The regulation of human factor XIIa by plasma proteinase inhibitors.

Studies of the inactivation of factor XIIa by plasma protease inhibitors in purified systems and in plasma were initiated to determine the relative importance of these inhibitors to the neutralization of factor XIIa. Factor XIIa was measured by the amidolysis of H-D-prolyl-L-phenylalanyl-L-arginine-p-nitroanilide dihydrochloride or by coagulant activity. C1 inhibitor (C1INH), alpha 2-antiplasmin (alpha 2AP), alpha 2-macroglobulin (alpha 2M), and antithrombin III (ATIII) inhibited factor XIIa with second-order rate constants of 2.2 X 10(5), 1.1 X 10(4), 5.0 X 10(3), and 1.3 X 10(3) M-1 min-1. Factor XIIa activity was not affected by alpha 1-proteinase inhibitor. Incubation of 125I-radiolabeled factor XIIa resulted in 1:1 stoichiometric complexes with C1INH (Mr 190,000), ATIII (Mr 125,000), and alpha 2AP (Mr 150,000 and 125,000) using sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Incubation of 125I-Factor XIIa with alpha 2M resulted in a component of Mr 85,000 on a reduced sodium dodecyl sulfate-polyacrylamide gel, indicating that a subunit of factor XIIa was covalently bound to a proteolyzed portion of alpha 2M. The relative effectiveness of each inhibitor at plasma concentrations was 61:2:3:1 for C1INH, alpha 2AP, alpha 2M, and ATIII, respectively. Kinetic studies of the inactivation of purified factor XIIa added to various plasmas containing different concentrations of C1INH verified the predictions from the purified systems. Gel filtration of radiolabeled factor XIIa incubated with plasma confirmed that factor XIIa-C1INH was the major complex. Analysis by sodium dodecyl sulfate-polyacrylamide gel electrophoresis indicated that the complexes in plasma had the same molecular size as those with purified inhibitors. C1INH functions as the predominant inhibitor of factor XIIa in plasma.

The Regulation of Human Factor XIIa by Plasma Proteinase Inhibitors* (Received for publication, July 17, 1984) Robin A. PixleyS, Marc Schapira4, and Robert W. ColmanS Studies of the inactivation of factor XIIa by plasma protease inhibitors in purified systems and in plasma were initiated to determine the relative importance of these inhibitors to the neutralization of factor XIIa. Factor XIIa was measured by the amidolysis of H-Dprolyl-L-phenylalanyl-L-arginine-p-nitroazilide dihydrochloride or by coagdant activity. C1 inhibitor (ClINH), a2-antiplasmin (aZAP), a2-macroglobulin (a2M), and antithrombin I11 (ATIII) inhibited factor XIIa with second-order rate constants of 2.2 X lo', 1.1 x lo4, 5.0 x lo3, and 1.3 x lo3 M" min". Factor XIIa activity was not affected by al-proteinase inhibitor. Incubation of '261-radiolabeled factor XIIa resulted in 1:1 stoichiometric complexes with ClINH (Mr 190,000), ATIII (Mr 125,000), and a2AP ( M , 150,000 and 125,000) using sodium dodecyl sulfate-polyacrylamide gel electrophoresis, Incubation of '"1-Factor XIIa with a2M resulted in a component of M, 85,000 on a reduced sodium dodecyl sulfate-polyacrylamide gel, indicating that a subunit of factor XIIa was covalently bound to a proteolyzed portion of a2M. The relative effectiveness of each inhibitor at plasma concentrations was 61:2:3:1 for ClINH, a2AP, a2M, and ATIII, respectively. Kinetic studies of the inactivation of purified factor XIIa added to various plasmas containing different concentrations of ClINH verified the predictions from the purified systems. Gel filtration of radiolabeled factor XIIa incubated with plasma confirmed that factor XIIa-ClINH was the major complex. Analysis by sodium dodecyl sulfate-polyacrylamide gel electrophoresis indicated that the complexes in plasma had the same molecular size as those with purified inhibitors. ClINH functions as the predominant inhibitor of factor XIIa in plasma.
Human factor XI1 (Hageman factor) (1) is a single-chain plasma protein of P-globulin mobility with a M , of 80,000 (2).
