Research Paper J Vase Res 1992:29:341-346
Department of Physiology, University of Virginia Health Sciences Center. Charlottesville, Va.. USA
KeyWords Vascular reactivity Arterioles Venules Endothelium Diffusion
A ccess of Blood-Borne Vasoconstrictors to the Arteriolar Smooth Muscle Abstract In vitro experiments have shown that luminally applied water-soluble vasoac tive materials have limited access to arteriolar smooth muscle cells, and as a result, the responses to such agents applied luminally are less than the responses to those applied adventitially. To determine the extent to which this ‘compartmentation’ influences arteriolar responsiveness to blood-borne wa ter-soluble vasoconstrictors in vivo, we applied phenylephrine, vasopressin and angiotensin 11 to arterioles in the hamster cheek pouch both by luminal perfusion, and by topical application to the arteriolar smooth muscle via micropipettes. The arterioles were about 2 orders of magnitude more sensitive to these water-soluble vasoconstrictors when they were applied topically than when they were applied luminally. In contrast, the arterioles were almost equally sensitive to the lipid-soluble ai-adrenoceptor agonist SKF 89748-A applied by either route. The venular wall appears to be much less effective as a barrier than the arteriolar endothelium. Phenylephrine and vasopressin both elicited large arteriolar constrictions when perfused through venules in close proximity to the arteriole, and these constrictions were larger than those observed when the drug was applied to the arteriole’s own lumen. Our obser vations confirm that the arteriolar endothelium can inhibit the direct access of water-soluble blood-borne agents to the arteriolar smooth muscle in vivo, and they suggest that the capillaries and venules could be the primary routes of access for water-soluble agents from the blood to the arteriolar smooth muscle.
Introduction Endothelial cells line the entire vasculature. Therefore, blood-borne vasoactive agents must pass through or be tween endothelial cells to reach vascular smooth muscle cells. There are three potential pathways from the blood to the arteriolar vascular smooth muscle: (1) diffusion directly across the arteriolar endothelium; (2) diffusion through the extravascular space after leaving the blood in
First received: December 31. 1991 Revised: March 9. 1992 Accepted: April 14. 1992
capillaries; and (3) passage through venular walls to nearby arterioles [1. 2], The relative contributions of the three pathways to the total delivery of blood-borne watersoluble agents from the blood to the arteriolar smooth muscle are unknown, but it is known that diffusion directly through the arteriolar endothelium is greatly restricted for water-soluble molecules. As a result, in iso lated arterioles water-soluble drugs can be hundreds of times less potent when applied to the intimal surface than
Michael J. Lew, PhD Baker Medical Research Inslitutc Commercial Road Prahran. Vic. 3181 (Australia)
© 1992 S. Karger AG, Basel 1018-1172/92/0294-0341 $2.75/0
Downloaded by: King's College London 137.73.144.138 - 1/16/2019 11:50:12 PM
Michael J. Lew Brian R. Duling
Methods Male golden hamsters (90-150 g body weight) were anesthetized with an intraperitoneal injection of pentobarbital sodium (80 mg/kg). Body temperature was maintained at 36-38 °C by radiative and con ductive heating. The animals breathed room air spontaneously through a tracheal cannula during the setup procedures, and 30% O: in nitrogen during the experiments. Supplemental anesthetic (4.5 mg/h) in saline (0.5 ml/h) was continuously infused through an intra peritoneal cannula to maintain anesthesia and normovolemia. The left cheek pouch was everted, opened and cleared of connective tis sue according to the method of Duling [6]. The cheek pouch was superfused at approximately 5 ml/min with a modified Ringer’s bicarbonate solution (pH 7.4) containing (m.W) NaCl. 132: KCI.4.7: CaClj, 2.0: MgSO.). 1.2; NaHCOs, 20. The superfusate was warmed to 37 °C and equilibrated with 5%COi in nitrogen. The preparations were mounted on the stage of a Zeiss ACM microscope and observed with a 50X long working distance objective (n.a. 0.6). Arterioles were selected for observation using resting diameter (25-60 pm) and
342
Lcw/Duling
image quality as selection criteria. Preparations without spontaneous vascular tone were not used. The image was transmitted by a video camera (Dage MTI) to a high resolution video monitor (Dage MTI) and the vessel diameter measured on the monitor screen using a modified Colorado Video Analyzer (model 321). The vessel diameter was recorded on a chart recorder (Gould Brush 260). Pipettes
Pipettes were pulled from 1.5 mm glass tubing (No. IBI50F-4. W-P Instruments, New Haven, Ct.. USA) using a vertical pipette puller (model 700C. David Kopf, Tujunga, Calif.. USA) and triple bevelled. Outside tip diameters were approximately 5 pm. The pipettes were mounted in a pipette holder (White Instrument Co.. Suitland. Md.. USA) attached to a Leitz micromanipulator. A fine polyethylene tube was advanced into the pipette to the start of its tapered tip. and drug solutions were injected into the pipette through this tube, exiting through a second port in the pipette holder. This arrangement allowed the solution in the pipette to be exchanged with a minimal dead space. After the solution in the pipette was changed, the pipette was pressurized for several minutes to force the unex changed solution from the tapered part of the pipette. Pressure was applied to the pipette manually with a 5-ml syringe. Protocols
Using the micropipcttes. concentration-response curves were con structed with the drugs applied selectively to the lumen or the outside of the arterioles. Application of the drugs to the outside of the arteri ole was accomplished by placing the pipette tip close (< 10 pm) to the arteriole and pressurizing the pipette. The ejection pressure was regu lated to eject fluid at a rate sufficient to displace the superfusion fluid in the vicinity of the pipette tip. as judged by distortion of the connec tive tissue around the arteriole during drug application. Luminal application was accomplished by impaling the arteriole several hundred microns upstream from the site of observation. The rate of injection was considered optimal when the perfusate almost com pletely replaced the flowing blood in the arteriole. The continued pas sage of a few red blood cells past the injection pipette indicated that the intra-arteriolar pressure was not elevated excessively by the injec tions. Both methods of drug application would have resulted in the drugs reaching the arteriole at a concentration that was similar to that in the pipette. Forabluminal application the fact that the ejection rate was high, and the pipette tip was placed close to the arteriole would have ensured that at least the central portion of the ejected stream of solution would have been unmixed with the superfusion solution. With the intraluminal application, the drug solution was observed to almost completely replace the blood in the arteriolar lumen. In cither case it is likely that the discrepancy between the nominal drug concen tration (that in the pipette) and the actual maximum concentration at the observ ed site of the arteriole was minimal. Drug application by either route was continued until a steady response level was obtained, typically about 20-30 s for outside application and 1-3 min for intravascular administration. The drug concentrations were always applied in ascending order, but the order of application to the outside or inside of the vessel was randomized. Because of the tachyphylaxis caused by concentrations of angioten sin II that caused large responses, the sequence of administration for angiotensin II was always 1 n.V/outside, 100 nM inside then 100 nM outside. The first two doses elicited relatively small responses, and induced little if any tachyphylaxis. None of the other drugs appeared to cause any tachyphylaxis.
