INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

THE BRONCHIAL CIRCULATION traverses and serves as primary nutrient source to airways, nerve bundles, hilar lymph nodes, and vasculature in the lung. It is important in the overall regulation of airway lining fluid balance, debris elimination, and conditioning of inhaled air (10). A central feature of this vasculature is its anastomosis with pulmonary veins just proximal to terminal bronchioles. Under normal conditions most intrapulmonary bronchial blood flow returns via these anastomotic connections to the left atrium, i.e., bronchopulmonary shunting (10). In congestive heart failure (CHF) the extent of such bronchopulmonary shunting is reduced (1, 3, 25, 37) and may contribute to bronchial hyperresponsiveness, or "cardiac asthma," often seen in CHF. In part, this may be caused by the heart failure-induced elevation in pulmonary venous pressure with a resultant increase in the delivery of bronchial blood flow to bronchial veins.

An alternative explanation, however, may relate to functional and/or structural changes in the bronchial vasculature per se in CHF. In the human pulmonary circulation, chronic pulmonary venous hypertension secondary to CHF is associated with structural and functional changes that include remodeling of smooth muscle mass, thickening of the endothelial basement membrane, and alterations in endothelium-dependent vasodilation (11). In a canine model of CHF induced by 4 wk of rapid ventricular pacing, our laboratory has previously shown that vasoconstrictor responses of the pulmonary vasculature to norepinephrine (NE), and to a lesser extent to ANG II, are enhanced (29, 34) although unrelated to structural alterations at this time point (33). Whether such adaptive changes occur in intrapulmonary bronchial vessels as a result of CHF is not known.

Earlier studies using isolated preparations of extrapulmonary bronchial arteries demonstrated that adrenergic vasoconstriction in these vessels is mediated predominantly by ₁-adrenoreceptors. In both bovine bronchial artery strips (2) and pig bronchial artery rings (24), only ₁-agonists caused vasoconstriction. Similar results were seen in isolated rings of canine extrapulmonary bronchial artery, where vasoconstriction induced by exogenous NE or release of endogenous neuronal NE was inhibited by the selective ₁-antagonist prazosin but not by the selective ₂-antagonist rauwolscine (27, 28). Furthermore, NE-mediated vasoconstriction of rat isolated extrapulmonary bronchial artery rings was enhanced by endothelium removal (38). Others (21-23, 35), using the canine model of pacing-induced CHF, have shown that endothelium-dependent, nitric oxide (NO)-related dilation of peripheral vessels is altered. However, we are not aware of any studies on the -adrenergic-mediated vasoconstriction of isolated intrapulmonary bronchial vessels, or its modulation by the endothelium, in either the normal canine lung or after the development of CHF. One cannot necessarily predict the sensitivity of the intrapulmonary segment of this vasculature from that obtained in extrapulmonary vessels.

Therefore, the present investigations were intended 1) to examine the -adrenergic vasoreactivity of canine intrapulmonary bronchial arteries, 2) to characterize adrenoreceptor subtypes involved in -adrenergic vasoconstriction induced by NE or phenylephrine (PE), and 3) to determine the effect of the endothelium and NO on -adrenergic vasoconstriction in the normal lung and that after 4 wk of pacing-induced heart failure. Finally, we sought to determine whether the effects of pacing-induced CHF were related to vessel branch order. Vessels were studied as cannulated and pressurized segments, to better simulate the physiological condition.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Right middle and cardiac lung lobes were isolated from normal canines (71 dogs) and those that had been chronically paced (240-245 beats/min) for 4 wk (55 dogs). The model of rapid ventricular pacing, the in vivo hemodynamic status, and pulmonary vascular function in caudal lung lobes in these animals have been previously reported (29, 34). Thus for the study reported here no new animals were required.

