Methods: We separated H-P complexes and non-heparin-protamine (non-H-P) complexes from heparinized defibrinated human plasma neutralized with protamine by chromatography and studied the changes in hemodynamics, airway pressure, and thromboxane B2 concentration after injection of H-P complexes or non-H-P complexes into seven goats. In addition, we studied whether these pulmonary responses were blocked in goats pretreated with cyclooxgenase inhibitor (Indomethacin, n = 5) or thromboxane synthetase inhibitor (OKY-046, n = 5).

Results: A very small dose of H-P complexes increased pulmonary arterial and peak airway pressures and was followed by thromboxane B2 release (from 12 [5.5-23] to 28 [16-44] mmHg, from 9.0 [7.5-15] to 12 [8-19] cmH2O, and from 0.85 [0.34-3.2] to 16.4 [1.4-39.3] ng *symbol* ml sup -1, respectively). On the other hand, animals that received non-H-P complexes showed no significant changes. Indomethacin totally blocked and OKY-046 partially blocked the increases in pulmonary arterial pressure and thromboxane B2 concentration.

Conclusions: H-P complexes play a major role in pulmonary hypertension after protamine reversal of heparin, and thromboxane A2 is a main mediator of the pulmonary hypertensive response to H-P complexes in goats.

Key words: Blood, anticoagulants: heparin. Blood, coagulation: heparin-protamine complexes; protamine. Lung(s), circulation: pulmonary hypertension. Metabolism, arachidonic acid: thromboxane B2.

PROTAMINE reversal of heparin-induced anticoagulation causes pulmonary and airway constrictive responses and thromboxane release in sheep and pigs. [1,2] The mechanism of this response has been suggested to be complex formation among heparin, a polyanion, and protamine, a polycation. The heparin-protamine (H-P) complexes activate complement and then initiate arachidonic acid metabolism, this leading to the generation of thromboxane A2 and, consequently, pulmonary vasoconstriction and bronchoconstriction. [3] 

However, to our knowledge, no one has determined whether H-P complexes are causative agents of pulmonary hypertension during protamine reversal of heparin. Several investigators have shown that H-P complexes, produced by mixing heparin and protamine in a syringe, caused pulmonary hypertension. [4,5] However, these H-P complexes were different from these formed by heparin reversal in vivo, because the interaction between heparin and protamine is affected by antithrombin III and other plasma proteins. [6] Thus, we separated H-P complexes that resembled those formed in vivo from heparinized defibrinated human plasma neutralized with protamine by chromatography according to the method of Shanberge et al., [6] and determined whether H-P complexes play a major role in pulmonary hypertension during protamine reversal of heparin. In addition, we confirmed that thromboxane was a main mediator of this pulmonary response to H-P complexes.

We prepared H-P complexes separated from heparinized defibrinated human plasma neutralized with protamine sulfate by gel filtration chromatography (Figure 1) according to the method described by Shanberge et al. [6] 

Buffers and Plasma. The reagent dilution buffer was 0.15 M NaCl, and the column elution buffer was 0.05 M tris (hydroxymethyl) aminomethane-HCl (Tris-HCl), 0.15 M NaCl, pH 7.4. Healthy human blood was collected from volunteers in polypropylene tubes containing 3.8% (0.38 g *symbol* 1 sup -1) sodium citrate 9:1 (volume: volume) and centrifuged for 20 min at 900g at 4 degrees Celsius. The supernatant plasma was drawn off the pellet with a polypropylene pipette.

Fractionation. Defibrinated plasma was obtained by incubating plasma for 10 min at 56 degrees Celsius and centrifuging twice for 10 min each time at 1,200 g at 4 degrees Celsius. Defibrinated plasma was heparinized with 25 IU porcine mucosal heparin-sodium (Novo Industry, Copenhagen, Denmark) and then neutralized with 0.166 mg protamine-sulfate (Shimizu, Shizuoka, Japan). The dosages of heparin and protamine were determined according to the method of Shanberge et al. [6] Sephadex G-200 (Pharmacia Fine Chemicals, Uppsala, Sweden) column was preequilibrated with 0.05 M Tris-HCl and 0.15 M NaCl at pH 7.4 and 4 degrees Celsius, and 5 ml heparinized defibrinated plasma neutralized with protamine was layered on the column and run. Fractions of 3 ml were collected.

Protein Determination. In each fraction, absorbance at 280 nm was determined using a spectrophotometer (UVIDEC-340, Nihon Bunko, Tokyo, Japan). H-P complexes were collected in the fraction with the first peak in absorbance, confirmed to be heparin-bound protamine complexes by the method of Shanberge et al. [6] We also defined other fractions, which included free heparin, free protamine, protein-bound heparin, protein-bound protamine, free protein and buffer solution as non-heparin-protamine (non-H-P) complexes.

