Author's current address V. A. Koivisto: Lilly Research Laboratory, Hamburg, Germany.
1The role of blood flow as a determinant of skeletal muscle glucose uptake is at present controversial and results of previous studies are confounded by possible direct effects of vasoactive agents on glucose uptake. Since increase in muscle blood flow can be due to increased flow velocity or recruitment of new capillaries, or both, it would be ideal to determine whether the vasoactive agent affects flow distribution or only increases the mean flow.
2In the present study blood flow, flow distribution and glucose uptake were measured simultaneously in both legs of 10 healthy men (aged 29 ± 1 years, body mass index 24 ± 1 kg m−2) using positron emission tomography (PET) combined with [15O]H2O and [18F]fluoro-2-deoxy-D-glucose (FDG). The role of blood flow in muscle glucose uptake was studied by increasing blood flow in one leg with sodium nitroprusside (SNP) and measuring glucose uptake simultaneously in both legs during euglycaemic hyperinsulinaemia (insulin infusion 6 pmol kg−1 min−1).
3SNP infusion increased skeletal muscle blood flow by 86 % (P < 0·01), but skeletal muscle flow distribution and insulin-stimulated glucose uptake (61·4 ± 7·5 vs. 67·0 ± 7·5 μmol kg−1 min−1, control vs. SNP infused leg, not significant), as well as flow distribution between different tissues of the femoral region, remained unchanged. The effect of SNP infusion on blood flow and distribution were unchanged during infusion of physiological levels of insulin (duration, 150 min).
4Despite a significant increase in mean blood flow induced by an intra-arterial infusion of SNP, glucose uptake and flow distribution remained unchanged in resting muscles of healthy subjects. These findings suggest that SNP, an endothelium-independent vasodilator, increases non-nutritive, but not nutritive flow or capillary recruitment.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
The role of insulin-stimulated blood flow as a determinant of glucose uptake has been studied mainly by increasing or decreasing blood flow with various vasoactive agents and then determining the impact of such an intervention on glucose uptake. Whilst some studies support the role of blood flow as a determinant of insulin-stimulated skeletal muscle glucose uptake (Baron et al. 1994, 1995), several studies contest it (Natali et al. 1994; Scherrer et al. 1994; Jamerson et al. 1996; Nuutila et al. 1996). Interpretation of these studies is, however, confounded by possible direct effects of vasoactive agents on glucose uptake since NG-monomethyl-L-arginine (L-NMMA) has been demonstrated to decrease glucose transport in isolated rat skeletal muscle preparations (Balon & Nadler, 1994) and adenosine to increase glucose transport in human adipocytes (Heseltine et al. 1995). Additionally, the metabolic effects of vasoactive agents may depend more on their effect on flow distribution, i.e. flow heterogeneity, than total flow (Clark et al. 1995), while capillary perfusion heterogeneity and net capillary exchange can be altered without changes in the mean flow into an organ (Duling & Damon, 1987; Clark et al. 1995). However, in all of the previous studies only the effect of vasoactive agent on mean blood flow, not flow distribution, has been studied.
Use of positron emission tomography (PET) combined with 15O-labelled water ([15O]H2O) and 2-[18F]fluoro-2-deoxy-D-glucose ([18F]FDG) allows quantification of both blood flow and glucose uptake directly in human skeletal muscle in vivo (Nuutila et al. 1996; Ruotsalainen et al. 1997). Recent refinements in reconstruction methods of PET images (Alenius & Ruotsalainen, 1997) enables pixel-by-pixel quantification of regional muscle blood flow, and measurement of absolute and relative dispersion of blood flow and the shape of its frequency distribution as indexes of flow heterogeneity. In this study we used these techniques to examine whether enhancement of blood flow by an endothelium-independent vasodilator sodium nitroprusside (SNP) changes flow distribution or augments insulin-stimulated glucose uptake in healthy subjects. Additionally, the effects of insulin and SNP infusions on distribution of blood flow between different tissue compartments of the femoral region were studied.