Drugs

Isolation of Vessels

To facilitate isolation and to minimize damage to the bronchial endothelium, a large pulmonary blood vessel was cannulated with polyethylene tubing and the vasculature filled passively with a mixture of gelatin (3.5%) and India ink (Kohl-I-Noor Radiograph, 3-5 drops/10 ml) in Krebs buffer (37°C). When the vasculature was filled, as indicated by blackening of the parenchyma and outflow of the gelatin-ink mixture, the lobe was placed in fresh Krebs buffer on ice to solidify the gelatin (~30 min). The lobe was then transferred to a dissecting chamber containing fresh cold buffer and kept over ice. Intrapulmonary bronchial vessels ranging from 83 to 1,032 µm in diameter (2-4 mm in length) were dissected with the aid of a stereozoom dissecting microscope. Vessels of uniform diameter or those with a smaller intact side branch (see Experimental Protocols) were isolated.

Vessels were immediately transferred to a suffusion chamber filled with Krebs-Ringer solution and cannulated with glass micropipettes. Uncannulated side branches were ligated with individual strands teased from 5-0 silk suture (Ethicon). One cannula was connected to a reservoir of Krebs-Ringer solution that could be moved vertically to set distending pressure. The other cannula was connected to a stopcock that could be opened to flush the lumen or closed to maintain a constant distending pressure. Once mounted in the suffusion chamber, vessels were continuously suffused (7 ml/min) with warmed Krebs-Ringer solution (37°C) aerated with 95% O₂-5% CO₂ (pH 7.35-7.45). Vessels were used only on the day of isolation to ensure viability, and they were normally mounted and suffusion was initiated within 2-3 h after the isolation of lung lobes. A stereozoom microscope (SMZ-2T, Nikkon) fitted with a calibrated eyepiece reticule was used to measure the diameters of cannulated vessels.

Mounted vessels were allowed to equilibrate at a distending pressure of 20 mmHg until contractile responses to 44.7 mM KCl had stabilized (0.5-2.5 h). Vessels were then challenged with 44.7 mM KCl over a range of transmural pressures to determine the distending pressure at which maximal constriction occurred (the optimal transmural pressure, P_TM). Vessels were allowed to equilibrate at each pressure for 10 min before suffusion with KCl. For each KCl challenge, the vessel was suffused with 44.7 mM KCl in Krebs-Ringer solution until the maximum contractile response was seen (<5 min), at which time the suffusate was switched to Krebs alone to allow relaxation. After the P_TM had been determined, 10 µM normetanephrine (to block extraneuronal uptake-degradation of NE), 5 µM propranolol (to inhibit -adrenoreceptors), 1 µM desipramine (to block neuronal reuptake of NE), and 5 µM ibuprofen or indomethacin (cyclooxygenase II inhibitors) were then added to the suffusate. Vessels were allowed to equilibrate with these antagonists (and others when used) for ~30 min before further study. Only vessels that exhibited stable constriction with KCl were accepted for study.

Experimental Protocols

Responses to -adrenergic agonists. Segments of bronchial vessels from control and paced animals were exposed to cumulative doses (10⁹ to 10³ M) of the -adrenergic agonists NE or PE. The inclusion of the -adrenergic blocker propranolol in the suffusate allowed us to specifically evaluate the -adrenergic responses of these intrapulmonary bronchial vessels. After these baseline measurements, subsets of these vessels were treated with specific ₁- or ₂-antagonists, or had the endothelium removed, as described below. Vessels used to study PE responses were initially pretreated with 3 × 10⁵ M PE until stable responses were obtained. PE pretreatment was done to remove the differences in maximum responses observed between initial and final dose curves in preliminary studies. In vessels exposed to NE, maximum responses were similar for initial and final curves without pretreatment.

Effects of ₁-adrenergic antagonism. In vessels from control (n = 16) and paced (n = 11) animals a second cumulative dose response to NE was determined in the presence of the ₁-blocker prazosin (10⁶ M). After initial exposure to NE, vessels were allowed to spontaneously return to baseline diameter (suffusion with Krebs buffer alone) and prazosin was added to the suffusate. After 30 min, dose responses to NE were repeated. In separate control (n = 11) and paced (n = 10) groups, dose responses to PE were determined again after treatment with prazosin, as described above.

