Skin blood flow increases in response to local heat due to sensorineural and nitric oxide (NO)-mediated dilation. It has been previously demonstrated that arteriolar dilation is inhibited with NO synthase (NOS) blockade. Flow, nonetheless, increases with local heat. This implies that the previously unexamined nonarteriolar responses play a significant role in modulating flow. We thus hypothesized that local heating induces capillary recruitment. We heated a portion (3 cm2) of the Pallid bat wing from 25°C to 37°C for 20 min, and measured changes in terminal feed arteriole (∼25 μm) diameter and blood velocity to calculate blood flow (n = 8). Arteriolar dilation was reduced with NOS and sensorineural blockade using a 1% (wt/vol) NG-nitro-l-arginine methyl ester (l-NAME) and 2% (wt/vol) lidocaine solution (n = 8). We also measured changes in the number of perfused capillaries, and the time precapillary sphincters were open with (n = 8) and without (n = 8) NOS plus sensorineural blockade. With heat, the total number of perfused capillaries increased 92.7 ± 17.9% (P = 0.011), and a similar increase occurred despite NOS plus sensorineural blockade 114.4 ± 30.0% (P = 0.014). Blockade eliminated arteriolar dilation (−4.5 ± 2.1%). With heat, the percent time precapillary sphincters remained open increased 32.3 ± 6.0% (P = 0.0006), and this increase occurred despite NOS plus sensorineural blockade (34.1 ± 5.8%, P = 0.0004). With heat, arteriolar blood flow increased (187.2 ± 28.5%, P = 0.00003), which was significantly attenuated with NOS plus sensorineural blockade (88.6 ± 37.2%, P = 0.04). Thus, capillary recruitment is a fundamental microvascular response to local heat, independent of arteriolar dilation and the well-documented sensorineural and NOS mechanisms mediating the response to local heat.
- precapillary sphincter
- skin blood flow
mechanisms that affect the transfer of heat include skin blood flow and the surface area available for heat dissipation (11). Previously, investigators have found marked increases in blood flow with the application of local heat in mammals (9, 11, 28, 29, 44, 53). Skin blood flow increases in a biphasic manner when locally heated (11, 28, 53). The initial increase in flow is primarily mediated by a local axonal reflex and is followed by a brief nadir (37). The second more sustained phase is believed to be primarily mediated by nitric oxide (NO) release and arteriolar dilation (28). The focus on increasing blood flow as a mechanism to dissipate heat has led to identifying the central roles of sensorineural and NO-mediated arteriolar dilation. The possible role of increased vascular surface area has not been addressed.
Although capillary recruitment has been studied extensively since the seminal frog mesenteric studies of August Krogh (30), the control of precapillary sphincters, bands of smooth muscle that encircle capillary branch points (8, 15), is still not well understood. Nonetheless, capillary recruitment has been associated with the ability of the microcirculation to alter vascular surface area (18). The majority of relative vascular resistance is at the arteriolar level during control conditions (42). However, factors playing a minor role during normal conditions, such as opening and closing of precapillary sphincters, may have a considerable impact on total vascular resistance when arteriolar resistance is greatly reduced, such as during local heating. In fact, capillary recruitment has been reported to decrease vascular resistance and increase microcirculatory blood volume, independent of upstream arteriolar activity (45, 51). An increase in blood volume in the capillary beds of the skin is evident with erythema (5), a common symptom associated with local and global overheating (48). Although increasing microvascular surface area has been identified as an ideal method to dissipate heat (40), laser Doppler fluximetry (LDF) studies are limited by their inability to quantify capillary recruitment.
The Pallid bat wing has long been used to study, noninvasively, the in vivo microcirculation in a controlled setting (15, 54, 55). This model has advantages over traditional skin blood flow studies, because investigators are able to pharmacologically manipulate and characterize the hemodynamics and structure of individual vessels of the thin, translucent wing without LDF. Furthermore, the microcirculation can be studied without the confounding effects of surgery, anesthesia, or drug delivery requiring invasive methods. The bat wing is similar to other mammalian microvascular beds; arterioles dilate with decreased pressure (the myogenic effect; 17, 55), deinnervation (23, 54), and increased metabolic factors (3, 6). The bat wing is also innervated by the autonomic and sensory nervous systems (19). Additionally, bat wing microvessels constrict with epinephrine and norepinephrine (36) and dilate with acetylcholine (49), serotonin (49), and increased blood flow (16).
