C3aS7- 77 , a C-Terminal Peptide, Causes
Thromboxane-dependent Pulmonary Vascular Constriction in Isolated Perfused Rat Lungs 1- 3
MELVIN L. MORGANROTH, SUZANNE O. SCHOENEICH, GERD O. TILL, PETER A. WARD, SUZANAH J. HORVATH, and M. MICHAEL GLOVSKY
Introduction SUMMARY Pulmonary hypertension occurs after the intravascular activation of complement. How-
Activation of the complement system ever, It Is unclear which activated complement fragments are responsible for the pulmonary vascumay be involved in the development of lar constriction. We investigated the 21-carboxy-terminal peptide of C3a (C3aS7-77) to see if it would pulmonary hypertension associated with cause pulmonary vascular constriction when infused Into Isolated buffer-perfused rat lungs. Injecthe adult respiratory distress syndrome tion of C3aS 7-77 (225 to 450 I1g)caused mean pulmonary arterial pressure (Ppa) to rapidly Increase. (1-3). In animal models, pulmonary However, the response was transient, with Ppa returning to baseline within 10 min of its administration. C3a57-77 also resulted in an increase in lung effluent thromboxane B2 (TXB 2 ) , concomitant with hypertension occurs after activation of the peak Increase in Ppa. C3a 57-77 did not affect the amount of 6-keto-PGF 1Q in the same effluent complement by endotoxin (4), zymosansamples. Indomethacin inhibited the C3a57-77"induced pulmonary artery pressor response and the activated plasma (5), protamine reversal associated TXB 2 production. Indomethacin also decreased lung effluent 6-keto-PGF,Q. The thromof heparin anticoagulation (6), or cobra boxane synthetase inhibitors CGS 13080 and U63,357 Inhibited the C3a57-77-induced pulmonary arvenom factor (7). Although activated tery pressor response and TXB 2 production without affecting 6-keto-PGF,Q. These inhibitors did complement fragments can cause vascunot inhibit pulmonary artery pressor responses to angiotensin II. Tachyphylaxis to C3aS7-77 occurred lar constriction (8), it is unclear which because a second dose of C3aS7-77 administered to the same lung failed to cause a pulmonary artery activated complement fragments (e.g., pressor response or TXB 2 production. The loss of the pressor response was not due to a C3a57-77C3a, C5a, membrane attack complex) are induced decrease in pulmonary vascular responsiveness because pressor responses elicited by responsible for the pulmonary hypertenangiotensin II were not altered by lung contact with C3a57-77' Thus, C3a57-77 caused thromboxanedependent pulmonary vascular constriction in isolated buffer perfused rat lungs. sion elicited after the systemic or local AM REV RESPIR DIS 1990; 141:296-300 activation of complement. The anaphylatoxin C3a could be responsible because it can constrict guinea pig parenchymal tissue and isolated vascular strips of some species in vitro (9-11). Coronary artery 69, and 71to 73. The terminal pentapep- in association with the vascular constricvasoconstriction also occurs after the in- tide that is necessary for spasmogenic ac- tion and whether inhibition of producfusion of purified human C3a into iso- tivity and the remainder of the amino tion of cyclooxygenase metabolites or lated guinea pig hearts (12). However, acid residues are identical in rats and TXB 2 , respectively, by the cyclooxygenase inhibitor indomethacin (19) or the thromthese studies were hampered by poten- humans. tial impurities in the C3a preparations. Our approach was to determine if boxane synthetase inhibitors U63,357 or Furthermore, the vasoconstrictive effect C3aS7-77 would directly cause pulmonary COS 13080 (20, 21) would inhibit the of C3a on an intact pulmonary vascula- vascular constriction in isolated buffer- vasoconstriction. ture has not been investigated. perfused rat lungs devoid of circulating C3a is a 77 residue protein with a spas- blood elements. Buffer-perfused lungs mogenic activity that is dependent on the were used in part to study the effect of carboxy-terminal arginine residue. The the exogenous C3a peptide because en- (Received in original form March 17, 1989 and in last fivecarboxy-terminal residues of C3a zymes in blood can inactivate the pep- revised form June 28, 1989) are identical in humans and in rats, and tide by removing the terminal arginine From the Pulmonary and Critical Care Divithis pentapeptide possesses spasmogen- residue (17). The ex vivo isolated lung sion, Department of Internal Medicine, and Departic activity (13). However, this pentapep- preparation also allows an assessment of ment of Pathology, University of Michigan, Ann tide has only 