Ann Thorac Surg 1996;61:1223-1230
© 1996 The Society of Thoracic Surgeons
Original Article: Cardiovascular
Genetically Engineered Serine Protease Inhibitor for Hemostasis After Cardiac Operations
Sunil K. Ohri, FRCS,
Rachel Paratt,
Jennifer M. Becket, BA,
John Brannan,
Beverley J. Hunt, MD,
Kenneth M. Taylor, FRCS
Cardiothoracic Unit, Department of Surgery, Royal Postgraduate Medical School, Hammersmith Hospital, London, and Research Haematology, Harefield Hospital, Middlesex, England
Accepted for publication December 28, 1995.
 |
Abstract
|
|---|
Background. The serine protease inhibitor aprotinin has been widely reported for its beneficial action in limiting blood loss after cardiopulmonary bypass (CPB). A potent human serine protease inhibitor known as protease nexin II or amyloid precursor protein has been recently isolated. A recombinant protein known as recombinant Kunitz protease inhibitor (rKPI; Scios Nova, Mountain View, CA) with sequence homology to the protease nexin II–amyloid precursor protein molecule has been manufactured.
Methods. Recombinant Kunitz protease inhibitor was assessed in an ovine model of CPB as a hemostatic agent after CPB. Sheep (n = 22) underwent CPB for 90 minutes. Two thoracic drains were sited and drain losses collected for a period of 3 hours after CPB. Wounds were subjectively assessed before closure for ``dryness'' using a visual analogue scale. Sheep were randomized to control (n = 8), aprotinin (n = 8), and rKPI (n = 6) groups.
Results. Control animals had a drain loss of 409.4 ± 39.4 mL/3 h, compared with 131.3 ± 20.3 mL/3 h for the aprotinin group and 163.7 ± 34.3 mL/3 h for the rKPI group (p= 0.16). Hemoglobin loss was 11.6 ± 3.6, 6.02 ± 2.1, and 4.6 ± 1.2 g/3 h for the control, rKPI, and aprotinin groups respectively (p = 0.25). The subjective analysis of the wounds at the end of CPB found aprotinin (1.25 ± 0.16; p < 0.05) and rKPI (1.17 ± 0.17; p < 0.05) animals to score significantly lower than control animals (2.63 ± 0.42).
Conclusions. On the basis of these in vivo findings, genetic modification may yield a more efficacious serine protease inhibitor with the inherent advantages of using a human-based protein.
|
Introduction
|
|---|
Blood loss after cardiopulmonary bypass (CPB) has both medical and financial implications. The inherent dangers of homologous blood transfusions have been highlighted, particularly in respect to transmission of the human immunodeficiency virus, although the costs of blood transfusions are not inconsiderable. Double-blind European studies have found that blood use may be reduced by 40% to 80% after CPB with aprotinin therapy [1–4]. The mechanisms involved in the development of the CPB bleeding diathesis are complex. The end result is excessive fibrinolysis and platelet dysfunction [5]. The importance of each component to the bleeding diathesis continues to be debated in the literature, but the ability of antifibrinolytic drugs such as 
-aminocaproic acid and tranexamic acid to limit CPB-induced blood loss suggests that the control of fibrinolysis may be the most important action of aprotinin [4].
Aprotinin belongs to a group of proteins known as the Kunitz protease inhibitors (KPIs), because they all share a Kunitz domain in their polypeptide structure, which confers an ability to inhibit serine proteases. Aprotinin is unusual in its ability to inhibit serine proteases as diverse as chymotrypsin to kallikrein, a coagulation factor that is central to the contact activation of CPB and links inflammation with blood coagulation [6, 7]. Although aprotinin was advocated as an antibleeding agent in open heart operations in the late 1960s, it was two decades later when the hemostatic benefit of this agent was recognized [3, 8, 9].