The bat wing has been previously established as a microvascular preparation exhibiting a vasodilatory response to heat stimuli. In fact, vascular diameter, vasomotion, and interstitial pressure all increase with the application of local heat (9, 29, 44, 52). Applying lidocaine to the bat wing to inhibit sensory nerves attenuates the increase in both arteriolar diameter and blood flow with local heat by ∼50%. With local heat, NO synthase (NOS) inhibition also results in a similar reduction of flow (∼50%), though it yields no significant increase in arteriolar diameter (53). Thus, the bat wing has been used to illustrate that a significant increase in arteriolar blood flow with local heat can occur without a corresponding change in arteriolar diameter (53). More importantly, the preparation allows for the visualization of individual capillary segments, as well as of individual vessel diameters. We therefore used the bat wing model to test the hypothesis that local heating induces capillary recruitment.
Animal care and experimental procedures were in compliance with protocols approved by the Texas A&M Institutional Animal Care and Use Committee. Two colonies of adult bats (25 total) were kept in one room that allowed free access to food, water, and flight opportunities (for exercise). Room lighting was automated to provide a 6:18-h light-dark cycle. Female bats were utilized for these studies because male bats were resistant to training. Bats were kept in a room with a constant temperature (∼26°C) throughout the year, to minimize the effects of an estrus cycle and maintain stable circulating estrogen and progesterone levels (41). Bats were trained to remain resting, unanesthetized, inside a plastic box with one wing extended outside of it (15, 55). The wing was visualized with an intravital microscope (Olympus BX61WI Fixed Stage Upright Microscope), and the image was captured via charge-coupled device camera (Panasonic KR222 S-Video Camera). The image was then relayed to a computer that analyzed and recorded it with specially designed diameter tracker software (LabView 7.1) (35). An optical Doppler velocimeter (model 4; A&M Health Systems) detected centerline red blood cell (RBC) velocity. The frequency response of the instrument is within the range of interest (0 Hz to 20 Hz), and proper instrument calibration allows measurement of RBC velocity using ×10 to ×60 objectives (7). Vessel diameter and RBC velocity were recorded at 30 frames/s. AVI videos were also recorded and converted to DVD format for postexperiment analysis.
Experimental procedures were similar to those employed previously (53). Briefly, bats were placed in a plastic box attached to a temperature-controlled heat plate (Olympus Tokai Hit Thermo plate). The experimental room was maintained at 25°C. The region of the wing touching the heat plate was restricted to the specific area of study (no larger than 3 cm2) between the fourth and fifth phalanges. Bat wing temperature was recorded at the end of every minute by using a laser thermometer (Raytek Raynger ST Pro). As the experiment began, heat plate temperature was set at 25°C for 10 min to establish baseline. Heat plate temperature was then increased from 25°C to 37°C for 20 min. The increase in temperature occurred over a period of 60–90 s, consistent with earlier work (53). The rate at which the bat wing was heated (∼0.1°C/s) is similar to that used in human studies (27, 37). This temperature range was selected to avoid disturbing the resting bats (53). The experiments were divided into three separate protocols consisting of a control and pharmacological blockade groups that were all performed in different experimental sessions and on different bats. An established experiment schedule ensured that each bat was used for experiments no more than once per week, and for no more than 1 h/experiment.
Arteriolar response to local heat.
The first protocol consisted of a set of experiments that examined the responses of terminal feed (i.e., ∼25 μm) arterioles to local heat. Arterioles were monitored with (n = 8) and without (n = 8) pharmacologic blockade consisting of a 2% (wt/vol) lidocaine plus 1% (wt/vol) l-NAME solution applied to a 3-cm2 area on both sides of the wing. Vessel diameters and RBC velocities were recorded during baseline at 25°C and with local heating at 37°C.
Capillary perfusion in response to local heat.