0.5070 of the activity of in- the effect of C3aS7- 77 on an intact pul- Arbor, Michigan; the Department of Internal Meditact C3a. The potency of synthetic monary vascular bed without exposing cine, University of Southern California, and the Department of Biology, Caltech, Los Angeles, carboxy-terminal peptides of C3a in- the exogenously administered peptide to California. creases as the length of the peptide in- inactivation by carboxypeptidase N presSupported by Grant No. HL-01595 from the creases (14, 15). A 21-residue carboxy- ent in blood. Because C3a is known to National Institutes of Health and by a grant from terminal peptide, C3aS7-77, has equivalent induce thromboxane production in vitro the American Heart Association of Michigan. 3 Correspondence and requests for reprints should activity to native C3a in vitro (14, 15). (18), we determined if C3as7-77-induced be addressed to Melvin L. Morganroth, M.D., PulRat C3aS7-77 has 71070 sequence homolo- vasoconstriction was dependent on throm- monary and Critical Care Division, Department gy with human C3aS7-77 (16). The differ- boxane production. We determined if of Internal Medicine, 3916 Taubman, Ann Arbor, ing amino acids are in positions 63, 66, thromboxane B2 (TXB 2 ) was produced MI 48109-0360. 1
2
296
297
C3a-INDUCED VASOCONSTRICTION IN ISOLATED RAT WNGS
We now demonstrate that the synthetic peptide, C3a S7- 71 causes thrornboxanedependent pulmonary vascular constriction in isolated buffer-perfused rat lungs. Methods An ex vivo model was used to assess pulmonary vascular constriction and cyclooxygenase metabolite formation after injection of C3aS7-77. Lungs were isolated from male LongEvans rats (weighing 225 to 325 g) anesthetized with pentobarbital (40 mg/kg intraperitoneally) as previously described (7, 19). Isolated lungs were ventilated via a tracheal cannula with an air gas mixture containing 210/0 O 2-5% CO 2 at a rate of 60 breaths/min using a Harvard animal respirator (Harvard Apparatus Co., South Natick, MA). Lungs were perfused with physiologic salt solution osmotically stabilized with Ficoll 70 (4 g/100 ml). The first 30 ml of lung effluent were discarded to eliminate circulating blood elements from the vascular space. This resulted in no detectable neutrophils in a 1:4 dilution of lung perfusate in 3% acetic acid. The lungs were perfused at a constant flow of 0.03 mlig rat body weight/min. Mean pulmonary artery pressure (Ppa) was measured continuously using a Statham transducer (Statham Instruments, Hato Rey, PR), plotted using a Grass recorder (Grass Instrument Co., Quincy, MA) and was less than 10 mm Hg at the outset of the experiment. Perfusion pressure was proportional to pulmonary vascular resistance because the flow for a given lung was held constant. Left atrial pressure was kept at zero em H 2 0 . The lungs and perfusate were kept at a temperature of 37° C. The lungs were perfused for 30 min to reach a stable perfusion pressure and temperature before initiation of the experimental protocols.
Administration of C3a Peptide A peptide consisting of the 21-carboxy-terminal amino acids (residues 57 to 77) of intact C3a was synthesized by stepwise solid-phase chemistry, as previously described (22, 23) and gave a single peak on reverse-phase high performance liquid chromatography. Briefly, the Merrifield solid-phase synthesis was employed in the study. Once synthesized, the protein was cleaved from the resin using hydrogen fluoride. C3aS7-77 was desalted by passage through ultragel A54 and eluted in veronalbuffered saline at pH 7.4. Two sequential series of pressor responses to intraarterial angiotensin II followed by C3aS7-77 were elicited (n = 11). In four isolated lungs, a third pressor response to angiotensin II was elicited after the second injection of C3a S7-77. Five to 10 min were allowed for recovery or return of perfusion pressure to baseline between angiotensin II and C3aS7-77induced pressor responses. Pressor responses to angiotensin II were elicited by close arterial bolus injection of 0.21lg of angiotensin II dissolved in 100 III of normal saline. Pres-
sor responses to C3aS7-77 were elicited by closed arterial bolus injection of 225 to 450 ug of C3aS7-n in 250 to 500 III of buffered saline. In a given lung, the same dose of C3aS7-77 was used for both injections. Larger doses of C3aS7-77 were not studied because of the limited quantities of material available. In control experiments, in a separate series oflungs (n = 4), buffer was administered between angiotensin II pressor responses rather than C3aS7-77.