Activation of the zymogen initially results in an active serine protease of identical molecular weight, containing two chains held together by disulfide bonds, designated factor XIIa (3). In plasma, factor XIIa hydrolyzes factor XI (4) and prekallikrein (5), resulting in conversion of these zymogens to active enzymes. Activation of factor XI results in the initiation of * This investigation was supported in part by National Institutes of Health Institutional Training Grant HL07248, American Heart Association Southeastern Pennsylvania Chapter Fellowship Award, National Institutes of Health Individual National Research Service Award HL06626, National Institutes of HeaIth Grant HL24365, Council for Tobacco Research Grant 1420, and Swiss National Science Foundation Grant 3.184.0.82. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked ''advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. intrinsic coagulation. Kallikrein can liberate bradykinin from high molecular weight kininogen (6) and may participate in surface-catalyzed fibrinolysis through the activation of plasminogen (7). Activated factor XI1 influences the extrinsic pathway of coagulation by activating factor VI1 (8) and activates the classical complement pathway by activating C1 (9, 10).
At least two plasma protease inhibitors have been reported to be capable of inhibiting factor XIIa: Ci inhibitor (ClINH') (11-13) and antithrombin I11 (ATIII) (13,14). In contrast, a2-antiplasmin (a2AP) (15), a*-macroglobulin (a2M) (13), and al-proteinase inhibitor (13) were not found to inhibit factor XIIa. However, quantitative comparisons of the potency of each inhibitor under similar experimental conditions have not been reported. Moreover, the only factor XIIa complex formed with a plasma protease inhibitor previously demonstrated is the factor XIIa-antithrombin I11 complex with M , 117,000 (14). We therefore studied the inhibition of factor XIIa by inhibitors in purified systems under identical conditions and present kinetic evidence that purified C i inhibitor is the most potent inhibitor of purified factor XIIa. We then conducted studies of factor XIIa in plasma and found that the theoretical inhibition rates expected from the purified systems agreed with those observed in plasma.

DISCUSSION
Factor XIIa has been reported in past investigations to be inhibited by two plasma protease inhibitors, ClINH (11-13) and ATIII (13,14). However, the relative importance of these inhibitors was not studied. Our results confirm that both inhibit factor XIIa, but ClINH is 2 orders of magnitude more potent. Since Stead et al. (14) reported that heparin accelerates ATIII inhibition of factor XIIa, we also studied the effect of heparin. At plasma concentrations of purified ATIII (5.8 PM) at 37 "c, the pseudo first-order rate constant of factor XIIa inactivation was 0.007 min", while with heparin (1.2 units/ml) it increased to 0.029 min". This 4-fold enhance-' The abbreviations used are: ClINH, C i inhibitor; ATIII, antithrombin 111; azM, al-macroglobulin; a2AP, a2-antiplasmin; SDS, sodium dodecyl sulfate; PAGE, polyacrylamide gel electrophoresis; HAE, hereditary angioedema.
Portions of this paper (including "Materials and Methods," "Results," and Figs. 1-5 and 7) are presented in miniprint at the end of this paper. Miniprint is easily read with the aid of a standard magnifying glass. Full size photocopies are available from the Journal of Biological Chemistry, 9650 Rockville Pike, Bethesda, MD 20814.
Request Document No. 84M-2195, cite the authors, and include a check or money order for $7.20 per set of photocopies. Full size photocopies are also included in the microfilm edition of the Journal that is available from Waverly Press. ment is small compared to the 171-fold advantage of ClINH over ATIII. In agreement with Chan et al. (13), we found that al-proteinase inhibitor does not inhibit factor XIIa. In contrast to previous investigations (13, 15), we found that both a2AP and a2M are inhibitors of factor XIIa and are more potent than ATIII but less potent than ClINH. These differences may have risen from the fact that previous investigators used a negatively charged surface, kaolin or glass, to activate factor XI1 to factor XIIa and therefore studied the inactivation of surface-bound factor XIIa which may exhibit different kinetics. Furthermore, in prior reports, residual factor XIIa activity was determined indirectly by activation of prekallikrein to kallikrein.
In this study, factor XIIa was activated in fluid phase by kallikrein. For studies with the purified inhibitors, residual factor XIIa activity was studied directly by using a synthetic chromogenic substrate, and the incubation of the enzyme and inhibitor was done under uniform conditions. Therefore, the potency of each inhibitor to inactivate factor XIIa can be quantified kinetically and compared in a purified system. As seen in Table I, the order of potency is ClINH > > azAP > azM > ATIII in a ratio of 171:84:1.