The second protocol consisted of a set of experiments that examined the capillary response to local heat and was divided into two separate groups, each with a 10-min baseline at 25°C followed by heating at 37°C for 20 min. For the first group (n = 8), the wing was imaged at ×100 magnification, providing a field of view of ∼600 μm ×480 μm populated by capillary segments 6–9 μm in diameter (5.0 ± 1.13 segments per field). To ensure that larger vessels do not obscure the visual field, a section of the wing was chosen that lacked arterioles or venules >15 μm in diameter. Perfused capillary segments were identified by a single column of moving RBCs contained within vessels originating from metarterioles or other capillaries. Capillary segments perfused with RBCs were counted at the end of every minute at both 25°C and 37°C. The second group involved a similar preparation; however, the 2% (wt/vol) lidocaine plus 1% (wt/vol) l-NAME solution was applied to a 3-cm2 area on both sides of the wing prior to placing it on the heat plate (n = 8). This allowed for pharmacological blockade of both the sensorineural and NO-mediated phases of the response to local heat (37).
Precapillary sphincter response to local heat.
The third protocol, performed separately from the previous two protocols, also contained control and pharmacological blockade groups. The first group had no pharmacologic treatment, and involved identifying a capillary segment originating from a metarteriole with a precapillary sphincter, viewed at ×600 magnification (n = 8). RBC velocity and video were recorded. The amount of time precapillary sphincters were open and the RBC velocities were examined before and after local heating. The second group in this protocol had 2% (wt/vol) lidocaine plus 1% (wt/vol) l-NAME solution applied to a 3-cm2 area on both sides of the wing prior to the experiment (n = 8). To minimize analysis time, we analyzed the final 20 s of every minute.
Data were pooled for each minute throughout the experiment and then averaged across the entire experimental subject population for that particular minute in the protocol. We compared baseline values to data collected during the final 10 min of local heating, which corresponded to steady state in our previous studies (53). Data are presented as means ± SE. Student's t-test, repeated-measures ANOVA, and linear regression were performed when appropriate to determine statistical significance, followed by Fisher's least significant difference post hoc statistical test. A P value of <0.05 was considered significant.
With local heating, bat wing temperatures (36.4 ± 1.8°C) correlated with heat plate temperatures (r2 = 0.901). Local heat significantly increased arteriolar diameter 35.6 ± 4.3% (P = 0.00002) and increased blood flow 187.2 ± 28.5% (P = 0.00003). Inhibition of sensorineural and NOS-mediated mechanisms eliminated arteriolar dilation with local heat (−4.5 ± 2.1%, Fig. 1). However, increases in calculated blood flow with local heat persisted (88.6 ± 37.2%, P = 0.003, Fig. 1), despite pharmacological blockade. NOS plus sensorineural blockade did significantly reduce the increases seen in blood flow with local heat by nearly one-half (187.2 ± 28.5% vs. 88.6 ± 37.2%, P = 0.04).
With an increase in local temperature from 25°C to 37°C, the number of perfused capillary segments per field of view increased from 5.0 ± 1.1 to 9.6 ± 1.4 (92.7 ± 17.9%, P = 0.011, Fig. 2). Despite the addition of 2% (wt/vol) lidocaine plus 1% (wt/vol) l-NAME solution, the number of perfused capillaries per field also increased from 6.2 ± 0.9 to 12.3 ± 1.9 (114.4 ± 30.0%, P = 0.014, Fig. 2).
Precapillary sphincters opened and closed cyclically during both baseline and during local heating. To analyze the precapillary sphincter activity, we quantified both the absolute time that sphincters opened and the relative percent of the cycle that sphincters were open. The time that precapillary sphincters remained open per cycle increased (81.6 ± 13.1%, P < 0.0001, Fig. 3) with local heat. This increase was sustained (94.0 ± 19.6%, P < 0.0001, Fig. 3) despite pharmacological blockade. Repeated-measures ANOVA revealed significant differences between baseline and heating values but not between control and blockade groups.
