INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Keywords: airway hyperreactivity; dry air challenge; eicosanoid mediator; heparin; hyperventilation-induced bronchoconstriction; peripheral airway resistance

Hyperventilation with cold, dry air causes airway narrowing in most asthmatic and in some normal individuals (1). This transient hyperventilation-induced bronchoconstriction (HIB) peaks between 2 and 10 min after hyperventilation or exercise ceases, and ends spontaneously 30 to 60 min later (4). A delayed or late-phase airway obstruction occurs in some asthmatic subjects from 3 to 13 h after exercise (7), although the underlying mechanism for its occurrence is controversial (13, 14). Detection of neutrophil chemotactic activity in plasma (9) and enhanced eosinophilic infiltration in bronchoalveolar lavage fluid (BALF) (6) suggests that an inflammatory process may contribute to an exercise-associated late-phase response in asthmatic subjects. Crimi and coworkers (6) concluded that fluctuations in lung function associated with airway inflammation, and not exercise per se, accounted for the delayed bronchoconstriction observed in their study. However, if hyperventilation with cold, dry air causes airway inflammation, then the hyperpnea that is intrinsically linked to exercise may be a stimulus resulting in delayed airway obstruction.

Late-phase peripheral airway obstruction and inflammation occurs reproducibly in a canine model of exercise-induced asthma (15). In this model, a bronchoscope is used to directly hyperventilate peripheral airways with dry gas at room temperature, bypassing the central airways. This dry air challenge (DAC) simulates the airway exposure that occurs during strenuous exercise in human subjects breathing subfreezing air, and results in similar thermal conditions in airways of similar size (16). The acute response to DAC in asthmatic human (17, 18) and canine peripheral airways (16) is well documented. The late-phase response occurs hours later, and coincides with granulocyte infiltration (15). It is unknown whether the bronchoconstrictive and proinflammatory mediators released during the acute response (19) contribute to development of the late-phase response.

Inhalation of heparin inhibits HIB in asthmatic human (22- 24) and in canine subjects (25). We previously reported that heparin reduced HIB in our canine model by ~ 50% to 60% (25). BALF mediator analyses revealed that inhaled heparin inhibited eicosanoid mediator production and release caused by DAC. We also reported that inhalation of heparin acutely increased leukocyte infiltration into airways, which is consistent with the chemoattractant activity previously reported for this drug (26, 27). In contrast to its acute activity, heparin inhibits antigen-induced late-phase cell infiltration at 24 h after an airway challenge in several animal models (28, 29). Although these observations were based on the study of different animal models, they suggest that the inflammatory/antiinflammatory activity of heparin is a time-dependent phenomenon.

The purpose of the present study was to determine whether inhaled heparin prevented hyperventilation-induced late-phase airway obstruction in dogs. Because heparin either attenuated or abolished acute hyperventilation-induced mediator release in our canine model (25), we hypothesized that heparin inhibits the development of the delayed response by inhibiting eicosanoid metabolism. In testing this hypothesis, we examined and compared hyperventilation-induced late-phase changes in peripheral airway function, mediator production, and leukocyte infiltration in vehicle- and heparin-treated sublobar airways.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Experimental Technique

Dogs were handled and maintained in accordance with the Policy and Procedures Manual published by the Animal Care and Use Committee of the Johns Hopkins University School of Public Health.

Anesthesia and instrumentation.Colony-bred male mongrel dogs (21.1 ± 1.2 [mean ± SE] kg, n = 6) were anesthetized with sodium thiopental (25 mg/kg, intravenously). Fentanyl citrate (25 µg, intravenously) was administered every 15 min, and thiopental (50 mg) was given supplementally as needed to maintain anesthesia. Depth of anesthesia was assessed from the canthal reflex, heart rate (HR), blood pressure, presence of spontaneous movements, and breathing. Dogs were intubated and mechanically ventilated (17 ml/kg) with room air; CO₂ was continuously monitored (Model LB-2; Beckman Inc., Anaheim, CA) and maintained at about 4.5% by adjusting respirator frequency. HR and mean arterial pressure (MAP) were recorded with a noninvasive blood pressure and HR monitor (Datascope Accutorr IA; Datascope Corp., Paramus, NJ). Body temperature was monitored with a rectal probe (YSI 402) and telethermometer (Yellow Springs Instrument Co., Yellow Springs, OH) and maintained with a warming pad.