The inhibition of factor XIIa by a2M deserves more analysis of the results because of its large molecular weight ( M , 725,000) and its ability to inhibit order enzymes in more than 1:l stoichiometry. The inhibition of factor XIIa by aZM was studied under pseudo first-order conditions with the inhibitor in excess. When the reciprocal plot of the data (Fig. 4, inset) is extrapolated to the ordinate, the intercept is negative. These kinetics may be attributed to azM's ability to inhibit more than 1 mol of enzyme for each mole of inhibitor, as has been found for trypsin and chymotrypsin (45,46). More than a 1:l stoichiometry for factor XIIa is indicated by the following consideration. When the relative concentration of aZM in Fig.  4 is twice the value indicated (or each mole of azM is able to inhibit 2 mol of factor XIIa), the line of the reciprocal plot in Fig. 4 is shifted to the left and a positive intercept is attained. If the line of the reciprocal plot is shifted to pass through the origin, the relative concentration of a2M is 1.6 times the value indicated in Fig. 4. This finding suggests that a2M can inhibit more than 1 mol of factor XIIIa/mol of a2M but is not conclusive evidence of 2:l stoichiometry. KI and k2 are still not determinable by this analysis.
Residual amidolytic activity using the chromogenic substrate H-D-prolyl-L-phenylalanyl-L-arginine-p-nitroanilide dihydrochloride in factor XIIa-a2M complexes after a long incubation was found to be less than 15% and was still linear with time (Fig. 4). This finding suggests but does not prove that factor XIIa complexed with azM is not available to interact with small substrates, in contrast to what has been shown for some other proteases (47,48). This paper presents the first description of the M, of the factor XIIa-ClINH complex. When radiolabeled factor XIIa was incubated with the purified inhibitors, enzyme-inhibitor complexes were observed with SDS-PAGE. Analysis of the mixtures resulting from the incubation of "'1-factor XIIa ( M , 80,000) with ClINH (MI 105,000) and ATIII (MI 62,000) demonstrated that radiolabeled complexes of, respectively, M , 190,000 and 125,000 were generated during factor XIIa inactivation (Fig. 6). The complex of factor XIIa with ATIII demonstrated in Fig. 6 agrees with the M , 117,000 complex reported by Stead et al. (14). Incubation of factor XIIa with aZAP ( M , 67,000) resulted in the formation of two complexes ( M , 150,000 and 125,000) not previously described. The appearance of secondary complexes with time has been shown for other enzyme-inhibitor reactions and is not unusual for this type of reaction (26). The complex ( M . 125,000) is attributed to proteolysis of the higher molecular weight complex ( M , 150,000). Factor XIIa was incubated with aZM ( M , 360,000 nonreduced dimers, M, 185,000 reduced). When labeled factor XIIa was reacted with purified a2M and the incubation mixture was subjected to SDS-PAGE in the absence of a reducing agent, a substantial part of the radioactivity did not enter the gel (data not shown). This observation is attributed to labeled factor XIIa bound to the unreduced aZM. On reduced SDS gel, the reaction resulted in an enzymeinhibitor complex at M, 85,000 (Fig. 6). This complex is attributed to the smaller molecular weight subunit of factor XIIa ( M , 32,000, see Fig. 1) bound to a proteolyzed fragment of the reduced aZM. Investigations of other enzymes inhibited by aZM have shown similar molecular weight products after incubation and analysis using reduced SDS-PAGE (49)(50)(51). Recently, a similar molecular weight moiety on SDS-PAGE has been shown after incubation of '251-labeled trypsin with The second-order rate constants determined in the purified systems are not a true indicator of the regulation of the activity of factor XIIa in a mixed system of inhibitors, as is found in plasma. The inhibition of factor XIIa in a mixed system would depend on the additive contribution of each inhibitor, which is dependent on its concentration and its second-order rate constant of inhibition. When the mean azM (52).

Comparison of the relative factor XIIa inhibitory activity and predicted relative effectiveness of Ci inhibitor, azantipiasmin, a2-macroglobulin, and antithrombin III in plasma
The constants k2, Kl, and k were determined under conditions of pseudo first-order kinetics as described under "Materials and Methods."  6). Samples were applied to a 5-15% acrylamide gradient gel with a 5% stacking gel. After electrophoresis, the gel was dried and exposed to an x-ray film.
concentration of these inhibitors in pooled plasma is considered, the ratio of the inhibitor potency toward factor XIIa of ClINH, azAP, azM, and ATIII is 61:2:31 (Table I). This analysis of the data in purified systems predicts that ClINH contributes 91% of the inhibition of factor XIIa in the presence of these other inhibitors (Table I).
