Amidolytic Assay of Human Factor XI in Plasma: Comparison With a Coagulant Assay and a New Rapid Radioimmunoassay

Amidolytic Assay of Human Factor XI in Plasma: Comparison With a Coagulant Assay and a New Rapid Radioimmunoassay By Cheryl F. Scott, Dipali Sinha, Frances S. Seaman, Peter N. Walsh, and Robert W. Colman The traditional coagulant assay for plasma factor Xl suffers from a relatively high coefficient of variation, the need for rare congenitally deficient plasma, and a poor correlation between precision and sensitivity. We have developed a simple functional amidolytic assay for factor XI in plasma using the chromogenic substrate PyrGIu-Pro-Arg-p-nitroanalide (S-2366). After inactivation of a1-antitrypsin, Cl inhibitor, and other plasma protease inhibitors with CHCI3, plasma was incubated with kaolin, in the absence of added calcium, which limited the enzymes formed to those dependent on contact activation. Soybean trypsin inhibitor was used to minimize the action of kallikrein on the substrate. Once the reaction was complete, corn trypsin inhibitor was used to inactive factor XIIa, the enzyme generated by exposure of plasma to negatively charged surfaces, which had activated the factor XI. The assay is highly specific for factor Xl. since plasma totally deficient in that zymogen yielded only 1 %-3% of the enzymatic activity in normal F ACTOR XI is one of four proteins comprising the contact activated system of blood coagulation, but it is the only component whose absence may result in a hemorrhagic disorder. Factor XI in plasma is activated in vitro by activated factor XII (factor XIIa). Factor XIIa arises from the interaction of factor XII zymogen with negatively charged surfaces, such as kaolin,2 ellagic acid,3 dextran sulfate,4 or sulfatides,5 and is amplified by kallikrein-mediated proteolysis in the presence of high molecular weight (HMW) kininogen,6 previously described as contact activation cofactor.7 The major plasma inhibitor of factor XIa is ct1-antitrypsin,8 which is responsible for 68% of plasma inhibitory activity.9 Antithrombin-III, in the presence or absence of heparin, is responsible for 16%, and Cl inhibitor and a2-plasmin inhibitor account for the remainder of the plasma inhibitory activity. Unlike what is observed for factor XIIa and kallikrein, plant inhibitors, such as soybean, corn, and From the Thrombosis Research Center and Hematology-Oncology Section, Department ofmedicine. Temple University School of Medicine, Philadelphia. PA. Supported in part by NIH Grants HL 14217. HL 25661. and HL 23465, and by Grants 1389 and l42ofrom the Councilfor Tobacco Research, Inc. Submitted May 9. 1983; accepted July 15. 1983. Address reprint requests to Dr. Robert W. Colman, Thrombosis Research Center, Temple University School of Medicine, 3400 North Broad Street, Philadelphia. PA 19140. Presented at the IX International Congress of Thrombosis and Haemostasis, July 5, 1983. Stockholm. Sweden. 41 1 984 by Grune & Stratton, Inc. 0006-4971/84/6301-0005$01.00/0 plasma under identical conditions. The requirements for complete conversion of factor XI to XIa in plasma within 60 mm were, respectively, factor XII, 0.6 U/mI, and high molecular weight kininogen. 0.2 U/mI. Prekallikrein was not an absolute requirement for complete activation but did accelerate the reaction. The intraassay coefficient of variation was 3.4%, and the mean of 35 normal plasmas was 1.00 U ± 0.24 SD. In addition, a new rapid radioimmunoassay was devised using staphylococcal protein A as the precipitating agent for a complex of factor XI antigen with monospecific rabbit antibody. The mean was 1.01 U ± 0.30 SD. The correlation coefficients for amidolytic versus coagulant and amidolytic versus radioimmunoassay were r = 0.95 for the former and 0.96 for the latter. Thus, a simple, accurate amidolytic assay and a radioimmunoassay have been devised for measuring factor Xl in plasma that correlate well with the coagulant activity of factor Xl, as determined in our laboratory. lima bean trypsin inhibitors, do not inhibit factor XIa substantially. Traditionally, factor XI is assayed by a modified activated partial thromboplastin time, using plasma deficient in factor XI as the substrate. #{176} Several reagents currently used for the activation of the contact factors in coagulant assays include kaolin, celite, or ellagic acid. Commercially prepared reagents for use with automated equipment contain an activating agent and mixed phospholipids. The I 98 1 College of American Physicians Survey, which compiled statistical data from 466 laboratories, revealed that there is poor correlation between precision and sensitivity of most commercial systems, indicating the need for improvement of the factor XI assay. Furthermore, since factor XI deficiency is variable from patient to patient, consistent substrate is difficult to obtain. Coagulation assays, in general, have a high intraassay coefficient of variation (as high as 54%), probably due to the lability of several of the coagulant proteins. Moreover, coagulant assays require laborious, precise timing of both the activation step and the recalcification step. An immunochemical assay of factor XI has been described, 2 but it is time-consuming, usually requiring 3 days for completion. Amidolytic assays for plasma prekallikrein4 3 and an important protease inhibitor of contact activation, Cl inhibitor, 4 have recently been developed. However, no amidolytic assay has been previously described for plasma factor XI, probably due to lack of a specific synthetic substrate for factor XIa. In this article, we describe the development of a new 42 Blood, Vol. 63, No. 1 (January), 1984: pp. 42-50