Effect of ₂-adrenergic antagonism. Similarly, in vessels from control (n = 10) and paced (n = 15) animals a second cumulative dose-response curve to NE was generated in the presence of the ₂-antagonist (yohimbine, 10⁷ M). Other than the antagonist used, the protocol remained the same as that described above for the ₁-adrenergic antagonist. Verification of the ₁-selective activity of PE was done with additional control vessels (n = 12) in which responses to PE were reevaluated after treatment with the selective ₂-antagonist rauwolscine (10⁷ M). Dose responses to PE in vessels from paced animals were not evaluated in the presence of an ₂-antagonist.

Effect of endothelium removal. In vessels from control (n = 9) and paced (n = 12) dogs the possible effect of an endothelium-derived vasoactive substance on vascular smooth muscle contraction was assessed by repeating the cumulative dose-response relation to NE after endothelium removal. Endothelial cells were removed by luminal perfusion of vessels with 1- to 2-ml boluses of air (16); usually this procedure was repeated three times. Vessels were allowed to equilibrate for 45-60 min at the end of which smooth muscle activity was checked with 44.7 mM KCl. The second dose-response relation was then generated. Endothelial function was evaluated before and after exposure to air by testing the ability of a single dose of ACh (between 10⁷ and 10⁶ M) to dilate vessels constricted with 3 × 10⁴ M NE. Results were used only if ACh-mediated dilation after air perfusion was 15%.

Effect of nitric oxide synthase inhibition. In subsets of vessels from control and paced dogs (n = 20 and 9, respectively), cumulative dose responses to NE were repeated in the presence of the nitric oxide synthase (NOS) inhibitor L-NAME (300 µM). After initial exposure to NE, vessels were allowed to return to baseline diameter by suffusion with Krebs-Ringer solution alone. L-NAME was then added to the Krebs, and vessels were allowed to equilibrate for an additional 20-30 min before repetition of dose responses to NE. In a separate subset of vessels (n = 11) from control dogs, a second, cumulative dose response to NE was generated in the presence of D-NAME (300 µM), the inactive enantiomeric isomer of L-NAME, as a negative control. In additional subsets of vessels obtained from control dogs, the ability of L-Arg (300 or 500 µM, n = 13 and 7, respectively) to reverse the effects of NOS inhibition (L-NAME, 300 µM) on NE-mediated constriction was assessed. L-Arg is the physiological substrate for NOS.

Effect of vessel branching order. To examine the effect of branching order on -adrenergic vasoreactivity, data were reanalyzed by comparing responses in main stem vessel segments located along the primary (1°) airway of the lung lobe to those in intact side branches of these segments that led directly to or were located on the next generation or secondary (2°) airway. These vessels were derived from the same groups as those described above, though responses were only compared in this manner when paired measurements in contiguous branches of the vessel could be obtained.

ACh and SNP bronchial vasodilation. KCl (18-30 mM) was used to preconstrict vessels isolated from control (n = 17) and paced (n = 13) dogs to evaluate vasodilatory responses to cumulative doses of ACh and SNP. Tone was set to equal ~70% of the response to 44.7 mM KCl. Responses to ACh and SNP were measured sequentially in each vessel, and the order was randomized among vessels. Vessels were washed with Krebs containing KCl to restore preconstrictor tone after initial dose responses and again after the second series of doses to confirm the stability of constrictor tone.

Time control experiments. Time controls for control and pace groups were performed, in separate groups of vessels, by obtaining responses to two sequential series of cumulative doses of NE or PE in the absence of an -adrenergic antagonist.

Data Analysis

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Overall baseline diameters, P_TM, and KCl responses. Vessels of similar size ranges were evaluated in the control (83-1,032 µm) and paced (108-761 µm) groups. P_TM was not significantly different between groups. In both groups, 2° branches were significantly more responsive to KCl than were 1° branches. Overall, vessels from paced dogs displayed responses to 44.7 mM KCl at P_TM that were significantly higher than in controls, a pattern observed at both 1° and 2° branches. These overall baseline data are summarized in Table 1.

