comparable effects of arteriolar and capillary stimuli on blood flow in rat skeletal muscle

Microvascular Research 53, 22–32 (1997)Article No. MR961988

Comparable Effects of Arteriolar and Capillary Stimulion Blood Flow in Rat Skeletal Muscle

Debra Mitchell, Jingcheng Yu, and Karel TymlJohn P. Robarts Research Institute and Departments of Medical Biophysics and Pharmacology and Toxicology,The University of Western Ontario, London, Canada, N6A 5C1

Received May 28, 1996

Although the capillary wall represents an active interface to those of capillary stimulations. When testing for thespeed of response in these two microvessels, the timebetween blood and tissue, the potential role of the capil-

lary in blood flow control has not been determined. The of noticeable VRBC change after NE (i.e., 10% from con-trol) was also similar. We concluded that (i) the ratgoals were (i) to establish the presence of the capillary

sensing and communication phenomenon (Dietrich and skeletal muscle capillary could respond to a variety oflocally applied materials and (ii) the capillary could haveTyml, Microvasc. Res. 43, 87–99, 1992) in mammalian

microvasculature and (ii) to determine the relative sensi- as profound an effect on microvascular flow as the arteri-ole. Thus capillary could have the potential to participatetivity of the capillary and the arteriole to locally applied

vasoactive agents. Using intravital video microscopy, in microvascular flow control. q 1997 Academic Press

norepinephrine (NE; 1007–3 1 1003 M), acetylcholine(ACh; 1004–1002 M), or bradykinin (BK; 1009–1003 M)was applied via micropipettes on capillaries (300 mm INTRODUCTIONdownstream from feeding arterioles) or on arterioles, atthe surface of the extensor digitorum longus muscle of

Our work in frog skeletal muscle yielded evidenceanesthetized rats. Red blood cell velocity (VRBC) in capil-that the capillary could sense vasoactive agents de-laries and arteriolar diameters was measured from videoposited locally on the tissue and communicate theirrecordings. The overall control VRBC and control diame-presence to the feeding arteriole to alter its tone (Die-ter were 190 mm/sec and 8.3 mm, respectively. NE ap-trich and Tyml, 1992a,b). We hypothesized that thisplied on the capillary caused a dose-dependent reductionsensing involved a receptor- or nonreceptor-medi-in VRBC (up to 100%, i.e., 0 mm/sec) via a constrictionated change in the membrane potential of the capil-of the feeding arteriole. Both ACh and BK applied onlary endothelium and/or pericyte and that thisthe capillary caused a dose-dependent increase in VRBC

change spread electrotonically along the capillary to(up to 115%) via arteriolar dilation. Based on two differ-the arteriole to mediate the upstream response (Songent approaches, these responses could not be explainedand Tyml, 1993). Recent reports that the capillary en-in terms of diffusion of agents from capillary to the arte-dothelium can depolarize (McGahren et al., 1996;riole. When testing for the relative sensitivity of theBeach et al., 1996) and that a local depolarization canarteriole and the capillary, application of NE and ACh

on arterioles caused VRBC and diameter responses similar be conducted from the arteriole to the capillary and

0026-2862/97 $25.00Copyright q 1997 by Academic Press

All rights of reproduction in any form reserved.22

AID MVR 1988 / 6z0b$$$161 01-07-97 02:43:59 mvrgs AP: MVR

Arteriolar and Capillary Effects on Blood Flow 23

along the arteriolar endothelium (Beach et al., 1996; Budreau, 1991). The EDL muscle surface microcircula-tion was then visualized via a Leitz ELR microscopeXia et al., 1995) support this hypothesis.

There is evidence for inter- and intraspecies heteroge- and a closed circuit television system (MTI camera, Pan-asonic WV5410 monitor, NV9240XD tape recorder).neity of endothelial/capillary function (Scott and Bick-

nell, 1993; Mehrke et al., 1991; Revest and Abbot, 1992). Three preparations of the muscle surface were used.The first allowed for visualization at low magnificationTherefore, it is not clear whether the mammalian capil-

lary can respond to a variety of agents as does the am- (0.781 1.05-mm field of view on the television monitor)with a 6.31 eyepiece and a 10X/0.22 NA objective. Aphibian capillary. Since our current concepts of micro-

vascular flow control are primarily based on mamma- plastic coverslip (18 1 18 mm) with a hole was placedon the EDL muscle and filled with degassed paraffinlian data, and since there is little information on

capillary responsiveness in mammalian tissue, the first oil, as described by Mitchell and Tyml (1996). A thintissue fluid layer (60 mm thick) was formed betweenobjective of the present study was to establish the key

features of the capillary sensing and communicating the muscle surface and the oil.The second preparation allowed for visualization atphenomenon in rat skeletal muscle.