Ten male volunteers (aged 29 ± 1 years, body mass index 24 ± 1 kg m−2, means ±s.e.m.) were studied. The subjects were healthy, as judged by history, physical examination and routine laboratory tests, and were taking no medication. The nature, purpose and potential risks of the study were explained to all subjects before they gave their written informed consent to participate. The study was performed according to the Declaration of Helsinki and approved by the ethical committee of Turku University Hospital.
The study design is shown in Fig. 1. All studies were performed after a 10–12 h overnight fast. Subjects were lying supine throughout the study. Three catheters were inserted, one in an antecubital vein for infusions of glucose and insulin and injections of [15O]H2O and [18F]FDG, another in the opposite radial artery for blood sampling, and a third in a femoral artery for infusion of sodium nitroprusside. The study for each subject consisted of a 90 min basal period (-90 to 0 min) and a 150 min normoglycaemic hyperinsulinaemic period (0–150 min). Blood flow was measured simultaneously in both legs using [15O]H2O and positron emission tomography (PET). After the basal blood flow measurement the first intra-arterial SNP infusion was started for 15 min (3.3 μg min−1 in the first 3 subjects, 6.7 μg min−1 in the following 7 subjects) and thereafter flow measurement was repeated. At 0 min an intravenous infusion of insulin (6 pmol kg−1 min−1) and at 90 min an intra-arterial SNP infusion were started, and continued to 150 min (Fig. 1). At 105 min blood flow measurements of both legs were repeated. Femoral muscle glucose uptake was measured immediately thereafter using [18F]FDG and PET (Fig. 1). Blood pressure (Hem-705C oscillometric blood pressure monitor; Omron Corp., Tokyo, Japan) and heart rate were recorded every 15–30 min during the insulin and SNP infusions. Blood samples for the measurements of serum insulin concentrations were taken as detailed below.
Measurement of the femoral region composition and localisation of different tissues
The patient lay supine while axial magnetic resonance (MR) images of both thighs were created in a 0.1 T MR imager (Merit, Picker Nordstar, Helsinki, Finland) equipped with a body coil using a T1-weighted gradient echo PS (partial saturation) 350/20 pulse sequence. The displayed matrix size was 256 pixels × 256 pixels and the slice thickness 10 mm. The images were analysed on the screen using the region-of-interest facility. The areas of femur, muscle tissue and subcutaneous fat were measured separately.
Production of positron emitting tracers and emission scanning
[15O]H2O (Nuutila et al. 1996; Ruotsalainen et al. 1997) and [18F]FDG (Hamacher et al. 1986) were produced as previously described. An eight-ring ECAT 931/08 tomograph (Siemens/CTI Corp., Knoxville, TN, USA) was used for scanning. The subject was positioned in the scanner with the femoral region within the gantry. Before emission scanning, a transmission scan for correction of photon attenuation in the body was performed for 15 min with a removable ring source containing 68Ge (total counts 15 × 106 to 30 × 106 in a plane). All data were corrected for dead time, decay and measured photon attenuation and reconstructed into a 128 pixel × 128 pixel matrix. The flow data were reconstructed with the use of a recently developed Bayesian iterative reconstruction algorithm with median root prior (the MRP method) with 150 iterations and a Bayesian coefficent of 0.3 (Alenius & Ruotsalainen, 1997). The glucose data were reconstructed using the traditional filtered-back-projection (FBP) method and Hann-filtering with a cut-off frequency of 0.5 cycles s−1.
Measurement of blood flow
For the flow studies [15O]H2O (30–45 mCi) was infused intravenously (30 s) and a dynamic scan, for 6 min, was started simultaneously. To obtain the input function, arterial blood was withdrawn with a pump at a speed of 6 ml min−1 from the radial artery, and the radioactivity concentration was measured using a two channel detector system (Scanditronix, Uppsala, Sweden), which was calibrated to the well counter (Bicron 3MW3/3, Bicron Inc., Newbury, OH, USA) and the PET scanner. The delay between the input curve and the tissue curve was solved by fitting (Ruotsalainen et al. 1997). Blood flow was calculated with a single compartment model and an autoradiographic method as previously described in detail (Nuutila et al. 1996; Ruotsalainen et al. 1997). Blood flow was calculated pixel by pixel into parametric flow images with a 200 s tissue integration time as previously described (Ruotsalainen et al. 1997). In the parametric flow images the smallest surface area after the MRP reconstruction was 5 mm × 5 mm × 7.3 mm (= 0.18 cm3) (Alenius et al. 1997).
