INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

THERE IS RENEWED INTEREST in the evidence that skeletal muscle microvascular perfusion is not always directly related to the total blood flow of muscle (19). Although there may be other explanations, such evidence is consistent with the concept of two separate circulations within muscle, one of which acts as a functional shunt by denying the opportunity for nutrient exchange to occur between the muscle cells and the blood constituents. Studies by Pappenheimer (29), Walder (38), Renkin (32), Barlow and colleagues (1), Hyman and co-workers (18), and Sonnenschein and Hirvonen (35) have all produced results that are consistent with this concept. The two circulations were referred to as "nutritive" and "nonnutritive" (2), where the latter was considered to serve as a series of functional vascular shunts for muscle. The anatomic description of the pathway for nonnutritive flow has been difficult to identify, although some of the vessels for this route appear to be closely associated with connective tissue (2, 14, 26).

Vessels in the nutritive pathway were considered to be those in direct contact with the skeletal muscle cells (17). In addition, estimations of blood flow for muscle at rest suggested that approximately half of the total flow was nutritive and thus capable of clearing intramuscular injected markers (17). More recent estimates suggest this proportion may be even lower (16, 36).

Research from our laboratory using the constant-flow-perfused rat hindlimb has led to the proposition that skeletal muscle metabolism (including oxygen uptake and lactate release as well as aerobic contractile performance) is controlled by vasoconstrictors that act to alter flow distribution within muscle (7, 9). One group of vasoconstrictors referred to as type A acts to increase oxygen uptake, lactate release (7), and contractile performance (31) by redirecting flow from a putative nonnutritive route to nutritive capillaries within muscle (26). A second group of vasoconstrictors referred to as type B has the opposite effect and acts to decrease oxygen uptake, lactate release (7), and aerobic contractile activity (13). These vasoconstrictors do not alter total flow to muscles but appear to redirect flow within the hindlimb muscles (26, 31) away from the nutritive route to the nonnutritive route or functional shunts. Taken together these findings suggest that vasoconstrictors control skeletal muscle metabolism by altering the proportion of nutritive to nonnutritive flow without affecting total flow. As indicated above, vessels in the nutritive pathway are considered to be those in direct contact with the skeletal muscle cells (17), and there is some indication that nonnutritive vessels are located in closely associated connective tissue, septa including perimysium (4, 25), and tendons (27). Indeed, direct evidence for these latter vessels has been obtained by Ley and colleagues (21), who showed them to be major routes for the shunting of leukocytes. In the present study we have investigated the laser Doppler flow (LDF) signal responses of surface macroprobes and implantable microprobes positioned on or in hindlimb skeletal muscles, respectively. The responses to changes in total flow and to vasoconstrictors that are known to alter hindlimb metabolism have been characterized.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Hindlimb Perfusions

1

1

1

1

Laser Doppler Flowmetry

Surface measurements. A small hole (~4-mm diameter) was made in the skin in the middle of the biceps femoris corresponding to point bf in Fig. 1. The hindlimb was then clamped by the foot so that the LDF probe (Perimed PF 2, operating wavelength of 632 nm) could be positioned over the center of the hole. For the measurement of red blood cell flux, the probe was placed vertically above and ~1 mm from the surface of the muscle. The probe comprised three fibers (each 800-µm diameter): one for illumination and two for detection. Settings on the detector unit were 4 kHz (gain setting 10) with a time constant of 3 s unless otherwise indicated. The signal (0-5 V) was continuously recorded on an IBM-compatible PC using a DI-190 I/O module and Windaq software. The data are expressed as volts.

View larger version (48K): [in this window] [in a new window]   Fig. 1.   General areas chosen for positioning of laser Doppler flow (LDF) probe either on or impaled into muscles of the perfused rat hindlimb. LDF macroprobe (3 × 800 mm) was always positioned over the biceps femoris (bf) through a hole in the skin. Fiber-optic microprobe (0.26-mm diameter) was inserted through a small incision in the skin in the approximate positions shown on either the tibialis (tib), vastus (vs), bf, or gastrocnemius (gs) muscles. Mechanism for insertion involved an initial right-angle puncture of the muscle surface, insertion for 1-2 mm, gentle rotation of the probe so that it is in parallel with the muscle axis, and further insertion for 3-6 mm. Apart from the general region indicated, positioning of the microprobes including precise site, initial depth, and final penetration was randomly chosen.