This study was approved by the subcommittee on animal research at Akita University School of Medicine. Experiments were performed in 24 goats of either sex. Anesthesia was induced by 30 mg *symbol* kg sup -1 thiopental sodium. After tracheal intubation, the lungs were ventilated mechanically (Servo-Ventilator 900B, Siemens Elema, Solna, Sweden) using a mixture of Oxygen²and air to maintain arterial Oxygen²tension at about 150 mmHg. The tidal volume was set to maintain end-expiratory CO²at 4.5 vol%. We maintained light anesthesia using 0.3-0.5% halothane. These animals were paralyzed by administration of 4 mg pancuronium bromide.

We placed polyethylene catheters into the right femoral artery for injection of H-P or non-H-P complexes and for blood sampling. We also placed a flow-directed thermodilution pulmonary artery catheter (7-French, Baxter, Irvine, CA) into the pulmonary artery through the right jugular vein to measure pulmonary arterial pressure and cardiac output by the thermodilution method. Goats were placed in the supine position for the duration of the experiment.

Baseline measurements of hemodynamic parameters, airway pressure, and blood samples were obtained from all animals with stable hemodynamic conditions at 1 h after surgical preparations.

Pulmonary Responses to Heparin-Protamine or Non-Heparin-Protamine Complexes. Goats were divided into two groups. The H-P complexes group (n = 7, weight 17 [13-28] kg) were injected with H-P complexes (1.2 ml *symbol* kg sup -1) obtained from one time fractionation into the femoral artery over 2.5 min. We measured pulmonary arterial, femoral arterial, and airway pressure continuously using calibrated pressure transducers (Baxter) and recorded them on a multichannel recorder (Nihon Kohden, Tokyo, Japan). We also measured right atrial pressure and pulmonary capillary wedge pressure intermittently. Cardiac output was measured by the thermodilution method using a cardiac output computer (COM 1, Baxter); 3 ml normal saline solution at 0 degree Celsius was injected into the right atrium and the mean values of duplicate measurements were calculated. Systemic vascular resistance was calculated by subtracting right atrial pressure from mean systemic arterial pressure and dividing by cardiac output and pulmonary vascular resistance was calculated by subtracting pulmonary capillary wedge pressure from mean pulmonary arterial pressure and dividing by cardiac output. We sampled blood 10 min before injection of H-P complexes and immediately after pulmonary arterial pressure was maximal after injection of H-P complexes for measurement of systemic arterial thromboxane B2 concentration. The non-H-P complexes group (n = 7, weight 16 [10-21] kg) were injected with non-H-P complexes (1.2 ml *symbol* kg sup -1) into the femoral artery over 2.5 min. Blood samples for assay of systemic arterial thromboxane B2 concentration were drawn at 10 min before and 4 min after injection. The measured parameters were the same as those in the H-P complexes group.

Effects of a Cyclooxygenase Inhibitor (Indomethacin) or Selective Thromboxane Synthetase Inhibitor (OKY-046, sodium (E)-3-[p-(1H-imidazol-1-yl-methyl)phenyl]-2-propenoate) on Pulmonary Responses to Heparin-Protamine Complexes. Goats were divided into three groups. We used the data from the H-P complexes group in the experimental protocol described in the section above as those of the control (no pretreatment) group. The indomethacin group (n = 5, weight 28 [14.5-32] kg) were infused with 5 mg *symbol* kg sup -1 indomethacin (Wako Pure Chemical Industries, Osaka, Japan) intravenously over 20 min before injection of H-P complexes (1.2 ml *symbol* kg sup -1). The OKY-046 group (n = 5, weight 12 [11.5-20] kg) were injected with 10 mg *symbol* kg sup -1 of OKY-046 (Ono Pharmaceutical, Osaka, Japan) intravenously and then infused at a rate of 100 micro gram *symbol* kg sup -1 *symbol* min sup -1 over 20 min before injection of H-P complexes (1.2 ml *symbol* kg sup -1). In both pretreated groups, the experimental protocol was, thereafter, the same as in the non-H-P complexes group in the experimental protocol described in the section above.

For determination of arterial concentrations of plasma thromboxane B2, the stable metabolite of thromboxane A2, 10 ml arterial blood was collected in glass test tubes containing ethylenediamine tetraacetic acid and 125 micro gram indomethacin and immediately centrifuged at 1,200g for 10 min at 4 degrees Celsius. The plasma was then aspirated and stored in polypropylene tubes at -70 degrees Celsius. We measured thromboxane B2 by tracing compounds and antisera according to standard radioimmunoassay procedures. [7] The lower limits of sensitivity of assays are < 1 pg/tube, and cross-reactivities of the antisera with major circulating prostanoids were < 1%.

Experimental data are presented as median and range. Values were compared between groups by the Mann-Whitney U test to indicate groups with significant differences when P < 0.05. Wilcoxon's U test was used to test for significant differences between values from two groups, and a P value less than 0.05 was accepted as significant. The time course data of pulmonary arterial pressure and airway pressure changes were compared between groups by one- or two-way analyses of variance, followed by Scheffe's test to indicate groups with significant differences when P < 0.05. Student's t test was used to test for significant differences between values from two groups, and a P value less than 0.05 was accepted as significant.