To our knowledge, the significance of the capillary higher magnification (0.154 1 0.207-mm field of view)with a 101 eyepiece and a 32X/0.40 NA objective. Theresponse in microvascular blood flow control has not

been addressed. Despite its potential ability to respond preparation was prepared as described in Mitchell andTyml (1996). This included placing the muscle onto ato vasoactive agents, the capillary could be much less

sensitive to agents than the arteriole, allowing for only coverslip, covering it with Saran wrap (Dow Chemical),and irrigating with warm saline (0.9% NaCl at 377). Aa minor role in blood flow control. The second objective

was to establish the relative sensitivity of the capillary small rectangle (Ç2 1 4 mm) was cut in the Saran toallow micropipette access and was filled with degassedand the arteriole to locally applied vasoactive agents.paraffin oil.

The third preparation was similar to the second prep-aration, but the EDL was superfused continuously and

METHODS unidirectionally with Krebs’ solution (377; rate: 2 ml/min; gassed with 95% N2/5% CO2; composition in mM:131.9 NaCl, 4.7 KCl, 20 NaHCO3, 1.2 MgSO4, and 2Animal PreparationCaCl2). It was visualized with the 6.31 eyepiece andthe 10X/0.22 NA objective (0.78 1 1.05-mm field ofAnimal protocols were approved by the University

Council on Animal Care at the University of Western view). In all three preparations, experiments were pre-ceded by a blood flow stabilization period (30 min) andOntario. Male Wistar rats (321 { 8 SE g; N Å 115; Har-

lan–Sprague–Dawley) were anesthetized with sodium continued for 4 hr on average.pentobarbital (65 mg/kg, ip; Somnotol, MTC Pharma-ceuticals) with supplemental injections (22 mg/kg) ev- Local Application of Materialsery 30–60 min, as needed. The rat was placed on thestage of an intravital microscope (Leitz, ELR) equipped Experiments were carried out on randomly chosen

surface microvessels of the EDL muscle. Up to sevenwith a heating pad (377C; American Medical System,K20C) and a heating lamp. A tracheotomy was per- microvessels (either capillaries or arterioles) were stim-

ulated in each EDL muscle. We used (i) the terminalformed to facilitate breathing. In some experiments, thecarotid artery blood pressure was measured via a Cobe portions of arterioles that fed capillaries and (ii) capil-

laries that ran parallel to the skeletal muscle fibers andpressure transducer. The animal was then placed onits side with one leg held by a clamp attached to the that were at least 500 mm in length. Figure 1 shows

the experimental arrangement for local stimulation ofmicroscope stage. The extensor digitorum longus (EDL)muscle was exposed as described previously (Tyml and microvessels. Arterioles were stimulated at Site 1 and

Copyright q 1997 by Academic PressAll rights of reproduction in any form reserved.


24 Mitchell, Yu, and Tyml

from the fluid layer. Based on the average EDL musclefiber thickness of 40 mm (Tyml et al., 1995), we estimatedthe distance between the muscle surface and a mi-crovessel to be in the range from 20 to 60 mm.

Pressure ejection. Glass pipettes (õ5 mm tip diame-ter) were backfilled with bradykinin (BK; Sigma) dis-solved in distilled water (dH20) or NE in DPBS stainedwith 0.5% Evan’s blue. Each pipette (mounted on amicromanipulator) was connected to Picospritzer II

FIG. 1. Schematic diagram of the experimental arrangement for lo- (General Valve Corp.), which ejected the drug from thecal stimulation of microvessels. In the majority of experiments, arteri-

pipette using pressurized air. We used the approach ofoles were stimulated at Site 1 and the VRBC responses were measuredDietrich (1989) and Mitchell and Tyml (1996) to ejectin capillary B, or capillaries were stimulated at Site 2 and the VRBC

responses were measured in capillary B. In several experiments, arte- and deposit small spherical droplets of drugs (Ç65 pl)rioles were stimulated at Site 1 and the arteriolar diameter responses on the muscle surface. The single pressure ejection con-measured at Site 1, or capillaries were stimulated at Site 2 and the fines the agent exposure to a small segment of the vas-arteriolar diameter responses measured at Site 1.

culature (Delashaw and Duling, 1991).

Evaluation of Microvascular Responsecapillaries at Site 2 (300 mm downstream from the feed-ing arteriole). We applied drugs locally either by ionto- The microvascular response to agents was evaluated

in the first and third preparations by means of the redphoresis (for charged drugs) or by pressure ejection(for uncharged drugs; Dietrich, 1989). Throughout this blood cell velocity (VRBC) measurements in the capillary

and in the second preparation by means of arteriolarpaper, the given concentrations are the concentrationsin the microelectrode or in the pressure-ejection pipette. diameter measurements.

