INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

THE PRO-INFLAMMATORY CHARACTERISTICS of tumor necrosis factor- (TNF-) have been documented extensively. Furthermore, numerous studies have demonstrated that these attributes contribute to the inflammatory conditions present in airways of asthmatic subjects. TNF- has been shown to activate inflammatory cells, upregulate adhesion molecules on endothelium and circulating leukocytes, increase the production of other cytokines (11, 23), and increase bronchial responsiveness (12, 28). This polypeptide is expressed primarily by alveolar and tissue macrophages, mast cells, and bronchial epithelial cells (3). Additionally, in most other airway cell systems studied, conditions simulating an inflammatory state result in expression of TNF- (11). Thus it is not surprising that TNF- has been recovered in greater quantities in the bronchoalveolar lavage fluid from symptomatic asthmatics compared with normal control subjects (4) and in allergic asthmatics late after antigen exposure (5). Perhaps less appreciated in the study of airways inflammation are the significant vasoactive properties of TNF-. Direct infusion of TNF- into the pulmonary vasculature of sheep resulted in an immediate and significant pulmonary vasoconstriction and a concomitant systemic vasodilation, both of which persisted for several hours (14). Regional vasoconstriction during TNF- infusion has been confirmed in coronary (13) and pial circulations (19). In cultured endothelial cells, TNF- exposure resulted in an increased secretion of endothelin-1 (ET-1; Ref. 6), which was accompanied by a corresponding increase in prepro-ET-1 mRNA transcript levels (17). ET-1 administration has been shown to result in vascular smooth muscle constriction due to its binding to the ET_A receptor (2). Pretreatment of an isolated heart preparation with a specific ET_A receptor antagonist resulted in a marked attenuation of TNF--induced coronary vasoconstriction (13). The airway circulation has been shown to be sensitive to exogenously administered ET-1, showing marked vasoconstriction (1). Whether release of TNF- by inflammatory cells within the airway causes secondary release of ET-1 by airway vascular endothelium has not been studied. Furthermore, whether this sequence results in significant vasoconstriction of the airway vasculature is not known. Given the involvement of TNF- in the overall coordination and maintenance of the inflammatory response in the asthmatic airway, the effects of TNF- on airway vascular dynamics may regulate inflammatory cell transit to the airway wall. It was the purpose of this study to determine the effects of infused TNF- on bronchial vascular resistance (BVR) and whether ET-1 was involved in the response. Although the hemodynamic factors influencing leukocyte kinetics through the pulmonary circulation have been studied (18), little is known regarding the consequences of changing hemodynamics on inflammatory cell recruitment to the airway wall. These experiments provide new information on the mechanisms by which a pro-inflammatory cytokine might alter the airway vasculature in a way that could influence subsequent inflammatory cell recruitment.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The study protocol was approved by the Johns Hopkins Animal Care and Use Committee. Anesthesia was induced in sheep (25-35 kg) with intramuscular ketamine (30 mg/kg) and subsequently maintained with intravenous pentobarbital sodium (20 mg · kg¹ · h¹). A tracheostomy was performed, the sheep were paralyzed with pancuronium bromide (2 mg iv), and the lungs were mechanically ventilated (10-12 ml/kg) at a rate (12-15 breaths/min) sufficient to maintain normal blood gases. A 5-cmH₂O positive end-expiratory pressure was applied. The left thorax was opened at the 5th intercostal space, and heparin (20,000 U) was administered. The esophageal and thoracic tracheal branches of the bronchoesophageal artery were ligated as previously described (25). The bronchial branch of the bronchoesophageal artery was isolated, cannulated, and perfused (0.6 ml · min¹ · kg¹) with autologous blood withdrawn from the descending aorta and pumped through a variable speed roller pump. BVR was defined as the bronchial artery inflow pressure/flow (24, 27).