Measurement of peripheral airway resistance.Two fiberoptic bronchoscopes (5.5 mm OD; Olympus BF Type P 10; Olympus Corporation of America, New Hyde Park, NY) were inserted through two airtight portals in the endotracheal tube used for ventilation and were wedged in left and right middle sublobar bronchi. Sublobar airway pressure (P_b) was measured via a polyethylene 90 (PE₉₀) catheter threaded through the suction port of the bronchoscope and connected to a pressure transducer (Statham; Gould Inc., Oxnard, CA). Compressed, dry 5% CO₂ in air at room temperature was delivered around the catheter and into the wedged sublobar segment at 200 ml/min. P_b was measured by stopping the ventilator during exhalation and allowing the unobstructed areas of the lung to equilibrate with atmospheric pressure at FRC. Under this condition, P_b decays to a plateau pressure greater than the atmospheric pressure surrounding the unobstructed lung, and RP = P_b/200ml/min = P_b/3.33 ml/s.

Airflow challenge.For DAC, bronchoconstriction was induced by temporarily increasing the flow rate of 5% CO₂ in dry air from 200 to 2,000 ml/min for 5 min.

Measurement of peripheral airway reactivity.Histamine diphosphate (25 µg/ml) was prepared daily for each experiment. The histamine solution was aerosolized in dry air with 5% CO₂, and was delivered at 200 ml/min for 15 s into a wedged sublobar segment via the bronchoscope. The PE₉₀ catheter was temporarily removed from the bronchoscope during the aerosol histamine challenge. Because intraairway mucus may act as a barrier and blunt airway reactivity to aerosolized agonists (30), histamine challenge (25 µg/ml, intravenously) was also done by injecting a 10 ml bolus into the cephalic vein. The postchallenge change in R_p (R_p) was recorded at 0.5, 2, 5, 10, and 15 min after either aerosol or intravenous administration of histamine. The suspicion of a barrier effect of intraairway mucus was subsequently confirmed in Protocol B, in which airway reactivity to intravenous histamine challenge on Day 5 was increased by ~ 2.5-fold over the reactivity on Day 1.

Bronchoalveolar lavage, differential cell counts, and mediator analyses.Three 20 ml aliquots of Hanks' buffered salt solution (37° C) were infused through the bronchoscope into a wedged sublobar bronchus, and each aliquot was recovered via gentle aspiration. The recovered BALF aliquots were pooled and stored at 4° C until the end of the experiment. At that time, BALF samples were centrifuged at 4° C for 10 min at 1,300 rpm. Ten microliters of supernatant were placed on a hemocytometer to determine total cell number. Differential cell counts were done on cytocentrifuged BALF samples prepared with a modified Wright-Giemsa stain. Trypan blue exclusion was used to evaluate cell viability.

BALF was concentrated with a Sep-Pak C₁₈ cartridge (Waters Associates, Milford, MA), eluted in 4 ml of methanol, and stored at 70° C. Aliquots of the eluate were analyzed as previously described (20), using commercially available enzyme-linked immunosorbent assay (ELISA) kits for prostaglandin (PG)D₂ (detection Limit: 3.2 pg/ml; Cayman Chemical Co., Ann Arbor, MI) PGE₂ (detection Limit: 14.9 pg/ml; Cayman), and leukotriene (LT)C₄, LTD₄, and LTE₄ (detection Limit: 0.2 ng/ml; Neogen Corp., Lexington, KY).

Experimental Protocol

Effect of heparin on the late-phase changes in baseline RP, BALF cell profiles, and mediator release resulting from DAC.A bronchoscope was wedged in a sublobar segment of the cardiac lobe (CL) in each dog, and baseline RP was measured. BW (vehicle control) was then aerosolized for 60 s into the CL (n = 6). The bronchoscope was then removed and two bronchoscopes were simultaneously wedged into the left middle lobe (LML) and the right middle lobe (RML), respectively. A map of the branching pattern of the airways from the origin of the specific lobar bronchus to the wedge position was recorded to allow identical bronchoscope placement at a later time. After recording of baseline RP, heparin was aerosolized into one sublobar location, and the other location was treated with BW. DAC was done in each segment at 60 min after aerosol treatment. RP was recorded at 0.5, 2, 5, 10, and 15 min after DAC. Each dog was reanesthetized 5 h after DAC. RP in each sublobar region was then measured again, after which each region was lavaged. Two weeks later the protocol was repeated, with the aerosol treatment reversed for the middle lobes (n = 12 lobes; six dogs).