Within-muscle measurements. A Moor Instruments Lab Server and Lab Satellite fitted with two P10M master probes was used with two P10s-TCG 260-µm slave probes, each fitted with TCG-fixed fiber. The operational wavelength of the laser light source was 780 nm. The fiber consisted of a single core of optic fiber (200-µm diameter) surrounded by a protective flexible outer sheath. The fiber was found to be sufficiently robust to be inserted into the muscle unaided by prior needle puncture. Measurements were made using the P0 setting (moorLAB V36 embedded software). This setting scales the raw laser Doppler flux output by a factor of 10-fold; we used it to improve the quality of the recorded signal. Subsequently, laser Doppler values were scaled to express the results in perfusion units (PU) by a factor from measurements made on the manufacturer's calibration fluid for each probe. Small incisions were made into the skin covering the midregion of any two of the tibialis, vastus, gastrocnemius, or biceps femoris (Fig. 1) of the perfused leg and the two probes were inserted. The procedure for insertion of each probe involved initial puncturing of the epimysium and muscle body to a depth of ~2 mm with the probe at right angles to the muscle surface, followed by rotation of the probe through 90° to be parallel to the longitudinal direction of the muscle fibers. The probe was then inserted 6 mm further and taped in place. This procedure avoided wounding and only 4 of the 97 sites were considered corrupted due to bleeding. Indeed, as pointed out by Oberg (28), by using very thin optical fibers (50-200 µm diameter) the trauma can be minimal with little disturbance to blood flow. After completion of the perfusion, the final placements were confirmed by surgical examination. In some animals the probes were repositioned in other muscles up to two additional times. This allowed assessment of as many as six different sites per hindlimb. The LDF signal (0-5 V) was continuously recorded on an IBM-compatible PC by using a DI-190 I/O module and Windaq software. The data are expressed as PUs to be consistent with the manufacturer's recommendations and to be distinguished from those of the surface probe, which differed in size and operating wavelength.

Perfusion pressure and oxygen uptake. Perfusion pressure was also continuously recorded using Windaq software. Perfusate entering and leaving the hindlimb was passed through an in-line A-Vox analyzer (A-Vox Systems, San Antonio, TX) and the signal was continously recorded also by using Windaq software.

Statistical analysis. Data were analyzed using Sigma Stat (Jandel Scientific). Comparison of basal signal strength for nutritive, nonnutritive, and mixed sites, unpaired analysis (Student's t-test) was used. For effects of vasoconstrictors and flow, paired analysis (Student's t-test) was used.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Surface Measurements

1

1

1

1

A time course from a typical experiment is shown in Fig. 2, where the LDF signal on the surface of the biceps femoris muscle vessels was measured as well as perfusion pressure and oxygen uptake for the hindlimb during successive injections (25 µl) of 300 pmol NE and 3 nmol 5-HT. NE (300 pmol), equivalent to a peak concentration of 15-50 nM (estimated from injection and perfusion flow rates), increased the LDF signal in association with increased perfusion pressure and stimulation of oxygen uptake. Values for LDF and oxygen uptake returned to basal ~3 min after NE injection. 5-HT (3 nmol), equivalent to a peak concentration of 150-500 nM, decreased the LDF signal in association with increased perfusion pressure and inhibition of hindlimb oxygen uptake (Fig. 2).

View larger version (14K): [in this window] [in a new window]   Fig. 2.   Typical trace of LDF signal from the surface macroprobe positioned at the center of the anterior end of the biceps femoris (A), arterial perfusion pressure (B), and oxygen uptake (O2, C) after injection of 300 pmol norepinephrine (NE, type A vasoconstrictor) and 3 nmol 5-HT (serotonin, type B vasoconstrictor) into the constant-flow-perfused rat hindlimb. Biological zero has not been subtracted.

Figure 3 shows a positive correlation (r = 0.909; P < 0.001) for change in oxygen uptake as a function of change in LDF signal from 10 experiments where the dose of injected NE ranged from 50 to 300 pmol (n = 5) and of 5-HT from 1 to 3 nmol (n = 5).