AMIDOLYTIC ASSAY OF PLASMA FACTOR Xl 43 functional amidolytic assay for plasma factor XI, using the synthetic substrate PyrGlu-Pro-Arg-pNA (5-2366). Although this substrate is not specific for factor XIa, our assay system allows only the activation and expression of factor XIa enzymatic activity. The methodology for inactivating inhibitors, activating factor XI, and preventing other enzymes from hydrolyzing the substrate is described. The specificity of the assay and the requirements for other proteins of the contact system are assessed in this report. In addition, we have developed a new, rapid radioimmunoassay (RIA) for factor XI that employs staphylococcal protein A as the second precipitating antibody. This RIA has been correlated with the coagulant and amidolytic assays. The results of this study allow us to recommend a practical and reproducible amidolytic assay and a radioimmunoassay for the measurement of purified factor XIa and plasma factor XI. Reagents MATERIALS AND METHODS Soybean trypsin inhibitor (SBTI), bovine serum albumin (crystalline), and tosyl-lys-chloromethylketone (TLCK) were purchased from Sigma Chemical Co., St. Louis, MO. Kaolin (acid washed) and chloroform were obtained from Fisher Scientific Chemical Co., (King of Prussia, PA), Inosithin (mixed soybean phospholipids) was purchased from American Concentrates, New York, NY. 5-2366 (PyrGlu-Pro-Arg-paranitroanilide. 2HC1) was furnished to us as a gift by AB KABI Peptide Research, M#{246}lndal, Sweden. It can be obtained in the United States from Helena Laboratories, Beaumont, TX. All other chemicals were reagent grade. Plasma Blood was obtained by clean venipuncture from normal healthy individuals after written, informed consent was obtained from each donor. Blood for all three assays was collected into plastic syringes containing 3.8% (9: 1 v/v) trisodium citrate. Plasma was frozen in l-ml aliquots at - 70#{176}Cimmediately after centrifugation. The prekallikrein-deficient plasma was kindly supplied by Dr. C. Abildgaard, Davis, CA. Plasma samples from two patients with severe liver disease were obtained from patients at Temple University Hospital. Reference plasma (pool of 20 normal donors) was purchased from George King Biomedicals, Overland Park, KS. Factor- XI-deficient plasma and kininogen-deficient plasma were donated directly to us. Factor-XII-deficient plasma was kindly supplied by Dr. Margaret Johnson, Wilmington, DE. Additional samples of factor-xi-deficient plasma were generously donated by Mr. George King (George King Biomedicals). Corn Trypsin Inhibitor Corn trypsin inhibitor was a generous gift of Patrick McDevitt and Dr. Edward P. Kirby ofthis institution. 5 The protein concentration was calculated using 1% E2,,m 2.0. Purified Factor XIa This was prepared according to Scott et al.9 Nine hundred milliliters of fresh-frozen plasma containing 50 sg/ml hexadimethrine bromide, 1 mm EDTA, and 0.02% sodium azide, was mixed, sequentially, with 3 batches of QAE-Sephadex (900-mI slurry in 20 mm Tris-Cl, ph 8.1). The material, which did not adsorb, was mixed with 1,000 ml SP-Sephadex equilibrated with 20 mmtris-cl, ph 8.1 (conductivity adjusted to less than 2 mmho), mixed for 30 mm, and filtered. The resin containing the adsorbed protein was poured into a column and eluted by a linear gradient between 0 and 0.13 M NaC1. The fractions containing factor XI coagulant activity were pooled, concentrated, and dialyzed in an Amicon concentrator (PM-30 membrane) versus 0.1 M sodium acetate, ph 5.3. The factor XI was eluted by a linear gradient from 0 to 0.35 M NaCI. The fractions containing factor XI coagulant activity were pooled and concentrated. Removal of IgG was accomplished by batch adsorption of 0.40 ml of factor XI (90 U/mI diluted I :2 with H2O) for each I ml of anti-igg-agarose that had been equilibrated in 0.1 M Tris-Cl, ph 8.0, containing 0.15 M NaCl. The sample was incubated for 5 mm prior to centrifugation for 2 mm at I 2,000 g. Yields between 40% and 55% ofthe original plasma concentration of factor XI were repeatedly observed. No prekallikrein or factor XII activity could be detected by coagulation assay. #{176} No IgG could be detected by radial immunodiffusion. The specific activity was 203 U/mg. Activation of purified factor XI was accomplished by incubating 2 x l0_6 M factor XI (260 sg/ml) with 100 M trypsin for 1.5 hr at 23#{176}C. The trypsin was then inactivated by incubation with TLCK (l0 M) at 37#{176}Cfor at least 3 hr. The half-life of trypsin amidolytic activity under these conditions was 4.5 mm. Factor XIa was a single component of mol wt I 50,000 on gel filtration and a single band on SDS gel electrophoresis of mol wt 160,000 in the absence of a reducing agent. 6 On reduction, two polypeptides of mol wt 50,000 and 30,000 were found, indicating complete activation to factor XIa. Purified Factor XI Human factor XI for the radioimmunoassay and for antiserum production was purified to homogeneity using the procedure of Bouma and Griffin)7 On SDS gel electrophoresis, the unreduced protein migrated as a single band at 160,000 daltons; whereas upon reduction, a single band appeared at 80,000 daltons. Purified factor XI was radiolabeled with DI, using the procedure of Bolton and Hunter. 