Effect of heparin on Histamine-induced bronchoconstriction.Two weeks after the previous study, the airway map was used to locate and wedge bronchoscopes into the same sublobar airways. After recording of baseline RP, we pretreated the LML and RML with either heparin or BW. Sixty min later, we measured airway reactivity (RP) to aerosolized histamine in both lobes. After baseline RP was reestablished, we recorded RP for intravenous histamine at the same times after challenge. After reestablishing baseline RP, we repeated the DAC in each location and recorded the postchallenge RP in a similar fashion. Dogs were reanesthetized 5 h after DAC. We then measured baseline RP again in both lung locations, and assessed airway reactivity sequentially, using an aerosol and then an intravenous histamine challenge (n = 6).

Statistical Analyses

Data for RP and RP were analyzed with Friedman's repeated measures analysis of variance on ranks, in conjunction with the Student-Newman-Keuls test for comparisons of individual treatment means. Before comparing data for cells and mediators recovered in BALF from hyperventilated sublobar airways, we used Wilcoxon's signed ranks test to confirm the absence of significant within-treatment differences in the data collected during Weeks 1 and 3 of the study. We used the Mann-Whitney rank sum test to compare cell and mediator from DAC for airways pretreated with either BW or heparin to vehicle control airways pretreated with BW. Spearman's rank correlation analysis (r_s) was used to determine whether the concentration of biochemical mediators in BALF was correlated with the magnitude of RP. Statistical significance was established at p < 0.05. Data are expressed as mean ± SE.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Effect of Heparin on HIB

Baseline RP was similar (p = 0.818) before treatment with either BW or heparin, at 0.67 ± 0.09 cm H₂O/ml/s and 0.70 ± 0.08 cm H₂O/ml/s (n = 12), respectively. Sixty minutes after heparin treatment, the baseline RP was 0.77 ± 0.09 cm H₂O/ml/s (n = 12). At that time, DAC increased RP by only ~ 41%, as compared with a 99% increase in airways that had been treated with BW (Figure 1). Hyperventilation-induced changes in RP after heparin treatment were significantly reduced at 2 and 5 min after DAC as compared with the changes at the same intervals after the vehicle control. Five hours later, baseline RP in heparin-treated/DAC (heparin/DAC) airways increased by ~ 22% (0.15 ± 0.04 cm H₂O/ml/s), which was not significantly greater (p = 0.631) than the initial baseline value of RP. In contrast, baseline RP in BW-treated/DAC (BW/DAC) airways increased by ~ 123% (0.63 ± 0.13 cm H₂O/ml/s), and was significantly increased (p < 0.05) as compared with its initial baseline value (Figure 1). This increase in baseline RP in BW-treated airways was significantly greater than that recorded in the same airways after pretreatment with heparin (p = 0.0001). Neither aerosol nor intravenous histamine significantly affected HR, MAP, or body temperature (Table 1).

View larger version (19K): [in this window] [in a new window]   Figure 1.   (A) Acute and delayed peripheral airway reactivity (RP) in response to DAC (2,000 ml/5 min) in airways pretreated with heparin (closed circles: 1,000 U/airway), or BW (open circles). A wedged airway pretreated with BW served as an unchallenged control (open squares). Data are expressed as the mean ± SE (n = 6). * p < 0.05 compared with baseline, p < 0.05 compared with heparin.

Effect of Heparin on BALF Cell Profiles and Mediator Release at 5 h after DAC

The volumes of BALF recovered from wedged control, BW/ DAC, and heparin/DAC sublobar segments were 43 ± 3.2 [mean ± SE] ml, 39 ± 2.0 ml, and 40 ± 1.8 ml, respectively. BALF recovered from airways subjected to BW/DAC and heparin/DAC tended to contain a greater (p = 0.052) total number of cells than that recovered from the wedged control (Figure 2). Macrophages and lymphocytes were not affected either by DAC or by treatment with heparin or BW. In contrast, neutrophils (p = 0.008), eosinophils (p = 0.029), and epithelial cells (p = 0.002) were all increased at 5 h after DAC as compared with the numbers of these cells in unchallenged control airways. Only eosinophil infiltration was inhibited by pretreatment with heparin (Figure 2; p = 0.02, n = 12).

