From the early part of the last century, since the era of Nobel Prize winners August Krogh and A.V. Hill, the mechanisms regulating muscle oxygen delivery have been one of the most studied areas in exercise physiology. This interest maintains focus because skeletal muscle blood flow is the most important single factor influencing muscle oxygenation. Recent methodologies that combine the study of blood flow with local metabolic measurements have revealed potential mechanisms involved in exercise-induced hyperemia. However, despite these advances, many aspects of blood flow regulation remain to be elucidated.
Apart from measures of mean muscle blood flow across an exercising limb, local mechanisms influence blood flow distribution resulting in heterogeneous perfusion during exercise (11). This phenomenon, both spatial and temporal in nature, can be simply defined as an unequal distribution of flow among blood vessels, muscle regions, and different muscles (4). Blood flow heterogeneity has been associated with the efficacy of oxygen delivery and, ultimately, tissue oxygenation and is therefore a functionally important parameter.
Numerous animal studies involving mostly the use of radioactive microspheres have shown that spatial blood flow distribution is heterogeneous both within and between skeletal muscles at rest and during exercise (11). However, until recently, studies on tissue blood flow heterogeneity in humans have been limited because of the need to use invasive techniques. Developments in imaging technologies have enabled the interrogation of human skeletal muscle, yielding data that support marked spatial heterogeneity similar to the observations in the various animal models. In addition, it has been demonstrated that large temporal variations in femoral artery blood flow occur across the muscle contraction cycle (13), but whether fluctuations in flow exist over multiple contraction cycles remains to be fully understood. With advances in noninvasive technology, blood flow heterogeneity and the mechanisms involved in its regulation can now be addressed in the human skeletal muscle.
This review aims to summarize methods previously used in the study of skeletal muscle blood flow heterogeneity in animals and to introduce new noninvasive techniques that can reasonably be used in the human model. Applications of these imaging techniques will also be presented, along with likely mechanisms responsible for heterogeneous blood flow distribution.
METHODS TO MEASURE MUSCLE BLOOD FLOW AND METABOLIC HETEROGENEITY
Since the majority of data revealing skeletal muscle blood flow heterogeneity have been generated from animal studies, it is important to introduce the methods that first pioneered this area of research. Several techniques have been used, of which the microsphere method has been the most often used. This method is based on the injection of radioactive, colored, or fluorescent microspheres into the circulation. After injection, microspheres are assumed to deposit in specific tissue regions in proportion to perfusion. The muscle is then dissected into small segments, and the blood flow in each region is quantified by the relationship between the quantity of microspheres injected and the number of microspheres trapped in a given segment of tissue (4). An index of flow heterogeneity, relative dispersion (RD), is calculated by dividing the standard deviation (SD) of flow values in those pieces of tissue by the mean of those values (4). Accuracy of this measurement parameter depends not only on blood flow, but also on the number of microspheres deposited in the region of interest and the accuracy of assessing the number of microspheres injected. Despite overall acceptance of this technique in the animal model, this method obviously cannot be applied to human studies because of its extremely invasive nature. It is also important to note that, although microsphere deposition reflects blood flow, no information is obtained regarding metabolic heterogeneity and must be combined with another method to study metabolism within the same muscle region.
Positron Emission Tomography (PET)
PET is an imaging technology that enables the study of different aspects of human physiology in a reasonably noninvasive manner. These include, for example, perfusion, metabolism, receptor density, and so on, in essentially all tissue regions of the human body. PET is based on the use of short-lived positron emitting radioisotopes such as oxygen-15 (half-life, 123 s). Tracers, such as 15O-H2O, are labeled with radioisotopes and then, in the case of 15O-H2O, applied to conduct muscle blood flow measurements. Specifically, the tracer is injected into circulation, and its concentration in the tissue over time is recorded with the PET scanner. Based on the kinetics of 15O-H2O, quantitative blood flow measurements can be derived. 15O-H2O is an ideal tracer for the tissue blood flow, because it is chemically inert and freely diffusible. In addition, the short half-life of 15O enables sequential perfusion measurements and the combination of perfusion measurements with other PET measurements such as glucose uptake and specific vascular receptor densities.