View larger version (14K): [in this window] [in a new window]   Fig. 3.   Relationship between peak change in oxygen uptake (O2) and LDF signal from the surface macroprobe for the constant-flow-perfused rat hindlimb. Values for LDF signal are from Fig. 2; , NE; , vehicle; , 5-HT.

Implantable Probes

View larger version (16K): [in this window] [in a new window]   Fig. 4.   Distribution of LDF signal changes from impaled microprobes. Values obtained after NE (A) or 5-HT (B) were expressed as a percentage of the basal signal and grouped in intervals of 10% starting from +5% and 5%.

Characteristic traces of NE-positive, NE-negative, and mixed sites are shown in Fig. 5. NE-positive sites showed LDF signal changes in parallel to oxygen uptake (i.e., increases and decreases with NE and 5-HT, respectively). NE-negative sites showed LDF signal changes opposite to oxygen uptake. For the total of 97 sites examined, 56.7% were found to be NE positive, 16.5% were NE negative, and 24.7% were mixed. Only 2.1% were corrupted due to bleeding. Table 1 also shows that basal LDF signal strength was greater at NE-negative sites than either NE-positive or mixed sites, and all three sites responded significantly to a doubling of flow rate from 4 to 8 ml/min. Although not shown by the data of Table 1, a breakdown of the data revealed that there were differences between the muscle types, with the basal signal from NE-positive sites in tibialis (24.7 ± 2.2; n = 28) similar to that of gastrocnemius (21.7 ± 4.5; n = 12) but greater than those in either biceps femoris (12.9 ± 2.3; n = 6; P < 0.01) or vastus (12.4 ± 2.8 PU; n = 9; P < 0.02).

View larger version (23K): [in this window] [in a new window]   Fig. 5.   LDF (measured in perfusion units, PU) tracings for NE-positive (A), NE-negative (B), and mixed (C) sites from impaled microprobes. Bolus injections of 3 nmol NE or 5-HT were made at 2 and 10 (or 12) min, respectively, during perfusion of the rat hindlimb at constant total flow and sites classified as described in the text. Perfusion pressure as well as arteriovenous oxygen difference for the whole hindlimb were also recorded. Tracings are representative of the groups for which n values are given in Table 1.

Table 2 summarizes the response of NE-positive sites to flow and vasoconstrictors, including 5-HT, NE, vasopressin, and ANG II. Whole hindlimb oxygen uptake as well as perfusion pressure are also shown. Increasing the pump flow rate progressively from 4 through 6-8 ml/min increased the LDF signal from 14.2 ± 1.0 to 22.8 ± 1.4 PU. Allowing for a biological zero of ~7.4, the increase due to flow represents a twofold increase. Whole hindlimb oxygen uptake and perfusion pressure increased by 29.4% and 65.3%, respectively. Three of the vasoconstrictors (NE, vasopressin, and ANG II) each increased the LDF signal in parallel to their effects on oxygen uptake. 5-HT inhibited both and in many cases the signal strength due to 5-HT was indistinguishable from biological zero for that site.

Sustained Infusions

View larger version (18K): [in this window] [in a new window]   Fig. 6.   LDF tracing from NE-positive (top left) and NE-negative (bottom left) sites using impaled microprobes after the constant infusion of 83 nM NE (solid bar). Hindlimbs were perfused at constant total flow and an infusion of a stock solution of 5 µM NE was introduced for the period shown. Representative tracings for perfusion pressure (bottom right) as well as arteriovenous oxygen difference (top right) for the whole hindlimb are also shown.

View larger version (19K): [in this window] [in a new window]   Fig. 7.   LDF tracing from NE-positive (top left) and NE-negative (bottom left) sites using impaled microprobes after the constant infusion of 750 nM 5-HT (solid bar). Hindlimbs were perfused at constant total flow and an infusion of a stock solution of 50 µM 5-HT was introduced for the period shown. Representative tracings for perfusion pressure (bottom right) and arteriovenous oxygen difference (top right) for the whole hindlimb are also shown.