View larger version (20K): [in this window] [in a new window]   Figure 2.   BALF cells recovered from wedged control airways pretreated with BW (open bars, n = 6) or from airways pretreated with either BW (closed bars, n = 12 from six dogs) or 1,000 U of heparin (crosshatched bars, n = 12 from six dogs) and subjected to DAC (2,000 ml/5 min). Bronchoalveolar lavage was done either 5 h after the bronchoscope was wedged (control) or 5 h after DAC ended. Mac = macrophages, Lym = lymphocytes, PMN = neutrophils, Eos = eosinophils, Epi = epithelial cells, Total = total nucleated cells. Data are expressed as mean ± SE. * p < 0.05.

Effect of Heparin on Hyperventilation-Induced Eicosanoid Mediator Release at 5 h after DAC

The collective concentration of LTC₄, LTD₄, and LTE₄ recovered in BALF from BW/DAC airways was markedly increased (p = 0.022) at 5 h after DAC as compared with the concentration for unchallenged control airways (Figure 3). Although the collective concentration of LTC₄, LTD₄, and LTE₄ recovered from heparin-treated airways at 5 h after DAC was increased (p = 0.002) over the respective control concentration, the increase was significantly less (p = 0.03) than that for BW/DAC airways (Figure 3). Pretreatment with aerosolized heparin abolished the hyperventilation-induced increase in PGD₂ (p = 0.03) as compared with the PGD₂ concentration in BW/DAC airways. In contrast, there was a tendency for the concentration of PGE₂ to increase in heparin-treated airways as compared with either BW/DAC (p = 0.06) or wedged control (p = 0.132, n = 6) airways.

View larger version (20K): [in this window] [in a new window]   Figure 3.   Eicosanoid mediators recovered from wedged control airways pretreated with BW (open bars, n = 6) or from airways pretreated with either BW (closed bars, n = 12 from six dogs) or 1,000 U of heparin (crosshatched bars, n = 12 from six dogs) and subjected to DAC (2,000 ml/5 min). Bronchoalveolar lavage was done either 5 h after the bronchosope was wedged (control) or 5 h after DAC ended. LTC4-E4 = LTC3, LTD4, and LTE4; PGD2 = prostaglandin D2; and PGE2 = prostaglandin E2. Data are expressed as mean ± SE. * p < 0.05.

Correlation between BALF cells and Mediators and HIB

Spearman's rank correlation analysis of pooled data (n = 30) revealed correlations between the percent increase in baseline RP that occurred between 0 h and 5 h after DAC and the concentrations of neutrophils (Figure 4A; r_s = 0.345, p = 0.062), eosinophils (Figure 4B; r_s = 0.384, p = 0.036), LTC₄, LTD₄, and LTE₄ (Figure 4C: r_s = 0.519, p = 0.003), and PGD₂ (Figure 4D: r_s = 0.423, p = 0.020) in BALF recovered at the latter time. Similar results were found when the percent change in baseline RP was correlated with either total granulocyte infiltration (i.e., numbers of neutrophils + eosinophils: r_s = 0.369, p = 0.045) or with the combined bronchoconstrictor mediators (i.e., LTC₄, LTD₄, and LTE₄ + PGD₂: r_s = 0.548, p = 0.002). Correlating the percent change in baseline RP with the ratio of bronchoconstrictive to bronchoprotective mediators (i.e., LTC₄, LTD₄, and LTE₄ + PGD₂/PGE₂) (31) resulted in an r_s value of 0.479 (p = 0.008). Correlations were also found between neutrophils and PGD₂ (r_s = 0.446, p = 0.014), and between eosinophils and PGD₂ (r_s = 0.446, p = 0.014) and LTC₄, LTD₄, and LTE₄ (r_s = 0.640, p < 0.001).

View larger version (18K): [in this window] [in a new window]   Figure 4.   Correlations between the increase in baseline RP between 0 h and 5 h after DAC and the concentrations of (A) neutrophils, (B) eosin-ophils, (C ) LTC4, LTD4, and LTE4, and (D) PGD2 in BALF recovered at 5 h. Open squares: wedged control airways pretreated with bacteriostatic water (n = 6); Open circles: airways pretreated with bacteriostatic water (n = 12 from six dogs) and subjected to DAC; closed circles: airways pretreated with 1,000 U of heparin (n = 12 from six dogs) and subjected to DAC.