Access from Blood to Arteriolar Smooth
Downloaded by: King's College London 137.73.144.138 - 1/16/2019 11:50:12 PM
when applied directly to the smooth muscle on the adven titial surface [3], While it has been shown that the lowpermeability of the arteriolar endothelium does influence the responses of arterioles to vasoactive agents in vivo [4], the magnitude of the effect is unclear. The effect of the endothelium on the potencies of water-soluble blood-borne agents in vivo could be less than that in vitro because of at least two factors. First, the isolated arteriole is suspended in a relatively large volume of free solution that can act as a diffusional sink for the molecules that traverse the endothelium. Such a sink could exaggerate the fall in concentration between the lumen and the smooth muscle cells [5] compared to the situation in intact tissues where the arterioles are sur rounded by tissue that will limit the rate of diffusion of the agonist away from the arteriolar surface. The second factor that distinguishes the in vivo situation from that in vitro is the presence of other routes of access by which the agonist might reach the smooth muscle. In the isolated arteriole, the only pathway from the lumen to the smooth muscle cells is directly across the arteriolar endothelium whereas transcapillary and transvenular routes arc avail able in vivo. The modes of drug administration used in our previous in vivo study [4] do not show the relative importance of the different pathways from the blood to the arteriolar smooth muscle. The aim of these experiments was to determine the rel ative intravascular and extravascular potencies of natural and artificial vasoconstrictors under controlled condi tions in which the delivery of blood-borne molecules to the arteriolar smooth muscle by all three routes (transarteriolar, transcapillary and transvenular) are possible.
In several experiments we measured the arteriolar responses to phenylephrine or vasopressin injeeted into the lumen of a venule that passed close ( < 20 pm) to the segment of the arteriole being ob served. These responses were compared to the responses to the same concentrations of agonist applied to the lumen of the same arteriole.
SKF 89748-A 40-,
Drugs
The following drugs were used in this study: pentobarbital Na (Nembutal. Abbot laboratories): phenylephrine HC1 (Sigma): SKF 89748-A (a gift from Smith Kline & French); angiotensin II acetate (Sigma): (arg8)-vasopressin acetate (Sigma). SKF 89748-A was stored refrigerated as a 5-m.V/ stock dissolved in dimethylsulfoxide. phen ylephrine was stored refrigerated as a 10-m.l/ stock dissolved in water, angiotensin II and vasopressin were stored in frozen aliquots at 0.1 mM. Each drug was diluted to the required concentrations in Ringer's solution on the day of the experiment. Dimethylsulfoxide was added to the phenylephrine dilutions to give the same concentra tion as was present in the SKF 89748-A dilutions.
SKF 89748-A concentration (log M)
1
Results
Phenylephrine Inira- and Periarteriolar Application Arteriolar constrictions to the lipid-soluble U|-adrenoceptor agonist SK.F 89748-A were similar with either intraluminal or adventitial application (fig. 1). The differ ence between the responses to 0.1 |jlM applied luminally and adventitially was small (about 20%) but significant (p < 0.05. paired t test). In contrast, the arteriolar responses to phenylephrine were strongly dependent of the route of phenylephrine application (fig. 2). With lumi nal application, more than 1 \iM phenylephrine was needed to elicit a constriction, but when applied to the adventitial surface of the vessels, that concentration of phenylephrine caused near-maximal responses. Thus lu minally applied phenylephrine was between 10 and 100 times less potent than adventitially applied phenyleph rine (fig. 2). Angiotensin II applied to the outside of the arterioles was suprathreshold at 1.0 nM and supramaximal at lOOnM. but 100 mV/ applied via the lumen caused re sponses of less than 50% (fig. 3). We were unable to apply
Phenylephrine concentration (log M)
2
Angiotensin II 40-,
5
? 3
30-
B
T
§
CD
I 20B
Inside
0
tered to the arterioles luminally (•) and adventitially (o) (n = 5). Error bars are SEM. Fig. 2. Concentration-response curves for phenylephrine admin istered to the arterioles luminally (•) and adventitially (o) (n = 5). Error bars are SEM. Fig. 3. Effect of angiotensin II administered luminally (•) and adventitially (o) on arteriolar diameter (n = 5). Error bars arc SEM.
1
10O
Rest
ih
—
-9
-8
Outside
r—
-7
Angiotensin II concentration (log M)
343
Downloaded by: King's College London 137.73.144.138 - 1/16/2019 11:50:12 PM
Fig. 1. Concentration-response curves for SKF 89748-A adminis
Vasopressin was about 100 times more effective in activating the arterioles from the adventitial side than from the lumen (fig. 4). Applied luminally. lOnM vaso pressin elicited responses of slightly less than 50%, but applied adventilially, only 1 nM was required for maxi mal responses. No dilation in response to any of the ago nists by either route of application was observed.