Relatively easy modeling of blood flow from tracer kinetics enables accurate measurements of blood flow in PET image voxels (volume elements) that represent tiny regions of the muscle (Fig. 1). RD is then calculated in the same manner as with the microsphere method (SD/mean). As with microspheres, the accuracy of the PET and [15O]-H2O method depends on the amount of the tracer in the muscle or, in turn, on the amount of flow and the amount of tracer injected. Within standard doses, the accuracy has been comparable to microspheres. The advantage over the microsphere method is its potential for use in humans and that it offers the measurements of metabolic indices (e.g., glucose uptake) in the same region of muscles.
A drawback of the PET method is its sensitivity to motion artifact such that measurements are difficult, but not impossible, to perform during dynamic exercise. This limitation has been overcome by using one-leg intermittent isometric exercise (Fig. 2A). With intermittent isometric exercise, a mode of dynamic muscle pumping can be performed in the PET scanner while keeping the limb relatively stable. Recently, PET has also been applied during dynamic exercise using a specially manufactured dynamometer (Fig. 2B). The limb is carefully strapped to the imaging table, significantly limiting movement in the area to be imaged. This model can be further enhanced by using a gated scanning mode, which performs the scan in short, relatively motionless time frames within each muscle contraction cycle.
Near-infrared Spectroscopy (NIRS)
NIRS is a noninvasive optical method typically applied for measuring tissue oxygenation, blood volume, and, in some cases, hemodynamics. Measurements of these parameters are based primarily on the absorption of light at wavelengths in the near-infrared (NIR) range (400-1000 nm) where oxygenated hemoglobin (oxy-Hb) and deoxygenated hemoglobin (deoxy-Hb) display different light absorption characteristics. Using light emission at wavelengths in the visible range (400-600), the different absorption spectra of oxy-Hb and deoxy-Hb explain the characteristic red color of fully oxygenated blood and the relatively bluish color of venous blood. In the NIR (700-900) range of the spectrum, their absorption spectra are distinctly different allowing greater precision in the separation and quantification of changes in these compounds by using an optode emission and receiving system incorporating several specific wavelengths in the NIR range. The isobestic point of 800 nm where both chromophores possess the same light extinction coefficients permits the calculation of [Hb] independent of oxygenation status. NIRS measures changes in tissue oxygenation in small arterioles, capillaries and venules of less than 200 um diameter. The microcirculation can be isolated because the light emitted into the larger vessels (arteries and veins) is almost completely absorbed by the larger relative molar concentration of Hb, and thus any detectable changes in absorption can be attributed to the microcirculation. The overlapping absorption spectra of muscle Mb and Hb make separation of these absorbers difficult, and this is a limitation of the technique. However, the contribution of Mb to light absorption changes is estimated to be ~10%, and recent findings suggest that during exercise, MbO2 displays a rapid reduction after the onset of exercise and remains relatively stable, which lends to the interpretation that light absorption changes with NIRS are attributable mainly to the oxygenation status of Hb. Nonetheless, the potential application of combined proton nuclear magnetic resonance and NIRS would allow for precise separation of these signals. Cytochrome oxidase (CtOx), the terminal enzyme in mitochondrial respiration, is also detectable by NIRS, but as the concentration is very small compared with Hb and Mb, variations in CtOx are often difficult to detect and are not often used for NIRS-related research. Yet, with recent modeling procedures, CtOx has been successfully quantified in the brain.
Near-infrared light is transmitted from an emitting optode to a light-detecting (receiving) optode after passage in a spherical pattern through tissue, the depth of which being approximately half the distance between the two optodes. Typically, most commercially available spectrophotometers provide optode separation distances between 4 and 5 cm. Most NIRS systems provide quantitative "changes" in Hb oxygen status because absolute [Hb] values require direct time-of-flight measures of photons to determine path length and account for scatter within heterogenous tissue medium; a necessary component of the Beer-Lambert law. Spectrophotometers with time-of-flight measures are available; however, they have had limited application to studies in exercise physiology. Nonetheless, some systems, such as spatially resolved spectroscopy, can provide an estimate of path length and have improved the quantification of chromophore concentrations.