8 The I-labeled factor XI was separated from the other iodination products by gel filtration using Sephadex G-25. The specific radioactivity of 251-factor-XI was approximately 500 cpm/ng. Antiserum Preparation Monospecific antibody to purified human factor XI was raised in rabbits by 3 weekly subcutaneous injections of purified human factor XI, 9 80-100 g, in a 1:1 emulsion with complete Freund s adjuvant followed by a booster injection of the same amount of antigen with incomplete Freund s adjuvant 7-10 days before bleeding. The antiserum formed a single precipitin band with purified factor XI, and at a dilution of 1:100, and abolished the clotting activity of factor XIa in a normal pooled plasma sample diluted 1:10. Preimmune rabbit serum was negative on all double immunodiffusion tests and did not inhibit factor XIa clotting activity in normal pooled plasma. Kinetic Studies of Purified Factor XIa on the Chromogenic Substrate S-2366 The Km for S-2366 was determined by using the substrate at various concentrations between 0.2 and I.6 mm and factor XIa at a final concentration of 0.6 nm. The buffer was 0.1 M Na phosphate, ph 7.6, containing 0.15 M NaC1 and I mm EDTA. The change in

44 scoi-r ET AL. absorbance at 405 nm was continuously recorded on a Gilford 240 spectrometer with a Gilford 601 5 chart recorder, and the initial rate measured at 37#{176}C. The molar extinction coefficient for pna at A5 is 9,800.#{176} The Km for the reaction between factor XIa and 5-2366 was 0.36 mm and the V,,,, (0.75 U/ml purified factor XIa) was 0.143 absorbance units/mm or 0.24 nmole/sec/ml. The K,,, was 2,000 sec, and the K,, to K,,, ratio was therefore 5.5 x 106 M sec_, if we assume two active sites for factor XIa. Amidolytic Assay of Plasma Factor XIa Activation One hundred and fifty microliters of plasma plus 150 MI CHC13, to inactivate plasma protease inhibitors,2 were mixed for I mm in a polypropylene microcentrifuge tube at 23#{176}C, followed by centrifugation at 1 2,000 g for 5 mm. One hundred microliters of the plasma (top layer) was then transferred to a 0.4-mi plastic microcentrifuge tube containing 10 il of 50 im soybean trypsin inhibitor, to prevent kallikrein action on the substrate. Ten microliters of kaolin* (20 mg/mlofo.l M Na phosphate, ph 7.6, containing 0.15 M NaCl and 1 mm EDTA) was then added. The sample was then intermittently mixed during the times indicated. Thirty microliters of the activated sample was then transferred to another 0.4-mi microcentrifuge tube containing 7.5.il ofcorn trypsin inhibitor (0.5 mg/mi) to inactivate factor XIIa and thereby prevent its action on the substrate. Assay Ten microliters of the activated plasma mixture was added to a prewarmed (37#{176}C)cuvette containing 270 zl 0.1 M Na phosphate, ph 7.6, 0.15 M NaCI, 1 mm EDTA, and 30 l of 5-2366 (11.5 mm). The final substrate concentration was 1. 15 mm. The change in absorbance was recorded at 405 nm (at a sensitivity of0.1-0.2 full scale and a chart speed of 2 cm/mm). Coagulant Assay offactor XI and XIa The coagulant activity of factor XI was determined by a modified activated partial thromboplastin time assay with factor-xi-deficient plasma as the substrate. #{176} One hundred microliters of factor-xideficient plasma was incubated with 100 s1 kaolin (5 mg/mi in saline), 100 p1 0.2% inosithin in 20 mm Tris-CI (ph 7.4) + 0. 1 5 M NaCI, and 10 il of plasma + 90 l of the abovementioned buffer for 5 mm at 37#{176}C. Then, 100 zl of 30 mm CaC12 was added to initiate clot formation. Factor XIa was assayed by substituting saline for kaolin and decreasing the incubation time, prior to recalcification, to 1 mm. One unit of factor XI is defined as the amount in 1 ml of normal pooled plasma. Radioimmunoassay Radioimmunoassay of factor Xl antigen was performed using the rabbit anti-factor-xl antibody as the primary antibody and protein A bacterial adsorbent as the second precipitating agent. Staphylococcus protein A (Staph A) can interact specifically with the Fc region of most mammalian immunoglobulins, including all rabbit lgg subclasses. A typical assay mixture, containing 25 il of a 1:20,000 dilution of rabbit anti-factor-xi antibody, 25 sl of purified factor XI or unknown sample, and 25 l of 25I-factor-XI, was incubated in an Eppendorf microcentrifuge tube with narrow bore extended tip, for 1.5 hr at 37#{176}C.Thirty microliters of a 10% suspension of Staph A was added to the tube, mixed thoroughly, and incubated another 30 mm at 23#{176}C.After the samples were centrifuged, the tips of the tubes containing the pellets were amputated and counted in a CG 4000 Intertechnique gamma counter. E C E 0 V 0.50 RESULTS Correlation of Purified Factor XIa Amidolytic and Coagulant Activity Ten microliters of purified factor XIa, at various dilutions, was assayed for amidolytic activity utilizing the substrate 5-2366. The same dilutions were also assayed for factor XIa coagulant activity, in the presence and absence of kaolin (Fig. I ). (No detectable differences were found between the two coagulant assays.) The correlation coefficient (r) was equal to 0.98. One unit factor XIa was found to hydrolyze 0.49.tmole S-2366/min/ml. Effect of Plasma Protease Inhibitors on Purified Factor XIa Amidolytic Activity in the Presence and Absence of Soybean Trypsin Inhibitor (SBTI) Since plasma protease inhibitors clearly have the potential to affect factor XI activation and factor XIa activity, we explored methods to obviate the