Three experimental protocols were performed and focused on the following: 1) the determination of the effects of TNF- on BVR; 2) the confirmation of ET_A receptor antagonism with BQ-123 after ET-1 challenge; and 3) the inhibition of TNF--induced bronchial vasoconstriction with BQ-123. The first series of experiments (n = 11 sheep) involved following a 180-min time course of BVR during and after TNF- infusion through a side port of the bronchial perfusion circuit during control (n = 6) or low (50% of control; n = 5) flow. Recombinant human TNF- (Sigma Chemical, St. Louis, MO) was purchased in 10-µg aliquots, diluted in 20 ml PBS on the day of the experiment and infused with an infusion pump (Harvard Apparatus, South Natick, MA). For control bronchial artery perfusion, TNF- infusion was set at 1 ml/min (10 µg total dose; 1 ml/min for 20 min). To match perfusate concentration during low bronchial artery perfusion (50% control), TNF- infusion was set at 0.5 ml/min (10 µg total dose; 0.5 ml/min for 40 min). Time control experiments were performed in additional sheep (n = 3) in which BVR was monitored over the course of 180 min. In the second series (n = 6), ET-1 (1 × 10⁹ M through 1 × 10⁶ M; American Peptide, Sunnyvale, CA) was infused (4 ml; 1 ml/min) through a side-port of the bronchial perfusion circuit. Bronchial artery pressure was measured continuously for 15 min after the infusion was complete. Then the subsequent higher dose of ET-1 was administered. BQ-123 (American Peptide), a selective ET_A receptor antagonist, was administered at a dose (3 × 10⁶ M) in excess of that previously shown to block ET-1-induced pulmonary arterial constriction (2), and the ET-1 dose-response relationship was determined. Because of the length of these experiments imposed by the sustained bronchial vascular constriction, two sheep were treated with ET-1 only, two sheep were pretreated with BQ-123 and then received ET-1, and two sheep received ET-1 followed by BQ-123 treatment and a second ET-1 dose response.

The third series of experiments was similar to the first in which the time course of BVR was measured over 180 min; however, the animals were first pretreated with BQ-123 (20 ml of 3 × 10⁶ M; 1 ml/min) before TNF- infusion (10 µg in 20 ml PBS at 1 ml/min).

Airways resistance. To determine whether TNF- and/or ET-1 delivered directly to the airway wall altered airway smooth muscle tone, conducting airways resistance was measured by the method of forced oscillation (10). A gas volume of ~30 ml was oscillated for 1.5 s at a frequency of 9 Hz after each tidal breath. Airway pressure was measured at a side arm of the tracheal cannula, and a flow signal was obtained from a pneumotachograph positioned between the oscillator and the cannula. Oscillatory signals were analyzed with an online computer that measures pressures at points of peak flow. An average resistance was obtained over 8-10 oscillatory cycles. Baseline airways resistance measured in this manner in anesthetized sheep typically results in a value of 1.0 to 2.0 cmH₂O · l¹ · s¹, which is close to the value reported by others (8).