VRBC was measured by the flying spot techniqueIontophoresis. Glass microelectrodes (õ1 mm tipdiameter) were backfilled with one of the following (Tyml and Ellis, 1982). VRBC con was the average VRBC 1–

2 min before drug application. After stimulation, thesubstances: (0)-norepinephrine hydrochloride (NE;Sigma, St. Louis, MO), acetylcholine chloride (ACh; microcirculatory view was recorded until the flow re-

turned to the control level, or at least to a level withinSigma), and atropine methyl bromide (Atr; Sigma). NEand Atr were dissolved in Dulbecco’s phosphate buffer {15% of VRBC con . Experiments where VRBC did not re-

turn to {15% VRBC con were not used. The lag time of asolution (DPBS, pH 7.4; Gibco, Burlington, ON, CatalogNo. 310-4040AJ). ACh was dissolved in phosphate response was defined as the time between agent appli-

cation and VRBC reaching{10% from VRBC con . The dura-buffer solution (PBS, pH 6.8; composition in mM: 138.9NaCl, 2.25 KCl, 1.75 KH2PO4, 1.4 Na2HPO4) to prevent tion of a response was defined as the time from the

onset of the response until the time when VRBC reacheddrug hydrolysis. Each microelectrode was mounted ona micromanipulator and electrically coupled to a mi- VRBC con , or a stable level within {15% of VRBC con .

VRBC test was defined as the velocity at its greatest changecroiontophoresis current programmer (World PrecisionInstruments, Model 260). A reference electrode was in- from VRBC con . Data were expressed as a percentage

change from VRBC con , DVRBC(%) Å 100% 1 (VRBC test 0serted into the gastrocnemius muscle of the same leg.Drugs were ejected by a /80 nA current for 10 sec. VRBC con)/VRBC con .

Arteriolar diameters (midwall to midwall) were mea-Current ejection started as soon as the microelectrodeentered the tissue fluid layer overlying the muscle sur- sured offline from the video screen (Mitchell and Tyml,

1996). The diameter resolution was {1 mm. The diame-face above the microvessel. The microelectrode did nottouch the muscle surface. At the end of the 10-sec stimu- ters were measured for 1–2 min prior to stimulation

(Dcon) and at the maximum change of the diameterlation period, the microelectrode was withdrawn away




within the first 3 min after stimulation (Dtest). We mea- on the agonist-induced remote response. If the responsewere due to diffusion from the capillary to the arteriole,sured arteriolar diameters at multiple sites along a 200-

mm length of the arteriole upstream from the branching then the arteriolar pretreatment with antagonist shouldattenuate the response more strongly than the capillaryinto capillary B. Reported diameters were from the site

of the maximal diameter change. Data were expressed pretreatment. ACh (10 mM) was applied at Site 2 andVRBC response in capillary B recorded. After this re-as a percentage change from Dcon , DD(%) Å 100% 1

(Dtest 0 Dcon)/Dcon . sponse, the muscarinic receptor antagonist Atr (30 mM)was applied at either Site 1 or Site 2. After 3.7 min(average), ACh (10 mM) was reapplied at Site 2 andExperimental ProtocolVRBC response recorded. In two additional experiments,we also tested for (i) effect of Atr on the direct arteriolarWe tested for (a) capillary sensing and communica-

tion phenomenon and (b) capillary versus arteriolar re- response to ACh and (ii) nonspecific effect of Atr atthe capillary. In the first experiment, VRBC response insponsiveness to locally applied agents.

(a) Capillary sensing and communication phenome- capillary B to ACh (10 mM) applied at Site 1 was com-pared to ACh response following a 3.7-min arteriolarnon. The phenomenon was defined here as a distant

retrograde effect of local capillary stimulation on the pretreatment with Atr (30 mM). In the second experi-ment, VRBC response in capillary B to NE (0.3 mM) ap-feeding arteriole diameter (Dietrich, 1989). Using the

first preparation, vasoactive agents (NE: 100 nM 0 30 plied at Site 2 was compared to response to NE (0.3mM, Site 2) following a 3.7-min capillary pretreatmentmM; ACh: 100 mM 0 10 mM; or BK: 1 nM 0 1 mM)

were applied locally to capillary B at Site 2. VRBC re- with Atr (30 mM) at Site 2.sponses were measured in capillary B (Fig. 1). In order (b) Capillary versus arteriole. To determine theto establish that these responses were mediated by the relative sensitivity of the capillary and the arteriole tofeeding arteriole, we measured arteriolar diameter re- locally applied agents, we used the first preparation tosponses using the second preparation in separate ani- apply NE (10 mM) or ACh (10 mM) to Site 1 or to Sitemals. 2. The VRBC response was measured in capillary B. Us-

The concern that the communication phenomenon ing the second preparation, the arteriolar diameter re-represents the effect of direct diffusion of agents from sponse was measured following NE (10 mM) or AChthe capillary to the arteriole has been addressed (Die- (10 mM) application to Site 1 or to Site 2.trich and Tyml, 1992a,b; Song and Tyml, 1993). We usedtwo strategies to assess the role of diffusion in the EDLpreparation. The first strategy involved superfusion of

Data Analysisthe muscle surface with Kreb’s solution, a procedurecommonly used to eliminate the effect of diffusion (Xia

