Platelet Factor 4 Injection Produces Acute Pulmonary Hypertension in the Awake Lamb 

Seventeen Suffolk lambs weighing 25–35 kg were used for our studies. The studies were approved and all lambs were cared for in accordance with the rules of the Massachusetts General Hospital Subcommittee on Research Animal Care. Using sterile techniques, an 8-Fr introducer was placed percutaneously into the jugular vein of each lamb after 1% xylocaine infiltration anesthesia. The lamb was then anesthetized via mask with 1–3% halothane (Halocarbon Laboratories, Hackensack, NJ) in Oxygen²and the trachea was intubated. The right femoral artery was cannulated and the catheter tip was advanced to the midthoracic aorta to sample blood and monitor systemic blood pressure (Psa). After awakening, a sterile 7-Fr flow-directed thermodilution catheter (American Edwards model 93A-131H-7F, Santa Ana, CA) was positioned with the tip in the pulmonary artery to measure mean pulmonary artery pressure (Ppa), balloon occlusion pressure (Pw), central venous pressure (Pv), thermodilution cardiac output (CO), and core blood temperature (T). The jugular venous catheter was used for fluid and drug administration. During the study, the lambs were housed in a Babraham veterinary cage with free access to food and water.

Lambs were not studied if they exhibited one or more of the following baseline criteria indicating preexisting illness: 1) pulmonary artery temperature > 40.0 degrees Celsius, 2) arterial WBC concentration > 12,000 or < 3,000/mm³, or 3) mean Ppa > 20 mmHg.

The Psa, Pv, Ppa, and Pw were measured and recorded continuously using calibrated Hewlett-Packard model 1280C pressure transducers and a Hewlett-Packard model 7754B recorder (Hewlett-Packard, Palo Alto, CA). Transducers were zeroed at the level of the right atrium and mean values measured at end expiration. Cardiac output was determined by thermodilution (Edwards Laboratories model 9520A cardiac output computer, Irvine, CA) using an average of three right atrial injections of 5 ml of O degrees Celsius saline at each sample time (-6, -1, 1, 2, 3, 5, 10, and 20 min).

At selected times (-6, -1, 1, 2, 3, 5, 10, and 20 min), 4 ml of arterial blood were collected in glass test tubes containing 0.5 ml of 15% EDTA and 100 micro gram of indomethacin, and were immediately transferred to ice. After centrifugation (2,000 g for 10 min at 4 degrees Celsius), plasma was aspirated and stored at -70 degrees Celsius in polypropylene tubes. Thromboxane Boron²analyses were performed by radioimmunoassay. [8].

Total blood white cell (WBC) counts were determined using a Coulter counter (model Zf; Coulter Electronics, Hialeah, FL) at -6, -1, 1, 2, 3, 5, and 10 min. Activated clotting time (ACT) was measured by a Hemochron 400 system (International Technidyne, Edison, NJ); heparin assay, prothrombin time, partial thromboplastin time, and fibrinogen were measured at -6, - 1, and 5 min by standard tests. [9–12] Arterial blood gas tensions and pH were measured with a polarograph (model M238; Corning Medical and Scientific, Medfield, MA) at -6, -1, 2, 3, 10, and 20 min.

Recombinant platelet factor 4 in a concentration of 4.56 mg/ml was supplied by the Repligen Corporation (Cambridge, MA). The preparation was free of endotoxin (less than 0.05 endotoxin units/mg by standard limulus assay). The carrier solution consisted of 10 mM sodium acetate with a pH of 5.0 and 150 mM sodium chloride. Porcine-derived heparin (1,000 U/ml; Elkins-Sinn, Cherry Hill, NJ) and protamine sulfate (10 mg/ml; Lilly, Indianapolis, IN) were obtained from commercial sources. Indomethacin capsules (25 mg/capsule; Merck Sharp Dohme, West Point, PA) were opened and the contents suspended in 50 ml of water before administration via an orogastric tube.

Each lamb was allowed to recover after surgery for 2 h before intravenous challenge. Two lambs were studied by anticoagulation with heparin (200 U/kg) intravenously at -5 min, followed by reversal with rPF4 (4 mg/kg) intravenously at 0 min. Subsequent intravenous doses of PF4 were reduced to 1.5 mg/kg because of the limited supply of rPF4. The remaining challenges were divided into five groups and administered at random:

- Group 1 (saline/rPF4, n = 5 challenges): normal saline (10 ml intravenously) at time -5 min, followed by rPF4 (1.5 mg/kg intravenously) at time 0 min.