Bronchial vasoreactivity to NE and PE. The maximal constriction, i.e., percent decrease in vessel diameter, induced by NE was not different between the control and paced groups as shown in Table 2. Normalized dose responses (Fig. 1) and ED₅₀ values for NE (Table 2) indicated that vessels in the paced group were slightly but significantly more sensitive (i.e., lower ED₅₀ values) to NE than were controls (P < 0.01). The paced group also had significantly lower ED₅₀ values for PE compared with control (P < 0.01). Furthermore, the ED₅₀ values for NE were significantly lower than those for PE in both the control (P < 0.01) and paced (P < 0.05) groups.

View larger version (20K): [in this window] [in a new window]   Fig. 1.   Contractile responses of canine isolated intrapulmonary bronchial arteries to cumulative doses of norepinephrine (NE, A) and phenylephrine (PE, B). A: each point of each curve represents the mean ± SE of 115 vessels from control dogs and 71 vessels from paced dogs, respectively. B: vessel numbers for each point from control and paced dogs are 23 and 10, respectively. Values are normalized relative to corresponding maximal responses to NE or PE.

Effects of ₁-adrenergic antagonism. ₁-Adrenoreceptor involvement in bronchial artery vasoconstriction mediated by NE or PE was assessed with the selective ₁-antagonist prazosin. In the control group, prazosin slightly enhanced the maximal response to NE compared with baseline in the same vessels (P = 0.05) and significantly increased the ED₅₀ (P < 0.05) as shown in Table 3. This is illustrated by the rightward shift in the dose-response relation (Fig. 2). In paired experiments with vessels from paced dogs, prazosin did not alter the maximum response to NE compared with NE alone but significantly increased the ED₅₀ (P < 0.01, Table 3, Fig. 2). Thus for both groups of vessels prazosin acted as a competitive inhibitor of NE-mediated vasoconstriction. Prazosin was also a competitive inhibitor of PE-mediated responses. Whereas the maximum response to PE in the control group was unaltered after prazosin, the ED₅₀ was significantly increased (P < 0.01) as indicated in Table 3. Similarly, in the paced group, prazosin did not affect maximum response to PE but significantly increased the corresponding ED₅₀. The rightward shift of the dose-response relation for PE in both groups after prazosin is shown in Fig. 3.

View this table: [in this window] [in a new window]   Table 3.   Maximum responses and ED50 values: paired measures before and after 1- or 2-antagonism

View larger version (20K): [in this window] [in a new window]   Fig. 2.   NE dose-response curves of canine isolated intrapulmonary bronchial arteries in absence and presence of 1-antagonist prazosin. Vessels from control (A; n = 16) and paced (B; n = 11) animals were first exposed to cumulative doses of NE in absence of prazosin, allowed to reequilibrate to baseline diameters, and then reexposed to doses of NE in presence of 106 M prazosin. Each point represents mean ± SE of results in all segments at a particular dose. All values were normalized to maximal response of corresponding curve.

View larger version (18K): [in this window] [in a new window]   Fig. 3.   PE dose-response curves of canine isolated intrapulmonary bronchial arteries in absence and presence of 1-antagonist prazosin. Vessels from control (A; n = 11) and paced (B; n = 10) animals were first exposed to cumulative doses of PE in absence of prazosin, allowed to reequilibrate to baseline diameters, then reexposed to PE in presence of 106 M prazosin. Individual responses represent mean ± SE of all values at a particular dose, each normalized to maximal response of corresponding curve.

Effects of ₂-adrenoreceptor antagonism. Although treatment with the ₂-antagonist yohimbine tended to decrease vessel diameter even before addition of exogenous NE, there were no significant differences between paired baseline and postyohimbine measurements (data not shown). Similarly, in vessels from control and paced animals the maximal responses to NE were not different before and after addition of yohimbine as shown in Table 3. The ED₅₀ for NE in controls was significantly lower in the presence of yohimbine than in its absence (P < 0.02), whereas the ED₅₀ for vessels from paced dogs was not significantly affected by the addition of yohimbine. This leftward shift of the NE dose relation in the control group is illustrated in Fig. 4. In control vessels (n = 12) rauwolscine, a selective ₂-antagonist, did not effect the maximum response or the ED₅₀ of PE, thus verifying that PE acts a selective ₁-agonist in this vasculature.