Aprotinin is extracted from bovine lung but is found in highest concentration in the mast cells of herbivores. Recently a ``human version'' of aprotinin or KPI has been recognized as a result of research of proteins controlling cerebral neuronal growth. Neuroscientists identified proteins that bound or linked with coagulation factors in the brain. These proteins were named protease nexins (PN); PN-I binds to thrombin, which is an inhibitor of neural growth. It was thus hypothesized that after cerebral injury, there is a leak of blood coagulation factors, which are inhibited by naturally occurring PNs [10]. This is plausible because cerebral capillaries lack the ability to manufacture thrombomodulin [11], which normally binds thrombin and activates protein C; activated protein C inhibits factors Va and VIIIa, thus limiting thrombosis in the microvasculature [12] (Fig 1
). A PN-II has also been isolated from the brain, which inhibits the action of XIa. This KPI has sequence homology to amyloid precursor protein, which forms the ß-amyloid deposits of Alzheimer's plaques [13, 14]. Further investigation has demonstrated the presence of PN-II in the alpha granules of platelets, but the brain as a tissue has the highest quantities of PN-II [15]. This human KPI has only 45% sequence homology to aprotinin (Fig 2), which has stimulated research into developing a genetically engineered KPI that may be more efficacious than bovine-derived KPI in limiting CPB-induced blood loss. Such a recombinant KPI (rKPI) has been produced using the Pichia pastoris yeast (SIBIA, San Diego, CA) and is a 61-amino acid polypeptide with more than 95% sequence homology at the Kunitz domain with human KPI (see Fig 2). We have assessed rKPI in an ovine model of CPB and compared its hemostatic potential with that of aprotinin.
View larger version (29K):
[in this window]
[in a new window]
|
Fig 1.. Coagulation and its control by protease nexins. (GpIb = glycoprotein Ib; HMWK = high molecular weight kininnogen; Pc = protein C; PGI2 = prostacyclin; PN-I and PN-II = protease nexin I and protease nexin II; Ps = protein S; tPA = tissue plasminogen activator; tPAI-1 and tPAI-3 = tissue plasminogen inhibitor 1 and 3.)
|
|
|
Material and Methods
|
|---|
In Vitro Assessment of Recombinant Kunitz Protease Inhibitor
Experiments to estimate the in vitro antikallikrein and antiplasmin effect of rKPI (Scios Nova, Mountain View, CA) and aprotinin (Bayer AG, Leverkusen, Germany) were undertaken. The approximate inhibitory potency against these two enzymes is expressed as inhibitory concentration in Table 1. Aprotinin was approximately three times more potent in vitro as an inhibitor of kallikrein than rKPI but more than 100 times more potent as a plasmin inhibitor. On the basis of these results rKPI was formulated to contain three times as much of the recombinant protein in the same volume of solution as aprotinin is supplied. Thus to administer any given dose of rKPI to achieve the same kallikrein inhibition, the same volume of drug was administered as aprotinin.
In Vivo Animal Studies
ANESTHESIA.
A prospective, randomized, double-blind study was undertaken in an ovine model of CPB. Sheep (n = 22) were randomized to three groups: control (n = 8), rKPI (n = 6), and aprotinin (n = 8). After cannulation of a foreleg vein, general anesthesia was induced with intravenous thiopentone. After endotracheal intubation, general anesthesia was maintained with a mixture of oxygen, nitrous oxide, and halothane; muscle relaxation was obtained with gallamine. All sheep had a stomach tube inserted to reduce the risk of gas-bloat syndrome. Cardiopulmonary bypass was instituted via right thoracotomy and insertion of a single two-stage right atrial cannula and carotid cannulation via neck cut-down. In addition a jugular vein was cannulated for venous sampling and the femoral artery was cannulated for continuous monitoring of arterial blood pressure. Needles were also inserted into the limbs for continuous recording of the electrocardiogram. A nasopharyngeal temperature probe was also sited and a bladder catheter inserted. All wounds were closed except for the thoracotomy to limit heat and fluid losses. Surgical hemostasis was undertaken in a standardized form in all animals as employed in the clinical setting.
CARDIOPULMONARY BYPASS PROTOCOL AND ANTICOAGULATION.
Before the institution of CPB, the animals were heparinized using 4 mg/kg of heparin intravenously. The activated clotting time was monitored using the Hemochron device with the activated clotting time kept greater than 750 seconds during CPB. At the end of CPB, the heparin was reversed with protamine sulfate to return the activated clotting time back to the preheparinization level.