Vasopressin
Intravenular Injections Where a venule crossed under or over the arteriolar segment being observed, we occasionally made injections of phenylephrine (1 ,uM n = 3. 10 \iM n = 1) or vasopressin (10 nM n = 3) into that venule and observed the responses of the arteriole. In every case, application of the drug to the lumen of the venule resulted in an arteriolar constric tion that was larger than the constriction that the same concentration of drug elicited from the arteriolar lumen (fig. 5). No change in the diameter of the venule was observed during the drug injections. Drug Application Site:
Venule
Discussion
Fig. 4. Concentration-response curves for vasopressin adminis tered to the arterioles intravascularly (•) and extravascularly (o) (n = 5). Error bars are SEM. Fig. 5. Arteriolar responses to phenylephrine (1 \iM n = 3 and 10 gM n = 1 clumped) or vasopressin (10 nM n = 3) applied via the lumen of a venule close to the arteriole (Venule), via the arteriolar lumen, and direct adventitial application.
any further concentrations of angiotensin II because of the occurrence of tachyphylaxis. Because the responses to I nM applied to the adventitia and 100 nM applied luminally were similar, it can be estimated that the angiotensin II potency differential for luminal and adventitial appli cation was about 100-fold.
344
Lew/Duling
Delivery o f Blood-Borne Agents to the Arteriolar Smooth Muscle If drugs in the blood had direct unrestricted access to the arteriolar smooth muscle cells by diffusion through the arteriolar endothelium, then transcapillary and transvenular pathways for the movement of drugs from the blood to the smooth muscle would not be important. That is because unrestricted diffusion through the endothelium is equivalent to the inside surface of the smooth muscle being in almost direct contact with the blood: unrestricted diffusion with a path of the thickness of the endothelium would be virtually instantaneous and would involve no significant decrease in concentration. With the smooth muscle cells exposed to the same concentration of drug as that in the arterial blood, the concentration gradients within the tissue would always be downwards towards the capillaries and venules. Thus no net movement of drug molecules from those vessels to the arteriolar smooth muscle would be possible. The data in this study, and that in our previous papers, suggest that those circumstances do not occur, the arteriolar endothelium does not allow unrestricted diffusion of all drugs from the blood to the smooth muscle cells [3, 4], and therefore the delivery of vasoactive agents from the blood to the arteriolar smooth muscle by the transcapillary and transvenular pathways is possible.
Access from Blood to Arteriolar Smooth Muscle
Downloaded by: King's College London 137.73.144.138 - 1/16/2019 11:50:12 PM
□
Transcapillary Delivery Capillaries are the main exchange vessel for delivery of blood-borne agents to the tissue, and capillary endothe lium is probably less tight than the arteriolar endothelium [9, 12, 13, 16. 17], All blood-borne vasoactive agents are able to reach the arteriolar smooth muscle by leaving the capillary blood and diffusing through the extravascular space. With bolus administration of agonists (as in our previous in vivo study [4]), delivery of the agonists to the arteriolar smooth muscle via the capillaries will always involve substantial dilution in the extravascular tissue. Even with extended infusions, the general tissue concen tration of a vasoconstrictor need not reach that present in the infusate because of losses of the vasoconstrictor from the tissue, either by degradation and uptake, which is common for endogenous vasoconstrictors, or by loss from the tissue into the superfusion solution in experiments such as these. The infusion durations that we used in these experi ments (1-3 min) were long enough to allow the attain
ment of steady response levels, and Ley and Arfors [12] found that accumulation of the water-soluble marker, flu orescein. in the extravascular space of the hamster cheek pouch started to slow after about 20 s of intra-arteriolar injection. Therefore it is reasonable to suppose that our agonist infusions were long enough to allow the agonists to approach steady-state concentration in the tissue. Nonetheless, the design of these experiments could have led to an underestimation of the periarteriolar concentra tion of agonist that would occur under more physiological circumstances by at least two mechanisms. The first is the loss of agonist into the nonphysiological diffusion sink provided by the superfusion solution previously men tioned. The second is that only those capillaries and ven ules fed by the arteriole that was impaled for drug injec tion were perfused with drug-containing solution. Any other capillaries or venules in the vicinity of the arteriole would have acted as diffusion sinks for the drug, rather than sources, thereby lowering the periarteriolar drug concentration. Under normal physiological circum stances. all of the arterioles, and thus all of the capillaries and venules, would be perfused with blood containing the vasoactive hormone. It should be noted that in our experi ments both of these mechanisms would have a larger effect on the delivery of agonist to the smooth muscle by the transcapillary and transvenular routes than by the direct transendothelial route. Transvenular Delivery The ability of norepinephrine to elicit arteriolar con strictions by leakage from venules to nearby arteriolar smooth muscle has recently been demonstrated by Tigno et al. [2], but they did not determine the relative threshold concentrations for norepinephrine applied to the arterio lar lumen and to the venular lumen. In our experiments we found that arterioles can be more sensitive to agonists in the lumen of nearby venules than to the same agonists in the arteriole’s own lumen (fig. 5). This finding is consis tent with that of Ley and Arfors [ 12] who showed that fluorescein (water-soluble, molecular weight 376 daltons) accumulates faster in the tissue around venules than in the tissue around arterioles when it is infused intra-arteriolarly in the hamster cheek pouch. The rapid movement of water-soluble molecules from the blood into the tissue around venules comes about because venular endothe lium is more permeable than the arteriolar endothelium [8, 12, 16, 17], Thus the venules can act as a source of vasoactive molecules for the arteriolar smooth muscle. The degree to which this can occur must be dependent on both the anatomical arrangement of the arterioles and
345
Downloaded by: King's College London 137.73.144.138 - 1/16/2019 11:50:12 PM
Passage through the Arteriolar Endothelium Small water-soluble molecules, such as phenylephrine, vasopressin and angiotensin II, move from the intravas cular space to the extravascular space predominantly by passage through the spaces between the endothelial cells [7], These spaces are a tiny fraction of the surface area of the arteriolar endothelium so the total flux through them is small [8-13]. Thus the concentration of water-soluble molecules on one side of the arteriolar endothelium can be vastly different from the concentration on the other side. This is illustrated by the water-soluble agonists in this study, which were present at the arteriolar smooth muscle at an average concentration about 100 times less than that applied to the other side of the endothelium (us ing the arteriolar responses as an indirect assay of the average concentration of agonists at the smooth muscle cells) (fig. 2-4). In contrast, lipid-soluble molecules can permeate the cell membranes of endothelial cells [7, 14, 15], making the entire surface area of the arteriolar endo thelium available for their movement from the lumen to the smooth muscle. Accordingly, SKF 89748-A adminis tered via the arteriolar lumen was present at the arteriolar smooth muscle cells at a concentration approaching that in the lumen (fig. 1). These data confirm that the direct passage of water-soluble vasoactive agents in vivo across the arteriolar endothelium is limited, and therefore indi cate that the transcapillary and transvenular routes may have important roles in delivery of vasoactive agents from the blood to the arteriolar smooth muscle cells.
