation (L-glucose is not recognized by glucose transporter proteins). Glucose-induced changes in the intracellular environment (for example, changes in pH by aerobic glycolysis) may alter intracellular signaling and promote active vasodilation.
Insulin itself appears to have relatively minor vasodilating effects at physiological concentrations in the fasting state, but a greater effect during the postprandial state (i.e. both insulin and glucose present in physiological concentrations). It has been proposed that insulin-mediated vasodilation may be an independent determinant of insulin-stimulated glucose uptake (11, 14, 33). Our results can be interpreted to support the concept that glucose uptake acts as a positive feedback regulating factor for insulin-mediated vasodilation. Therefore, we hypothesize that reduced insulin-stimulated glucose uptake may not only result from but may also exacerbate impaired insulin-mediated vasodilation, and this may be relevant to altered vascular tone in insulin-resistant states such as hypertension. This would reconcile two apparently conflicting observations: in patients with noninsulin-dependent diabetes mellitus, insulin-mediated vasodilation is blunted with respect to controls when glucose levels are clamped at euglycemia (5.2 mmol/L) (34), but insulin-mediated vasodilation is maintained when each patient’s own basal fasting glucose level is maintained by a higher rate of glucose infusion (35).
In conclusion, these double blind, controlled studies 1) confirm that insulin alone at physiological concentrations is a weak vasodilator, and 2) demonstrate that concomitant infusion of a physiological dose of D-glucose significantly augments this response. These findings may explain discrepancies in the results of previously reported studies examining the vascular effects of insulin in local and systemic models of hyperinsulinemia. The present experimental set-up may create an experimental model appropriate for clarifying the mechanisms underlying the direct vasodilating effects of insulin, because systemic effects of insulin are avoided and local conditions are more physiological, mimicking the postprandial state. The present data provide strong evidence for the hypothesis that insulin exerts a direct, locally mediated, vasodilating effect that is amplified stereospecifically by concomitant tissue glucose uptake and metabolism. Further mechanistic studies investigating the interaction between glucose uptake and different regulators of vascular tone in endothelial/vascular smooth muscle tissue are required in normal subjects and patients with insulin-resistant states.
|
Footnotes |
|---|
Received December 17, 1997.
Revised March 4, 1998.
Accepted March 11, 1998.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
vasoreactivity in obese
hypertensive or lean normotensive men. Hypertension. 22:584–590.2-adrenergic pathway in human. Hypertension. 24:429–438.This article has been cited by other articles:
 |
|
 |
|
  R. Muniyappa, M. Montagnani, K. K. Koh, and M. J. Quon Cardiovascular Actions of Insulin Endocr. Rev., August 1, 2007; 28(5): 463 - 491. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. S. Perlstein, M. Gerhard-Herman, N. K. Hollenberg, G. H. Williams, and A. Thomas Insulin Induces Renal Vasodilation, Increases Plasma Renin Activity, and Sensitizes the Renal Vasculature to Angiotensin Receptor Blockade in Healthy Subjects J. Am. Soc. Nephrol., March 1, 2007; 18(3): 944 - 951. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Sjostrand, S. Gudbjornsdottir, L. Strindberg, and P. Lonnroth Delayed Transcapillary Delivery of Insulin to Muscle Interstitial Fluid After Oral Glucose Load in Obese Subjects Diabetes, January 1, 2005; 54(1): 152 - 157. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S Jadhav, J Petrie, W Ferrell, S Cobbe, and N Sattar Insulin resistance as a contributor to myocardial ischaemia independent of obstructive coronary atheroma: a role for insulin sensitisation? Heart, December 1, 2004; 90(12): 1379 - 1383. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. Randriamboavonjy, J. Schrader, R. Busse, and I. Fleming Insulin Induces the Release of Vasodilator Compounds From Platelets by a Nitric Oxide-G Kinase-VAMP-3-dependent Pathway J. Exp. Med., February 2, 2004; 199(3): 347 - 356. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. G. Clark, M. G. Wallis, E. J. Barrett, M. A. Vincent, S. M. Richards, L. H. Clerk, and S. Rattigan Blood flow and muscle metabolism: a focus on insulin action Am J Physiol Endocrinol Metab, February 1, 2003; 284(2): E241 - E258. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Rask-Madsen, N. Ihlemann, T. Krarup, E. Christiansen, L. Kober, C. Nervil Kistorp, and C. Torp-Pedersen Insulin Therapy Improves Insulin-Stimulated Endothelial Function in Patients With Type 2 Diabetes and Ischemic Heart Disease Diabetes, November 1, 2001; 50(11): 2611 - 2618. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Mather, M. Laakso, S. Edelman, G. Hook, and A. Baron Evidence for physiological coupling of insulin-mediated glucose metabolism and limb blood flow Am J Physiol Endocrinol Metab, December 1, 2000; 279(6): E1264 - E1270. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. J. Cleland, J. R. Petrie, M. Small, H. L. Elliott, and J. M. C. Connell Insulin Action Is Associated With Endothelial Function in Hypertension and Type 2 Diabetes Hypertension, January 1, 2000; 35(1): 507 - 511. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Ueda, J. R. Petrie, S. J. Cleland, H. L. Elliott, J. M.C. Connell, C. Cardillo, and J. A. Panza Vasodilator Response to Local Hyperinsulinemia • Response Hypertension, December 1, 1999; 34 (6): e12 - e13. [Full Text] [PDF]
|
|
|
|
|
|
N. J. Brown, J. V. Gainer, C. M. Stein, and D. E. Vaughan Bradykinin Stimulates Tissue Plasminogen Activator Release in Human Vasculature Hypertension, June 1, 1999; 33(6): 1431 - 1435. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Zierler Whole body glucose metabolism Am J Physiol Endocrinol Metab, March 1, 1999; 276(3): E409 - E426. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. J. Cleland, J. R. Petrie, S. Ueda, H. L. Elliott, and J. M. C. Connell Insulin-Mediated Vasodilation and Glucose Uptake Are Functionally Linked in Humans Hypertension, January 1, 1999; 33(1): 554 - 558. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