Combined NIRS and Indocyanine Green (ICG) Dye
Recently, it has become possible to quantify blood flow in discrete regions of the microcirculation using NIRS in combination with a light-absorbing tracer (3). ICG is a dye with a peak light absorption in human blood in the NIR range (800 nm) and has been used routinely for measuring cardiac output and limb blood flow by photodensitometry. It is ideal for repeated measurements because of its nontoxicity and rapid clearance by the liver. By combining NIRS and ICG, blood flow can be calculated by applying basic tracer principles: the rate of accumulation of a tracer in a given tissue is equal to its rate of inflow minus its rate of outflow. If the ICG is introduced rapidly and its rate of accumulation is measured over time, blood flow can be measured as a ratio of the tracer accumulated to the quantity of tracer introduced over a given time (Fig. 3).
NIRS is an interesting tool for investigating both blood flow and local tissue oxygenation during dynamic exercise because it allows for continuous and variable frequency data collection. Information can be simultaneously obtained regarding microvascular blood flow and local tissue metabolic status, providing multifaceted insight into the coupling between oxygen demand and delivery. The technique has been well correlated with established blood flow techniques (133Xe washout, dye dilution) and has been used to quantify progressive changes in tissue perfusion induced by incremental exercise, allowing for the assessment of perfusion variations between tissue regions (2). This technique does not require limb stability for measurements, therefore making it ideal for a variety of exercise modes including dynamic exercise protocols.
A drawback of the NIRS/ICG method for measuring blood flow heterogeneity is the tissue sampling size; the tissue area measured (depending on distance between emitting and receiving optode) and the blood flow and volume itself (which can range between ~1 and 10 mL of vascular volume) are considerably larger than the size of microvascular units. In this case, discrete regions of heterogeneity may not be detected, or the degree of heterogeneity observed will be the sum of flow to multiple microvascular units. Moreover, unless numerous optodes are used, a clear picture of blood flow distribution within or between muscles may not be fully appreciated.
Magnetic Resonance Imaging (MRI) and Spectroscopic Imaging
MRI is an application of nuclear magnetic resonance. It is based on the body's natural magnetic properties, which are used to produce detailed images from any part of the body. Dynamic MRI with the arterial spin-labeling technique has recently been used to quantify skeletal muscle perfusion in humans (5). With this method, the MR images are dynamically alternated between tag and control images (5). In tag images, the magnetization of arterial blood is inverted before image acquisition, and in the control images, it is not. From the subtraction of these magnetic resonance images in a dynamic time series, perfusion within each voxel can be calculated (5). This allows the calculation of perfusion heterogeneity within and between the muscles similarly as with PET images but without the need for radioactive tracers. Magnetic resonance spectroscopic imaging can be used to measure phosphocreatine depletion, providing an index of muscle oxygen consumption. This, combined with arterial spin-labeling technique, has enabled studies of the local association between blood flow and oxygen metabolism (14).
HETEROGENEITY IN BLOOD FLOW AND METABOLISM BETWEEN AND WITHIN MUSCLES
As in the previous section, it remains important to briefly review the results obtained from work in animals. These studies, starting in the 1980s, have shown that there is marked variation in muscle blood flow both within and between muscles (11). As previously reviewed by Laughlin and Armstrong (10) and Laughlin et al. (11), exercise influences blood flow heterogeneity both between and within skeletal muscles in different animal species. At rest, blood flow seems to be preferentially distributed to active oxidative muscle fibers, and during exercise, the increase in blood flow seems to be directly related to the percentage of fast-twitch, oxidative-glycolytic fibers and to muscles composing mostly of these fibers. In contrary, blood flow to muscles composed of mainly fast-twitch glycolytic muscle fibers is even decreased at low exercise intensities. As blood flow does increase in these muscles at higher intensities, it seems that at these lower exercise intensities, blood flow is not increased simply because these fibers are not recruited during low-intensity exercise. Thus, fiber-type distribution and recruitment of different fibers at different exercise intensities and/or durations seem to be the main reasons for heterogeneous blood flow distribution within and between skeletal muscles in different animal species (10,11).