Measurement of 6-keto-PGF l n and TXB 2 In the same isolated lung preparations, 1 ml of lung effluent was obtained immediately before the first injection of C3aS7-77, coincident with peak increase in Ppa caused by the first injection of C3aS7 - 77, immediately before the second injection of C3aS7-n , and after the second injection of C3a S7-77. The lung effluent was immediately added to a tube containing indomethacin (10 ug/ml) to prevent cyclooxygenase product formation after sample collection. After extraction, using a C18 sep pac, TXB 2 , the stable metabolite of thromboxane All and 6-keto-PGF 1n , the stable metabolite of prostacyclin, were measured by radioimmunoassay in duplicate as previously described (24). The TXB 2 radioimmunoassay utilized an antibody with a sensitivity of 10 pg/ml and negligible cross-reactivity for other eicosanoids. The 6-keto-PGF tn radioimmunoassay utilized an antibody with a sensitivity of 18 pg/ml and negligible cross-reactivity for other eicosanoids. Recoveries were measured in each sample and averaged 81.5 ± 1.4%.
Administration of Cyclooxygenase or Thromboxane Synthetase Inhibitors In other lungs, either a cyclooxygenase inhibitor, indomethacin (5 ug/ml, n = 5), or thromboxane synthetase inhibitors CGS 13080 (1 ug/ml, n = 5) (Ciba-Geigy, Summit, NJ) or U63,357 (5 ug/rnl, n = 6) (Upjohn, Kalamazoo, MI) were added to the lung
perfusate. Indomethacin was dissolved in normal saline containing sodium carbonate and added to the perfusate, at physiologic pH, 25 min prior to the first injection of angiotensin II. CGS 13080 and U63,357 were dissolved in normal saline and added to the perfusate 15min prior to the first injection of angiotensin II. Two series of angiotensin II pressor responses separated by a single injection of 360 to 450 ug of C3aS7-77 were elicited as described above. Lung effluent samples (l ml) were obtained from each lung prior to the injection of C3aS7-n and coincident with the peak increase in pressure after the injection of C3aS7-n for the measurement of TXB2 and 6-keto-PGF l n as described above.
Statistical Analysis Values are expressed as mean ± standard error of the mean. A paired t test was used when responses in the same lung were compared. One-way analysis of variance and the Student Newman-Keuls multiple comparison test were used when more than two groups were compared. A p value < 0.05 was accepted as significant.
Results
Effect of C3aS 7 - 77 on Pulmonary Arterial Pressure and Vascular Responsiveness The first intraarterial injection of C3a S7- n (225 to 450 ug) caused Ppa to rapidly increase (figure 1). The peak increase in Ppa (5.6 ± 1.1 mm Hg, n = 11) occurred less than 3 min after its administration. Ppa then returned to baseline within 10 min. The peak increase in Ppa (0.4 ± 0.2 mm Hg, n = 11) after a second injection of the same dose of C3as 7 - n was markedly reduced (p < 0.05). This loss of response to C3a S7- 71 is illustrated in figure 1 by a Ppa tracing from a representative experiment. A dose of 0.2 ug of angiotensin
ANGIOTENSIN II
1
30 PA PRESSURE
20
[mm Hg]
10
ANGIOTENSIN II
!
cla
C3a
+
0
o
10
20
30
40
TIME[MINUTES] Fig. 1. A representative tracing showing the effects of the sequential injections of angiotensin II and C3a S7 - n on mean pulmonary arterial (PA) pressure in an isolated buffer-perfused rat lung. The times of injection of angiotensin II and C3a S7 - n are indicated by the arrows. C3a s 7- n-induced vascular constriction was not dependent on the prior administration of angiotensin II because PA pressure still increased 5.3 ± 1.2 mm Hg (n = 3) when only C3a S 7-11 wa's administered.