Biological Zero

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

This work represents the second (see also Ref. 19) study where there is a major divergence between total flow and LDF measurements in muscle and raises the question as to how this might occur. We believe that two issues have an important bearing: one concerning what LDF probes actually detect, two being the influence of vessel architecture. For LDF measurements it is generally thought that movement of the blood cells causes a frequency shift with some of the photons also backscattered or absorbed by the tissue. Doppler broadened and backscattered light carries information about the movement and concentration of the cells. LDF probes are thus considered to measure "tissue perfusion." Although this has not been fully defined, Leahy and colleagues (20) propose that tissue perfusion is the product of local speed (nonvectorial) and the concentration of moving cells. However, it is far from clear whether the product derives from equal quantitative input from each or whether one or another has a dominant effect. Equal input would mean that if total (volume) flow remained constant then changes in vessel diameter, distribution of flow from few to many vessels, or the distribution of flow from short to tortuous vessels should not affect the signal from an LDF probe. To put it another way, the relationship between LDF signal and total flow should always hold. The findings in this present study which clearly show a divergence between total flow and LDF signal have thus necessitated an evaluation of the relationship between vessel architecture and LDF signal. Therefore in a separate study (5) we have constructed models of different vessel architectures from commercially available polymer tubes of a size such that even with complex formations they could be fitted comfortably within the field of measurement of a standard LDF probe. We examined the effect of tube architecture on LDF signal using polymer tubes of 250-, 100-, and 50-µm internal diameter. At 3% hematocrit the LDF signal was linear for each of the three tube sizes from 10 to 80 µl/h. The signal strength was greatest from the smallest tube and least from the largest tube even though volume flow was the same in each. For a single tube (100 µm) that doubled back on itself twice to cross the field of measurement three times, the LDF signal at any flow (10-80 µl/h, hematocrit 3%) was approximately threefold greater than that for the same tube crossing the field of measurement once. The effect of progressively switching flow (constant at 120 µl/h, hematocrit 9%) from five tubes to one tube in a manifold of five tubes (100 µm) gave rise to a progressive increase in signal. Overall, the data underscore the predominant role of cell speed (nonvectorial) rather than velocity or cell number in determining LDF signal. If these models can be considered representative of the microvasculature of tissues in vivo, then they may help explain a recent report in skeletal muscle where changes in total (volume) flow failed to correlate with the LDF signal in vivo (19) and with the present data in perfused rat hindlimb. In the present experiments (which were conducted at constant flow) NE increased the LDF signal at the majority of sites and increased overall hindlimb oxygen uptake. A decrease in vessel diameter due to NE-mediated vasoconstriction of the vessels under the probe could account for the increase in the LDF signal. However, although this might explain the effects of NE to increase the LDF signal, it could not account for the effects of 5-HT which in the hindlimb also caused vasoconstriction, but at the same LDF sites as where NE was stimulatory, caused a decrease in LDF signal and a decrease in overall hindlimb oxygen uptake. From the tube-manifold studies where the LDF signal decreased when flow was switched from one to five identical tubes, it would seem likely that any vasoconstrictor that acts to decrease capillary flow by derecruiting the number of capillaries perfused without affecting total flow would increase the LDF signal. Thus 5-HT, which decreases oxygen uptake and presumably muscle capillary flow, might be expected to increase, not decrease, the LDF signal.

One explanation from the perfused tube studies (5) that might account for the effects of NE and 5-HT to increase and decrease the LDF signal, respectively, is that blood flow is redirected to (NE) or from (5-HT) tortuous multiple-pass capillaries much as were used in the model described above. This would mean that the two networks comprise on the one hand the long tortuous capillaries that could be considered as nutritive, and on the other the relatively short single-pass capillaries (possibly nonnutritive). Switching the blood flow from one to the other by site-specific vasoconstrictors might then account for the changes in metabolism reported herein. Type A vasoconstrictors such as NE are known to constrict at the region of terminal arterioles. The rise in pressure may be sufficient to overcome the intrinsic resistance of the longer capillaries allowing flow. This would account for the previously observed increase in red blood cell washout and newly recruited vascular space detected by FITC dextran that accompanied NE-mediated vasoconstriction (26). In contrast, type B vasoconstrictors such as 5-HT are thought to vasoconstrict higher in the vascular tree and dilate the arterioles (see Ref. 7 and references therein) thereby lowering the pressure in the vicinity of the capillary networks. Under these conditions flow in the lower-resistance short nonnutritive capillaries would be favored. Whereas the short nonnutritive capillaries are near to the muscle fibers, it is also possible that some are contained within the connective tissue that surrounds the bundles of muscle fibers (the epimysium). As such these would be sufficiently insulated from muscle cells to be functionally nonnutritive. From our previous studies it is clear that some of the nonnutritive connective tissue vessels are located in tendons (2, 14, 27) and are therefore outside the present regions of LDF measurement which are located more centrally.