Effect of Heparin on Hyperventilation-Induced AHR

Pretreatment with either aerosolized heparin or BW did not affect (p > 0.05) histamine-induced bronchoconstriction (Figure 5). In airways treated with BW, reactivity to histamine was enhanced (p = 0.001) at 5 h after DAC (Figure 5A). In contrast, RP to histamine did not change (p > 0.05) in airways treated with heparin (Figure 5B). The initial response to intravenous histamine challenge was similar to that produced by aerosol challenge (p > 0.05). However, airway reactivity to intravenous histamine challenge in BW-treated airways was markedly enhanced when compared either with the response to aerosolized histamine (p = 0.04) or with the response to intravenous histamine in heparin-treated airways (p < 0.001) (Figure 5).

View larger version (18K): [in this window] [in a new window]   Figure 5.   Peripheral airway reactivity (RP) to either aerosol (A and B) or intravenous (C and D) histamine challenge before (open circles) and 5 h after (closed circles) hyperventilation with dry air in bronchi pretreated with either bacteriostatic water (A and C) or heparin (B and D). Vertical bar = histamine challenge. Data are expressed as mean ± SE. * p < 0.05 compared with baseline, p < 0.05 compared with heparin; n = 6.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Peripheral airway hyperventilation in dogs damages the bronchial mucosa (32, 33), stimulates mast cell degranulation (32, 33), and results in the release of biochemical mediators (21). Bronchial epithelial cells and mast cells are capable of producing a variety of chemotactic factors, including eotaxin (34), interleukin (IL)-5 (35), IL-8 (36), and LTB₄ (37). Theoretically, these chemotactic mediators may be responsible for the influx of granulocytes that typically occurs within 5 h after canine peripheral airways are challenged with dry air (15, 33). These inflammatory mediators may also contribute directly to the late-phase airway obstruction (Figure 1) and bronchial hyperreactivity (Figure 5) that characterize our canine model.

Pretreatment with inhaled heparin reduced hyperventilation-induced eicosanoid mediator release (Figures 3 and 4), eosinophil infiltration (Figures 2 and 4), and peripheral airways obstruction (Figure 1), and abolishes the airway hyperreactivity (Figure 5) that accompanies these late-phase events. The finding that heparin inhaled at 1 h before DAC inhibited the acute and late-phase responses by ~ 40% and 50% (median = 80%), respectively (Figure 1), suggests that heparin either inhibits the initiation of early-phase pathways that are necessary for the development of late airways obstruction, or interferes with pathways initiated at both times. Garrigo and colleagues (22) found that heparin inhaled either at 1 h or 3 h (but not at 6 h) before exercise inhibited bronchoconstriction in asthmatic subjects. Diamant and coworkers reported that multiple doses of unfractionated heparin tended to reduce the early response and significantly reduced the late response to inhaled antigen in asthmatic subjects, but did not reduce airway reactivity to histamine (38). In contrast, Ahmed and associates (39) reported that only ultralow-molecular-weight (ULMW) heparin attenuated antigen-induced acute and late-phase responses and hyperrreactivity in sheep. A study of dose, heparin plasma activity, and the terminal half-life time of intravenously administered unfractionated heparin in dogs (40), found that heparin activity decreases by 75% at 1 h, 94% at 2 h, and 99.6% at 4 h after infusion. These data therefore suggest that pretreatment with heparin is more likely to affect the early to middle stages of the cascade that produces late-phase changes in peripheral airway function, rather than to affect critical late-phase events. However, this scenario does not preclude the possibility that heparin may localize on endothelial cells at concentrations sufficient to influence cell movement and activation, and may do so throughout the entire posthyperventilation period.

Heparin exhibits numerous nonanticoagulant properties; it inhibits complement activation (41), mast-cell degranulation (42), histamine release (43), L- and P-selectin (44), neutrophil chemotaxis (45), late-phase infiltration of granulocytes (28, 29), and T-cell function (46). Potential modes of action for heparin include electrostatic neutralization of polycationic mediators (47), blockade of inositol triphosphate-mediated calcium release (48, 49), and interference with interactions between very late-acting antigen (29) and either vascular cell adhesion molecule (50, 51) or fibronectin (51, 52). Although heparin may inhibit cell infiltration by binding a variety of proteins directly involved in the recruitment of granulocytes (53), it can also inhibit specific postreceptor second messenger signals, such as the late-phase activation of mitogen activated protein kinase (MAPK) (56). Pharmacologic interference with p38 MAPK inhibits interleukin (IL)-5 synthesis (57) and IL-8 protein and messenger RNA expression (58), suggesting that this signaling pathway can modulate late-phase cell infiltration (57).