Measurement of glucose uptake
For the [18F]FDG study, 4 mCi [18F]FDG was injected intravenously over 2 min and dynamic scanning for 30 min was started (12 × 15 s, 4 × 30 s, 3 × 60 s, 1 × 120 s, 4 × 300 s, n= 6 subjects). Blood samples for measurement of plasma radioactivity were withdrawn once during each frame time. The radioactivity concentration in plasma samples was measured with the well counter calibrated to PET. The three-compartment model of [18F]FDG kinetics was used as described previously (Patlak & Blasberg, 1985; Gambhir et al. 1989; Nuutila et al. 1992). Plasma and tissue time-activity curves were analysed graphically to measure the fractional rate of tracer phosphorylation (Ki). The rate of the glucose uptake (rGU) is obtained by multiplying Ki by the plasma glucose concentration ([Glc]p) divided by a lumped constant term (LC): rGU = ([Glc]p/LC)Ki. Based on recent human data, a lumped constant value of 1.2 for skeletal muscle was used (Utriainen et al. 1998; Peltoniemi et al. 1999). Glucose extraction (arteriovenous glucose difference) was calculated using Fick's equation by dividing muscle glucose uptake by muscle blood flow.
Regions of interest
Regions of interest (ROIs) were drawn in muscle compartments, in bone and fat, around the great vessels, and over the whole cross-sectional limb. Muscle ROIs were drawn in the posterior, anterolateral and anteromedial muscular compartments of the thigh on four transaxial slices in both legs. The localisation of the muscle compartments was verified by comparison of the flow images with the transmission image and the MR image. ROIs for bone and bone marrow were drawn in the transmission image as a circular form covering the whole bone area. The thickest area of subcutaneous fat was located posterolaterally in the MR images and an elliptical subcutaneous ROI was drawn there. A small ROI around the femoral artery was localised in the flow image. The ROI covering for whole limb cross-section was determined according to a phantom study. For determination of the edge of the limb cross-section a phantom study was performed (Raitakari et al. 1996). The edge of the phantom could be formed with pixels of the value of 30 % of the maximum in the transmission image (photon attenuation image), consistent with brain phantom studies performed previously in our laboratory (Eronen et al. 1995). Identical ROIs were used for flow and glucose uptake studies.
Heterogeneity analysis of blood flow
Distribution of blood flow was constructed from the pixel-by-pixel images. ROIs were drawn on the muscle compartments as described above and used for heterogeneity analysis. Muscle ROI comprised on average 65.2 ± 11.7 cm3. The standard deviation (s.d.) of flow values was used to characterise the absolute dispersion of flow and the coefficient of variation (c.v.) was used as a measure of the relative dispersion of flow or the true flow heterogeneity (Duling & Damon, 1987; Vicini et al. 1997). The c.v. for blood flow was calculated by dividing the s.d. by its respective mean value (s.d./mean flow) (Duling & Damon, 1987). However, it is possible that the coefficient of variation remains constant in response to an intervention, while the shape of the distribution changes. The latter would also indicate redistribution of blood flow (Duling & Damon, 1987). This possibility was evaluated by visually analysing the shape of the histograms depicting relative flow.
Whole body glucose uptake
Whole body glucose uptake was determined independently of the PET measurements using the euglycaemic hyperinsulinaemic clamp technique, as described previously (DeFronzo et al. 1979). During hyperinsulinaemia normoglycaemia was maintained using a variable rate of infusion of 20 % glucose based on arterial plasma glucose concentrations (Kadish et al. 1968). Serum insulin concentrations were measured at 30 min intervals during hyperinsulinaemia (Kuzuay et al. 1977).
Results are given as means and standard errors of the mean (s.e.m.). Analysis of variance for repeated measures (ANOVA) was used. When significant interaction was found, analyses were continued by comparing means using Student's paired t test. A probability level of 0.05 was considered significant. Associations between selected variables were studied by calculating Pearson's correlation coefficients. Calculations were made using the SAS statistical analysis system (SAS Institute Inc., Gary, NC, USA).