venules, and on the ability of the activating agonist to elicit responses that propagate along the arteriolar tree [2, 18], Because the arteriolar endothelium reduces the direct access of water-soluble blood-borne molecules to the arte riolar smooth muscle, this mechanism could have an important role. Consequences The fact that blood-borne vasoactive agents can reach their sites of action by the transcapillary and transvcnular pathways rather than (or in addition to) directly across the arteriolar endothelium has an important consequence for vascular responsiveness. It leads to the possibility that the responsiveness of a vascular bed, or segment of a vascular bed. might be determined by its anatomical arrangement. The effectiveness of the transcapillary pathway would be dependent on the number of capillaries present in the tis sue, and the effectiveness of the transvenular pathway would be dependent on the degree to which the tissue had arteriole-venule pairing. This means that the responsive ness of a vascular bed (or part of a vascular bed) can be designed in a way that would influence the responsiveness to vasoconstrictors in general, a mechanism that would stand in addition to the distribution of specific receptors [4].
Conclusion In this study we have demonstrated that the arteriolar smooth muscle of the hamster cheek pouch is partially isolated from the blood by the arteriolar endothelium, and therefore the delivery of blood-borne vasoactive agents to the arteriolar smooth muscle cells is. at least in part, the result of movement of these agents through the capillary' and vcnular walls into the extravascular tissue space. Regional differences in the permeability of the endothelium (arteriolar, capillary or venular), or in the anatomical arrangement of the capillaries and venules with respect to the arterioles, could be important factors in determining regional vascular reactivity to blood-borne agents.
Acknowledgments We thank Dr. J.P. Hieble of Smith, Kline & French for supplying the SKF 89748-A. and Mr. and Mrs. H. Caplan for their continuing support. We also thank D.N. Damon for his expert advice and assis tance. This work was supported by National Heart, Lung and Blood Institute grants HL-19242 and HL-12792. and by American Heart Association, Virginia Affiliate grant VA-89-F-14.
1 Hester RL. Guyton AC: Venular-arteriolar dif fusion of vasocactive metabolites. FASEB J 1989;3:A 1388. 2 Tigno XT. Lev K. Pries AR. Gaehtgens P: Venulo-arteriolar communication and propa gated response - a possible mechanism for local control of blood (low. Pflügers Arch Phar macol 1989:414:450-456. 3 Lew MJ, Rivers RJ. Duling BR: Arteriolar smooth muscle responses are modulated by an intramural diffusion barrier. Am J Physiol 19B9;257:H 10—H16. 4 Lew MJ, Duling BR: Arteriolar reactivity in vivo is influenced by an intramural diffusion barrier. Am J Physiol 1990:259:1I574-H58I. 5 Hynes MR. Duling BR: C a '' sensitivity of iso lated arterioles from the hamster cheek pouch. Am J Physiol 1991 ;260:H355-H361. 6 Duling BR: The preparation and use of the hamster cheek pouch for studies of the microcirculation. Microvasc Res 1973:5:423-429.
7 Rcnkin EM: Multiple pathways of capillary permeability. Circ Res 1977:41:735-743. 8 Simionescu M. Simionescu N. Palade GE: Seg mental differentiations of cell junctions in the vascular endothelium. The microvasculalure. J Cell Biol 1975;67:863-885. 9 Simionescu N, Simionescu M, Palade GE: Structural basis of permeability in sequential segments of the microvasculature of the dia phragm. Microvasc Res 1978:15:17-36. 10 Okuda T, Yamamoto T: The ultrastructural basis of the permeability of arterial endothe lium to horseradish peroxidase. Cell Tissue Res 1983;231:1 17-128. 11 Mink D. Schiller A. Kriz W. Taugner R: Intercndolhclial junctions in kidney vessels. Cell Tissue Res 1984;236:567-576. ’ 12 Ley K. Arfors K-E: Segmental differences of microvascular permeability for FITC-dextrans measured in the hamster cheek pouch. Micro vasc Res 1986:31:84-99.
13 Bundgaard M: The three-dimensional organi zation of tight junctions in a capillary endothe lium revealed by serial-section electron micros copy. J Ullrastruct Res 1984:88:1 —17. 14 Pappenheinter JR. Renkin EM. Borrero LM: Filtration, diffusion and molecular seiving through peripheral capillary membranes. A contribution to the pore theory1of capillary' per meability. Am J Physiol 1951;167:13-46. 15 Rcnkin EM: Capillary permeability to lipidsoluble molecules. Am J Physiol 1952:168: 538-545. 16 Olescn S-P: Electrical resistance of arterioles and venules in the hamster cheek pouch. Acta Physiol Scand 1985:123:12 1-126. 17 Rous P. Gilding BM. Smith F: The gradient of vascular permeability. J Exp Med 1930:51: 807-830. 18 Delashaw .1B, Segal SS, Duling BR: Drugs with different mechanisms of action induce similar conducted responses in arterioles. FASEB J 1988;2:A 1861.