problem. Purified factor XIa lost only 8% of its amidolytic activity in 60 mm (Fig. 2, curve a) when incubated in buffer. However, when purified factor XIa was incubated with normal native plasma, where inhibitors were present at a ratio of > 1,000: 1 (Fig. 2, curve b), 65% ofthe factor XIa activity was inhibited within 10 mm, and the initial inactivation obeyed pseudo-firstorder kinetics. However, when plasma was treated with CHCI32 prior to incubation with purified factor XIa (Fig. 2, curve c), only 10% of the amidolytic activity was lost at 10 mm, indicating that most of the factor XIa plasma inhibitors were inactivated by U) 0.25 a. Factor XI CoojIont C?MTY. U/mI #{149}Micronized kaolin, prepared by Dr. James Brown of Helena Laboratories, Beaumont TX, can be used at a final concentration of 0.45 mg/ml. Its two main advantages over kaolin are that it remains in suspension for a longer period of time and it causes less optical interference. Fig. 1. Correlation of purified factor XIa amidolytic and coagu- Iant activity. Purified factor Xla. at various dilutions. was assayed directly for amidolytic activity. The same dilutions were also assayed for factor Xla coagulant activity. in the absence of kaolin (r = 0.98, slope 1.02. intercept = 2.0).

AMIDOLYTIC ASSAY OF PLASMA FACTOR XI 45 a x 9 C 1 TIME Fig. 2. Effect of plasma protease inhibitors on purified factor XIa amidolytic activity. Purified factor Xla (2.5 nm) was incubated at 23 C with either 0.1 M Na phosphate, 0.1 5 M NaCI, 1 mm EDTA. and 0.1 % polyethylene glycol 6000 (curve a), native plasma (curve b), native plasma and 4.5 sm SBTI (curve d). CHCI3-treated plasma (curve c), or CHCI3-treated plasma and 4.5 M 5BTI (curve e). At the times indicated, 1 0.d was assayed for residual amidolytic activity (see Materials and Methods). The data are expressed as percent original factor Xla activity. CHC13. Since substantial amounts of kallikrein would be generated concomitant with the formation of factor XIa in a plasma assay, we explored the possibility of including SBTI in the incubation mixture to inhibit the kallikrein formed. However, we found that SBTI, after partially inhibiting purified factor XIa, affected the rate and extent of inhibition of purified factor XIa by plasma protease inhibitors. Thus, when 4.5.tM SBTI was added to native plasma (Fig. 2, curve d) and the plasma incubated with purified factor XIa, an initial 21% decrease in amidolytic activity was observed followed by a diminished inactivation rate, as compared to native plasma (curve b). Finally, when 4.5.tM SBTI was added to CHC13-treated plasma prior to incubation with factor XIa (Fig. 2, curve e), there was an initial loss of 2 1 %, but no further fall in activity was observed over the 60-mm period. It therefore appears that SBTI not only inhibits kallikrein activity, but, after a rapid but partial inactivation of factor XIa, affords protection against further inhibition by plasma protease inhibitors. The combination of CHC13 treatment and SBTI resulted in factor XIa in plasma that was stable for at least 1 hr at approximately 80% of its initial activity. (mm) the presence of SBTI. When plasma was incubated with kaolin in the presence of SBTI to inhibit kallikrein (Fig. 3, curve a), an initial increase in amidolytic activity was observed over the first 10 mm, with maximal activity attained at 35 mm, followed by a decrease in activity of about 20% by 80 mm. When the plasma was treated with CHC13, prior to incubation with kaolin and SBTI, a more complete activation was observed, which was maximal by 20 mm, without a subsequent decrease in activity for at least 80 mm (Fig. 3, curve b). If SBTI was omitted from the CHC13- treated plasma, prior to incubation with kaolin, twice as much amidolytic activity resulted (data not shown). This activity corresponded to 0.35 U/mI kallikrein, when assayed with the amidolytic substrate S-2302. Therefore, since kallikrein can hydrolyze 5-2366, SBTI must be included in the incubation mixture in order to prevent substantial amounts of kallikrein in the assay. The actual contribution of kallikrein toward the amidolytic activity of 5-2366, after including SBTI, ranged from 2% to 5% of the total (data not shown). Therefore, in all subsequent experiments, plasma protease inhibitors were inactivated with CHC13 prior to activation of factor XI, and SBTI was included to stabilize the factor XIa as well as to minimize the contribution of kallikrein. To ascertain the specificity of this assay, we applied it to several samples of plasma from individuals with severe factor XI deficiency (< I 0% of normal). Only 2%-12% of normal pooled plasma amidolytic activity was formed, which correlated well with the coagulant activity of the sample (0%-9% of normal). To further x 2 t a 5 II 80 60 40 20 TIME (mm) b. 0. Activation of Plasma Factor Xl in Native and Chloroform-Treated Plasma in the Presence of Soybean Trypsin Inhibitor We then assessed the effect of plasma protease inhibitors on the activation of factor XI in plasma in Fig. 3. Activation of plasma factor Xl in native and CHCI3- treated plasma in the presence of SBTI. Native plasma (X) (1 50 Ml). 1 5 ILl Of 50 MM SBTI, and 1 5 Ml of 20 mg/mi kaolin were incubated. At the times indicated. 30 MI was transferred to a tube containing 7.5 ill of corn trypsin inhibitor (1 mg/mi). CHCI3-treated plasma (#{149}) was incubated under similar conditions. The data are expressed as percent of total activatable factor Xl.