Acute Exercise in Humans
In concordance with animal studies, human studies have also shown marked spatial heterogeneity in muscle blood flow. Ament et al., using PET (1), and Frank et al., using MRI (5), were the first two groups to demonstrate that there is marked variation in blood flow between and within muscles during exercise in humans. Unfortunately, only SD (and not RD) values were reported, making it impossible to make any conclusions about the level of heterogeneity regardless of the mean blood flow. In more recent PET studies, RD values have been reported, revealing that in both isometric and dynamic exercise, the degree of heterogeneity observed depends largely on the muscle studied (6,9). In the quadriceps femoris muscle group, blood flow was least heterogeneous in the vastus intermedius and medialis and most heterogeneous in the vastus lateralis and rectus femoris during intermittent isometric exercise (6). During dynamic exercise, the vastus lateralis has had markedly higher heterogeneity than the other three muscles (9). Intermittent isometric exercise decreased heterogeneity in the former two muscles, whereas it has increased heterogeneity in the latter two. There is also marked variation between muscles so that mean blood flow in the exercising quadriceps femoris muscle group is markedly higher in vastus intermedius and medialis than in vastus lateralis or rectus femoris both during isometric and dynamic exercise (6,9). This is most likely a result of different recruitment patterns of these muscles, which is likely further explained by variations in relative proportions of different fiber types within these muscles.
PET studies have shown blood flow distribution between muscles only in transaxial direction, in the middle parts of the muscles. With the utilization of NIRS, this distribution can be measured also in axial direction (along the muscle). Recent studies applying this method have shown marked heterogeneity in this direction between and within the different muscles of the quadriceps femoris muscle group. These studies have shown that blood flow is higher in proximal compared with distal parts of the muscles.
Studies applying ultrasound Doppler to the estimation of muscle blood flow have shown that there is marked variation in femoral and brachial artery blood flow in humans during the muscle contraction cycle. Blood flow in femoral artery has been shown to be highest during relaxation phase and lowest during contraction, falling even to zero or negative at highest peak force (13). These changes in arterial flow may also be responsible for a portion of the variations observed in muscle perfusion, but this issue has not been addressed because of lack of applicable methodology. However, future work with MR imaging could potentially resolve this concern, and temporal variations in muscle perfusion and metabolism over the contraction cycle could be quantified. It is also currently unknown whether muscle blood flow fluctuates at both rest and during exercise with lower frequencies (i.e., vasomotion) during the course of minutes or hours. Some animal studies suggest that, at least in heart and lungs, spatial pattern of regional perfusion during resting conditions is quite stable over time. This suggests that the basis for blood flow heterogeneity could be the vascular structure, although physiological control cannot be excluded because, for example, exercise has been shown to cause changes in flow heterogeneity. The potential mechanisms causing the heterogeneity will be dealt with in detail later in this review.
Effects of Exercise Training in Humans
Effects of endurance training on muscle or limb blood flow have been studied extensively. However, because of methodological differences, clear consensus has not been obtained as to whether training has effects on muscle blood flow. In most of the studies, training-induced changes in muscle blood flow (mainly quadriceps femoris) at the same absolute submaximal intensity have revealed a decrease or no change (12). In studies in which muscle oxygen consumption has also been measured, it has been at the same level before and after training. Thus, depending on changes in perfusion, oxygen extraction has been either slightly elevated or unchanged. Endurance training also seems to have an effect on blood flow heterogeneity within muscles. In recent PET studies, it was shown that blood flow distributes less heterogeneously within the muscles in endurance-trained compared with untrained men. Further analysis has also demonstrated that the differences in perfusion heterogeneity, as observed in voxels of PET images, are also present in microvascular units. This suggests that there are microvascular adaptations in human skeletal muscle that encourage more homogeneous flow distribution. In contrast to perfusion heterogeneity within muscles, no differences have been observed in perfusion heterogeneity between the quadriceps femoris muscles during one-legged intermittent isometric exercise in trained and untrained men.