346
Lew/Duling
Access from Blood to Arteriolar Smooth Muscle
Downloaded by: King's College London 137.73.144.138 - 1/16/2019 11:50:12 PM
References
Department of Physiology, University of Virginia Health Sciences Center. Charlottesville, Va.. USA
KeyWords Vascular reactivity Arterioles Venules Endothelium Diffusion
A ccess of Blood-Borne Vasoconstrictors to the Arteriolar Smooth Muscle Abstract In vitro experiments have shown that luminally applied water-soluble vasoac tive materials have limited access to arteriolar smooth muscle cells, and as a result, the responses to such agents applied luminally are less than the responses to those applied adventitially. To determine the extent to which this ‘compartmentation’ influences arteriolar responsiveness to blood-borne wa ter-soluble vasoconstrictors in vivo, we applied phenylephrine, vasopressin and angiotensin 11 to arterioles in the hamster cheek pouch both by luminal perfusion, and by topical application to the arteriolar smooth muscle via micropipettes. The arterioles were about 2 orders of magnitude more sensitive to these water-soluble vasoconstrictors when they were applied topically than when they were applied luminally. In contrast, the arterioles were almost equally sensitive to the lipid-soluble ai-adrenoceptor agonist SKF 89748-A applied by either route. The venular wall appears to be much less effective as a barrier than the arteriolar endothelium. Phenylephrine and vasopressin both elicited large arteriolar constrictions when perfused through venules in close proximity to the arteriole, and these constrictions were larger than those observed when the drug was applied to the arteriole’s own lumen. Our obser vations confirm that the arteriolar endothelium can inhibit the direct access of water-soluble blood-borne agents to the arteriolar smooth muscle in vivo, and they suggest that the capillaries and venules could be the primary routes of access for water-soluble agents from the blood to the arteriolar smooth muscle.
Introduction Endothelial cells line the entire vasculature. Therefore, blood-borne vasoactive agents must pass through or be tween endothelial cells to reach vascular smooth muscle cells. There are three potential pathways from the blood to the arteriolar vascular smooth muscle: (1) diffusion directly across the arteriolar endothelium; (2) diffusion through the extravascular space after leaving the blood in
First received: December 31. 1991 Revised: March 9. 1992 Accepted: April 14. 1992
capillaries; and (3) passage through venular walls to nearby arterioles [1. 2], The relative contributions of the three pathways to the total delivery of blood-borne watersoluble agents from the blood to the arteriolar smooth muscle are unknown, but it is known that diffusion directly through the arteriolar endothelium is greatly restricted for water-soluble molecules. As a result, in iso lated arterioles water-soluble drugs can be hundreds of times less potent when applied to the intimal surface than
Michael J. Lew, PhD Baker Medical Research Inslitutc Commercial Road Prahran. Vic. 3181 (Australia)
© 1992 S. Karger AG, Basel 1018-1172/92/0294-0341 $2.75/0
Downloaded by: King's College London 137.73.144.138 - 1/16/2019 11:50:12 PM
Michael J. Lew Brian R. Duling
Methods Male golden hamsters (90-150 g body weight) were anesthetized with an intraperitoneal injection of pentobarbital sodium (80 mg/kg). Body temperature was maintained at 36-38 °C by radiative and con ductive heating. The animals breathed room air spontaneously through a tracheal cannula during the setup procedures, and 30% O: in nitrogen during the experiments. Supplemental anesthetic (4.5 mg/h) in saline (0.5 ml/h) was continuously infused through an intra peritoneal cannula to maintain anesthesia and normovolemia. The left cheek pouch was everted, opened and cleared of connective tis sue according to the method of Duling [6]. The cheek pouch was superfused at approximately 5 ml/min with a modified Ringer’s bicarbonate solution (pH 7.4) containing (m.W) NaCl. 132: KCI.4.7: CaClj, 2.0: MgSO.). 1.2; NaHCOs, 20. The superfusate was warmed to 37 °C and equilibrated with 5%COi in nitrogen. The preparations were mounted on the stage of a Zeiss ACM microscope and observed with a 50X long working distance objective (n.a. 0.6). Arterioles were selected for observation using resting diameter (25-60 pm) and
342
Lcw/Duling
image quality as selection criteria. Preparations without spontaneous vascular tone were not used. The image was transmitted by a video camera (Dage MTI) to a high resolution video monitor (Dage MTI) and the vessel diameter measured on the monitor screen using a modified Colorado Video Analyzer (model 321). The vessel diameter was recorded on a chart recorder (Gould Brush 260). Pipettes
Pipettes were pulled from 1.5 mm glass tubing (No. IBI50F-4. W-P Instruments, New Haven, Ct.. USA) using a vertical pipette puller (model 700C. David Kopf, Tujunga, Calif.. USA) and triple bevelled. Outside tip diameters were approximately 5 pm. The pipettes were mounted in a pipette holder (White Instrument Co.. Suitland. Md.. USA) attached to a Leitz micromanipulator. A fine polyethylene tube was advanced into the pipette to the start of its tapered tip. and drug solutions were injected into the pipette through this tube, exiting through a second port in the pipette holder. This arrangement allowed the solution in the pipette to be exchanged with a minimal dead space. After the solution in the pipette was changed, the pipette was pressurized for several minutes to force the unex changed solution from the tapered part of the pipette. Pressure was applied to the pipette manually with a 5-ml syringe. Protocols
Using the micropipcttes. concentration-response curves were con structed with the drugs applied selectively to the lumen or the outside of the arterioles. Application of the drugs to the outside of the arteri ole was accomplished by placing the pipette tip close (< 10 pm) to the arteriole and pressurizing the pipette. The ejection pressure was regu lated to eject fluid at a rate sufficient to displace the superfusion fluid in the vicinity of the pipette tip. as judged by distortion of the connec tive tissue around the arteriole during drug application. Luminal application was accomplished by impaling the arteriole several hundred microns upstream from the site of observation. The rate of injection was considered optimal when the perfusate almost com pletely replaced the flowing blood in the arteriole. The continued pas sage of a few red blood cells past the injection pipette indicated that the intra-arteriolar pressure was not elevated excessively by the injec tions. Both methods of drug application would have resulted in the drugs reaching the arteriole at a concentration that was similar to that in the pipette. Forabluminal application the fact that the ejection rate was high, and the pipette tip was placed close to the arteriole would have ensured that at least the central portion of the ejected stream of solution would have been unmixed with the superfusion solution. With the intraluminal application, the drug solution was observed to almost completely replace the blood in the arteriolar lumen. In cither case it is likely that the discrepancy between the nominal drug concen tration (that in the pipette) and the actual maximum concentration at the observ ed site of the arteriole was minimal. Drug application by either route was continued until a steady response level was obtained, typically about 20-30 s for outside application and 1-3 min for intravascular administration. The drug concentrations were always applied in ascending order, but the order of application to the outside or inside of the vessel was randomized. Because of the tachyphylaxis caused by concentrations of angioten sin II that caused large responses, the sequence of administration for angiotensin II was always 1 n.V/outside, 100 nM inside then 100 nM outside. The first two doses elicited relatively small responses, and induced little if any tachyphylaxis. None of the other drugs appeared to cause any tachyphylaxis.