Sheep received nonpulsatile perfusion using a Stockert pump-oxygenator (Stockert Instrumente GmbH, Munich, Germany) for 90 minutes. The first 45 minutes of CPB involved cooling to 28°C with rewarming to 37°C in the second 45 minutes. A crystalloid pump prime consisting of 2.0 L of Hartmann's solution with 25 mmol of bicarbonate was used. A bubble oxygenator without arterial line filtration was used, and alpha-stat pH methodology was employed for the management of arterial acid-base control. A flow-prioritized CPB protocol was used, with animals receiving a cardiac index of 2.4 L•min-1•m-2. The same output from the pump-oxygenator was maintained during both the hypothermic and rewarming phases of CPB. Vasoactive drugs were employed only if the mean arterial blood pressure exceeded the range 40 to 90 mm Hg; otherwise, the flow rate was maintained by adding volume (NaCl, 0.9%) to the extracorporeal circuit.
DRUG DOSING.
Drug dose was determined by the weight of the sheep. For a 70-kg animal, 2 x 106 KIU (280 mg) was given intravenously following induction of anesthesia before CPB. The pump-oxygenator was primed with a further 2 x 106 KIU (280 mg) and 500,000 KIU/h (70 mg/h) was infused during CPB. Thus for a 70-kg animal a total dose of 4.75 x 106 KIU (665 mg) was administered for 90 minutes of CPB. Recombinant KPI, which had been formulated as previously described, was given as an equal volume as calculated for aprotinin. Control animals received an equal volume of 0.9% NaCl.
BLOOD SAMPLING PROTOCOL.
Blood was sampled at the following times:
- Immediately after jugular venous line cannulation
- Immediately after infusion of drug/saline solution but before heparinization
- Before CPB, after heparinization
- 5 minutes on CPB
- 30 minutes on CPB
- 45 minutes on CPB
- 90 minutes on CPB
- 10 minutes after CPB
- 30 minutes after CPB
- 1 hour after CPB
- 2 hours after CPB
- 3 hours after CPB
HEMATOLOGIC ASSESSMENT.
Fibrinolysis was monitored by measuring antiplasmin activity and the euglobulin clot lysis time. Kallikrein inhibition was measured using a chromogenic assay (Channel Diagnostics, Walmer, UK). Antiplasmin activity was measured using an assay (Instrumentation Laboratories, Warrington, UK) that measured the combined effects of 
2-antiplasmin and the pharmacologic agents rKPI and aprotinin. The euglobulin clot lysis time was determined using the method described by Machin and Mackie [16].
Hemostatic activation was measured by assaying the levels of thrombin-antithrombin complexes using an enzyme-linked immunosorbent assay kit (Behring, Maidenhead, UK). Antithrombin III levels were measured by a chromogenic factor assay (Instrumentation Laboratories) using an automated coagulation laboratory (ACL 300 Research; Instrumentation Laboratories). Platelet function was assessed using whole blood platelet aggregometry (Chrono-Log Corp, Havertown, PA) with collagen, adenosine diphosphate, and ristocetin as platelet agonists (Lab Medics Ltd, Stockport, UK).
ASSESSMENT OF BLOOD LOSS.
Before closure of the thoracotomy two drains (28-gauge) were sited in the right hemithorax and mediastinum. A subjective assessment of the dryness of the surgical field and wounds was made before closure of the skin and recorded by means of a visual analogue scale (dry = 1, oozy = 2, very oozy = 3, and wet = 4). The drains were connected to an underwater drainage system, and suction was applied at a force of 10 kPa. The drain bottle was heparinized to determine not only the volume of tissue fluid and blood lost after CPB but also the hemoglobin drain losses. Drain losses were collected for a total of 3 hours after CPB.
Statistical Analysis
Values are expressed as mean ± standard error of the mean. Analysis of variance was undertaken to determine if there was a significant difference between the three groups (Kruskal-Wallis test). Only if this was significant (p < 0.05) was a modified unpaired t test undertaken to determine if there was a significant difference between any two mean values. Comparisons between paired samples have been undertaken using the Wilcoxon-signed rank test. Two-tailed p values have been calculated, and p values less than 0.05 have been considered to be statistically significant.
|
Results
|
|---|
Platelet Aggregation
Platelet aggregation was reduced for adenosine diphosphate and collagen (Fig 3a, b) but increased for ristocetin (Fig 3c) in all three groups after heparinization. One hour after CPB, ADP- and collagen-induced aggregation had not recovered in any of the three groups, but ristocetin-induced aggregation was higher in the aprotinin group 1 hour after CPB compared with postinduction levels (p < 0.05) and compared with the control (p < 0.05) and rKPI groups (p < 0.05) (see Fig 3c).