Statistical analysis. All data are presented as the mean values ± SE. The maximum increases in BVR at 180 min during low and control flows were compared using an unpaired t-test. The changes in BVR after TNF- administration and over the course of 180 min were analyzed using a one-way repeated measures analysis of variance followed by the post hoc Duncan's multiple range test. Paired and unpaired t-tests were used to compare single time points. The ET-1 concentration that significantly altered airways resistance and peak inspiratory pressure was determined by a one-way analysis of variance and Duncan's multiple range test. A P value of 0.05 was accepted as significant. Two-tailed P values were used, except in the case when ET_A inhibition was expected to result in an attenuated response, then the one-tailed P value was used.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Baseline bronchial artery pressure was 112 ± 4 mmHg for the 25 sheep studied. This pressure was obtained during perfusion at the control flow (18 ± 1 ml/min), which had been set based on sheep body weight (30 ± 1 kg). Individual baseline pressure and flow values resulted in a calculated average BVR (pressure/flow) of 6.3 ± 0.2 mmHg · ml¹ · min¹. Mean systemic arterial pressure for the group of sheep studied was 95 ± 3 mmHg. Peak inspiratory pressure was 16 ± 1 cmH₂O, and baseline airways resistance (n = 16) was 2.1 ± 0.2 cmH₂O · l¹ · s¹. The maximum change in BVR at 180 min resulting from the administration of TNF- during control (n = 6) or low flow (n = 5) perfusion of the bronchial vasculature did not differ from each other (P = 0.988); therefore, the results from all animals of this protocol were grouped together. Figure 1 demonstrates the average time course of BVR during and after TNF- administration. TNF- infused directly into the bronchial vasculature resulted in an early vasodilation. By 20 min into the infusion, BVR decreased significantly (P < 0.05) to 87% of baseline. However, the decrease in BVR reversed, and by 120 min, BVR was significantly increased by an average of 26% compared with the start of infusion (P < 0.05). The enhanced tone was maintained throughout the remainder of the experiment. PBS alone delivered at these low infusion rates had no effect on bronchial vascular tone. These observations can be compared with the control experiments (n = 3) in which BVR did not change over the time course of a 180-min time period. Systemic arterial pressure in sheep exposed to TNF- averaged 93 ± 4 mmHg and decreased progressively over time to 69 ± 5 mmHg at 180 min (P = 0.003). Initial peak inspiratory pressure and airways resistance averaged 17.1 ± 0.9 cmH₂O and 1.9 ± 0.2 cmH₂O · l¹ · s¹ (n = 6), respectively. Neither index of airway smooth muscle tone was altered by TNF- infusion throughout the 180-min time course of measurement (P > 0.6).

View larger version (13K): [in this window] [in a new window]   Fig. 1.   Changes in bronchial vascular resistance (BVR) over time in 11 sheep during and after tumor necrosis factor- (TNF-) infusion. Results are averaged for 20-min and 40-min infusions of TNF- (10 µg total dose). An early vasodilation was observed followed by a sustained constriction. *P < 0.05.

ET-1 responsiveness was confirmed in six additional sheep. Dose-response information is shown in Fig. 2A. As expected, a substantial increase in BVR was observed with increasing concentrations of ET-1. Furthermore, increases in BVR after the two doses of ET-1 that bracketed the level of vasoconstriction observed after TNF- are presented in Fig. 2B, before and after BQ-123 administration. A significant attenuation of the increase in BVR after 10⁸ M ET-1 and BQ-123 pretreatment was observed (P = 0.004). The change in BVR after the administered dose of BQ-123 and 10⁷ M ET-1 tended to decrease (P = 0.07). With regard to the two indexes of airway smooth muscle tone, only at a dose of 10⁶ M ET-1 was there a significant increase in peak inspiratory pressure (45 ± 3 cmH₂O) and airways resistance (6.7 ± 1.7 cmH₂O · l¹ · s¹; P = 0.01). The increase in airways resistance at this dose was significantly attenuated by BQ-123 pretreatment (4.8 ± 1.9 cmH₂O · l¹ · s¹; P = 0.04) but not the increase in inspiratory pressure (40 ± 7 cmH₂O; P = 0.27).

View larger version (8K): [in this window] [in a new window]   Fig. 2.   A: dose-response effects of endothelin-1 (ET-1) on BVR. Results are presented as the average change from baseline (means ± SE; n = 4 sheep). B: changes in BVR before (solid bars) and after BQ-123 (open bars) at the two doses of ET-1 that resulted in a constriction of a magnitude that bracketed the TNF- induced response. *P < 0.05.