The percent of cycle time that the precapillary sphincters remained open with heat also increased from 56.1 ± 5.8% to 88.4 ± 3.4% (P = 0.0002, Fig. 4). The increase in percent time open was sustained with pharmacological blockade (50.6 ± 7.0% to 84.7 ± 6.5%, P = 0.003, Fig. 4). Again, there was a significant difference between baseline and heating groups, but not between control and blockade groups.
RBC velocity through the capillary sphincters increased 63.3 ± 3.3% with heat (Fig. 5). The percent increase in velocity with heat was not statistically different from the control group with pharmacological blockade (52.5 ± 7.0%, P = 0.22, Fig. 5).
Our results indicate that local heat does indeed induce capillary recruitment in the bat wing. First, the number of perfused capillaries increased within the first few minutes of local heating (Fig. 2). Second, the time that precapillary sphincters remained open increased (Figs. 3 and 4). Finally, the relative increase in arteriolar blood flow was greater than the relative increase in capillary blood flow, indicating that blood was dispersed into multiple capillaries. In each of these results, increases in capillary perfusion occurred in the absence of feed arteriolar dilation due to NOS and sensorineural blockade, consistent with our previous work (53). Since the combination of sensorineural and NOS inhibition had no measurable effect on capillary perfusion, we did not separately investigate their effects on capillary recruitment. Not only does capillary recruitment increase vascular surface area, it may be responsible for a considerable proportion of the biphasic flow response to local heating. We could perform these novel studies since the thermoregulatory microcirculation of the Pallid bat wing enables visualization of a physiological phenomenon that is not accessible in intact human skin studied with LDF.
The importance of capillary recruitment with local heat.
A simple calculation illustrates the substantial role that capillary recruitment plays in the response to local heat. Without pharmacological intervention, terminal feed arterioles dilate 35% and blood flow increases ∼190% with local heat. With NOS and sensorineural blockade, however, arteriolar dilation is completely blocked, and yet RBC velocity still significantly increases. The increase in capillary number (∼100%, Fig. 2) with local heat and pharmacological blockade is similar to the sustained increase in blood flow observed in the upstream arterioles after inhibiting dilation (∼90%, Fig. 1). Thus, it can be inferred that roughly half the increase in flow is due to capillary recruitment.
While our experiments provide multiple pieces of evidence for capillary recruitment, several observations could be misinterpreted. For instance, arterial-venous (A-V) shunting in response to local heat may contribute to the increased flow response without altering capillary perfusion. Our observations support the work of Davis (15), however, who did not find evidence of A-V shunts in the Pallid bat wing. Similarly, venous behavior with heat may also affect the blood flow response and capillary perfusion. The response of venular networks to local heat has not yet been characterized. It also remains possible that the number of perfused capillaries, precapillary sphincter opening time, and capillary RBC velocity increased only in the field of view, but decreased in areas outside of it. However, care was taken not to introduce selection bias by choosing a particular field of view for study. Although we used accepted pharmacological doses (37, 53, 57), it was also possible that the blockade was not complete. Nonetheless, pharmacological blockade did abolish the response to touch, as well as the response of the terminal feed arterioles to heat (Fig. 1) and did not significantly alter baseline values for perfused capillaries (Fig. 2) or precapillary sphincter activity (Figs. 3 and 4). This is consistent with reports that lidocaine does not increase the number of perfused capillaries without local heat (4, 22).
Relevance to skin blood flow in other mammals.
The largest limitation of this study is, of course, the relevance of results for the response to local heating in other mammals. As described in detail in the introduction, the behavior of the bat wing microvasculature is similar to more common animal models (3, 16, 17, 19, 23, 36, 49, 54, 55). More importantly, the bat wing has been previously shown to have a thermoregulatory microcirculation (29, 44) and has exhibited blood flows similar to those reported in human skin, including a well-defined biphasic flow response to local heat (53). The size of the region of tissue studied, the pharmacological agents used, and the rate at which the bat wing was heated is consistent (37) with those previously reported in human LDF studies (28). Our results indicated that the activity of precapillary sphincters increases within the first minute of applying local heat (Figs. 3 and 4), leading us to believe that capillary recruitment is involved in both the first and second phases in biphasic flow response. Although we have previously reported an initial peak in flow within 1–5 min of locally heating the bat wing (53), this study indicates that a corresponding peak in capillary recruitment is notably absent. Accordingly, capillary recruitment may play a minimal role in the initial increase in blood flow with local heating. Furthermore, capillary recruitment has a salutary effect of increasing microvascular surface area, which can help dissipate heat, or help transfer heat from tissues to protect against local thermal injury.