View larger version (31K):
[in this window]
[in a new window]
|
Fig 3. . Platelet aggregometry using adenosine diphosphate (ADP), collagen, and ristocetin as agonist. (CPB = cardiopulmonary bypass; rKPI = recombinant Kunitz protease inhibitor.)
|
|
Fibrinolysis
No significant difference in kallikrein inhibition was noted after induction of anesthesia in the three groups (Fig 4). In the aprotinin group, after administration of the drug, there was an increase in kallikrein inhibitory activity from a baseline value of 109.4% ± 4.6% to 134.1% ± 12.6%, which was significantly greater than the control group (p < 0.05). After heparinization and during CPB this difference was maintained and persisted after CPB for the duration of the study period. There was no significant difference between the control and rKPI groups in the pre-CPB period, but during CPB, a difference was established between control and rKPI sheep: at the end of CPB (90 minutes), control animals had a mean kallikrein inhibition of 51.7% ± 9.1% compared with 89.0% ± 9.7% in the rKPI group (p < 0.05). However, after CPB, when drug infusion was discontinued, the difference between rKPI and control sheep was lost (see Fig 4).
In the control group the antiplasmin activity fell once on CPB due to hemodilution. However, the antiplasmin activity rose significantly in the aprotinin- and rKPI-treated groups due to the antiplasmin activity of the pharmacologic agents being picked up in the assay (Fig 5).
View larger version (42K):
[in this window]
[in a new window]
|
Fig 5. . Antiplasmin activity. (CPB = cardiopulmonary bypass; rKPI = recombinant Kunitz protease inhibitor.)
|
|
The euglobulin clot lysis times were very similar in all three groups before CPB (Fig 6). The most significant reduction in the euglobulin clot lysis time occurred after heparinization. During CPB, the aprotinin group had a higher euglobulin clot lysis time than either the control or rKPI groups. After CPB there no significant differences between the three groups (see Fig 6).
View larger version (40K):
[in this window]
[in a new window]
|
Fig 6. . Euglobulin clot lysis time. (CPB = cardiopulmonary bypass; rKPI = recombinant Kunitz protease inhibitor.)
|
|
Coagulation
In all three groups, the antithrombin III levels were unchanged before CPB, but during CPB and after CPB the levels were reduced as a result of hemodilution (Fig 7). Hemostatic activation during CPB was shown by the increase in thrombin-antithrombin III complexes (Fig 8). All three groups had an increase in thrombin-antithrombin levels during CPB. The lowest levels were observed in the aprotinin group.
View larger version (42K):
[in this window]
[in a new window]
|
Fig 7. . Antithrombin III. (CPB = cardiopulmonary bypass; rKPI = recombinant Kunitz protease inhibitor.)
|
|
View larger version (45K):
[in this window]
[in a new window]
|
Fig 8. . Thrombin-antithrombin complexes. (CPB = cardiopulmonary bypass; rKPI = recombinant Kunitz protease inhibitor.)
|
|
Blood Loss
The control group had the highest drain loss at 409 ± 39.4 mL/3 h, compared with 163.7 ± 34.3 mL/3 h for rKPI and 131.3 ± 20.3 mL/3 h for aprotinin (Fig 9), but this was not statistically significant (p = 0.16). Hemoglobin loss in the 3-hour post-CPB period showed a similar profile to total drain losses: the control group had a hemoglobin loss of 11.6 ± 3.6 g/3 h, rKPI 6.02 ± 2.1 g/3 h, and aprotinin 4.6 ± 1.2 g/3 h (p = 0.24). The subjective bleeding score was lowest for the rKPI (1.17 ± 0.17; p < 0.05) and aprotinin (1.25 ± 0.16; p < 0.05) groups compared with the control group at 2.63 ± 0.42 (Fig 10).
View larger version (42K):
[in this window]
[in a new window]
|
Fig 9. . Drain loss after cardiopulmonary bypass in 3-hour period. (rKPI = recombinant Kunitz protease inhibitor.)
|
|
View larger version (28K):
[in this window]
[in a new window]
|
Fig 10. . Hemoglobin (HB) loss in drains over 3-hour period after cardiopulmonary bypass the bleeding score. (OR = operating room; rKPI = recombinant Kunitz protease inhibitor.)
|
|
|
Comment
|
|---|
This study has shown that in an ovine model of CPB there is a nonsignificant reduction in bleeding in the rKPI group compared with controls. This reduction was not as marked as that of the aprotinin-treated group. This could reflect on our inability to achieve equivalent in vivo antikallikrein effect for rKPI as compared with aprotinin.