In the present study a distinct heterogeneity of sites was identified from LDF microprobes placed at random in various muscles in the perfused rat hindlimb. The heterogeneity was not detected with the much larger surface probe unless the probe was moved to regions where tendon vessels were apparent. Of the sites examined with the microprobes, either of three characters was displayed when the hindlimb was injected with bolus amounts of the vasoconstrictors NE or 5-HT. Some responded with an increased LDF signal from NE and a decreased LDF signal from 5-HT and were designated as "NE positive." Those responding with a decreased LDF signal from NE and an increased LDF signal from 5-HT were designated as "NE negative." Sites failing to respond to either NE or 5-HT were deemed as "mixed." Because the vasoconstrictors NE and 5-HT have been previously described as increasing or decreasing muscle nutritive flow based on their respective effects to stimulate or decrease muscle metabolism (7), we propose that the NE-positive sites are located in the nutritive vascular route. Similarly, because NE and 5-HT have been previously shown to decrease or increase, respectively, putative nonnutritive flow in tibial tendon vessels of the biceps femoris (27), the NE-negative sites reflect the positioning of the probe in the nonnutritive vascular route. A mixed site could represent a location where nutritive and nonnutritive sites are juxtaposed so a positive response from one is obscured by a negative response of similar magnitude from the other. Indeed, close inspection of mixed traces (e.g., Fig. 5C) shows LDF signal fluctuation at the points when NE and 5-HT were injected. It is unlikely that mixed sites were where there was no blood flow, as increasing the flow from 4 to 8 ml/min significantly increased the LDF signal at these sites (Table 1) and the basal LDF signal was greater than biological zero.

In addition to their differing character with respect to responses to the vasoconstrictors NE and 5-HT, two other differences were detected. First, the proportion of sites assessed favored NE-positive over NE-negative sites by a factor of 56.7% to 16.5%, or ~3.5:1. The ratio may be closer to 2.4:1 if the assumption is made that mixed sites comprise equal proportions of NE-positive (nutritive) and NE-negative (nonnutritive) components. Second, NE-negative sites in general showed a higher basal LDF signal than NE-positive sites (Table 1). Together, these additional properties suggest that nonnutritive vessels are relatively fewer than nutritive.

Anatomic models of the muscle microvasculature depicted by Lindbom and Arfors and their colleagues (4) as well as those of Myrhage and Eriksson (25) provide a possible explanation for the close coexistence of nutritive and nonnutritive sites within the muscle body. Detailed drawings of the microvasculature of the rabbit tenuissimus muscle (4) show feed arteries that branch to supply transverse arterioles that in turn cross the muscle body (which is flat in this muscle) to first supply terminal arterioles and capillaries of the muscle and then end in vessels supplying the connective tissue. Our own recent studies suggest that these vessels nourish attendant adipocytes (11) and thus can only be considered nonnutritive for the more metabolically active muscle cells. Furthermore, these connective tissue vessels and their adipocytes are seen in close proximity of the muscle nutritive capillaries (8). Thus two vascular networks operate in parallel and a number of studies using intravital microscopy have shown that relative flow in the two networks can be influenced by various physiologically relevant agents and conditions (4, 22-24) that redistribute flow consistent within the two regions representing the nutritive and nonnutritive routes of muscle (10). For the tenuissimus muscle, the region containing nutritive capillaries is ~3,000-µm wide and somewhat less for the nonnutritive capillaries (23). Thus an LDF probe of 200-µm diameter could be placed to detect exclusively one or the other route. For muscles other than the tenuissimus that are cylindrical, the same vascular unit is observed (25); but now the transverse arteriole radiates out from the center of the muscle fibril and the connective tissue vessels are contained in the perimysium. The dimensions are similar but because of the cylindrical nature of the perimysium the probability of positioning a 200-µm probe to measure nonnutritive flow exclusively would be less than for the tenuissimus and the probability of receiving a mixed signal would be greater. The data we have obtained in the present study and the conclusions we have drawn from the perfused polymer tube studies are not inconsistent with this vascular arrangement.