298
MORGANROTH, SCHOENEICH, TILL, WARD, HORVATH, AND GLOVSKY
D PRE C3a INJECTION •
12
POST C3a INJECTION * p
Thromboxane-dependent Pulmonary Vascular Constriction in Isolated Perfused Rat Lungs 1- 3
MELVIN L. MORGANROTH, SUZANNE O. SCHOENEICH, GERD O. TILL, PETER A. WARD, SUZANAH J. HORVATH, and M. MICHAEL GLOVSKY
Introduction SUMMARY Pulmonary hypertension occurs after the intravascular activation of complement. How-
Activation of the complement system ever, It Is unclear which activated complement fragments are responsible for the pulmonary vascumay be involved in the development of lar constriction. We investigated the 21-carboxy-terminal peptide of C3a (C3aS7-77) to see if it would pulmonary hypertension associated with cause pulmonary vascular constriction when infused Into Isolated buffer-perfused rat lungs. Injecthe adult respiratory distress syndrome tion of C3aS 7-77 (225 to 450 I1g)caused mean pulmonary arterial pressure (Ppa) to rapidly Increase. (1-3). In animal models, pulmonary However, the response was transient, with Ppa returning to baseline within 10 min of its administration. C3a57-77 also resulted in an increase in lung effluent thromboxane B2 (TXB 2 ) , concomitant with hypertension occurs after activation of the peak Increase in Ppa. C3a 57-77 did not affect the amount of 6-keto-PGF 1Q in the same effluent complement by endotoxin (4), zymosansamples. Indomethacin inhibited the C3a57-77"induced pulmonary artery pressor response and the activated plasma (5), protamine reversal associated TXB 2 production. Indomethacin also decreased lung effluent 6-keto-PGF,Q. The thromof heparin anticoagulation (6), or cobra boxane synthetase inhibitors CGS 13080 and U63,357 Inhibited the C3a57-77-induced pulmonary arvenom factor (7). Although activated tery pressor response and TXB 2 production without affecting 6-keto-PGF,Q. These inhibitors did complement fragments can cause vascunot inhibit pulmonary artery pressor responses to angiotensin II. Tachyphylaxis to C3aS7-77 occurred lar constriction (8), it is unclear which because a second dose of C3aS7-77 administered to the same lung failed to cause a pulmonary artery activated complement fragments (e.g., pressor response or TXB 2 production. The loss of the pressor response was not due to a C3a57-77C3a, C5a, membrane attack complex) are induced decrease in pulmonary vascular responsiveness because pressor responses elicited by responsible for the pulmonary hypertenangiotensin II were not altered by lung contact with C3a57-77' Thus, C3a57-77 caused thromboxanedependent pulmonary vascular constriction in isolated buffer perfused rat lungs. sion elicited after the systemic or local AM REV RESPIR DIS 1990; 141:296-300 activation of complement. The anaphylatoxin C3a could be responsible because it can constrict guinea pig parenchymal tissue and isolated vascular strips of some species in vitro (9-11). Coronary artery 69, and 71to 73. The terminal pentapep- in association with the vascular constricvasoconstriction also occurs after the in- tide that is necessary for spasmogenic ac- tion and whether inhibition of producfusion of purified human C3a into iso- tivity and the remainder of the amino tion of cyclooxygenase metabolites or lated guinea pig hearts (12). However, acid residues are identical in rats and TXB 2 , respectively, by the cyclooxygenase inhibitor indomethacin (19) or the thromthese studies were hampered by poten- humans. tial impurities in the C3a preparations. Our approach was to determine if boxane synthetase inhibitors U63,357 or Furthermore, the vasoconstrictive effect C3aS7-77 would directly cause pulmonary COS 13080 (20, 21) would inhibit the of C3a on an intact pulmonary vascula- vascular constriction in isolated buffer- vasoconstriction. ture has not been investigated. perfused rat lungs devoid of circulating C3a is a 77 residue protein with a spas- blood elements. Buffer-perfused lungs mogenic activity that is dependent on the were used in part to study the effect of carboxy-terminal arginine residue. The the exogenous C3a peptide because en- (Received in original form March 17, 1989 and in last fivecarboxy-terminal residues of C3a zymes in blood can inactivate the pep- revised form June 28, 1989) are identical in humans and in rats, and tide by removing the terminal arginine From the Pulmonary and Critical Care Divithis pentapeptide possesses spasmogen- residue (17). The ex vivo isolated lung sion, Department of Internal Medicine, and Departic activity (13). However, this pentapep- preparation also allows an assessment of ment of Pathology, University