We previously reported that inhalation of heparin resulted in acute infiltration of leukocytes into canine peripheral airways (25), a phenomenon consistent with the finding in other studies that heparin can exhibit chemoattractant activity for monocytes and neutrophils (26, 27). We believe that this is the first time that the dual nature (i.e., the time-dependent inflammatory/antiinflammatory activity) of heparin has been reported in the same animal model. The latter antiinflammatory activity may reflect the ability of heparin to strongly bind cytokines that are chemotactic for granulocytes, including IL-5 and IL-8 (53, 54). However, heparin reduces hyperventilation-induced granulocytic inflammation in dogs, primarily by inhibiting eosinophil influx (Figures 2 and 4). These data confirm our previous observation (25) that heparin is a relatively specific antagonist of eosinophil trafficking stimulated by a nonantigenic agent in the canine lung. Although heparin may interfere with hyperventilation-associated IL-5, but not with IL-8, activity in canine peripheral airways, it seems more likely that some other mechanism of action (such as the polyanionic nature of heparin) accounts for this antieosinophil activity.

Preliminary experiments using pharmacologic interventions have suggested that hyperventilation-induced late-phase airway obstruction depends on cyclooxygenase and lipoxygenase activity during or immediately after DAC (59). Our findings in the present study, of reduced levels of LTC₄, LTD₄, LTE₄, and PGD₂ in BALF recovered from airways pretreated with heparin and subjected to DAC are consistent with this concept (Figure 3). The tendency for PGE₂ to increase (p = 0.06) after DAC in heparin-treated airways may also account for some of the inhibitory efficacy of heparin. In fact, both heparin and PGE₂ inhibit DNA synthesis in human airway smooth-muscle cells (60), and this potential inhibition of mediator synthesis is consistent with the concept described earlier. In conjunction with the previously described pharmacokinetics of heparin in dogs (40), and with our previous work showing that heparin inhibits the release of eicosanoids immediately after DAC (25), our current results suggest that inhaled heparin interferes with development of the late-phase response by inhibiting the eicosanoid mediator production and release that normally occur during the acute response to DAC.

Inhaled heparin abolished peripheral airway hyperreactivity to histamine (Figure 5), regardless of the route of administration of histamine. When compared with airway reactivity before DAC, airway reactivity to aerosolized and intravenous histamine at 5 h after DAC increased by 0.47 cm H₂O/ml/s (Figure 5a) and by 1.2 cm H₂O/ml/s (Figure 5c), respectively. These values conceal an average sevenfold increase in RP (SE = 3.8, range: 0.7 to 25.4 cm H₂O/ml/s; median = 3.66 cm H₂O/ml/s; n = 6) when intravenous is compared with aerosol histamine challenge, and suggest that intravenous challenge is a more sensitive test for detecting changes in peripheral airways reactivity. This difference in sensitivity may reflect intraairway mucus and its accumulation during the late phase, which may blunt airway reactivity to aerosolized agonists, probably via a barrier effect (30). Alternatively, the pattern of distribution associated with different routes of drug administration could result in disparate airway responses. Although aerosol and intravenous delivery of methacholine (MCh) produced similar changes in whole-lung resistance in rats, intravenous administration of MCh caused greater airway constriction than did aerosol delivery of Mch (61). However, this effect decreased with decreasing airway size (61), raising questions about the relevance of this observation to our peripheral airway model.

In conclusion, inhaled heparin inhibits late-phase peripheral airway obstruction and hyperreactivity caused by hyperventilation with dry air. Although the exact mechanism of the antiasthmatic activity of heparin is unknown, its efficacy may result directly from inhibition either of the eosinophil infiltration or of the eicosanoid mediator release that accompanies late-phase changes in peripheral airway function. The latter effect may represent the unknown antiinflammatory mechanism referred to by Ahmed and coworkers (39), and may in part be responsible for the inhibitory activity of ULMW heparin. In contrast to sheep, in which only ULMW heparin inhibits antigen-induced late-phase AHR (39), in canine peripheral airways, unfractionated heparin completely abolishes the late-phase response to hyperventilation with dry air. Similar but less impressive effects have been reported in antigen challenged asthmatic subjects treated with multiple doses of unfractionated heparin (38), suggesting that airway sensitivity, airway location, and dosing may account for these apparent interspecies differences.