In five additional sheep, the ET_A receptor antagonist BQ-123 (3 × 10⁶ M; 20 ml at 1 ml/min) was administered as an infusion directly into the bronchial artery. BVR was unaltered from baseline (both = 6.0 ± 0.2 mmHg · ml¹ · min¹) after BQ-123 administration (P = 0.96). TNF- was administered during control flow perfusion of the bronchial artery. The effects of TNF- on BVR after BQ-123 pretreatment are compared with the changes in BVR in the five sheep in which TNF- was delivered in the same manner (1 ml/min) under control flow conditions and are presented in Fig. 3A as a percentage of baseline values. The changes from baseline BVR at the two time points in which TNF- previously caused a significant increase in BVR are presented in Fig. 3B, as well as the changes observed after BQ-123 pretreatment and TNF- challenge. A significant reduction in the vasoconstriction was observed at both 120 min (P = 0.028) and 180 min (P = 0.014).

View larger version (10K): [in this window] [in a new window]   Fig. 3.   A: percentage of baseline BVR after 20-min infusion of TNF- (10 µg total dose) without (; n = 5 sheep) and with BQ-123 pretreatment (; n = 5 sheep). B: changes in BVR after TNF- infusion at final time points without (solid bars) and with BQ-123 pretreatment (open bars). *P < 0.05.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The results of this study demonstrate that recombinant human TNF- exerts a substantive modulating influence on the bronchial vasculature in sheep. An early vasodilation was observed, which reversed slowly over time, and by 120 min, bronchial vascular tone was significantly elevated over baseline. A number of studies have been reported that examined the effects of a range of TNF- doses on both regional and total systemic vascular responses. Overall, the total systemic response appears to be related to the balance of the predominant vasoactive factors released by TNF- treatment, which include ET-1, NO, and cyclooxygenase products (9, 14, 15, 20). Additionally, regional release and responsiveness to these factors appear to contribute to specific vascular beds displaying a predominance of vasodilation or vasoconstriction. When treated with TNF-, both splanchnic (21) and skeletal muscle arterioles (9) exhibited a decreased resistance, whereas the vasoconstrictor properties of this cytokine predominated for coronary (13), pial (19), and pulmonary circulations (14). In the present study, temporal differences in the bronchial vascular responsiveness to TNF- infusion were observed. Although both an early bronchial vasodilation and a late sustained vasoconstriction were observed, additional experiments were performed to focus on the mechanism by which TNF- exerted its vasoconstrictor response. The rationale for this choice was based on the overall objective of determining how TNF- may modify the airway vasculature from a pro-inflammatory perspective. An increased BVR during the constant arterial pressure conditions, which might typify the intact in vivo state, would result in a decrease in blood flow to the airway wall. Hogg and colleagues (18) demonstrated that leukocyte sequestration in the lung was significantly increased during low flow conditions. Given the importance of TNF- in coordinating inflammatory changes in the airway, the mechanism responsible for the vasoconstriction observed in the first series of experiments was pursued.

Marsden and Brenner (17) have demonstrated that in cultured bovine aortic endothelial cells, TNF- induced a time- and dose-dependent release of ET-1. The rate of increased secretion was maximal over 1-8 h and coincided with an increase in prepro-ET-1 mRNA transcript levels. Furthermore, Klemm and colleagues (13) demonstrated that in rats 15 min after TNF- infusion, there was in an increase in the circulating levels of ET-1. Coronary vasoconstriction after TNF- infusion in an isolated rat heart preparation was largely prevented by pretreatment with an ET_A receptor antagonist (13). Therefore, in the present study, ET-1 appeared to be a likely candidate for the secondary mediator responsible for the observed TNF--induced bronchial vascular constriction. Previous work by Barman and coworkers (1) showed that, in a canine preparation, a single 40-µg intravenous dose of ET-1 resulted in a 50% reduction in systemic blood flow to large airways. To put in context the available ET-1 in the airway wall, Mariassy and colleagues (16) evaluated the relative amounts of constitutively expressed ET-1 of bronchial epithelial cells compared with endothelium. They demonstrated that endothelial cells secreted seven times the amount of immunoreactive ET-1 compared with epithelium derived from the large bronchi of normal sheep (16). Thus the ET-1 dose-response data (Fig. 2A) confirms and extends the studies of Barman et al. (1). Furthermore, the finding that a relevant dose of an ET_A receptor antagonist could significantly attenuate the increase in BVR 180 min after the infusion of TNF- (Fig. 3) strongly suggests a role for ET-1 in the vasoconstrictor response.