Data were expressed as the mean { SE. Data wereet al., 1995; Frame and Sarelius, 1995). Using the thirdanalyzed using Student’s t test, the Wilcoxon Signedpreparation, we directed the superfusate away from theRanks test, and the Mann–Whitney test. The null hy-feeding arteriole and the collecting venule. Either DPBSpothesis (DVRBC Å 0% or DD Å 0%) was rejected ifor 3 mM NE (both stained with 0.5% Evan’s blue dye)statistical significance (Põ 0.05) was reached. Kruskal–was pressure-ejected (8–10 psi, 5 msec) onto Site 2 andWallis tests were performed to determine if a responsethe response measured in capillary B. The Evan’s bluewas dose-dependent. ANOVAs were performed amongdye allowed visual confirmation that (i) material wasthe VRBC con in different experiments to ensure that theejected from the pipette and (ii) the bulk of the materialcontrol velocities came from the same population. Inwas swept away from the capillary, the arteriole, andthis paper, n indicates the number of blood vessels usedthe venule.while N indicates the number of animals. Statistics wereThe second strategy was based on the effect of antag-

onists, applied either on the capillary or on the arteriole, calculated using n, unless otherwise stated.




reductions, while BK and ACh caused VRBC increases, inall visible capillaries fed by the same terminal arteriole,rather than in the stimulated capillary alone. The paral-lel VRBC changes in these capillaries were matched byarteriolar constrictions and dilations, respectively (Fig.3). For a particular set of NE stimulations (n Å 13),we were able to measure simultaneously VRBC in thestimulated capillary and diameter in the feeding arteri-ole. All 13 stimulations resulted in constriction (left bar,Fig. 3) and in subsequent capillary VRBC reduction(DVRBC Å 090 { 5%). For vasodilative stimuli, in gen-eral, the degree of VRBC increase was proportional tothe degree of dilation (compare DVRBC and DD for both1002 M ACh and 1004 M BK in Figs. 3 and 4). Together,these data indicated that the diameter change elicitedFIG. 2. Effect of locally applied norepinephrine (NE; 1005 M), acetyl-by stimulation at Site 2 was the major cause for thecholine (ACh; 1002 M), and bradykinin (BK; 1004 M) on red blood cell

velocity (VRBC) in capillary B (Fig. 1). These agents were locally ap- VRBC change in capillary B.plied at Site 2 at time Å 0 sec (dotted line). Figure 4 summarizes the DVRBC dependencies on NE

in the range 1007–3 1 1003 M, on ACh in the range1004–1002 M, and on BK in the range 1009–1003 M. All

RESULTS responses were dose-dependent.Figure 5 (left and middle bars) shows that capillary

Atr blocked the DVRBC response to capillary ACh. Ap-Control Experimentsplication of Atr itself did not change VRBC (DVRBC Å

Current-ejected PBS (DVRBC Å 20 { 11%; VRBC con Å200 { 47 mm/sec; n Å 5), pressure-ejected PBS (DVRBC

Å 016 { 17%; VRBC con Å 272 { 50 mm/sec; n Å 9),DPBS (DVRBC Å 038 { 22%; VRBC con Å 133 { 29 mm/sec; n Å 7), and dH2O (DVRBC Å 22 { 24%; VRBC con Å223 { 40 mm/sec; n Å 12) did not cause a significantchange in VRBC (P ú 0.05). The control prestimulationVRBC’s (i.e., a 200-mm/sec level; Figs. 4–6, legends) werewithin the range reported for rat skeletal muscle (Potteret al., 1993; Renkin, 1984). The VRBC had therefore thecapacity for a large change (e.g., up to a 1800-mm/seclevel; Tyml and Cheng, 1995). The mean blood pressurewas 116 { 7 mmHg (N Å 10) and did not change withlocal arteriolar or capillary stimulations.

Capillary Sensing and Communication PhenomenonFIG. 3. Effect of NE, ACh, and BK on arteriolar diameters. All agentsFigure 2 shows an example of VRBC responses in capil-were locally applied on capillary at Site 2 (Fig. 1). The values given

lary B to NE, BK, and ACh applied at Site 2 (Fig. 1). In are the percentage changes from the control diameters (Dcon). Theagreement with our study in the frog muscle (Dietrich Dcon was 8.5 { 0.4 mm. n is the number of arterioles. *Significant

change in diameter compared to 0%.and Tyml, 1992a), we observed that NE caused VRBC




Capillary versus Arteriole

Figure 6 shows that for both ACh and NE therewere no significant differences between DVRBC re-sponses or between arteriolar diameter responses,when agents were applied either on capillaries or onarterioles (i.e., unpaired stimulations). Although thisfigure suggests that arteriolar stimulations tended toresult in stronger responses than capillary stimula-tions, there was no consistency in this trend as theranges of responses overlapped considerably. Refer-ring to Fig. 4, the present concentrations of ACh (1002