- Group 2 (saline/carrier, n = 5 challenges): normal saline (10 ml intravenously) at time -5 min, followed by rPF4 diluent buffer (carrier without rPF4, 10 ml intravenously) at time 0 min.

- Group 3 (saline/protamine, n = 5 challenges): normal saline (10 ml intravenously) at time -5 min, followed by protamine (2 mg/kg intravenously) at 0 min.

- Group 4 (indomethacin/saline/rPF4, n = 5 challenges): indomethacin pretreatment (10 mg/kg via nasogastric tube 2 h before the experiment), saline (10 ml intravenously) at time -5 min, followed by rPF4 (1.5 mg/kg intravenously) at 0 min.

- Group 5 (heparin/protamine, n = 5 challenges): heparin (200 U/kg intravenously) at time -5 min, followed by protamine (2 mg/kg intravenously) at 0 min.

Each lamb underwent a maximum of three challenges with a 2-h recovery period between each challenge. The order of the challenges was determined at random. Indomethacin-pretreated lambs underwent only one challenge.

All data are expressed as mean plus/minus SEM. Differences over time within or among study groups were evaluated by ANOVA for repeated measures as implemented in the SAS statistical program (Version 6.0; SAS Institute Inc., Cary, NC). Depending on the variable, different sets of planned comparisons were tested individually with significance declared at an alpha level of 0.05 after a Bonferroni correction.

In the two initial studies, reversal of heparin anticoagulation with rPF4 led to severe acute pulmonary hypertension (peak Ppa 40 and 33 mmHg).

Intravenous administration of 1.5 mg/kg rPF4 alone (group 1) increased the Ppa from 17.2 plus/minus 0.7 mmHg (mean plus/minus SEM) at baseline to 31.2 plus/minus 2 mmHg at 1 min (P < 0.05, Figure 1). The degree of peak Ppa hypertension was comparable with the magnitude of the acute pulmonary hypertension in the heparin-protamine reversal reaction in group 5 (Figure 1, P = NS peak values differ). Neither injection of the carrier alone (group 2) nor intravenous injection of protamine alone without heparin pretreatment (group 3) produced pulmonary hypertension. Pretreatment with indomethacin (group 4) completely blocked the pulmonary vasoconstriction after the injection of PF4 alone (Figure 1).

Reversal of heparin anticoagulation with rPF4 produced a marked elevation of plasma thromboxane Boron²levels concomitant with the episode of transient acute pulmonary vasoconstriction. Plasma TxB²levels increased from < 0.25 ng/ml to 50 and 30 ng/ml in the two lambs studied. The injection of rPF4 alone (group 1) was associated with increased plasma TxB²levels from 0.06 plus/minus 0.02 ng/ml to 3.96 plus/minus 1.21 ng/ml at 1 min (Figure 2). This was comparable with plasma TxB²levels at 1 min in the heparin protamine neutralization animals (group 5, Figure 2, P = NS values differ). Pre-treatment with indomethacin (group 4) completely prevented the increase of arterial plasma TxB²levels (Figure 2). Injection of carrier (group 2) or protamine alone (group 3) did not increase plasma TxB²levels.

Injection of rPF4 (group 1) led to transient systemic vasoconstriction (a 75% increase of SVR from the baseline level at 1 min) with brief, mild systemic hypertension (10% at 1 min, data not shown). The cardiac output was transiently decreased by about 40% at 1 min. These systemic hemodynamic changes were completely prevented by pretreatment with indomethacin (group 4). The systemic hemodynamic measurements in all the other groups did not change significantly during challenge (data not shown).

The circulating leukocyte concentration and Pa^O2did not change in animals given intravenous rPF4 alone (data not shown). In both studies of intravenous PF4 reversal of heparin anticoagulation, PF4 was found to rapidly and completely reverse heparin's prolonging effect on the activated clotting time, prothrombin, and partial thromboplastin time, without any change of plasma fibrinogen levels (data not shown).