View larger version (21K): [in this window] [in a new window]   Fig. 4.   NE dose-response curves of canine isolated intrapulmonary bronchial arteries in absence and presence of 2-antagonist yohimbine. Vessels from control (A; n = 10) and paced (B; n = 15) animals were first exposed to cumulative doses of NE in absence of 2-antagonism, allowed to reequilibrate to baseline diameters, and reexposed to doses of NE in presence of 107 M yohimbine. Individual responses represent mean ± SE of all values at a particular dose, as noted previously.

Effects of endothelium removal. Successful endothelium removal was verified by the loss of ACh-mediated vasodilation in both groups. A single dose of ACh dilated NE-preconstricted vessels by 93.9 ± 2.7% before and 7.2 ± 4.8% after luminal air perfusion in the control group compared with 93.2 ± 3.0 and 5.4 ± 4.4% dilation, respectively, in the paced group. After endothelium removal in the control group, the overall maximum response and sensitivity, as measured by the ED₅₀, to NE significantly increased (P < 0.01). In contrast, in the paced group, there was no difference in maximum response to NE before and after endothelium removal (Table 4). Furthermore, the NE ED₅₀ in the paced group was significantly increased after endothelium removal (P < 0.01). The normalized dose-response relationships for NE before and after endothelium removal are shown in Fig. 5.

View larger version (21K): [in this window] [in a new window]   Fig. 5.   NE dose-response curves of canine isolated intrapulmonary bronchial arteries before and after endothelium (Endo) removal. Vessels with intact endothelium from control (A; n = 9) and paced (B; n = 12) dogs were initially exposed to cumulative doses of NE. Endothelium was then removed (luminal air perfusion, see METHODS), and vessels were allowed to reequilibrate in Krebs-Ringer solution for 45-60 min or until stable responses to 44.7 mM KCl were seen. Vessels were then reexposed to cumulative doses of NE. Endothelium removal was verified by observing a lack of dilatory response to acetylcholine (ACh) in vessels preconstricted with NE (see METHODS for details).

Effect of NOS Inhibition. L-NAME, somewhat similar to yohimbine, decreased vessel diameter before exposure to NE but to a greater extent in control vs. paced groups (9.6 ± 2.2 and 0.9 ± 0.9%, respectively, P < 0.02). Thus NE-mediated diameter changes were calculated relative to the baseline diameter after L-NAME. The dose-response curves are shown in Fig. 6. In the control group L-NAME did not significantly affect the maximum response to NE but markedly increased the sensitivity to NE. Overall, ED₅₀ fell significantly (P < 0.01, Table 4), indicating that NO attenuates contractile responses to NE in intrapulmonary bronchial arteries of the normal dog. In the paced group, maximum responses to NE were not different from the corresponding vessel set in the controls and were also unaffected by L-NAME. However, in contrast to the control group, ED₅₀ values measured in the paced group before and after L-NAME were not different. Thus modulation of NE-mediated vasoconstriction of bronchial arteries by NO was no longer apparent after pacing-induced CHF in the dog.

View larger version (18K): [in this window] [in a new window]   Fig. 6.   Effect of nitric oxide synthase (NOS) inhibition on NE intrapulmonary bronchial vasoconstriction. A subset of isolated intrapulmonary bronchial arteries, in which constrictor responses to cumulative doses of NE (109 to 3 × 104 M) were determined, were allowed to regain passive, baseline tone by suffusion with Krebs-Ringer solution after which NOS inhibitor N-nitro-L-arginine methyl ester (L-NAME, 300 µM) was added to suffusate. After 30 min a second cumulative dose response to NE was determined in presence of L-NAME. Response at each concentration of NE represents mean ± SE of n = 20 and n = 9 vessels isolated from control (A) and paced (B) dogs, respectively. Individual responses are normalized to maximum response of corresponding curve.