46 SCOTF ET AL. hl7z, E 00204060810 Factor Xl Coogulant Activity. U/M Fig. 4. Effect of diluting normal plasma with factor-xldeficient plasma. Normal pooled plasma was mixed with factor- Xl-deflcient plasma at various ratios and then activated and assayed as in Materials and Methods. The same dilutions were assayed for coagulant activity. The data are expressed as percent of normal factor Xl activity (r = 0.99, slope 0.97, intercept - 0.01). validate the assay, normal pooled CHC13-treated plasma was mixed with CHC13-treated factor-xideficient plasma at various ratios. The mixed plasma samples were then incubated with kaolin for 60 mm and assayed as in Fig. 3. The mixtures of normal and factor-xi-deficient plasma (Fig. 4) resulted in amidolytic activity that was proportional to the contribution of the factor XI from normal plasma. One amidolytic unit of factor XIa in this plasma assay hydrolyzes 0.38 smole S-2366/min/ml. This result indicates that kallikrein or any other plasma protease did not substantially contribute to the amidolytic activity under these conditions (see Fig. 7C). The increased ratio of amidolytic to clotting activity in the purified factor XIa (Fig. 1) compared to plasma factor XI (Figs. 4 and 8A) is due to the fact that SBTI, which is employed in the plasma assay, results in a decrease of approximately 20% of the factor XIa amidolytic activity. SBTI was not added when assaying purified factor XIa (Fig. 1). Radioimmunoassay for Factor Xl A representative standard curve for the assay shows that bound 251-factor-XI is inversely proportional to the logarithm of the concentration of normal pooled plasma added (Fig. 5). This relationship is linear between 0.005 U/ml and 0.05 U/ml. Test materials were diluted as necessary to give values that fell within this range. Similar standard curves were obtained when purified factor XI was assayed instead of normal pooled plasma. In contrast, the addition of plasma severely deficient in factor XI did not appreciably decrease binding of radiolabeled factor XI to the antibody. To further validate the assay, various amounts of normal pooled plasma or purified factor XI were added to factor-xi-deficient plasma and then assayed both by coagulant and radioimmunoassay (Fig. 6). The titers of factor XI coagulant activity were plotted against those of factor XI antigen, and the regression line was calculated by the method of least squares. The values for reconstitution of factor-xi-deficient plasma with purified factor XI and the factor XI in normal pooled plasma were colinear with r = 0.99. Activation Requirements for Factor XI Amidolytic Activity Factor XII Since factor XIIa is the enzyme that cleaves factor XI to XIa, it is obvious that decreases in factor XII could influence this assay. To determine the requirement for this enzyme, factor-xii-deficient plasma and factor-xi-deficient plasma were mixed, in various ratios, subsequent to chloroform treatment. Each mixture was activated with kaolin (Fig. 7A), as in Materials and Methods. At various times, a portion of the mixture was transferred to a tube containing corn trypsin inhibitor to stop the reaction. The samples were assayed as in Materials and Methods, and the values were corrected for dilution of the factor XI by the factor-xi-deficient plasma. Factor-XI-deficient plasma was used as the diluent in order to maintain normal concentrations of all other plasma proteins. Activation of factor XI did not occur in the absence of factor XII. At least 0.6 U/ml of factor XII was 50 40 a, IC 24 C 20,.6.2 08 * 04 /0 ooi 0005 001 005 010 Factor XI Antigen. U/mI 0 0.4 O 12 16 20.4 Factor XI Coogulont Activity. U/mI Fig. 5. Standard curve for radioimmunoassay obtained with different dilutions of normal pooled plasma. Abcissa represents units/milliliter of factor Xl added (1 ml of normal pooled plasma contains 1 U of factor Xl). The ordinate represents percentage of bound radioactivity relative to the total. Fig. 6. Relationship of factor Xl coagulant activity and level of antigen as determined by radioimmunoassay. Factor-Xl-deflcient plasma was reconstituted with normal pooled plasma () or purified factor Xl (0) and assayed by both methods (see text) (r - 0.99. slope = 1.03. intercept = 0.01).