M) and of NE (1005 M) were submaximal since theplateaus of the concentration–response curves werenot yet reached. For this reason, we interpreted thedata of Fig. 6 to indicate that, at least for concentra-tions near the submaximal concentration, the re-FIG. 4. Effect of increasing concentrations of NE, ACh, or BK ap-sponses to capillary and arteriolar stimuli were simi-plied at Site 2 on VRBC in capillary B. The values given are the percent-

age change from control VRBC (VRBC con). The overall VRBC con was 183{ lar. Finally, Fig. 7 shows that there was no significant8 mm/sec. The durations of responses to increasing NE concentrationswere 186 { 36, 123 { 11, 174 { 50, and 311 { 35 sec, to increasingACh concentrations 10 { 3, 33 { 6 and 90 { 11 sec, and to increasingBK concentrations 195 { 51, 347 { 179, 451 { 73, 544 { 262, and 591{ 166 sec, respectively. All responses were dose-dependent by theKruskal–Wallis test. For each agent, the number of capillaries at aparticular concentration ranged from 6 to 19. *Significant change inVRBC compared to 0%.

3.3 { 5.1%, n Å 6). Figure 5 (left and right bars) alsodemonstrates that Atr applied on the arteriole did notchange the DVRBC response to capillary ACh. ArteriolarAtr pretreatment significantly attenuated the responseto direct arteriolar ACh application. DVRBC in capillaryB was reduced from 134 { 29 (n Å 12) to 60 { 25% (nÅ 7). With regard to nonspecificity of Atr, capillary Atrpretreatment did not affect the response to capillary NEapplication. DVRBC in capillary B remained unchangedat 096 { 3% (n Å 5).

FIG. 5. Effect of local pretreatment of capillary and of arterioleIn the Krebs’ superfusion experiment, the bulk of with atropine (Atr; 30 mM) on VRBC response in capillary B. ACh (10

mM) was applied alone at Site 2 (left bar) and was then reapplied atboth NE and DPBS ejected onto a capillary was visu-Site 2 either after Atr was applied at the same Site 2 (middle bar)alized to be swept away from the capillary, arteriole,or after Atr was applied to the arteriole (Site 1; right bar). Theand venule in less than 1 sec. Ejected NE caused aVRBC con for all stimulations was 239 { 28 mm/sec. The durations

significant reduction in VRBC (DVRBC Å 0100 { 0%, of responses were 169 { 29, 59 { 20, and 200 { 34 sec, respectively.nÅ 7) while ejected DPBS did not change VRBC (DVRBC n is the number of capillaries. *Significant change in VRBC compared

to 0%. #Significant difference compared to response to ACh alone.Å 0 { 0%, n Å 4).




difference in lag times between VRBC responses to cap-illary NE (Site 2) and arteriolar NE (Site 1).

DISCUSSION

The aims of the study were to establish the capillarysensing and communication phenomenon in the ratEDL muscle and to compare the sensitivity of the capil-lary and the arteriole to locally applied vasoactiveagents. We found that (i) the rat muscle capillary couldsense a variety of materials and (ii) surprisingly, thecapillary could be as sensitive to agents as the arteriole.

A major concern with the phenomenon is that it mayrepresent a direct arteriolar response to agents that dif-fuse from the capillary to the arteriole, without anycapillary involvement. Our recent studies in the am-phibian model provided evidence against diffusion, aswell as against neural conduction and venous–arterialcross-talk (Dietrich and Tyml, 1992a,b). For example,diffusion could not explain the fact that capillary dam-age abolished the remote arteriolar response and thatstimuli equidistant from the arteriole resulted in thisresponse only when applied on the capillary (Dietrichand Tyml, 1992a).

Our present data based on the two strategies wereconsistent with the frog studies. The first strategy in-volved unidirectional superfusion, which is commonlyused to eliminate the possible effect of diffusion (Xia et FIG. 6. Effect of ACh and NE applied to the supplying arteriole

(Site 1) and to the capillary (Site 2; Fig. 1). (A) VRBC responses inal., 1995; Frame and Sarelius, 1995). Our superfusioncapillary B for arteriolar and capillary stimulations. (B) Arteriolarexperiment demonstrated that capillary NE caused thediameter responses for arteriolar and capillary stimulations.VRBC response, indicating that diffusion was not respon-VRBC con and Dcon for all stimulations were 185 { 9 mm/sec and 8.6

sible for this response. Our second strategy yielded data { 0.6 mm, respectively. For ACh, the durations of VRBC responses(Fig. 5) that were also inconsistent with diffusion. If the were 90 { 11 sec (capillary) and 109 { 46 sec (arteriole). For NE,

the durations of VRBC responses were 123 { 11 sec (capillary) andresponse to capillary ACh (Fig. 5, left bar) were due to173{ 29 sec (arteriole). n is the number of microvessels. *Significantdiffusion from the capillary to the arteriole, then thechange in VRBC compared to 0%. There were no differences betweenblocking effect of capillary Atr (Fig. 5, middle bar)responses to arteriolar and capillary stimulations.

would have to occur at the arteriole, via Atr diffusionfrom the capillary to the arteriole. However, thisblocking effect could not occur, since arteriolar Atr hadno effect on the capillary ACh response (Fig. 5, right that a mechanism other than diffusion was responsible

for the response.bar). This figure demonstrates that the site of ACh sens-ing was at the capillary rather than at the arteriole and In addition to the outcome of the two strategies, the




sistent with the lack of difference between lag times.Thus, results of Fig. 7 could not be explained by diffu-sion.