CAUSE OF HETEROGENEITY IN MUSCLE BLOOD FLOW AND METABOLISM
Vascular Architecture and Blood Flow Heterogeneity
During the early stages of microvascular research, not only was it assumed that blood flow was evenly distributed throughout metabolically active muscle (as demonstrated by units of blood flow in mL/100 g of tissue/min), but it was also believed that the behavior of similarly classified microvessels is uniform across the vascular bed of tissue. As clearly demonstrated by Sweeney and Sarelius (15), despite similar anatomical and functional characteristics, the microvessels are a significant source of heterogeneous blood flow responses.
Contrary to the historic belief that the individual capillary is the basic unit of functional control within the microcirculation, recent work has elucidated that blood flow distribution in striated muscle is the result of the recruitment of groups of capillaries that originate from a single common terminal arteriole. Vasoactive stimuli acting on the terminal arteriole result in either increased or diminished flow through the whole network of associated capillary groups.
Studies investigating capillary network architecture found a significant variation in structure at points of known differences in oxygen inflow with a higher ratio of capillary groups per microvascular unit in areas with higher flow. This suggests that certain locations within the muscle have larger regions of control and that the activation of a single microvascular unit at these points may perfuse more capillary cross-sectional area. Moreover, the terminal arterioles supplying the various capillary modules have also been documented to respond differently from each other in a spatially organized fashion.
Vascular resistance has been found to increase with each subsequent branch from a proximal to distal direction along the main arteriole. Higher resistance would contribute to the observed variations in cell flux and oxygen flow into each terminal arteriole with lower levels of perfusion evident in the more distal vessels.
The anatomical arrangement of capillaries has also been shown to vary according to the types of muscle fiber that they feed. Capillaries associated with oxidative fibers are narrower and shorter in length. They also display a higher degree of tortuosity than those associated with glycolytic fibers. This anatomical structure is believed to provide shorter oxygen diffusion distances from vessel to fiber mitochondria and an overall increase in capillary cross-sectional area. Therefore, in spatial heterogeneity, microvascular function and flow rates through individual segments can be understood to be dependent on their architecture and location within the vascular network (15).
Metabolism and Blood Flow Heterogeneity
In addition to the influence of architecture and anatomy on microvascular function, physiology can also partly explain muscle blood flow heterogeneity. One such explanation is heterogeneous metabolism: this suggests that the primary cause for heterogeneous blood flow would be the metabolic needs of the tissue and that the vasculature, in turn, adapts accordingly (8). Ample evidence from myocardial tissue in animals suggests that there is a close association between regional blood flow and substrate or oxygen utilization. In skeletal muscle, in contrast, the association between regional blood flow and metabolism is less obvious. In resting muscle, no correlation (r ~ 0.1) has been observed between regional blood flow and glucose uptake in both animals and humans. Low metabolic rate and blood flow are probably the best explanations for this. In the exercising muscle, the correlation is much better (r ~ 0.5) but still not strong enough for local glucose uptake and its heterogeneity to provide a full explanation for the marked level of blood flow heterogeneity within the muscle. However, glucose uptake from blood is not the major source of energy during exercise, and this may also explain the poor association between glucose uptake and blood flow in skeletal muscle.
Oxygen, instead, is the most important substance that muscles require during exercise, and the supply of oxygen is highly dependent on blood flow, and thus, they would be expected to correlate. In a recent study, the association between local blood flow and oxygen consumption within muscles was measured with the MRI techniques described previously in this review (14). Interestingly, despite the unexpectedly high degree of blood flow heterogeneity in relation to the voxel size, the results from this study suggest that local blood flow and oxygen metabolism may not be so well matched between different regions in the exercising muscle (Fig. 4). In concordance, studies using NIRS have shown that O2 saturation, reflecting the balance between oxygen delivery and consumption, is heterogeneous within and between muscles. This supports the concept that there may be a spectrum of matching between muscle oxygen delivery and demand and, therefore, variations in O2 extraction. With PET, we have also shown large individual variation in correlation between perfusion and oxygen uptake in different quadriceps femoris muscles (7). This also suggests that there is at least some degree of mismatch between oxygen delivery and use in human skeletal muscle. Unfortunately, it is currently impossible, or at least the results would not be accurate enough, to calculate oxygen consumption or extraction in the voxels using PET and to correlate them at the voxel level in PET images.