Aprotinin has been shown to reduce the requirement of homologous blood transfusions in patients undergoing CPB, particularly in high-risk patients who are undergoing redo operations and patients with endocarditis [17]. However, concerns exist about the safety of such an agent. First, it is a bovine protein and thus there is a recognized risk of allergic reaction to its use in humans. Second, as it alters hemostasis perioperatively, it may increase the risk of thrombotic complications postoperatively. Bidstrup and associates [18] reported no increase in graft thrombosis using magnetic resonance imaging to investigate the patency of grafts, although this does not address the problem of microvascular thromboses. Finally, aprotinin is excreted renally, and there has been evidence of renal tubular damage in hypothermic animals [19] and some evidence that a similar problem may arise clinically in profound hypothermic circulatory arrest [20]. Clearly the advantage of using rKPI is that it is a human-based protein and should be less likely to provoke an allergic response. Moreover, it is possible to genetically manipulate the protein structure to engineer a molecule with defined inhibitory effects.
The mechanism of action of aprotinin is much debated. Currently it is widely accepted that it has profound antifibrinolytic effects, most markedly an antiplasmin effect. The efficacy of aprotinin in lower doses than those used in the high-dose regimen suggest that its anti-kallikrein effect is not as important as its antiplasmin effect. It seems doubtful that aprotinin has any major effect on platelet function or number. In this study similar hemostatic effects were seen in the aprotinin- and rKPI-treated groups in that both had increased antikallikrein and antiplasmin effects with a nonsignificant reduction in hemostatic turnover reflected by thrombin–antithrombin III complexes. In all these parameters rKPI had a lesser effect. This probably reflects our inability to achieve equivalence in antikallikrein activity in vivo as determined by in vitro calculations. Also rKPI intrinsically has less antiplasmin activity. Finally, there was no significant difference in platelet function as measured by whole blood aggregometry, although there was a improvement in ristocetin-induced aggregation after CPB with aprotinin.
A problem in using the ovine model is that aprotinin may reduce bleeding by different mechanisms from the human model. Aprotinin may have less ability to reduce bleeding in sheep than humans; if the data for its efficacy in humans were extrapolated, then sheep should have a greater reduction in blood loss than was observed in this study. Second, aprotinin also appears to be a potent antiinflammatory agent in the ovine model because the sheep appear to be losing less tissue fluid than controls from cut surfaces perioperatively. This underlines the importance of measuring hemoglobin loss as well as fluid losses when studying perioperative bleeding.
This study suggest that rKPI may be a useful agent in the control of post-CPB bleeding. Moreover, this approach holds promise for the future, because further genetic engineering of this protein may produce a more potent and desirable hemostatic agent.
|
Footnotes
|
|---|
Address reprint requests to Mr Ohri, Cardiothoracic Unit, Department of Surgery, Royal Postgradulate Medical School, Hammersmith Hospital, London W12 OHS, England.