Access from Blood to Arteriolar Smooth
Downloaded by: King's College London 137.73.144.138 - 1/16/2019 11:50:12 PM
when applied directly to the smooth muscle on the adven titial surface [3], While it has been shown that the lowpermeability of the arteriolar endothelium does influence the responses of arterioles to vasoactive agents in vivo [4], the magnitude of the effect is unclear. The effect of the endothelium on the potencies of water-soluble blood-borne agents in vivo could be less than that in vitro because of at least two factors. First, the isolated arteriole is suspended in a relatively large volume of free solution that can act as a diffusional sink for the molecules that traverse the endothelium. Such a sink could exaggerate the fall in concentration between the lumen and the smooth muscle cells [5] compared to the situation in intact tissues where the arterioles are sur rounded by tissue that will limit the rate of diffusion of the agonist away from the arteriolar surface. The second factor that distinguishes the in vivo situation from that in vitro is the presence of other routes of access by which the agonist might reach the smooth muscle. In the isolated arteriole, the only pathway from the lumen to the smooth muscle cells is directly across the arteriolar endothelium whereas transcapillary and transvenular routes arc avail able in vivo. The modes of drug administration used in our previous in vivo study [4] do not show the relative importance of the different pathways from the blood to the arteriolar smooth muscle. The aim of these experiments was to determine the rel ative intravascular and extravascular potencies of natural and artificial vasoconstrictors under controlled condi tions in which the delivery of blood-borne molecules to the arteriolar smooth muscle by all three routes (transarteriolar, transcapillary and transvenular) are possible.
In several experiments we measured the arteriolar responses to phenylephrine or vasopressin injeeted into the lumen of a venule that passed close ( < 20 pm) to the segment of the arteriole being ob served. These responses were compared to the responses to the same concentrations of agonist applied to the lumen of the same arteriole.
SKF 89748-A 40-,
Drugs
The following drugs were used in this study: pentobarbital Na (Nembutal. Abbot laboratories): phenylephrine HC1 (Sigma): SKF 89748-A (a gift from Smith Kline & French); angiotensin II acetate (Sigma): (arg8)-vasopressin acetate (Sigma). SKF 89748-A was stored refrigerated as a 5-m.V/ stock dissolved in dimethylsulfoxide. phen ylephrine was stored refrigerated as a 10-m.l/ stock dissolved in water, angiotensin II and vasopressin were stored in frozen aliquots at 0.1 mM. Each drug was diluted to the required concentrations in Ringer's solution on the day of the experiment. Dimethylsulfoxide was added to the phenylephrine dilutions to give the same concentra tion as was present in the SKF 89748-A dilutions.
SKF 89748-A concentration (log M)
1
Results
Phenylephrine Inira- and Periarteriolar Application Arteriolar constrictions to the lipid-soluble U|-adrenoceptor agonist SK.F 89748-A were similar with either intraluminal or adventitial application (fig. 1). The differ ence between the responses to 0.1 |jlM applied luminally and adventitially was small (about 20%) but significant (p < 0.05. paired t test). In contrast, the arteriolar responses to phenylephrine were strongly dependent of the route of phenylephrine application (fig. 2). With lumi nal application, more than 1 \iM phenylephrine was needed to elicit a constriction, but when applied to the adventitial surface of the vessels, that concentration of phenylephrine caused near-maximal responses. Thus lu minally applied phenylephrine was between 10 and 100 times less potent than adventitially applied phenyleph rine (fig. 2). Angiotensin II applied to the outside of the arterioles was suprathreshold at 1.0 nM and supramaximal at lOOnM. but 100 mV/ applied via the lumen caused re sponses of less than 50% (fig. 3). We were unable to apply
Phenylephrine concentration (log M)
2
Angiotensin II 40-,
5
? 3
30-
B
T
§
CD
I 20B
Inside
0
tered to the arterioles luminally (•) and adventitially (o) (n = 5). Error bars are SEM. Fig. 2. Concentration-response curves for phenylephrine admin istered to the arterioles luminally (•) and adventitially (o) (n = 5). Error bars are SEM. Fig. 3. Effect of angiotensin II administered luminally (•) and adventitially (o) on arteriolar diameter (n = 5). Error bars arc SEM.
1
10O
Rest
ih
—
-9
-8
Outside
r—
-7
Angiotensin II concentration (log M)
343
Downloaded by: King's College London 137.73.144.138 - 1/16/2019 11:50:12 PM
Fig. 1. Concentration-response curves for SKF 89748-A adminis
Vasopressin was about 100 times more effective in activating the arterioles from the adventitial side than from the lumen (fig. 4). Applied luminally. lOnM vaso pressin elicited responses of slightly less than 50%, but applied adventilially, only 1 nM was required for maxi mal responses. No dilation in response to any of the ago nists by either route of application was observed.