Mechanisms regulating capillary recruitment.
Despite great strides in the study of capillary recruitment in general, much work remains to elucidate the mechanisms of capillary recruitment in response to local heat. Initial studies of capillary recruitment in skeletal muscle focused on adrenergic stimuli as primary mediators (24, 32, 56). Previous skin blood flow studies have mostly discounted the roles of adrenergics (12, 37), as well as prostanoids (21), in the response to local heat. These mechanisms appear to be more closely associated with whole body thermoregulation, as opposed to the strikingly different response to local heat. Although reports have discounted a sympathetic contribution to the biphasic response to local heat (43), a recent study reports that the response to slower and more sustained local heating may be influenced by adrenergic mechanisms (25). Investigators have also implicated NO as playing a role in capillary recruitment in other microcirculatory beds (14, 33, 50). Our data indicate that NOS inhibition has the greatest effect on upstream arteriolar dilation, and no significant effect on capillary recruitment (Figs. 1–5). This is consistent with previous work in hamster cheek pouch indicating that NO has disparate impact on the different portions of a microcirculatory network. In particular, NO has a greater effect on vessels with more smooth muscle (arterioles) than those with limited smooth muscle (i.e., metarterioles) (20). It appears from our results (Figs. 3 and 4) that precapillary sphincters, separating metarterioles and nonmuscular capillaries, are relatively insensitive to NOS inhibition. Besides arising from different mechanisms, the discrepancy with other reports may lie in the preparation or methodology of estimating capillary recruitment in human subjects (46). It also remains possible that a different isoform of NOS, heat shock proteins (34), or some downstream NO-dependent mechanism (57) could act to increase capillary perfusion. ATP-sensitive K+ channels have also been shown to play a role in capillary recruitment (14). To our knowledge, there have been no studies implicating a role of K+ATP channels in the flow response to local heat. Figures 3–5 suggest that the opening of precapillary sphincters is a proximate cause for capillary recruitment and thus a primary site of action.
Clinical implications of the biphasic flow response and capillary recruitment.
Many investigators have attempted to utilize skin blood flow as a measure of peripheral vascular health in a variety of pathologies (2, 26, 31, 38). Diabetes has been a main focus in many of these studies and is characterized as a disease of the microvasculature, including endothelial dysfunction, decreased capillary density, and vascular smooth muscle deficiencies (13). Previously, the biphasic response to local heat has been proposed as a screening tool for endothelial dysfunction and poorly functioning smooth muscle (10). Our results suggest that altered flow response in human skin may also be indicative of global capillary recruitment deficiencies resulting from diabetes (13) or renal failure (47). The flow response to local heat may thus provide a diagnostic indicator of the progression of peripheral microvascular dysfunction in patients afflicted with hyperglycemia, insulin resistance, and then frank diabetes. Capillary recruitment is also involved in the microvascular response to insulin, and causes a biphasic flow response in mesenteric microcirculation (39). This response has been linked to the involvement of K+ATP channels (1), and is also deficient in the early stages of insulin resistance. Thus, K+ATP channels may play a role in both insulin sensitivity and capillary recruitment.
In summary, our results provide the first direct evidence of capillary recruitment accompanying the biphasic flow response to local heat in a mammalian microvasculature. Interestingly, this response is independent of sensorineural and NO-dependent mechanisms, previously thought to be the sole mediators of the response to local heating in a thermoregulatory microcirculation. These results may provide investigators an additional means to develop the response to local heat as a diagnostic tool to evaluate the health of the peripheral microcirculation.
Portions of this work were supported by grants National Heart, Lung, and Blood Institute Grant K25-HL-070608, American Heart Association Grant AHA-05651164, and Centers for Disease Control Grant CDC-623086.
We would like to thank Dr. Cristine Heaps for insightful comments.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 the American Physiological Society