Glucose and insulin concentrations
Plasma glucose concentration was 5.8 ± 0.4 mmol l−1 basally and 5.6 ± 0.9 mmol l−1 during hyperinsulinaemia. Serum insulin concentrations averaged 37 ± 7 pmol l−1 in the fasting state and 456 ± 144 pmol l−1 during hyperinsulinaemia.
Composition of femoral regions measured with MRI
On average 72 ± 3 % of the femoral region composition was muscle tissue, 25 ± 3 % subcutaneous fat and skin, 2.5 ± 0.3 % bone and bone marrow, and 0.5 ± 0.02 % large vessels.
Muscle blood flow
Muscle blood flow was comparable in the two legs before SNP and insulin infusions (difference 0.2 ± 0.2 ml (100 g muscle)−1 min−1, catheterised vs. control leg). SNP increased muscle blood flow by 86 % from 2.2 ± 0.3 to 4.1 ± 0.8 ml (100 g)−1 min−1 (P < 0.05). During hyperinsulinaemia SNP infusion increased muscle blood flow by 72 % (from 2.5 ± 0.4 to 4.3 ± 0.7 ml (100 g)−1 min−1, P < 0.05). Muscle blood flow was not significantly increased by insulin infusion. During SNP and insulin infusions it was comparable to the flow during SNP infusion alone.
Heterogeneity of muscle blood flow
Examples of frequency distributions of absolute flows from one subject are shown in Fig. 2. In response to SNP absolute dispersion of skeletal muscle blood flow increased significantly from 0.7 ± 0.1 to 1.8 ± 0.6 ml (100 g)−1 min−1 (P < 0.01). During hyperinsulinaemia SNP infusion induced a comparable increase to SNP infusion alone in absolute dispersion (from 0.7 ± 0.1 to 1.5 ± 0.5 ml (100 g)−1 min−1, P < 0.01). Relative dispersion of muscle blood flow remained unchanged during both SNP infusion and SNP and insulin infusions (0.3 vs. 0.3 vs. 0.3, control vs. SNP vs. SNP and insulin, Fig. 2).
During hyperinsulinaemia skeletal muscle glucose uptake was unchanged in the SNP-infused leg when compared to the control leg (60.7 ± 7.9 vs. 56.4 ± 5.9 μmol kg−1 min−1, not significant, n.s.). Skeletal muscle glucose uptake covered 80 % of the whole femoral region glucose uptake in the control leg and 84 % in the SNP-infused leg. Muscle glucose extraction was 38 % lower in the SNP-infused than in the control leg (1.6 ± 0.3 vs. 2.5 ± 0.3 mmol l−1, P < 0.01). Whole body glucose uptake (31.4 ± 2.3 μmol kg−1 min−1) correlated closely with skeletal muscle glucose uptake both in the control leg (r= 0.96, P < 0.01) and in the SNP-infused leg (r= 0.85, P < 0.05). Examples of flow and glucose uptake images are shown in Fig. 3.
Distribution of blood flow between different leg tissues
In the basal state mean blood flow in the femoral region was similar in the two legs (2.3 ± 0.6 and 2.4 ± 0.6 ml (100 g)−1 min−1, catheterised vs. control leg, n.s.). Muscle tissue accounted for most of the total blood flow in both legs (77 ± 4 % in the control leg and 78 ± 4 % in the SNP-infused leg, n.s., Fig. 4, Table 1). Intrafemoral SNP infusion both with and without simultaneous insulin infusion induced significant and comparable flow increments in muscle, fat and bone tissues without changing the distribution of blood flow between these tissues (Table 1).