Effects of vessel size and branching order. The relative responsiveness of 1° and 2° vessel segments differed somewhat between control and paced groups. As noted earlier, in both groups, 2° vessel segments had significantly greater maximum responses to KCl than did the mainstem segments (Table 1). A similar pattern was observed with NE in both groups, whereas maximum response to PE in 2° segments was only significantly greater in the control group (P < 0.01). The number of paired PE-treated segments in the paced group was small, and the relative differences as a function of branch order did not attain statistical significance. Nonetheless, 2° segments were overall more responsive to constrictor stimuli than were their corresponding, paired mainstem parent segments. In contrast to these results, the sensitivity of 2° segments from either group to NE or PE, as measured by the ED₅₀ values, was not different than that of the parent segment (Table 5). No consistent branch order-related pattern was seen with respect to effects of endothelium removal or inhibition of NOS with L-NAME in either group.

View this table: [in this window] [in a new window]   Table 5.   Effect of vessel branching order on -adrenergic vasoreactivity

ACh- and SNP-induced vasodilation. Initial constrictor tone for ACh or SNP responses did not differ between groups (Table 6). In both groups SNP-induced maximum dilation was greater than that for ACh (P < 0.05 in each group). The ED₅₀ for ACh, however, was less than that for SNP, indicating that in both groups vessels were more sensitive to ACh stimulation. Maximum dilation (P < 0.05) and sensitivity (i.e., ED₅₀, P = 0.05) to ACh were depressed in the paced vs. control group. SNP-mediated vasodilation and sensitivity were similar for the two groups (Fig. 7). Thus differences in vasodilator response of the normal and paced groups are likely caused by diminished endothelial dilatory function after pacing in the dog as smooth muscle responses to NO were similar for the two groups.

View larger version (16K): [in this window] [in a new window]   Fig. 7.   Effect of pacing-induced congestive heart failure on ACh- (A) and sodium nitroprusside (SNP)-mediated (B) dilation of intrapulmonary bronchial arteries. Vessels were preconstricted to submaximal tone with 18-30 mM KCl in presence of ibuprofen (5 µM) to inhibit cyclooxygenase II. After stabilization of vascular tone, sequential cumulative dose responses to ACh and SNP (109 to 3 × 104 M for both substances) were determined. Sequence order of responses to ACh or SNP was randomized between dogs for both groups. Vessels were allowed to return to preconstrictor tone between initial and final dose curves by suffusion with Krebs containing preconstrictor concentration of KCl and ibuprofen. At each concentration of ACh and of SNP, paired responses represent means ± SE of n = 17 and n = 13 vessels from control and paced dogs, respectively.

Time control experiments. In separate groups of vessels, ED₅₀ values for two sequential cumulative dose-response curves for NE or PE were compared for vessels from both control and paced dogs. In controls, the ED₅₀ values for the initial and final dose-response curves for NE (n = 5, 6.53 ± 0.42 vs. 6.36 ± 0.49 M) and PE (n = 5, 5.30 ± 0.18 vs. 5.60 ± 0.24 M) were similar (P > 0.10 for both agonists). In vessels from paced dogs, ED₅₀ for the initial and final dose-response curves were also not different for NE (n = 4, 6.46 ± 0.29 vs. 6.53 ± 0.42 M) and PE (n = 4, ± 4.92 ± 0.37 vs. 4.60 ± 0.47 M, P > 0.30 for both agonists). Because mean ED₅₀ values for time controls were different from those presented earlier for NE and PE (Table 2) in both groups, and because data from vessels used for time controls were not included in results reported in the preceding sections, a separate analysis was made in which time control data were included with other data. This analysis did not reveal any change in the results, i.e. ED₅₀ values remained lower in the paced vs. control group for both NE and PE.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Previous studies in several species have demonstrated that the bronchial circulation in vivo or rings from extrapulmonary bronchial arteries constrict in response to adrenergic stimulation (9, 20, 24, 26, 27, 38). Our studies indicate that -adrenergic constriction in the canine intrapulmonary bronchial arterial bed is mediated predominantly by ₁-adrenoreceptors, in agreement with these earlier findings. However, our studies provide the first in depth in vitro investigation of the -adrenergic vasoreactivity of the intrapulmonary portion of this vasculature, and also describe an unique in vitro procedure that allows direct comparison of contiguous and successive branch orders of the arterial vasculature. Furthermore, our results suggest a role for endothelium-dependent vasodilation that opposes direct adrenergic stimulation of contractile activity in intrapulmonary bronchial arteries. Finally, we have observed that pacing-induced CHF is associated with modestly altered ₁-mediated vasoconstriction and attenuation of endothelium-dependent vasodilation.