AMIDOLYTIC ASSAY OF PLASMA FACTOR Xl 47 100 E 16 14 A 1.2 ctoff,ia.t $0 a) a a) C a) (9 a x 0 U a 0 I- 0 I00 80 60 40 20 I00 80 60 40 20 --... p 40 60 TIME (mm) 08 ).. 06 04 02,.6 I.4 2. 1.0 0.8. 06! :: 0. S 0 02 0.4 06 08 1.0.2.4.6 Coagulant ActMty. U/mI Fig. 7. Activation requirements for factor Xla amidolytic activity. (A) Factor-Xl-deficient plasma was mixed with factor-xlideficient plasma at various ratios and activated. as in Materials and Methods, in order to assay the contribution of factor XII toward the activation of factor Xl. (B) Factor-XI-deficient plasma mixed with HMW kininogen-deficient plasma. (C) Factor-Xldeficient plasma mixed with prekallikrein-deflcient plasma. The data are expressed as percent of activatable factor XI. (A) 0.06 U/mI, (#{149}) 0.4 U/mI, (X) 0.2 U/mI, () 0.1 U/mI, (0) factor in incubation mixture. required in order to ensure full activation of the factor XI in plasma if a 60-mm incubation time was employed. However, with as little as 0.1 U/ml factor XII, 75% of the maximal factor XIa activity evolved during that time. HMW Kininogen When HMW kininogen-deficient plasma was incubated under the same conditions as in Fig. 7A, activation of factor XI did not occur (Fig. 7B), as previously reported.23 HMW kininogen-deficient plasma was mixed with factor-xi-deficient plasma in various ratios. At least 0.2 U/ml HMW kininogen was required in order to ensure full activation of factor XI in 60 mm, but even 0. 1 U/ml gave 85% of the maximal activation at 60 mm. Prekallikrein Kallikrein accelerates the activation offactor XII on a negatively charged surface by a feedback reaction and, thus, indirectly increases the rate of formation of factor XIa during contact activation. Therefore, the rate of activation of factor XII in prekallikreindeficient plasma is slow. Prekallikrein-deficient O 02 04 6 08 0 12.4 1.6 Factor XI Antigen, U/mI Fig. 8. (A) Correlation of factor Xl amidolytic and coagulant activity. Factor Xl amidolytic activity was assayed for amidolytic and coagulant activity (see Materials and Methods) in plasma from 35 normal individuals (s), 2 patients with liver disease (0), and 6 patients with less than 10% factor Xl (X) (r - 0.95). (B) Correlation of antigenic and amidolytic factor XI. Radioimmunoassay was performed on the samples from Fig. 8A (r - 0.96). (#{149}) Normal, (0) liver disease, (X) factor-xl-deflcient. plasma was mixed with factor-xi-deficient plasma under the same conditions as Fig. 7A. Even in the absence of prekallikrein, 89% of the factor XI was activated in 60 mm (Fig. 7C). Thus, unlike factor XII or HMW kininogen, prekallikrein was not an absolute requirement for generating factor XIa amidolytic activity, as expected. However, the presence of prekallikrein had an accelerating effect on the rate of activation of factor XI due to the feedback-amplification of factor XII activation.22 This result implied that kallikrein activity was still expressed in the presence of SBTI. In separate determinations, however, the concentration of kallikrein, as determined by amidolytic assay, was 0.008-0.048 U/ml, corresponding to the 0.1-0.6 U/ml added prekallikrein (Fig. 7C), which generated less than 5% of the possible amidolytic activity. Correlation of Factor XI Amidolytic and Coagulant Activity in Plasma Factor XI amidolytic activity was determined in 35 normal plasma samples using normal pooled plasma

48 SCOTr ET AL. (20 donors) as a reference. Factor XIa activity ranged from 0.62 to 1.39 U/ml, with a mean of 1.00 ± 0.24 SD. The coagulant activity, determined on the same aliquot of each plasma, showed a range of 0.60-1.45 and a mean of0.99 ± 0.30. In addition, determinations were performed on plasma samples from two patients with liver disease as well as five patients with factor XI deficiency (Fig. 8A). The correlation coefficient between the amidolytic activity and the coagulant activity was r = 0.95. Based on multiple determinations on the pooled normal plasma on at least 5 separate occasions, the intraassay coefficient of variation for the amidolytic assay was 3.4%. The plasma from normal individuals assayed for factor XI antigen showed a range of 0.56-1.61 U/ml (mean 1.01 ± 0.30). When the same samples from Fig. 8A were assayed for antigenic reactivity (Fig. 8B) and compared with the values obtained for amidolytic activity, the correlation coefficient (r) was equal to 0.96, similar to that of coagulant activity and amidolytic activity. DISCUSSION We present two new assays for factor XI in plasma, one functional, utilizing a tripeptide amide as substrate, and the other antigenic, a rapid (1-day) RIA using Staph A as the second precipitating agent. Both of these assays correlate closely with each other as well as with the conventional coagulant assay, as performed in our laboratory. 