To confirm the predictions calculated from the purified inhibition reactions, purified factor XIIa was added to various plasmas with different concentrations of ClINH present. The rate of inhibition was measured and compare to the predicted rate based on the pure reactions. The data from these experiments (Fig. 7) closely fit the theoretical lines calculated for the sum of the Kob (Table I), providing evidence that ClINH is the inhibitor that is primarily responsible for the regulation of factor XIIa. At pharmacological concentrations of heparin (0.6-1.6 units/ml), no change in the rate of factor XIIa inhibition was observed (53).
Evidence for enzyme-inhibitor complexes formed in plasma is shown in the Sephadex G-200 gel filtration experiments (Figs. 8 and 9). Labeled factor XIIa incubated with either purified ClINH or normal plasma gave an identical sieving profile, again indicating that complexing with ClINH is the predominant inhibitor in plasma. On examination of the complex peak by SDS-PAGE (Fig. 9), the ascending limb and peak tube contained mainly a MI 190,000 complex similar to that found with purified factor XIIa and ClINH (Fig. 6). Lower concentrations of factor XIIa complexes with aZM, aZAP, and ATIII were also identifiable (Fig. 9). The apparent M, for the factor XIIa-ClINH complex of 300,000 by Sephadex G-200 gel filtration (Fig. 8) exceeds the predicted value of 190,000 as observed for the complex by SDS-PAGE (Figs. 6 and 9). ClINH has been reported to be a molecule with high asymmetry (54) and was found to gel-filter at MI 220,000 (44). These findings indicate that factor XIIa (Mr 80,000 On a sieving gel) and ClINH (MI 220,000 on a sieving gel) complex in a manner which results in a molecule retaining high asym- total factor XI1 (factor XI1 + factor XIIa) was chromatographed in buffer alone. The solid line represents the percent total counts/min in each fraction. The dotted line represents the calculated factor XIIa. The concentration of factor XIIa was estimated as 63% of the total by scanning a reduced radioautogram (see "Materials and Methods"). B, the same preparation of '251-labeled total factor XI1 was incubated at 37 "C for 15 min with purified ClINH. The counts/min in the factor XI1 peak (tubes 32-60) representing factor XIIa were calculated by multiplying by 0.63. The fractions representing the complex (tubes 10-31) were not adjusted since factor XI1 zymogen is not known to complex with protease inhibitors. C, the same preparation of labeled total factor XI1 was incubated at 37 "C for 10 min with normal human plasma. The calculations are the same as in B. The arrows indicate the peak fractions of internal standards of IgM (M, 900,000), IgG (M, 160,000), and human serum albumin (HSA; M, 68,000) obtained from plasma or purified proteins. metry and thus a high frictional ratio on a sieving column. On denaturation in SDS, this asymmetry is lost, which results in an apparent molecular weight that is lower by PAGE. Similar observations have been reported for other enzymeinhibitor complexes (44,55).
Hedner and Martinsson (56) have described another human inhibitor, a partially purified MI 75,000 plasma protein that inhibits urokinase-or streptokinase-induced activation of plasminogen and also inhibits the clotting activity of activated bovine factor XII. Since the inhibitory actidty of plasma can be predicted from the rate constants of C1 inhibitor, anti-  Fig. 8C, were incubated in SDS a t 37 "C for 2 h and then applied to an acrylamide slab gel (3 mm thick, 7% running gel, 3.5% stacking gel). thrombin 111, a*-antiplasmin, and a*-macroglobulin, this inhibitor does not significantly inhibit factor XIIa in normal plasma.
Similar findings suggesting the primacy of C i inhibitor have been observed in a recent study of the inhibition of factor XIIf (57). Since factor XIIa and XIIf have virtually identical light chains, this observation indicates that the heavy chain does not play an important role in inhibiting factor XIIa in plasma.
Deficiency of the major plasma inhibitor of factor XIIa, C i inhibitor, gives rise to the disease hereditary angioedema (HAE). Consistent with the known ability of C i inhibitor to inactivate C i , patients with HAE are known to have in vivo activation of C1 as assessed by diminished levels of its substrates, C2 and C4. Recently, we have reported evidence for prekallikrein activation during abdominal and laryngeal attacks of HAE (58), consisting of a decrease in functional prekallikrein without changes in prekallikrein antigen. In addition, during attacks of HAE, a decrease in the substrate of kallikrein, high molecular weight kininogen, was noted. Such a reaction would form bradykinin which might contribute to the attacks of edema. The conclusion that bradykinin is important in the pathogenesis of HAE has been reached by Since activated factor XI1 can cleave and activate both C1 and prekallikrein, the decrease in C i inhibitor in HAE might make an individual vulnerable to even small degrees of factor XI1 activation.