of Michigan, Ann tide has only 0.5070 of the activity of in- the effect of C3aS7- 77 on an intact pul- Arbor, Michigan; the Department of Internal Meditact C3a. The potency of synthetic monary vascular bed without exposing cine, University of Southern California, and the Department of Biology, Caltech, Los Angeles, carboxy-terminal peptides of C3a in- the exogenously administered peptide to California. creases as the length of the peptide in- inactivation by carboxypeptidase N presSupported by Grant No. HL-01595 from the creases (14, 15). A 21-residue carboxy- ent in blood. Because C3a is known to National Institutes of Health and by a grant from terminal peptide, C3aS7-77, has equivalent induce thromboxane production in vitro the American Heart Association of Michigan. 3 Correspondence and requests for reprints should activity to native C3a in vitro (14, 15). (18), we determined if C3as7-77-induced be addressed to Melvin L. Morganroth, M.D., PulRat C3aS7-77 has 71070 sequence homolo- vasoconstriction was dependent on throm- monary and Critical Care Division, Department gy with human C3aS7-77 (16). The differ- boxane production. We determined if of Internal Medicine, 3916 Taubman, Ann Arbor, ing amino acids are in positions 63, 66, thromboxane B2 (TXB 2 ) was produced MI 48109-0360. 1
2
296
297
C3a-INDUCED VASOCONSTRICTION IN ISOLATED RAT WNGS
We now demonstrate that the synthetic peptide, C3a S7- 71 causes thrornboxanedependent pulmonary vascular constriction in isolated buffer-perfused rat lungs. Methods An ex vivo model was used to assess pulmonary vascular constriction and cyclooxygenase metabolite formation after injection of C3aS7-77. Lungs were isolated from male LongEvans rats (weighing 225 to 325 g) anesthetized with pentobarbital (40 mg/kg intraperitoneally) as previously described (7, 19). Isolated lungs were ventilated via a tracheal cannula with an air gas mixture containing 210/0 O 2-5% CO 2 at a rate of 60 breaths/min using a Harvard animal respirator (Harvard Apparatus Co., South Natick, MA). Lungs were perfused with physiologic salt solution osmotically stabilized with Ficoll 70 (4 g/100 ml). The first 30 ml of lung effluent were discarded to eliminate circulating blood elements from the vascular space. This resulted in no detectable neutrophils in a 1:4 dilution of lung perfusate in 3% acetic acid. The lungs were perfused at a constant flow of 0.03 mlig rat body weight/min. Mean pulmonary artery pressure (Ppa) was measured continuously using a Statham transducer (Statham Instruments, Hato Rey, PR), plotted using a Grass recorder (Grass Instrument Co., Quincy, MA) and was less than 10 mm Hg at the outset of the experiment. Perfusion pressure was proportional to pulmonary vascular resistance because the flow for a given lung was held constant. Left atrial pressure was kept at zero em H 2 0 . The lungs and perfusate were kept at a temperature of 37° C. The lungs were perfused for 30 min to reach a stable perfusion pressure and temperature before initiation of the experimental protocols.
Administration of C3a Peptide A peptide consisting of the 21-carboxy-terminal amino acids (residues 57 to 77) of intact C3a was synthesized by stepwise solid-phase chemistry, as previously described (22, 23) and gave a single peak on reverse-phase high performance liquid chromatography. Briefly, the Merrifield solid-phase synthesis was employed in the study. Once synthesized, the protein was cleaved from the resin using hydrogen fluoride. C3aS7-77 was desalted by passage through ultragel A54 and eluted in veronalbuffered saline at pH 7.4. Two sequential series of pressor responses to intraarterial angiotensin II followed by C3aS7-77 were elicited (n = 11). In four isolated lungs, a third pressor response to angiotensin II was elicited after the second injection of C3a S7-77. Five to 10 min were allowed for recovery or return of perfusion pressure to baseline between angiotensin II and C3aS7-77induced pressor responses. Pressor responses to angiotensin II were elicited by close arterial bolus injection of 0.21lg of angiotensin II dissolved in 100 III of normal saline. Pres-
sor responses to C3aS7-77 were elicited by closed arterial bolus injection of 225 to 450 ug of C3aS7-n in 250 to 500 III of buffered saline. In a given lung, the same dose of C3aS7-77 was used for both injections. Larger doses of C3aS7-77 were not studied because of the limited quantities of material available. In control experiments, in a separate series oflungs (n = 4), buffer was administered between angiotensin II pressor responses rather than C3aS7-77.