Norepinephrine-mediated contraction of isolated canine intrapulmonary bronchial arteries indicates that -adrenoreceptors are present postsynaptically on smooth muscle cells in this vasculature. In vessels from normal animals, NE-mediated vasoconstriction was substantially inhibited by blockade of ₁-adrenoreceptors with the selective ₁-antagonist prazosin (5). The selective ₁-agonist phenylephrine (20) also induced constriction of intrapulmonary bronchial arteries, which was inhibited by prazosin but not by the selective ₂-antagonist rauwolscine. Thus it is evident that ₁-adrenoreceptors mediate substantial vasoconstriction in this vasculature. Similarly, Corboz et al. (8) reported that the ₁-adrenoreceptor was the predominant subtype involved in NE or PE-mediated vasoconstriction of small arterioles of the rat tracheal microvasculature.

In contrast to the effects of ₁ blockade, pretreatment with the nonselective ₂-antagonist yohimbine enhanced, rather than diminished, the sensitivity to NE-mediated constriction of vessels isolated from normal animals, an observation also made by Zschauer et al. (38). Blockade of ₂-receptor binding by yohimbine presumably increases binding of NE at more efficacious ₁-receptors (12, 13) and thus increases sensitivity to this nonselective agonist (20). Although selective ₂-blockade had no effect on NE-mediated constriction of canine extrapulmonary bronchial vessels (27), our results indirectly suggest that ₂-adrenoreceptors are present in the intrapulmonary portion of this vasculature. Nonetheless, without the use of selective ₂-agonists, we cannot confirm a direct role for ₂-adrenoreceptors in NE-mediated vasoconstriction of canine isolated intrapulmonary bronchial arteries.

Other mechanisms, aside from a possible shift between -subtype binding sites, could account for the enhanced sensitivity to NE that is observed after yohimbine in vessels from normal dogs. Non-₂-adrenoreceptor effects of yohimbine (e.g., antagonism of ₁-adrenergic, dopamine, serotonin receptors and monamine oxidase, agonist activity at serotonin receptors) do not likely explain our observations because these effects occur at doses substantially higher (>10⁵ M) (17) than that used in our studies (10⁷ M). Yohimbine also antagonizes presynaptic ₂-adrenoreceptors, which inhibit neuronal NE release, and possibly increases NE concentration at the postjunctional receptor (i.e., ₁) site. Although basal, i.e., unstimulated, release of NE from presynaptic nerve terminals has been demonstrated in various smooth muscle preparations (19, 36), the quantities of NE released are thought to be insufficient to evoke postjunctional receptor activity (36). Thus the vasoconstriction that we observed with yohimbine even in the absence of NE (mean = 9.5% diameter change) is probably not caused by modulation of neuronal NE release. A third, and more plausible, explanation is that yohimbine inhibits endothelial ₂-receptors (32) that release nitric oxide (NO) on activation by NE (7). We were able to demonstrate that endothelium removal or inhibition of NOS with L-NAME results in augmented sensitivity and responsiveness (endothelium removal only) to NE in vessels from normal dogs. These results were similar to those seen with yohimbine and to those observed in endothelium-denuded rabbit bronchial arteries (38). Our results suggest that there is an endothelium-derived factor(s) that, in normal canine intrapulmonary bronchial arteries, opposes direct -adrenoreceptor-mediated vasoconstriction.