|
References
|
|---|
- Fraedrich G, Weber C, Bernard C, et al. Reduction of blood transfusion requirement in open heart surgery by administration of high doses of aprotinin-preliminary results. Thorac Cardiovasc Surg 1989;37:89–91.[Medline]
- Dietrich W, Spannagl M, Jochum M, et al. Influence of high-dose aprotinin treatment on blood loss and coagulation patterns in patients undergoing myocardial revascularization. Anesthesiology 1990;73:1119–26.[Medline]
- Harder MP, Eijsman L, Roozendaal KJ, et al. Aprotinin reduces intraoperative and postoperative blood loss in membrane oxygenator cardiopulmonary bypass. Ann Thorac Surg 1991;51:936–41.[Abstract]
- Blauhut B, Gross C, Necek S, et al. Effects of high-dose aprotinin on blood loss, platelet function, fibrinolysis, complement, and renal function after cardiopulmonary bypass. J Thorac Cardiovasc Surg 1991;101:958–67.[Abstract]
- Harker LA. Bleeding after cardiopulmonary bypass. N Engl J Med 1986;314:446–8.[Medline]
- Fritz H, Wunderer G. Biochemistry and applications of aprotinin, the kallikrein inhibitor from bovine organs. Drug Res 1983;33:479–94.[Medline]
- Laurel MTP, Ratnoff OD, Everson B. Inhibition of the activation of Hageman factor (factor XII) by aprotinin (Trasylol). J Lab Clin Med 1992;119:580–5.[Medline]
- Bidstrup B, Royston D, Sapsford RN, Taylor KM. Reduction in blood loss and blood use after cardiopulmonary bypass with high dose aprotinin (Trasylol). J Thorac Cardiovasc Surg 1989;97:364–72.[Abstract]
- Havel M, Teufelsbauer H, Knobl P, et al. Effect of intraoperative aprotinin administration on postoperative bleeding in patients undergoing cardiopulmonary bypass operation. J Thorac Cardiovasc Surg 1991;101:968–72.[Abstract]
- Marx J. A new link in the brain's defenses. Science 1992;256:1278–80.[Free Full Text]
- Maruyama I, Bell CE, Majerus PW. Thrombomodulin is found on endothelium of arteries, veins, capillaries, and lymphatics, and on syncytiotrophoblasts of human placenta. J Cell Biol 1985;101:363–71.[Abstract/Free Full Text]
- Housholder GT. The role of protein C in congenital and acquired thrombotic disorders. J Oral Maxillofac Surg 1988;46:781–7.[Medline]
- Glenner GG, Wong CW. Alzheimer's disease: initial report of the purification and characterisation of a novel cerebrovascular amyloid protein. Biochem Biophy Res 1984;120: 885–90.
- Masters CL, Simms G, Weinmann NA, et al. Amyloid plaque core protein in Alzheimer's disease and Down's syndrome. Proc Natl Acad Sci USA 1985;82:4245–9.[Abstract/Free Full Text]
- Van Nostrand WE, Schmaier AH, Farrow JS, Cunningham DD. Protease nexin-II (amyloid beta-protein precursor): a platelet alpha-granule protein. Science 1990;248:745–8.[Abstract/Free Full Text]
- Machin SJ, Mackie IJ. An account of laboratory techniques. In: Chanarian I, ed. Laboratory haematology. New York: Churchill Livingstone, 1989;263–399.
- Bidstrup BP, Harrison J, Royston D, Taylor KM, Treasure T. Aprotinin therapy in cardiac operations: a report on use in 41 cardiac centers in the United Kingdom. Ann Thorac Surg 1993;55:971–6.[Abstract]
- Bidstrup BP, Underwood SR, Sapsford RN, Streets EM. Effect of aprotinin (Trasylol) on aorta-coronary bypass graft patency. J Thorac Cardiovasc Surg 1993;105:147–53.[Abstract]
- Fischer JH. Effects of Trasylol on the kidneys: dependence on temperature and dose. In: Dudziak R, Reuter H, Kirchhoff P, Schumann F, eds. Proteolysis and proteinase inhibition in cardiac and vascular surgery. Stuttgart: Schatauer Verlag, 1985;127–35.
- Sundt TM, Kouchoukos NT, Saffitz JE, et al. Renal dysfunction and intravascular coagulation with aprotinin and hypothermic circulatory arrest. Ann Thorac Surg 1993;55:1418–24.[Abstract]
This article has been cited by other articles:
|
|
|
|
 
S. K Ohri, R. Parratt, T. White, J. Becket, J. J Brannan, B. J Hunt, and K. M Taylor
A genetically engineered human Kunitz protease inhibitor with increased kallikrein inhibition in an ovine model of cardiopulmonary bypass
Perfusion,
May 1, 2001;
16(3):
199 - 206.
[Abstract]
[PDF]
|
|
|
|
|
|
|
I. Nagahiro, T. White, M. Yano, C. H.R. Boasquevisque, M. Hiratsuka, J. D. Cooper, and G. A. Patterson
Recombinant kunitz protease inhibitor ameliorates reperfusion injury in rat lung transplantation
Ann. Thorac. Surg.,
August 1, 1998;
66(2):
351 - 355.
[Abstract]
[Full Text]
[PDF]
|
|
|
Top
Abstract
Introduction