Table 1. Blood flow rates (in ml (100 g tissue)−1 min−1) in the basal state (Control) and during infusion of insulin, SNP and SNP plus insulin in the crosssectional thigh region (Whole leg), femoral muscles, bone (including bone marrow) and subcutaneous fat
The primary aim of this study was to establish whether vasodilatation induced by an endothelium-independent vasodilator, SNP, changes skeletal muscle blood flow distribution or glucose uptake. We found that intra-arterial infusions of SNP increased mean muscle blood flow and absolute, but not relative, flow dispersion. Despite an SNP-induced significant increase in mean blood flow, insulin-stimulated glucose uptake remained unchanged in resting muscles of healthy subjects. Additionally, flow distribution between different tissues of the femoral region, i.e. between muscle, fat and bone tissues, remained unchanged during intrafemoral SNP infusion with and without simultaneous insulin infusion.
Use of PET and [15O]H2O enables quantification of low flow rates, e.g. flow in resting skeletal muscle (Ruotsalainen et al. 1997), and presentation of blood flow in pixel-by-pixel parametric images. In resting skeletal muscle, the relation between tissue radioactivity and blood flow is linear after a bolus injection of [15O]H2O (Ruotsalainen et al. 1997). Therefore the calculated flow in the volume of measurement represents true average flow. The MRP method is a new reconstruction algorithm that reduces noise and improves image quality without compromising spatial resolution (Alenius & Ruotsalainen, 1997). The better image quality enables quantification of absolute and relative muscle flow distributions as indices of flow heterogeneity.
Flow heterogeneity can be defined simply as an uneven distribution of flow amongst perfused vessels (Duling & Damon, 1987). The physiological significance of flow heterogeneity has been linked with efficiency of tissue oxygenation and capillary exchange of diffusible solutes (Ellis et al. 1994). The possibility of distinguishing between flow distribution (flow heterogeneity) and mean skeletal muscle blood flow is of importance because capillary perfusion heterogeneity and net capillary exchange can be altered without changing the mean flow into an organ (Duling & Damon, 1987; Clark et al. 1995). Furthermore, the metabolic effects of vasoactive agents might depend more on their effect on flow heterogeneity, than total flow (Clark et al. 1995). The standard deviation of blood flow is used as an index of absolute flow dispersion, while the relative dispersion is used as a true measure of flow heterogeneity and is thought to reflect the extent to which some vessels receive more and some less flow than their appropriate fraction of total (Duling & Damon, 1987).
In the present study we measured relative blood flow dispersion directly in skeletal muscle in the basal state and during an intrafemoral SNP infusion with and without simultaneous physiological insulin infusion in healthy humans. Relative blood flow dispersion and normalised blood flow distribution histograms remained unchanged suggesting that the fractional distribution of muscle blood flow remained unchanged as the mean blood flow increased. Values of relative flow dispersion of this study are of the same magnitude as those previously found in humans using PET technique (Utriainen et al. 1997; Vicini et al. 1997) or in animals using labelled microspheres (Duling & Damon, 1987; Iversen et al. 1989). While relative dispersion of blood flow remained unchanged absolute dispersion of blood flow increased during SNP infusions with and without simultaneous insulin infusion. This finding could be interpreted to reflect the increment of mean flow since absolute dispersion changes as a function of mean flow. The smallest surface area that can be quantified after the MRP reconstruction process is 5 mm × 5 mm × 7.3 mm (0.18 cm3) (Alenius et al. 1997) and thus the values of flow heterogeneity in this study measure heterogeneity within a region of that size and not at capillary level. However, the PET methodology has advantages compared to some previous approaches, since it is suitable for use in humans and no invasive manipulation of tissue or innervation is required.
Both in the basal state and during SNP and insulin infusions muscle tissue accounted for most (about 80 %) of the total blood flow of the femoral region. This finding is consistent with results of our previous study (Raitakari et al. 1996), where skeletal muscle blood flow was also found to be the major location for insulin-stimulated blood flow and to relate to the whole leg flow values. In the present study blood flow of subcutaneous fat accounted for about 15 % of the total limb blood flow and subcutaneous flow values were comparable to previously reported values measured using 133Xe (2.4 ml (100 g)−1 min−1; Jansson & Lönnroth, 1995). In PET images thin skin may be difficult to distinguish from subcutaneous fat and thus blood flow of skin might have been slightly underestimated in the present study. However, since skin constituted only 5 % of the total cross-sectional area of the tissues of the femoral region, the possible underestimation of skin blood flow had only a minor effect on blood flow values of the whole femoral region.