Vasopressin
Intravenular Injections Where a venule crossed under or over the arteriolar segment being observed, we occasionally made injections of phenylephrine (1 ,uM n = 3. 10 \iM n = 1) or vasopressin (10 nM n = 3) into that venule and observed the responses of the arteriole. In every case, application of the drug to the lumen of the venule resulted in an arteriolar constric tion that was larger than the constriction that the same concentration of drug elicited from the arteriolar lumen (fig. 5). No change in the diameter of the venule was observed during the drug injections. Drug Application Site:
Venule
Discussion
Fig. 4. Concentration-response curves for vasopressin adminis tered to the arterioles intravascularly (•) and extravascularly (o) (n = 5). Error bars are SEM. Fig. 5. Arteriolar responses to phenylephrine (1 \iM n = 3 and 10 gM n = 1 clumped) or vasopressin (10 nM n = 3) applied via the lumen of a venule close to the arteriole (Venule), via the arteriolar lumen, and direct adventitial application.
any further concentrations of angiotensin II because of the occurrence of tachyphylaxis. Because the responses to I nM applied to the adventitia and 100 nM applied luminally were similar, it can be estimated that the angiotensin II potency differential for luminal and adventitial appli cation was about 100-fold.
344
Lew/Duling
Delivery o f Blood-Borne Agents to the Arteriolar Smooth Muscle If drugs in the blood had direct unrestricted access to the arteriolar smooth muscle cells by diffusion through the arteriolar endothelium, then transcapillary and transvenular pathways for the movement of drugs from the blood to the smooth muscle would not be important. That is because unrestricted diffusion through the endothelium is equivalent to the inside surface of the smooth muscle being in almost direct contact with the blood: unrestricted diffusion with a path of the thickness of the endothelium would be virtually instantaneous and would involve no significant decrease in concentration. With the smooth muscle cells exposed to the same concentration of drug as that in the arterial blood, the concentration gradients within the tissue would always be downwards towards the capillaries and venules. Thus no net movement of drug molecules from those vessels to the arteriolar smooth muscle would be possible. The data in this study, and that in our previous papers, suggest that those circumstances do not occur, the arteriolar endothelium does not allow unrestricted diffusion of all drugs from the blood to the smooth muscle cells [3, 4], and therefore the delivery of vasoactive agents from the blood to the arteriolar smooth muscle by the transcapillary and transvenular pathways is possible.
Access from Blood to Arteriolar Smooth Muscle
Downloaded by: King's College London 137.73.144.138 - 1/16/2019 11:50:12 PM
□
Transcapillary Delivery Capillaries are the main exchange vessel for delivery of blood-borne agents to the tissue, and capillary endothe lium is probably less tight than the arteriolar endothelium [9, 12, 13, 16. 17], All blood-borne vasoactive agents are able to reach the arteriolar smooth muscle by leaving the capillary blood and diffusing through the extravascular space. With bolus administration of agonists (as in our previous in vivo study [4]), delivery of the agonists to the arteriolar smooth muscle via the capillaries will always involve substantial dilution in the extravascular tissue. Even with extended infusions, the general tissue concen tration of a vasoconstrictor need not reach that present in the infusate because of losses of the vasoconstrictor from the tissue, either by degradation and uptake, which is common for endogenous vasoconstrictors, or by loss from the tissue into the superfusion solution in experiments such as these. The infusion durations that we used in these experi ments (1-3 min) were long enough to allow the attain
ment of steady response levels, and Ley and Arfors [12] found that accumulation of the water-soluble marker, flu orescein. in the extravascular space of the hamster cheek pouch started to slow after about 20 s of intra-arteriolar injection. Therefore it is reasonable to suppose that our agonist infusions were long enough to allow the agonists to approach steady-state concentration in the tissue. Nonetheless, the design of these experiments could have led to an underestimation of the periarteriolar concentra tion of agonist that would occur under more physiological circumstances by at least two mechanisms. The first is the loss of agonist into the nonphysiological diffusion sink provided by the superfusion solution previously men tioned. The second is that only those capillaries and ven ules fed by the arteriole that was impaled for drug injec tion were perfused with drug-containing solution. Any other capillaries or venules in the vicinity of the arteriole would have acted as diffusion sinks for the drug, rather than sources, thereby lowering the periarteriolar drug concentration. Under normal physiological circum stances. all of the arterioles, and thus all of the capillaries and venules, would be perfused with blood containing the vasoactive hormone. It should be noted that in our experi ments both of these mechanisms would have a larger effect on the delivery of agonist to the smooth muscle by the transcapillary and transvenular routes than by the direct transendothelial route. Transvenular Delivery The ability of norepinephrine to elicit arteriolar con strictions by leakage from venules to nearby arteriolar smooth muscle has recently been demonstrated by Tigno et al. [2], but they did not determine the relative threshold concentrations for norepinephrine applied to the arterio lar lumen and to the venular lumen. In our experiments we found that arterioles can be more sensitive to agonists in the lumen of nearby venules than to the same agonists in the arteriole’s own lumen (fig. 5). This finding is consis tent with that of Ley and Arfors [ 12] who showed that fluorescein (water-soluble, molecular weight 376 daltons) accumulates faster in the tissue around venules than in the tissue around arterioles when it is infused intra-arteriolarly in the hamster cheek pouch. The rapid movement of water-soluble molecules from the blood into the tissue around venules comes about because venular endothe lium is more permeable than the arteriolar endothelium [8, 12, 16, 17], Thus the venules can act as a source of vasoactive molecules for the arteriolar smooth muscle. The degree to which this can occur must be dependent on both the anatomical arrangement of the arterioles and
345
Downloaded by: King's College London 137.73.144.138 - 1/16/2019 11:50:12 PM
Passage through the Arteriolar Endothelium Small water-soluble molecules, such as phenylephrine, vasopressin and angiotensin II, move from the intravas cular space to the extravascular space predominantly by passage through the spaces between the endothelial cells [7], These spaces are a tiny fraction of the surface area of the arteriolar endothelium so the total flux through them is small [8-13]. Thus the concentration of water-soluble molecules on one side of the arteriolar endothelium can be vastly different from the concentration on the other side. This is illustrated by the water-soluble agonists in this study, which were present at the arteriolar smooth muscle at an average concentration about 100 times less than that applied to the other side of the endothelium (us ing the arteriolar responses as an indirect assay of the average concentration of agonists at the smooth muscle cells) (fig. 2-4). In contrast, lipid-soluble molecules can permeate the cell membranes of endothelial cells [7, 14, 15], making the entire surface area of the arteriolar endo thelium available for their movement from the lumen to the smooth muscle. Accordingly, SKF 89748-A adminis tered via the arteriolar lumen was present at the arteriolar smooth muscle cells at a concentration approaching that in the lumen (fig. 1). These data confirm that the direct passage of water-soluble vasoactive agents in vivo across the arteriolar endothelium is limited, and therefore indi cate that the transcapillary and transvenular routes may have important roles in delivery of vasoactive agents from the blood to the arteriolar smooth muscle cells.