5-2366, when used with factor XIa, displays a Kcat to Km ratio of 5.5 x 106 M sec, making this chromogenic substrate very suitable for human factor XIa. This ratio is higher by an order of magnitude than the two substrates we have used previously for assay of purified factor XIa.24 Furthermore, the (2,000 sec ) is higher than any chromogenic substrate known. Under the conditions used, normal plasma factor XI was fully activated in 20 mm, and this activity remained stable for at least 80 mm (Fig. 2, curve b), due to prior inactivation of plasma protease inhibitors by CHC13 and partial protection by SBTI. This stability should facilitate automation, which is a major advantage of this type of assay. This amidolytic assay appears to be specific for factor XI by the following criteria: ( 1 ) purified factor XIa showed similar activity to that in normal plasma; (2) mixtures of normal plasma with factor-xideficient plasma revealed a precise linear correlation ofamidolytic with coagulant activity, dependent on the contribution of factor XI from normal plasma (Fig. 3); (3) factor-xi-deficient plasma showed virtually no activity, despite normal prekallikrein, HMW kininogen, and factor XII levels; and (4) a close correlation was found between the amidolytic activity, the coagulant assay, and the antigenic determination of factor XI. Consideration of the details of the assay help account for its usefulness. Plasma protease inhibitors destroy -75% of the factor XIa activity in 1 5 mm. However, SBTI, which initially inhibited 21% of the activity, appeared to slow down subsequent inhibition, since only -40% of the factor XIa was destroyed in 15 mm. Chloroform treatment inactivated most of the plasma protease inhibitors of factor XIa (less than 15% of factor XIa activity inhibited in 15 mm). Destruction of plasma protease inhibitors by CHCI3 permits unopposed formation of factor XIa. In addition to the protective effect of SBTI on factor XIa activity, it was necessary to include SBTI in the incubation mixture to inactivate most of the kallikrein that would have formed, although the small amount of residual kallikrein had an accelerating effect. The dual role of SBTI in this assay allowed the development of an activation procedure whereby the formed factor XIa was stable for long periods of time. This protective effect of SBTI has also been noted for thrombin25 and enzymes not inhibited by this plant inhibitor. The time of 60 mm for the activation step was selected to ensure full activation of the factor XI. Since 0.6 U/ml factor XII and 0.2 U/ml HMW kininogen were required for full activation of factor XI in 60 mm (Fig. 7), a low value for factor XI could reflect a deficiency of one of these contact coagulation proteins. Therefore, it is recommended that the finding of decreased factor XI amidolytic activity be further evaluated by coagulant assays for factor XII and high molecular weight kininogen. Alternatively, addition of normal plasma to the patient plasma in equal amounts prior to the assay would provide these proteins. This step would also allow detection of naturally occurring antibody to factor XI, such as that recently reported.26 The amidolytic assay that we have developed is easy to perform, requires a small amount ofsample, and can be adapted to an automated system. Since the activity during the activation process plateaus between 20 and at least 120 mm (data not shown), many samples can be activated at the same time without the need for laborious, precise timing of the activation step, as is required for the coagulant assay, allowing the activated factor XI to be assayed spectrophotometrically, at leisure. Furthermore, this substrate can be employed for purified factor XIa or partially purified factor XI. In the case of partially purified material, however, adequate amounts of contact activation factors must be included, and inhibitors must be included to inactivate any unwanted proteases. The amidolytic activity in the 35 normal plasma samples approximated a normal distribution, with a

AMIDOLYTIC ASSAY OF PLASMA FACTOR XI 49 range from 0.62 to 1.39 U/ml of the normal reference plasma, similar to the values observed for both coagulant and antigenic activity. The assay detected low concentrations of factor XI in two patients with liver disease. These results agreed well with values reported previously using an immunoassay. 2 The antigenic activity was also commensurately reduced. Excellent correlation (r = 0.95-0.96) between coagulant, amidolytic activity, and antigenic levels reinforces our confidence in both new assays. The assays ofcoagulation factor XI previously available include functional coagulant assays #{176}or immunochemical determinations. 2 The coagulant assays require either plasma genetically deficient in factor XI or normal plasma that has been adsorbed with celite. The latter may not be specific, since the procedure variably removes concentrations of factor XII and HMW kininogen. The former assay has a disadvantage, since it utilizes rare plasma with various degrees of factor XI deficiency. An assay employing a synthetic chromogenic substrate has intrinsic advantages, since these substrates are more homogeneous as well as stable. The interlaboratory coefficient of variation for coagulant assays for plasma factor XI ranges from 27.9% to 101.4%, while the intraassay coefficient of variation for our assays is 3.4%. The previously described RIA is time-consuming, requiring 3 days for completion of each set. Our RIA, which employs Staph A as the second precipitate, only requires 1 day to complete a set of 10 triplicate determinations. With our assay system, a low amidolytic value could be assayed the same day for antigenic reactivity of factor XI in order to distinguish an abnormal protein or an inhibitor of factor XI from a true deficiency. Since patients with mild deficiencies rarely have spontaneous hemorrhage, they are primarily at potential risk during or after surgical procedures. A preoperative assay, which can reproducibly detect 30% or less of normal activity, could minimize the risk. These two new assays, alone or combined, should prove valuable in detecting and evaluating both genetic and acquired deficiencies of plasma factor XI. ACKNOWLEDGMENT We would like to thank Drs. Lief Aurell and Petter Friberger of KABI VITRUM for providing the synthetic substrate 5-2366, essential to develop this assay, and for the generous contribution of KABI VITRUM which allowed C.F.S. to present this work at the IX International Congress on Thrombosis and Haemostasis in Stockholm, Sweden, July 3-8, 1983. 