This study demonstrates that rPF4, administered intravenously at a dose of 1.5 mg/kg into the awake lamb, produces an acute and marked elevation of pulmonary vascular resistance and pulmonary artery pressure. The magnitude of these increases was similar to that measured after protamine reversal of heparin anticoagulation in the lamb. These findings are surprising in view of the recently reported efficacy and hemodynamic safety of rPF4 reversal of heparin anticoagulation in the rat. [7].

Injection of rPF4 alone produced a marked and rapid increase of arterial plasma TxB²concentration that was concomitant with the increase of Ppa and PVR. This reaction appears similar to that occurring after heparin-protamine reversal in lambs. It appears that, in the case of heparin and protamine, it is the interaction of the polyanions (e.g., heparin) and polycations (e.g., protamine) to form complexes that activate complement via the classical pathway. In the case of rPF4, however, the reaction occurs without prior heparin induced anticoagulation and, therefore, is likely to be caused by the rPF4 molecule itself. The buffer solution used as a carrier for rPF4 did not provoke this adverse drug reaction in the lamb.

The degree of acute pulmonary vasoconstriction produced by rPF4 is comparable with that of the heparin-protamine reaction and is associated with a significant depression of cardiac output. However, hypoxemia and neutropenia did not occur after injecting 1.5 mg/kg rPF4, indicating that complement was not activated by the dose of rPF4 that we injected.

Inhibiting the production of thromboxane A²by pretreating the lambs with the cyclooxygenase inhibitor indomethacin prevented both the increase of plasma TxB²and pulmonary vasoconstriction after rPF4 injection. The precise cellular source of thromboxane in these drug reactions remains uncertain. Pulmonary artery hypertension induced by the heparin-protamine reaction has been documented in platelet depleted lambs, [13] as well as in cat lungs perfused with acellular dextran. [14] It has been suggested that the resident pulmonary intravascular macrophages produce thromboxane either directly or via stimulation of one or more other cell types. [15] There are large numbers of pulmonary intravascular macrophages in sheep. [16] These cells may release vasoactive mediators after acute stimulation with a foreign protein. Sheep and other artyodactylae, for example, are significantly more responsive to the intravascular infusion of liposomes than rats. [17] Rats and humans have far fewer pulmonary intravascular macrophages and, thus, may not be sensitive to intravenous rPF4 injection. However, some humans may have greater numbers of pulmonary intravascular macrophages, and they may well react adversely to the intravenous injection of PF4. Catastrophic pulmonary vasoconstriction after heparin-protamine neutralization in humans is a sporadic adverse response and, thus, many patients will have to be studied to reliably document whether this reaction occurs with heparin-rPF4 neutralization.

It is possible that the structural differences between human and lamb PF4 may be responsible for the marked stimulation of the arachidonic acid cascade and thromboxane release that we demonstrated in the lamb. Human rPF4 contains 70 amino acid residues and has a molecular weight of 7,800. [18] Complete amino acid sequences for human, rat, bovine, and rabbit PF4 have been described. The nonhuman PF4 proteins have additional amino acids at the N terminus that may make them differ immunologically. [19] Platelet factor 4 has considerable amino acid homology with several endogenous peptide cytokines, including connective tissue-activating peptide (Ctap)-III and their relatives IP-10 and interleukin 8. [19] Recombinant PF4 has been shown to bind to interleukin-8 receptors and activate neutrophils when its N-terminus is modified. [20] However, we did not find a leukopenia after rPF4 injection. It is possible that the sequence and structural differences between human and sheep PF4 may be responsible for the thromboxane release that we demonstrated in the lamb. Because we did not measure transient leukopenia after injection, it is unlikely that complement activation is directly involved in this ovine reaction.

This study demonstrates that intravenous injection of human rPF4 into the lamb causes the consistent release of large quantities of thromboxane into plasma and produces severe pulmonary vasoconstriction. The effects of rPF4 injection into other species, including primates and, especially, humans, should be carefully investigated in large numbers of patients before rPF4 is clinically adopted to reverse heparin anticoagulation.

The authors thank Ralph D'Agostino, A.M., for statistical assistance; Malahat Kavosi, for technical support; and Ed Lowenstein, M.D., and Theodore Maione, Ph.D., for suggestions and critical review of the study.

*Myers JA, Gray GS, Peters DJ, Grimaila RJ, Hunt AJ, Maione TE, Mueller WT: Expression and purification of active recombinant platelet factor 4 from a cleavable fusion protein. Protein Expression and Purification 2:136–143, 1992.