TNF- infusion significantly decreased BVR during and early after the completion of infusion. Potent vasodilators, such as NO and cyclooxygenase products, have been shown to be released in response to TNF- infusion (9, 15, 20, 21). Although the mechanism responsible for bronchial vasodilation was not investigated, ET_A blockade tended toward enhanced vasodilation (Fig. 3A). Thus it is likely that, as in other vascular beds, it is the balance of vasoactive modulators released by cytokine stimulation that will determine the overall vascular response. A potential complicating factor related to the possible sheer-induced release of endothelial-derived substances such as NO influenced the study design. Because the degree of vasoconstriction observed in the initial control flow experiments might have been attenuated by the higher sheer stress of these imposed flow conditions, additional studies were performed with a reduced flow yet maintaining a constant TNF- concentration in the blood perfusate. However, this concern proved to be unwarranted since the degree of vasoconstriction at the 180-min time point did not differ between the control flow and low flow groups. Thus the data from all these studies were grouped together (Fig. 1).

Exogenously administered TNF- has been shown also to have a significant effect on airway smooth muscle. Wheeler and colleagues (28) showed that intravenous infusion of TNF- (10 µg/kg) acutely increased airways resistance in sheep. Furthermore, airways responsiveness to inhaled histamine was increased 6 h after intravenous TNF- administration. In another study, Kips et al. (12) demonstrated that baseline airways resistance was not altered after a 30-min exposure to aerosolized TNF- (1 µg/ml) in rats. However, airways responsiveness to intravenous 5-hydroxytryptamine was increased 90 min after the aerosol challenge (12). In the present study, direct delivery of a lower dose of TNF- to the airway smooth muscle via the bronchial circulation showed that no bronchoconstriction occurred. Although no tests of airway smooth muscle reactivity were performed, it is possible that the bronchial vascular constriction observed in this study at the later time points could contribute to airways hyperresponsiveness. The airway circulation has been shown to modulate agonist-induced tone by its ability to provide an essential clearance function (7, 25). Thus an increase in vascular resistance and decrease in airway perfusion might contribute to the enhanced responsiveness observed in the two previous studies. As shown by others (22) and as expected, ET-1 significantly increased airway smooth muscle tone and pretreatment with an ET_A receptor antagonist attenuated the increase in airways resistance (22).

The BVR was defined in a manner that differed from previous studies using this experimental model where it was calculated from the results of a complete pressure-flow relationship and estimation of the effective downstream pressure (26). Because of the potential for sheer-induced factors to influence bronchial vascular tone, vascular resistance in the present study was estimated in the simplest manner. Thus at constant flow, either control flow or 50% flow, BVR was defined as the inflow pressure/flow. It seems unlikely that an increase in the downstream pressure for the bronchial vasculature could be responsible for the increased bronchial artery pressure and increased BVR reported, since there were no obvious systemic effects of the low delivered dose of TNF-. Furthermore, although left atrial pressure was not measured in these experiments, Wheeler and colleagues (28) showed that it was not altered from baseline 3 h after an intravenous infusion of a much larger dose of TNF- (10 µg/kg) in sheep.

In summary, the results of this study demonstrate that the bronchial circulation is responsive to the direct, intravascular infusion of the cytokine TNF-. BVR decreased early after the infusion of TNF- but reversed and remained elevated 2 h after the start of infusion. Vasoconstriction could be prevented by administration of an ET_A receptor antagonist, thus suggesting a role for ET-1 in the observed vasoconstriction. Given the role for TNF- in leukocyte recruitment, changes in bronchial vascular dynamics may contribute to the overall inflammatory state of the airway.