In the present study blood flow in one leg was increased by an endothelium-independent vasodilator SNP and thereafter glucose uptake was quantified simultaneously on both legs. Despite a significant increase in mean muscle blood flow, glucose uptake remained unchanged. When glucose extraction was calculated by dividing skeletal muscle glucose uptake by muscle blood flow, we found glucose extraction to be 36 % lower in the SNP-infused than in the control leg. Thus, in resting skeletal muscle the increased glucose supply due to increased blood flow was compensated for by decreased glucose extraction. Taken together with the finding of unchanged relative flow dispersion, i.e. true flow heterogeneity, these findings suggest that vasodilatation induced by SNP increased non-nutritive flow. Furthermore, reduced glucose extraction with simultaneously increased flow and glucose delivery might reflect an autoregulatory mechanism tending to maintain the glucose influx at levels appropriate for the needs of resting muscle cells. Earlier studies performed using indirect calorimetry (DeFronzo et al. 1981) or nuclear magnetic resonance spectroscopy (Shulman et al. 1990) have suggested that the majority of peripheral glucose disposal during hyperinsulinaemia is accounted for by muscle glycogen synthesis. The current results using PET and [18F]FDG confirm these findings.
Previous data regarding the role of blood flow in regulating insulin-stimulated glucose uptake are controversial. In studies where either bradykinin (Nuutila et al. 1996) or adenosine (Natali et al. 1994) has been used intra-arterially to increase blood flow, insulin-stimulated glucose uptake has remained unchanged. A significant increase in blood flow induced by an intra-arterial infusion of bradykinin did not change insulin-stimulated glucose uptake in normal (Nuutila et al. 1996) or in obese subjects (Laine et al. 1998). Similarly, a twofold increase in forearm blood flow by adenosine did not enhance insulin-stimulated glucose uptake in hypertensive subjects (Natali et al. 1994). A decrease in insulin-stimulated blood flow by an intra-arterial infusion of L-NMMA had no effect on insulin-stimulated whole body glucose uptake in normal subjects when studied by Scherrer et al. (1994), while Baron et al. (1995) reported that a decrease in blood flow by an intra-arterial infusion of L-NMMA decreased insulin stimulated leg glucose uptake. However, conclusions from studies using L-NMMA or adenosine to modulate blood flow should be interpreted with caution, as it has been demonstrated that L-NMMA decreases glucose transport in isolated rat skeletal muscle preparations (Balon & Nadler, 1994) and that adenosine increases glucose transport in human adipocytes (Heseltine et al. 1995). Bradykinin has been demonstrated to have no effects on glucose uptake in isolated myocytes (Rosen et al. 1983). Recently, in an experimental study of isolated rat skeletal muscle preparations SNP was found to increase rates of glucose oxidation and decrease glycogen synthesis (Young & Leighton, 1998). However, if rates of glucose oxidised and used for glycogen synthesis reported in that study are summed up to present glucose uptake, SNP increased glucose uptake rates by only 6 % in that experimental study. In the present study glucose uptake, not glucose oxidation nor glycogen synthesis was quantified using [18F]FDG and PET, and found to be unchanged despite a significant flow increment induced by SNP. Thus, at present the majority of studies argue against the hypothesis that changes in blood flow regulate insulin stimulated glucose uptake.
In conclusion, we examined the effect of an intrafemoral infusion of SNP, an endothelium-independent vasodilator, on leg blood flow, flow distribution and glucose uptake. The main results of this study were, that SNP induced nearly a twofold elevation in mean muscle blood flow, while true flow heterogeneity and insulin stimulated glucose uptake remained unchanged in resting muscles of healthy humans. These findings suggest that SNP increases non-nutritive, but not nutritive flow or capillary recruitment. Additionally, during hyperinsulinaemia muscle tissue accounted for two-thirds of blood flow of the femoral region.
We thank the technicians in the Cyclotron/PET Centre at the University of Turku for their superb assistance. Financial support from Turku University Foundation, Helsinki University Research Foundation, Finnish Academy of Science, Paulo Foundation, Yrjö Jahnsson Foundation and Novo Nordisk Fonden is acknowledged.