The VRBC lag times of Fig. 7 were longer than thediameter lag times reported for a direct arteriolar stimu-lation in the hamster cheek pouch preparation (Delas-haw and Duling, 1991). This difference could be ex-plained in terms of differences in preparation used, inagent applied, and/or in type of microvessel stimu-lated. Recently McGahren et al. (1996) reported thatmembrane potential responses to KCl measured di-rectly in hamster cheek pouch capillaries were slowerthan those in arterioles. Similarly, Rivers (1996) showeda 4-sec lag time to methacholine in these arterioles. Re-garding our EDL muscle model, lag times associatedwith NE application were consistent with times re-quired by NE to reach the stimulated microvessel. As-suming that the agent had to diffuse approximately 100

FIG. 7. Response times, expressed in terms of lag times, to localmm (60 mm fluid thickness plus average 40 mm surfacecapillary (Site 2, Fig. 1) and arteriolar stimulations (Site 1) with NEto microvessel distance) to reach the microvessel, and(1005 M). For both stimulations, VRBC was measured in capillary B.

Lag time was measured as the time elapsed between the beginning using the diffusion model described in the Appendix,of the NE stimulus and the time when poststimulation VRBC reached the required diffusion time to the microvessel was com-90% of VRBC con . n is the number of microvessels. There was no differ- puted to be about 5 sec, i.e., comparable to values inence in lag times between the two vessels.

Fig. 7.Referring to our own study (Dietrich and Tyml,

1992a) the present lag times were much shorter thanthose reported for the frog sartorius muscle (i.e., 20–30lack of difference between lag times to capillary NE

and to arteriolar NE (Fig. 7) was clearly inconsistent sec). This is consistent with the shorter lag time in therat tibialis anterior muscle than in the frog muscle re-with diffusion. If the remote response were due to dif-

fusion, then the necessary diffusion time between Sites ported in the same study. A possible explanation forthis could be a faster NE effect at the capillary due to2 and 1 would have to be matched by the difference

between the two lag times. Using the Einstein relation the higher temperature in the warm-blooded than inthe cold-blooded animal. In general, blood flow re-x2 Å 2Dt (x is distance, D is diffusion coefficient, and t

is time; for review see Kutchai, 1990) we computed sponses to physiological stimuli were found to beslower in amphibian than in mammalian muscle (Tyml,that NE diffusion from Site 2 to Site 1 would require

substantial time, i.e., 118 sec. In this computation x Å 1987, 1991).Comparing our previous frog muscle data and the300 mm while D was estimated as 0.38 1 1009 m2/sec

(i.e., that of sucrose of MW 302 as D for NE was not present rat data, there were both similarities and differ-ences in the magnitudes of responses. In both musclesavailable; Dietrich, 1989). It is possible, however, that

this model is erroneous since the ‘‘leading edge’’ of the there were dose-dependent VRBC responses to NE andto ACh. ACh responses were significantly attenuateddiffusion front may arrive at Site 1 much sooner than

predicted by the root mean square time of the Einstein by Atr, indicating that in both species the responseswere receptor mediated. The magnitudes of ACh-in-relation. For this reason we have employed another

model that incorporates this front (Appendix). The duced VRBC increases were comparable for similar stim-uli, i.e., 84% increase for 30 mM in the frog (Song andnewly computed diffusion time of 44 sec was still incon-




Tyml, 1993) and 92% increase for 10 mM in the rat (Fig. In conclusion, we demonstrated that the capillarysensing and communication phenomenon could not4). However, the rat was more sensitive to NE. For

example, 3 mM applied iontophoretically caused an be explained in terms of diffusion. The present study(i) established the presence of the phenomenon in the86% reduction in VRBC in the frog (Dietrich and Tyml,

1992a), while only 0.01 mM caused a comparable reduc- rat EDL muscle and (ii) demonstrated for the firsttime that the capillary in this muscle could be as sen-tion in the rat (i.e., 82%, Fig. 4). At present, we are not

able to explain the larger sensitivity in the rat, as rele- sitive to local vasoactive agents as the arteriole. Weinterpret the data to indicate that the capillary hasvant parameters (e.g., receptor density on capillaries)

in either species are not known. With regard to BK, the the potential to participate in microvascular bloodflow control.dose dependency to this agent (Fig. 4) was reported