venules, and on the ability of the activating agonist to elicit responses that propagate along the arteriolar tree [2, 18], Because the arteriolar endothelium reduces the direct access of water-soluble blood-borne molecules to the arte riolar smooth muscle, this mechanism could have an important role. Consequences The fact that blood-borne vasoactive agents can reach their sites of action by the transcapillary and transvcnular pathways rather than (or in addition to) directly across the arteriolar endothelium has an important consequence for vascular responsiveness. It leads to the possibility that the responsiveness of a vascular bed, or segment of a vascular bed. might be determined by its anatomical arrangement. The effectiveness of the transcapillary pathway would be dependent on the number of capillaries present in the tis sue, and the effectiveness of the transvenular pathway would be dependent on the degree to which the tissue had arteriole-venule pairing. This means that the responsive ness of a vascular bed (or part of a vascular bed) can be designed in a way that would influence the responsiveness to vasoconstrictors in general, a mechanism that would stand in addition to the distribution of specific receptors [4].
Conclusion In this study we have demonstrated that the arteriolar smooth muscle of the hamster cheek pouch is partially isolated from the blood by the arteriolar endothelium, and therefore the delivery of blood-borne vasoactive agents to the arteriolar smooth muscle cells is. at least in part, the result of movement of these agents through the capillary' and vcnular walls into the extravascular tissue space. Regional differences in the permeability of the endothelium (arteriolar, capillary or venular), or in the anatomical arrangement of the capillaries and venules with respect to the arterioles, could be important factors in determining regional vascular reactivity to blood-borne agents.
Acknowledgments We thank Dr. J.P. Hieble of Smith, Kline & French for supplying the SKF 89748-A. and Mr. and Mrs. H. Caplan for their continuing support. We also thank D.N. Damon for his expert advice and assis tance. This work was supported by National Heart, Lung and Blood Institute grants HL-19242 and HL-12792. and by American Heart Association, Virginia Affiliate grant VA-89-F-14.
1 Hester RL. Guyton AC: Venular-arteriolar dif fusion of vasocactive metabolites. FASEB J 1989;3:A 1388. 2 Tigno XT. Lev K. Pries AR. Gaehtgens P: Venulo-arteriolar communication and propa gated response - a possible mechanism for local control of blood (low. Pflügers Arch Phar macol 1989:414:450-456. 3 Lew MJ, Rivers RJ. Duling BR: Arteriolar smooth muscle responses are modulated by an intramural diffusion barrier. Am J Physiol 19B9;257:H 10—H16. 4 Lew MJ, Duling BR: Arteriolar reactivity in vivo is influenced by an intramural diffusion barrier. Am J Physiol 1990:259:1I574-H58I. 5 Hynes MR. Duling BR: C a '' sensitivity of iso lated arterioles from the hamster cheek pouch. Am J Physiol 1991 ;260:H355-H361. 6 Duling BR: The preparation and use of the hamster cheek pouch for studies of the microcirculation. Microvasc Res 1973:5:423-429.
7 Rcnkin EM: Multiple pathways of capillary permeability. Circ Res 1977:41:735-743. 8 Simionescu M. Simionescu N. Palade GE: Seg mental differentiations of cell junctions in the vascular endothelium. The microvasculalure. J Cell Biol 1975;67:863-885. 9 Simionescu N, Simionescu M, Palade GE: Structural basis of permeability in sequential segments of the microvasculature of the dia phragm. Microvasc Res 1978:15:17-36. 10 Okuda T, Yamamoto T: The ultrastructural basis of the permeability of arterial endothe lium to horseradish peroxidase. Cell Tissue Res 1983;231:1 17-128. 11 Mink D. Schiller A. Kriz W. Taugner R: Intercndolhclial junctions in kidney vessels. Cell Tissue Res 1984;236:567-576. ’ 12 Ley K. Arfors K-E: Segmental differences of microvascular permeability for FITC-dextrans measured in the hamster cheek pouch. Micro vasc Res 1986:31:84-99.
13 Bundgaard M: The three-dimensional organi zation of tight junctions in a capillary endothe lium revealed by serial-section electron micros copy. J Ullrastruct Res 1984:88:1 —17. 14 Pappenheinter JR. Renkin EM. Borrero LM: Filtration, diffusion and molecular seiving through peripheral capillary membranes. A contribution to the pore theory1of capillary' per meability. Am J Physiol 1951;167:13-46. 15 Rcnkin EM: Capillary permeability to lipidsoluble molecules. Am J Physiol 1952:168: 538-545. 16 Olescn S-P: Electrical resistance of arterioles and venules in the hamster cheek pouch. Acta Physiol Scand 1985:123:12 1-126. 17 Rous P. Gilding BM. Smith F: The gradient of vascular permeability. J Exp Med 1930:51: 807-830. 18 Delashaw .1B, Segal SS, Duling BR: Drugs with different mechanisms of action induce similar conducted responses in arterioles. FASEB J 1988;2:A 1861.
346
Lew/Duling
Access from Blood to Arteriolar Smooth Muscle
Downloaded by: King's College London 137.73.144.138 - 1/16/2019 11:50:12 PM
References