Injection of H-P complexes caused a significant increase in pulmonary arterial pressure and end inspiratory airway pressure. Recovery was complete within 30 min. Pulmonary arterial pressure increased by approximately 130%, 5 (2-6) min after injection of H-P complexes and pulmonary vascular resistance at the peak of pulmonary arterial pressure increased to a value about 10-fold greater than the baseline. Cardiac output was significantly decreased by 38% compared with baseline. Plasma thromboxane B2 concentrations increased significantly approximately 19-fold after injection of H-P complexes. Systemic arterial, right atrial and wedge pressures did not change significantly. Injection of non-H-P complexes caused little changes in pulmonary arterial, systemic arterial and end inspiratory airway pressures. Plasma thromboxane B2 concentrations were also unchanged (Table 1and Figure 2).

In the indomethacin group, all of the effects of H-P complexes were blocked. In the OKY-046 group, the increases in pulmonary and femoral arterial pressure, systolic vascular resistance, and plasma thromboxane B2 concentration were attenuated significantly compared with the control group. The decrease in cardiac output were also attenuated significantly, but the other measured variables were unchanged (Table 2).

Our study demonstrates that the H-P complexes alone cause thromboxane release followed by pulmonary arterial hypertension and that other moieties that form upon protamine administration do not in the goat. Pulmonary arterial pressure increased markedly, by more than double; this increase, combined with a transient decrease in cardiac output, caused the pulmonary vascular resistance to increase to ten times the control level after injection of H-P complexes. However, non-H-P complexes had no such effects. In addition, we also clarified directly that thromboxane plays a major role in the pulmonary hypertensive responses, because cyclooxygenase inhibitors totally blocked, and selective thromboxane synthetase inhibitors partially blocked these responses. The results may indicate participation by other yet unidentified mediators of pulmonary hypertension. Previous investigations have also shown that thromboxane A2 is the main mediator of pulmonary hypertensive responses to protamine reversal of heparin in pigs and sheep. [8-10] 

Several studies have previously demonstrated that the injection of H-P complexes formed in vitro causes pulmonary arterial hypertension and increases in thromboxane concentrations in pigs and isolated cat lungs. [4,5] However, the H-P complexes used in previous studies were formed by mixing heparin and protamine without human plasma in a syringe. Thus, these H-P complexes would be different from those formed in vivo. Anti-thrombin III and other plasma proteins mediate the neutralization of heparin by protamine. [6] Thus, these previous studies have not presented clear evidence that H-P complexes play a major role in pulmonary hypertensive responses after protamine reversal of heparin in vivo. The H-P complexes separated from heparinized defibrinated human plasma neutralized with protamine sulfate by gel filtration used in this study, however, were very similar to those formed in vivo after protamine reversal of heparin. [6] 

Small quantity of human albumin will co-elute with H-P complexes. Injecting human albumin into goat may produce transient pulmonary hypertension. But the fraction of non-H-P complexes contains far more quantity of human albumin than the fraction of H-P complexes, because the peak in absorbance at 280 nm at the fraction of non-H-P complexes is higher than at the fraction of H-P complexes. If human albumin in the fraction of H-P complexes would produce pulmonary hypertension, non-H-P complexes also would produce pulmonary hypertension. However, non-H-P complexes did not produce pulmonary hypertension. Thus, we believe human albumin in the fraction of H-P complexes do not affect pulmonary response.

The quantity of H-P complexes required to cause the pulmonary hypertensive response has not yet been determined. We injected a part of solution obtained from one time fractionation. So, the maximal amount of H-P complexes that we injected was less than 25 IU of heparin and 0.166 mg protamine, even if all of the heparin and protamine changed into H-P complexes in vitro. We previously studied pulmonary hypertensive responses after injection of 2 mg *symbol* kg sup -1 of protamine followed by 2 mg *symbol* kg sup -1 (200 IU *symbol* kg sup -1) of heparin in goats. [11] In the current study, if we had administered these agents according to our previous protocol we would have used 37.2 mg protamine followed by 3,720 IU heparin, because the average weight of the goats was 18.6 kg. The amounts of H-P complexes injected in the current study were thus much less than those that would have been formed in vivo during protamine reversal of heparin in our previous study. However, pulmonary hypertensive responses were almost the same between our two studies (13.7 plus/minus 5.36 mmHg increase in the previous study, and 14 (8-28) mmHg increase in the current study). Morel et al. reported that pulmonary hypertension after protamine reversal depended on the amount of circulating H-P complexes. [12] Thus, our results suggest that amounts of circulating H-P complexes during protamine reversal of heparin anticoagulation were very low and that most of the H-P complexes might be bound to the vascular wall. In fact, both heparin and protamine are apt to bind to the surface of vascular endothelial cells. [13,14] Further study concerning this point is therefore required.

We conclude that H-P complexes play a major role in pulmonary hypertension after protamine reversal of heparin in the goats, and that thromboxane A2 is the main mediator of pulmonary hypertensive responses to H-P complexes in vivo.

The authors are grateful to Yoshitugu Tobe for his technical assistance in animal preparations and to Yoko Masaki for separation of heparin-protamine complexes.