here for the first time.The major finding of the present study was the appar-

ent comparable effect of local arteriolar and capillarystimulus on arteriolar diameter and capillary blood APPENDIXflow (Fig. 6). We believe that this finding may be im-portant in our understanding of the coupling betweenvasoactive metabolites generated within tissue and the In order to estimate the diffusion time of NE fromarteriolar tone. Based on models of diffusion of metabo- Site 2 to Site 1 (Fig. 1) within the 60-mm-thick tissuelites described above, the diffusion process may be slow fluid layer (i.e., between muscle surface and oil), thein relation to the speed of vasomotor response (Segal following diffusion model was considered. Based on aand Duling, 1989). Consequently, metabolites gener- temperature distribution model in a semiinfinite solidated near an arteriole could reach the arteriole faster exposed to a sudden temperature step (Inpropera andand in higher concentration than those generated far DeWitt, 1981), the diffusion model was described inaway. This could result in a spatially weighted effect terms of C(x, t), such thatof metabolites on the arteriolar tone. The conceptualproblem arises that, if microvascular flow control relied [C(x, t) 0 Cs]/[Ci 0 Cs] Å erf [x/2(Dt)0.5], (1)solely on diffusion, then such spatial effect might notbe desirable. Metabolic flow control could attend more where x is the distance along capillary B (Fig. 1), x Åto the needs originating from a sleeve of tissue around 0 is Site 2, x Å 300 mm is Site 1; t is time, t Å 0 is timethe arteriole than to the needs from the tissue as a of NE application at Site 2; C(x, t) is concentration ofwhole. NE at distance x and time t; Cs is C(0, t), i.e., NE

Data from Fig. 6 indicate that there may not be such a concentration at Site 2; Ci is C(x, 0), i.e., initial NEproblem. The magnitude of arteriolar response might be concentration at Site 1; D is diffusion coefficient ofindependent of the location of the local stimulus. This NE; and erf is Gaussian error function (Inpropera andrelative independence (i.e., over 300 mm distance) is con- DeWitt, 1981).sistent with the reported length constant of 1.5 mm found The model included the following assumptions:for the arteriolar smooth muscle/endothelial layer (Xiaand Duling, 1995; Xia et al., 1995). In the presence of such • NE is not lost to muscle or to oil, and it is not

degraded or bound to receptor.a relatively large length constant, attenuation of commu-nicated signal over 300 mm might have not been detect- • Site 2 is a continuous source of NE and does not

decrease with time.able in our present set-up. Since the capillary network hasbeen shown to integrate different stimuli into an overall • Arteriole at Site 1 lies in the tissue fluid layer.

To model diffusion time between Sites 2 and 1 withinresponse (Song and Tyml, 1993), it is possible that thecapillaries can participate in metabolic coupling between the fluid layer, we used the left hand side of Eq. (1)

andthe arteriole and the tissue as a whole.




Cs Å 1002 mol/m3 (i.e., 1005 M NE, as in Fig. 7), REFERENCES

Ci Å 0 mol/m3,Beach, J. M., McGahren, E. D., and Duling, B. R. (1996). Electrical sig-

nals are conducted between capillaries and arterioles. FASEB J. 10,andA55 [Abstract].

Delashaw, J. B., and Duling, B. R. (1991). Heterogeneity in conductedC(300 mm, t) Å 1004 mol/m3

arteriolar vasomotor response is agonist dependent. Am. J. Physiol.260 (Heart Circ. Physiol. 29), H1276–H1282.

Dietrich, H. H. (1989). Effect of locally applied epinephrine and nor-(i.e., 1007 M NE; this is a concentration less than theepinephrine on blood flow and diameter in capillaries of rat mesen-ED50 that causes constriction in superfused arterioles;tery. Microvasc. Res. 38, 125–135.Joh et al., 1991; Wu and Benoit, 1994) to compute erf w

Dietrich, H. H., and Tyml, K. (1992a). Microvascular flow response toÅ 0.9900, where localized application of norepinephrine on capillaries in rat and

frog skeletal muscle. Microvasc. Res. 43, 73–86.Dietrich, H. H., and Tyml, K. (1992b). Capillary as a communicatingw Å x/2(Dt)0.5. (2)

medium in the microvasculature. Microvasc. Res. 43, 87–99.Frame, M. D., and Sarelius, I. H. (1995). L-arginine-induced conducted

From a table of the Gaussian error function (Inpropera signals alter upstream arteriolar responsiveness to L-arginine. Circ.and DeWitt, 1981), the corresponding value of w was Res. 77, 695–701.1.16. Using this value of w, Eq. (2), and Inpropera, F. P., and DeWitt, D. P. (1981). Fundamentals of Heat Trans-

fer, Chapter 5, pp. 204–206, 796. Wiley, Toronto.Joh, T., Granger, D. N., and Benoit, J. N. (1991). Intestinal microvascu-x Å 300 1 1006 m,

lar responsiveness to norepinephrine in chronic portal hyperten-sion. Am. J. Physiol. 260, H1135–H1143.