There are relatively few studies deploying intramuscular LDF microprobes (e.g., see Ref. 28), and none has addressed the issue of heterogeneity of response to vasoconstrictors. There are however a number of studies where larger LDF probes have been used on the surface of skinned muscle similar to the one we report herein. The larger surface probe records frequency changes from much larger volumes of tissue (in many cases approaching 1 mm³) and therefore penetrates considerably below the surface. Of particular interest are studies of muscle blood flow where total and regional flow have been manipulated by pharmacological agents. The findings suggest that the surface LDF probes predominantly detect flow we would characterize as nutritive. In one such study (15), adenosine was infused into anesthetized rabbits. Mean arterial blood pressure decreased and there was an increase in flow heterogeneity, a decrease in the local oxygen consumption (vastus medialis), and a decrease in LDF signal in the same region on the contralateral leg. The authors concluded that adenosine had caused a markedly reduced capillary flow with increased tissue oxygenation. Our studies using the constant-flow-perfused rat hindlimb show that vasodilators such as adenosine strongly oppose the effect of type A (cf. type B) vasoconstrictors such as NE and redirect flow from the nutritive to the nonnutritive route (6).

In a study by Kuznetsova and colleagues (19), the muscle-surface LDF signal was recorded from the biceps femoris muscle of anesthetized rats that had been injected with chlorisondamine chloride, an autonomic blocking agent that blocks nicotinic ganglionic transmission, as well as the ₁-blocker atenolol. Total muscle blood flow was assessed with radioactive microspheres. Infusion of ANG II or phenylephrine increased the LDF signal without affecting total muscle blood flow. In a separate series of experiments, isoproterenol decreased the LDF signal despite a large increase in total blood flow. These findings are consistent with our observations in the constant-flow-perfused rat hindlimb where ANG II and phenylephrine are both type A vasoconstrictors and increase oxygen uptake by increasing nutritive flow. Furthermore, we have found isoproterenol to be a vasodilator that opposes type A vasoconstriction and redirects flow from the nutritive to the nonnutritive route; thus oxygen uptake is inhibited (6).

In skeletal muscle of the rat, the LDF signal has been shown to correlate well with other measurements of flow such as radioactive microspheres or electromagnetic flowmetry, each of which measures volume flow (34). However, as pointed out by Kuznetsova and colleagues (19), the linear correlation may not apply if tissue perfusion, estimated by LDF, changes without any significant change in volume flow to an organ. In fact the principal finding made by these authors was that certain types of vasoconstrictors (as outlined above) were able to cause a dissociation between volume of blood flow and LDF signal. Kuznetsova and co-workers (19) were of the view that changes in the LDF signal accompanying changes in vascular tone were best explained by changes in red blood cell velocity. Thus agents such as isoproterenol that decrease the LDF signal in the light of increased total blood flow do so by decreasing red blood cell velocity, the result of a decreased arteriolar-venular pressure gradient. From the present studies we would argue that the LDF probe recorded a signal from the predominantly nutritive region, and phenylephrine and ANG II increased while isoproterenol decreased blood delivery to this region. The LDF signal reflected the extent of perfusion of the tortuous nutritive network as predicted from the perfused polymer tube studies (5).

In our perfused hindlimb system, total flow was maintained constant and we have previously shown (using microspheres) that agents such as 5-HT do not alter total flow between muscles of high and low oxidative capacity, or between muscle and nonmuscle tissue (30). The fact that 5-HT infusion results in a marked decrease in oxygen uptake, a change in flow pattern (26), and an increase in flow in connective tissue and tendon vessels (27) suggests that 5-HT acts to redirect flow from nutritive capillaries to nonnutritive vessels, which may include tendon vessels. Takemiya and Maeda (36) reported that at rest blood flow in the tendon of the tibialis anterior, gastrocnemius, and soleus exceeds that of the same muscles by approximately twofold. In addition, they showed that NE or exercise decreased tendon vessel flow and that an exercise-mediated decrease in tendon vessel flow occurred in conjunction with increased muscle flow. Together their findings also imply that blood flow can be switched from either of the two routes to match demand.