CONCLUSIONS AND PERSPECTIVES FOR THE FUTURE
Although heterogeneity of blood flow and metabolism are currently described in both animal and human skeletal muscle, there is still lack of a complete explanation as to why flow is heterogeneous. There seems to be a weaker correlation between local perfusion and metabolism in skeletal muscle than in other tissues, in particular, the heart. This can potentially be explained by the different metabolic profiles within these tissues, but noteworthy is that the only metabolic parameters to which blood flow heterogeneity has been related are the uptake of glucose and oxygen, and this was reported only in a few studies. Clearly, more work is required to fully explain how closely local metabolism and blood flow are associated at the micro level of tissue and what factors regulate perfusion with respect to metabolism.
1. Ament, W., J. Lubbers, G. Rakhorst, W. Vaalburg, G.J. Verkerke, A.M. Paans, and A.T. Willemsen. Skeletal muscle perfusion measured by positron emission tomography
during exercise. Pflugers Arch.
2. Boushel, R., H. Langberg, J. Olesen, M. Nowak, L. Simonsen, J. Bulow, and M. Kjaer. Regional blood flow
during exercise in humans measured by near-infrared spectroscopy
and indocyanine green. J. Appl. Physiol.
3. Boushel, R., and C.A. Piantadosi. Near-infrared spectroscopy
for monitoring muscle oxygenation. Acta Physiol. Scand.
4. Duling, B.R., and D.H. Damon. An examination of the measurement of flow heterogeneity in striated muscle. Circ. Res.
5. Frank, L.R., E.C. Wong, L.J. Haseler, and R.B. Buxton. Dynamic imaging of perfusion in human skeletal muscle during exercise with arterial spin labeling. Magn. Reson. Med.
6. Kalliokoski, K.K., J. Kemppainen, K. Larmola, T.O. Takala, P. Peltoniemi, A. Oksanen, U. Ruotsalainen, C. Cobelli, J. Knuuti, and P. Nuutila. Muscle blood flow
and flow heterogeneity during exercise studied with positron emission tomography
in humans. Eur. J. Appl. Physiol.
7. Kalliokoski, K.K., J. Knuuti, and P. Nuutila. Relationship between muscle blood flow
and oxygen uptake
during exercise in endurance-trained and untrained men. J. Appl. Physiol.
8. Karch, R., F. Neumann, B.K. Podesser, M. Neumann, P. Szawlowski, and W. Schreiner. Fractal properties of perfusion heterogeneity in optimized arterial trees: a model study. J. Gen. Physiol.
9. Laaksonen, M.S., K.K. Kalliokoski, H. Kyrolainen, J. Kemppainen, M.Teras, H. Sipila, P. Nuutila, and J. Knuuti. Skeletal muscle blood flow
and flow heterogeneity during dynamic and isometric exercise in humans. Am. J. Physiol. Heart Circ. Physiol.
10. Laughlin, M.H., and R.B. Armstrong. Muscle blood flow
during locomotory exercise. Exerc. Sport Sci. Rev.
11. Laughlin, M.H., R.J. Korthuis, D.J. Duncker, and R.J. Bache. Control of blood flow
to cardiac and skeletal muscle during exercise. In: Handbook of Physiology-Section 12: Exercise: Regulation and Integration of Multiple Systems
, edited by L.B. Rowell and J.T. Shepherd. New York: Oxford University Press, 1996, p. 705-769.
12. Lawrenson, L., J. Hoff, and R.S. Richardson. Aging attenuates vascular and metabolic plasticity but does not limit improvement in muscle VO(2) max. Am. J. Physiol Heart Circ. Physiol.
13. Radegran, G. Ultrasound Doppler estimates of femoral artery blood flow
during dynamic knee extensor exercise in humans. J. Appl. Physiol.
14. Richardson, R.S., L.J. Haseler, A.T. Nygren, S. Bluml, and L.R. Frank. Local perfusion and metabolic demand during exercise: a noninvasive MRI method of assessment. J. Appl. Physiol.
15. Sweeney, T.E., and I.H. Sarelius. Spatial heterogeneity in striated muscle arteriolar tone, cell flow, and capillarity. Am. J. Physiol.