and Kutchai, H. C. (1990). Cellular membranes and transmembrane trans-port of solutes and water. In Principles of Physiology (R. M. Berneand M. N. Levy, Eds.), Chapter 1, pp. 7–8. Mosby, St. Louis.D Å 0.38 1 1009 m2/sec

McGahren, E. D., Beach, J. M., and Duling, B. R. (1996). In vivo capil-lary endothelium depolarizes in response to KCl. Microcirculation

(i.e., that of sucrose of MW 302 as D for NE was not 3, 97 [Abstract].available; Dietrich, 1989), t was computed to be 44 sec. Mehrke, G., Pohl, U., and Daut, J. (1991). Effects of vasoactive agonists

on the membrane potential of cultured bovine aortic and guineaTo compute diffusion time between the pipette tippig coronary endothelial cells. J. Physiol. London 439, 277–299.and the microvessel, we used the same values and rela-

Mitchell, D., and Tyml, K. (1996). Nitric oxide release in rat skeletaltionships as above, except for x Å 100 1 1006 m (60 mmmuscle capillary. Am. J. Physiol. 260 (Heart Circ. Physiol. 39), H1696–

fluid layer thickness plus 40 mm average muscle surface H1703.to microvessel distance). Accordingly, t was computed Potter, R. F., Dietrich, H. H., Tyml, K., Ellis, C. G., Cronkwright, J., and

Groom, A. C. (1993). Ischemia-reperfusion induced microvascularto be 4.9 sec.dysfunction in skeletal muscle: Application of intravital video mi-croscopy. Int. J. Microcirc. Clin. Exp. 13, 173–186.

Renkin, E. M. (1984). Control of microcirculation and blood-tissueexchange. In Handbook of Physiology (American Physiological Soci-ACKNOWLEDGMENTSety, Eds.), Vol. 4, pp. 627–687.

Revest, P. A., and Abbott, N. J. (1992). Membrane ion channels ofendothelial cells. Trends Pharmacol. Sci. 13, 404–407.

Rivers, R. J. (1996). Late onset response caused by nitroprusside andThe authors thank Jay Cronkwright, Livio Rigutto, and Drs. WenTao, Linong Cheng, Stephen Sims, and John T. Hamilton for their adenosine. Microcirculation 3, 103 [Abstract].

Scott, P. A., and Bicknell, R. (1993). The isolation and culture of micro-technical help and useful suggestions. We are grateful to Dr. ChrisEllis, Dr. Craig Miller, and Mr. Ryon Bateman for their help with the vascular endothelium. J. Cell Sci. 105, 269–273.

Segal, S. S., and Duling, B. R. (1989). Conduction of vasomotor re-diffusion model. The study was supported by the Heart and StrokeFoundation of Ontario and the Medical Research Council of Canada sponses in arterioles: A role for cell-to-cell coupling? Am. J. Physiol.

256 (Heart Circ. Physiol. 25), H838–H845.(salary support for K.T.).




Song, H., and Tyml, K. (1993). Evidence for sensing and integration nique as a television method for measuring red cell velocity inmicrovessels. Int. J. Microcirc. Clin. Exp. 1, 145–155.of biological signals by the capillary network. Am. J. Physiol. 265

(Heart Circ. Physiol. 34), H1235–H1242. Tyml, K., Mathieu-Costello, O., and Noble, E. (1995). Microvascularresponse to ischemia, and endothelial ultrastructure, in disusedTyml, K. (1987). Red cell perfusion in skeletal muscle at rest and after

mild and severe contractions. Am. J. Physiol. 252, H485–H493. skeletal muscle. Microvasc. Res. 49, 17–32.Xia, J., and Duling, B. R. (1995). Electromechanical coupling and theTyml, K. (1991). Heterogeneity of microvascular flow in rat skeletal

muscle is reduced by contraction and by hemodilution. Int. J. Mi- conducted vasomotor response. Am. J. Physiol. 269 (Heart Circ. Phys-iol. 38), H2022–H2030.crocirc. Clin. Exp. 10, 75–86.

Tyml, K., and Budreau, C. H. (1991). A new preparation of rat extensor Xia, J., Little, T. L., and Duling, B. R. (1995). Cellular pathways ofthe conducted electrical response in arterioles of hamster cheekdigitorum longus muscle for intravital investigation of the micro-

circulation. Int. J. Microcirc. Clin. Exp. 10, 335–341. pouch. Am. J. Physiol. 269 (Heart Circ. Physiol. 38), H2031 –H2038.Tyml, K., and Cheng, L. (1995). Heterogeneity of red blood cell veloc-

ity in skeletal muscle decreases with increased flow. Microcircula- Wu, Z. Y., and Benoit, J. N. (1994). Vascular NE responsiveness inportal hypertension: Role of portal pressure and portosystemiction 2, 181–193.

Tyml, K., and Ellis, C. G. (1982). Evaluation of the flying spot tech- shunting. Am. J. Physiol. 266, H1162–H1168.



comparable effects of arteriolar and capillary stimuli on blood flow in rat skeletal muscle

Documents