Pacing-induced CHF caused significant alterations in the -adrenergic vasoreactivity of isolated intrapulmonary bronchial arteries. The sensitivity of -adrenergic vasoconstriction, predominantly mediated by the ₁-subtype as discussed above, was enhanced for both NE and PE after pacing. However, maximum vasoconstrictor response to these agonists was not affected. These results agree, in part, with other reports on experimental CHF and -adrenergic vasoreactivity in other systemic vascular beds. In a study that used a similar canine model of pacing-induced CHF, both the sensitivity to and the maximum tension developed in response to NE, PE and epinephrine in dorsal pedal artery and saphenous vein rings were increased (14, 15). Teerlink et al. (31), on the other hand, demonstrated increased responsiveness but unaltered sensitivity to NE in the thoracic artery of rats subjected to coronary ligation that led to heart failure. Differences in the experimental model and/or species may explain the differing results of these two studies. The second major finding of our work was that after pacing the endothelium no longer attenuates adrenergic vasoconstrictor activity in this vasculature. Furthermore, we found that ACh-mediated, endothelium-dependent vasodilation is diminished but the smooth muscle response to exogenous NO remains normal after pacing. This suggests that the endothelium is the cellular site of dysfunction during depressed vasodilation in the paced group, as observed in other peripheral vascular beds after pacing-induced CHF (21-23, 35).

Altered ₁-vasoreactivity in the intrapulmonary bronchial arterial network during CHF is coincident with that in the pulmonary vasculature. Using isolated, blood-perfused lung lobes, Townsley and co-workers (34) demonstrated that NE-mediated pulmonary arterial and venous vasoconstriction was enhanced after pacing-induced CHF in dogs, and that this was abolished by ₁-antagonism. It is not unexpected that altered vasoreactivity is present in both lung vasculatures because there are extensive anastomotic connections between the bronchial and pulmonary networks (3, 6, 10, 37). Thus structural and/or functional vascular remodeling caused by elevated pulmonary venous pressures might explain the enhanced bronchial arterial -adrenergic responsiveness as previously suggested for the pulmonary vasculature (34). The present studies suggest, however, that pacing-induced alterations in ₁-vasoconstriction may be due, at least in part, to a concomitant decrease in the ability of endothelium-derived vasodilatory factors to effectively attenuate ₁-adrenoreceptor-mediated vasoconstriction.

The preparation used in these studies allowed, in some vessels, concurrent examination of main stem arteries (i.e., 1° arteries) and their intact side branches (i.e., 2° arteries) such that direct comparisons could be made. The smaller-sized 2° vessel segments were significantly more responsive to KCl, and to -agonists, than were their paired 1° segment counterparts, indicating that overall contractile ability was proportionately greater in 2° segments. These results are in agreement with previous theoretical and experimental work, which indicates that contractility increases with successive branching along the arterial tree within a vasculature and is optimal at the level of resistance vessels (4, 12, 18). However, few of the alterations incurred by pacing-induced CHF were related to branching order. Generally, the potency of -agonists was increased after pacing at both 1° and 2° branches. Furthermore, the loss of endothelium- or NO-dependent relaxation after pacing was seen equally at both branch orders.

Results of our investigation of canine isolated intrapulmonary bronchial arteries can be summarized as follows. NE causes a dose-dependent constriction of isolated canine intrapulmonary bronchial arteries, an effect that is mediated primarily via ₁-adrenoreceptors. Pacing-induced CHF leads to enhanced sensitivity to ₁-adrenoreceptor activity. An endothelium-derived factor is implicated in a mechanism of vasodilation that opposes -adrenergic vasoconstriction under normal conditions. After pacing this mechanism is apparently lost, which may account for the enhanced sensitivity to -agonists. Finally, although overall results are indicative of a heterogeneity of responses along the canine intrapulmonary bronchial arterial tree, vascular segments of different branch order were equally affected by the pacing-induced and CHF-related alterations in the -adrenergic vasoreactivity of this vasculature. Although a number of questions remain to be resolved in future work, the present findings nonetheless have identified several key alterations underlying enhanced -adrenergic vasoreactivity in intrapulmonary bronchial arteries after pacing-induced CHF.