The proportion by which vasoconstrictors were able to alter muscle metabolism during constant infusion (oxygen uptake was stimulated by 69.5% by NE and inhibited by 29.5% by 5-HT) was similar to the amount that they changed flow, which was indicated by the LDF signal at the NE-positive sites (NE = +51%; 5-HT = 61%). It would therefore seem likely that redistribution of flow between the nutritive (NE positive) route and nearby connective tissue (such as the perimysium, epimysium, or tendon), albeit the nonnutritive (NE negative) route, could account for the observed changes in metabolism as argued previously (7, 10).

In this study we have defined nutritive flow in terms of increases in metabolism, specifically oxygen uptake. We and a number of other groups have shown that oxygen uptake by the perfused rat hindlimb is dependent on total flow (3). In the present study when flow was increased under basal conditions (no vasoconstrictor) from 4 to 8 ml/min, oxygen uptake for the whole hindlimb increased from 45.6 ± 1.9 to 60.9 ± 1.5 (P < 0.05). Thus the perfused rat hindlimb responds with increased oxygen uptake if nutritive flow is increased by either an increase in total flow or switching flow from the nonnutritive route. However, it is important to note that changes in oxygen uptake should not always be considered as a surrogate indicator of changes in nutritive flow; there may be circumstances, for example, when metabolism is increased without a change in capillary recruitment (nutritive flow). One such condition would be when uncoupling of respiration occurs. We also know that changes in oxygen uptake by the perfused rat hindlimb are not the result of direct metabolic effects of the vasoconstrictors acting on skeletal muscle. Thus if the vasoconstriction is blocked by the simultaneous infusion of a vasodilator, both the increase in pressure and oxygen uptake are blocked (12). Similarly, 5-HT shows none of the effects seen in perfused hindlimb muscle when added to isolated incubated muscle, and the effects in perfusion are blocked by vasodilators (30). Overall, nutritive flow might be defined more generally with regard to microvascular flow that is spatially related to a metabolically active compartment in muscle and facilitates nutrient exchange. In the perfused hindlimb, limitation of nutritive flow (for example, by 5-HT) shunts microvascular flow away from a metabolically active compartment resulting in decreased oxygen exchange between the muscle and the perfusate. If oxygen availability becomes limiting in some regions this may secondarily affect fuel metabolism. In addition, blood flow coursing through a muscle bed itself may engender some work on the muscle by mechanisms not yet defined, which alters oxygen consumption. Clearly, oxygen consumption is one particular manifestation of microvascular recruitment that is prominently affected in the perfused hindlimb.

Finally, LDF probe dimensions may be important in determining the nature of the signal received. The LDF surface probe used in this study had a detector surface area of ~1 mm² (two optical fibers of 800-µm diameter each), and when placed on the surface it detected a signal from a volume of ~1 mm³ of tissue. When positioned on the center of the anterior end of the biceps femoris, the signal appeared to be only of one kind, responding positively to NE and negatively to 5-HT. This contrasts with measurements made using a much smaller probe (0.03 mm²) inserted in the muscle; here LDF signals were heterogeneous, with some responding as above (i.e., NE positive), some responding in an opposite manner to the above (i.e., NE negative), and some failing to respond (mixed). With the smaller probe it was also noted that the NE-positive sites outnumbered the other two by ~3:1. Thus it appears likely that the larger probe receives signal from a mixture of sites that convey a character that is predominantly NE positive. Alternatively, NE-negative sites may not be located near the surface of the muscle, although for various reasons alluded to above this is unlikely.

In summary, when positioned randomly in the body of a number of hindlimb muscles, LDF microprobes identify sites that differ in their response to vasoconstrictors. This heterogeneity is not visible to larger probes on the muscle surface. Over half of the sites show properties consistent with a nutritive role for muscle metabolism. Nonnutritive sites are present but at a lower proportion, and nearly a quarter of the sites are likely to represent a mixture of both nutritive and nonnutritive. Nonnutritive sites show a higher basal signal.