Measurement of 6-keto-PGF l n and TXB 2 In the same isolated lung preparations, 1 ml of lung effluent was obtained immediately before the first injection of C3aS7-77, coincident with peak increase in Ppa caused by the first injection of C3aS7 - 77, immediately before the second injection of C3aS7-n , and after the second injection of C3a S7-77. The lung effluent was immediately added to a tube containing indomethacin (10 ug/ml) to prevent cyclooxygenase product formation after sample collection. After extraction, using a C18 sep pac, TXB 2 , the stable metabolite of thromboxane All and 6-keto-PGF 1n , the stable metabolite of prostacyclin, were measured by radioimmunoassay in duplicate as previously described (24). The TXB 2 radioimmunoassay utilized an antibody with a sensitivity of 10 pg/ml and negligible cross-reactivity for other eicosanoids. The 6-keto-PGF tn radioimmunoassay utilized an antibody with a sensitivity of 18 pg/ml and negligible cross-reactivity for other eicosanoids. Recoveries were measured in each sample and averaged 81.5 ± 1.4%.
Administration of Cyclooxygenase or Thromboxane Synthetase Inhibitors In other lungs, either a cyclooxygenase inhibitor, indomethacin (5 ug/ml, n = 5), or thromboxane synthetase inhibitors CGS 13080 (1 ug/ml, n = 5) (Ciba-Geigy, Summit, NJ) or U63,357 (5 ug/rnl, n = 6) (Upjohn, Kalamazoo, MI) were added to the lung
perfusate. Indomethacin was dissolved in normal saline containing sodium carbonate and added to the perfusate, at physiologic pH, 25 min prior to the first injection of angiotensin II. CGS 13080 and U63,357 were dissolved in normal saline and added to the perfusate 15min prior to the first injection of angiotensin II. Two series of angiotensin II pressor responses separated by a single injection of 360 to 450 ug of C3aS7-77 were elicited as described above. Lung effluent samples (l ml) were obtained from each lung prior to the injection of C3aS7-n and coincident with the peak increase in pressure after the injection of C3aS7-n for the measurement of TXB2 and 6-keto-PGF l n as described above.
Statistical Analysis Values are expressed as mean ± standard error of the mean. A paired t test was used when responses in the same lung were compared. One-way analysis of variance and the Student Newman-Keuls multiple comparison test were used when more than two groups were compared. A p value < 0.05 was accepted as significant.
Results
Effect of C3aS 7 - 77 on Pulmonary Arterial Pressure and Vascular Responsiveness The first intraarterial injection of C3a S7- n (225 to 450 ug) caused Ppa to rapidly increase (figure 1). The peak increase in Ppa (5.6 ± 1.1 mm Hg, n = 11) occurred less than 3 min after its administration. Ppa then returned to baseline within 10 min. The peak increase in Ppa (0.4 ± 0.2 mm Hg, n = 11) after a second injection of the same dose of C3as 7 - n was markedly reduced (p < 0.05). This loss of response to C3a S7- 71 is illustrated in figure 1 by a Ppa tracing from a representative experiment. A dose of 0.2 ug of angiotensin
ANGIOTENSIN II
1
30 PA PRESSURE
20
[mm Hg]
10
ANGIOTENSIN II
!
cla
C3a
+
0
o
10
20
30
40
TIME[MINUTES] Fig. 1. A representative tracing showing the effects of the sequential injections of angiotensin II and C3a S7 - n on mean pulmonary arterial (PA) pressure in an isolated buffer-perfused rat lung. The times of injection of angiotensin II and C3a S7 - n are indicated by the arrows. C3a s 7- n-induced vascular constriction was not dependent on the prior administration of angiotensin II because PA pressure still increased 5.3 ± 1.2 mm Hg (n = 3) when only C3a S 7-11 wa's administered.
298
MORGANROTH, SCHOENEICH, TILL, WARD, HORVATH, AND GLOVSKY
D PRE C3a INJECTION •
12
POST C3a INJECTION * p