1. Ratnoff OD, Davie EW, Mallet DL: Studies on the activation of Hageman factor: Evidence that activated Hageman factor in turn activates plasma thromboplastin antecedent. J Clin Invest 40:803, I 961 2. Margolis J: The interrelationship of coagulation of plasma and release ofpeptides. Ann NY Acad Sci 104:133, 1963 3. Ratnoff DO, Crum JD: Activation of Hageman factor by solutions ofellagic acid. J Lab Clin Med 63:359, 1964 4. Kluft C: Determination of prekallikrein in human plasma: Optimal conditions for activation of prekallikrein. J Lab Clin Med 91:83, 1978 5. Tans G, Griffin JH: Properties of sulfatides in factor-xiidependent contact activation. Blood 59:69, 1982 6. Griffin JH, Cochrane CG: Mechanisms for the involvement of high molecular weight kininogen in surface-dependent reactions of Hageman factor. Proc NatI Acad Sci USA 73:2554, 1976 7. Schiffman 5, Lee P: Partial purification and characterization ofcontact activation cofactor. J Clin Invest 56:1082, 1975 8. Heck LW, Kaplan AP: Substrates of Hageman factor: I. Isolation and characterization of human factor XI (PTA) and inhibition of the activated enzyme by a1-antitrypsin. J Exp Med 140:1615, 1974 9. Scott CF. Schapira M, James HL, Cohen AB, Colman RW: The inactivation of factor XIa by plasma protease inhibitors. J Clin Invest 69:844, 1982 10. Proctor RR, Rapaport SJ: The partial thromboplastic time with kaolin: A simple screening test for first stage plasma clotting deficiencies. Am J Clin Pathol 35:212, 1961 I 1. Pearson RW, Triplett DA: Factor XI assay results in the CAP Survey ( 198 1 ). Am J Clin Pathol 78:6 15, 1982 12. Saito H, Goldsmith AH Jr: Plasma thromboplastin antecedent (PTA factor Xl): A specific and sensitive radioimmunoassay. Blood 50:377, 1977 REFERENCES 13. Fisher CA, Schmaier AH, Addonizio VP, Colman RW: Assay of prekallikrein in human plasma: Comparison of amidolytic esterolytic, coagulation and immunochemical assays. Blood 59:963, I 982 14. Schapira M, Silver LD, Scott CF, Colman RW: New and rapid functional assay for Cl-inhibitor in human plasma. Blood 59:719, 1982 I 5. Kirby EP, McDevitt PJ: The binding of bovine factor XII to kaolin. Blood 61:652, 1983 16. Laemmli UK: Cleavage of structural proteins during assembly of the head of bacteriophage T4. Nature 227:680, 1970 17. Bouma BN, Griffin JH: Human blood coagulation factor XI: Purification properties and mechanism of activation by activated factor XII. J Biol Chem 252:6432, 1977 18. Bolton AE, Hunter WM: The labeling of proteins to high specific radioactivities by conjugation to a 251-containing acyling agent. Application to the radioimmunoassay. Biochem J 133:529, I 973 19. Lipscomb MS, Walsh PN: Human platelets and factor XI: Localization in platelet membranes of factor XI-Iike activity and its functional distinction from plasma factor XI. J Clin Invest 63:1006, I 977 20. Friberger P: Chromogenic peptide substrates. Their use for the assay of factors in the fibrinolytic and plasma kallikrein-kinin systems. Scand J Clin Lab Invest (Suppl) 162:1, 1982 21. Colman RW, Mattler L, Sherry 5: Studies on the prekallikrein (kallikreinogen) kallikrein enzyme system of plasma. II. Evidence relating the kaolin-activated arginine esterase to plasma kallikrein. J Clin Invest 48:23, 1969 22. Schiffman 5, Pecci R, Lee P: Contact activation of factor XI: Evidence that the primary role of contact activation cofactor (CAC) is to facilitate the activation of factor XII. Thromb Res 1:319, 1977 23. Weiss AS, Gallin JI, Kaplan AP: Fletcher factor deficiency.

50 SCOTT CT AL. Abnormalities of coagulation, fibrinolysis, chemotactic activity and kinin generation attributable to absence of prekallikrein. J Clin Invest 53:622, 1974 24. Scott CF. Schapira M, Colman RW: Effect of heparin on the inactivation rate of human factor XIa by antithrombin III. Blood 60:940, 1982 25. Lanchantin GF, Friedmann JA, Hart DW: Interaction of soybean trypsin inhibitor with thrombin and its effect on prothrombin activation. J Biol Chem 244:865, 1969 26. Stern DM, Nossel HL, Owen J: Acquired antibody to factor XI in a patient with congenital factor XI deficiency. J Clin Invest 69:1270, 1982

1984 63: 42-50 Amidolytic assay of human factor XI in plasma: comparison with a coagulant assay and a new rapid radioimmunoassay CF Scott, D Sinha, FS Seaman, PN Walsh and RW Colman Updated information and services can be found at: http://www.bloodjournal.org/content/63/1/42.full.html Articles on similar topics can be found in the following Blood collections Information about reproducing this article in parts or in its entirety may be found online at: http://www.bloodjournal.org/site/misc/rights.xhtml#repub_requests Information about ordering reprints may be found online at: http://www.bloodjournal.org/site/misc/rights.xhtml#reprints Information about subscriptions and ASH membership may be found online at: http://www.bloodjournal.org/site/subscriptions/index.xhtml Blood (print ISSN 0006-4971, online ISSN 1528-0020), is published weekly by the American Society of Hematology, 2021 L St, NW, Suite 900, Washington DC 20036. Copyright 2011 by The American Society of Hematology; all rights reserved.