Dec 27, 1988 - Sigma Chemical Co. Human fibrinogen was cleaved with CNBr (i g offibrinogen and 1.3 g of CNBr in 100 mL of. 700 g/L formic acid reagent) for ...
CLIN. CHEM. 35/5, 787-793 (1989)
Standardization of Methods for Measuring Plasminogen Activator Inhibitor Activity in Human Plasma WayneL. Chandler,1 Swee-ChIn Loo,1 Son
V.
Nguyen,1Gottfrled Schmer,1 andJohn R. Stratton2
We have standardized the measurement of plasminogen activator inhibitor type 1 (PAl-i) activity in plasma. One-chain tissue-type plasminogen activator (t-PA; EC 3.4.21.31; final activity, 5 mt. units/mL) was incubated with plasma (final dilutions 1:4 to 1:40) in phosphate buffer (pH 7.4, ionic strength = 0.15) for 15 miri at37 “C, followed by acidification and measurement of residual t-PA activity by art amidolytic method. The PAl-i activity assay was 98% specific for PAl-i activity in samples from both pregnancy and nonpregnancy, and varied linearly with added plasma volume when the percent inhibition of t-PA was between 8% and 50%. For the standardized method, analytical recovery was 93 ± 5%, the detection limit was 1.6 arbitrary units per milliliter (1 arb. unit of PAl-i activity = inhibition of 1 mt.unit of t-PA activity), and total imprecision was 10.2 (SD 0.7) arb. units/mL (CV = 7%, n = 20). The average PAl-i activity in 10 healthy individuals drawn between 0800 and 1000 hours was 23.9 ± 15.4 arb. units/mL. Compared with the standardized assay, two of three previously described assays underestimated PAl-i activity in plasma by 77% and 85%, respectively.
AddItional Keyphrases: pregnancy
.
clotting
.
reference in-
teival
individuals should be similar for different assays, yet reported activities have ranged from a low of 0.6 to a high of 24.5 arb. units/mL, a more than 40-fold difference (2, 3). The purpose of this study was not to develop a new assay for PM-i activity, but instead to standardize the current methods based on the use of t-PA (in contrast to urokinase assays). To do this, we evaluated the effects of t-PA type (one-chain vs two-chain) and concentration, pH and time of incubation, plasma dilution, and type of anticoagulant. On the basis of these results we propose here a standardized assay of PM-i activity and evaluate its specificity, recovery, detection limit, and imprecision. Finally, we compare results by this standardized assay with those by three previously described assays (2-4).
Materials and Methods Human Subjects
Studies with human subjects were carried out according to the principles of the Declaration of Helsinki. Informed consent was obtained from all participants, and the study was approved by the University of Washington Human Subjects Review Committee. Materials
Plasminogen activator inhibitor type 1 (PM-i) is a plasma protein that rapidly and specifically inhibits tissue-type plasminogen activator (t-PA; EC 3.4.21.31) and urokinase (1). In theory, measurement of PM-i activity in plasma should be relatively simple: an excessof pure t-PA is added to plasma, the mixture is incubated to allow the PM-i and tPA to react, and the residual t-PA activity is determined. The difference between the initial and final t-PA activity values is equal to the PM-i activity. One arbitrary unit of PM-i activity is defined as the amount of PM-i that inhibits 1 international unit of t-PA activity under the specified analytical conditions. Seemingly simple, this assay has proven difficult to standardize. For instance, what constitutes an “excess”of added t-PA activity? How long and under what conditions should the t-PA and PM-i be incubated? How should the t-PA activity be measured? What type of t-PA should be used, one-chain or two-chain? Several different assays have been proposed that produce widely varying results. For example, the estimated average activity of PM-i in plasma from healthy nonpregnant
‘Department of LaboratoryMedicineand “Division of Cardiology, Department of Medicine, University of Washington;and” Seattle Veterans Administration Medical Center, Seattle, WA. Address for correspondence: Department of LaboratoryMedicine, SB-b, University of Washington,Seattle, WA 98195. “Nonstandard abbreviations:HIJVE, human umbilical vein endothelium; PAl-i, plasminogen activator inhibitor type 1; PBS, phosphate-buffered isotonic saline; SDR, standard deviation of the regression; SDS-PAGE, sodium dodecyl sulfate/polyacrylamidegel electrophoresis; and t-PA, tissue plasminogenactivator. Received December 27, 1988; accepted February 13, 1989.
Cyanogen bromide (CNBr), bovine albumin (98% pure, salt free), and Triton X-iOO surfactant were from Sigma Chemical Co., St. Louis, MO. The chromogenic substrate D-
valyl-L.-phenylalanyl-L-lysyl-p-nitroarnline (S-2390) was obtained from Helena Diagnostics, Beaumont, TX. All other materials not described below were reagent or analytical grade. Plasminogen. Human glu-plasminogen (product no. 400, lot no. G0187 and (30287) was from American Diagnostica Inc., Greenwich, CT. The specific activity of the plasminogen preparation was approximately 4.7 units/mg, where 1 unit = the plasminogen activity in 1 mL of pooled normal plasma (5). Sodium dodecyl sulfate/polyacrylamide gel electrophoresis (SDS-PAGE) of reduced and nonreduced samples of the plasminogen preparation demonstrated a single band, Mr 90 000. There was no evidence of lys-plasminogen bonds. Fibrinogen and CNBr-cleaved fibrinogen. Human fibrinogen (96% clottable, product no. F4883, lot no. 67F-9461) was from Sigma Chemical Co. Human fibrinogen was cleaved with CNBr (i g of fibrinogen and 1.3 g of CNBr in 100 mL of 700 g/L formic acid reagent) for 17 h at room temperature. This was followed by dialysis twice for 4 h against 100 volumes of distilled water at 25 #{176}C and once overnight at 4#{176}C against 100 volumes of 75 mmol/L Tris acetate buffer (pH 8.2 at 37 #{176}C) containing i g of Triton X-i00 per liter (6). Tissue plasminogen activator. One-chain melanoma-derived t-PA (product no. 115, lot no. 7375, specific activity 670 000 mt. units/mg), two-chain melanoma-derived t-PA (product no. 116, lot no. 1047c, 750 000 mt. units/mg), and polyclonal goat anti-melanoma t-PA IgG (product no. 387, lot no. 363) were from American Diagnostica Inc. SDS-PAGE of reduced samples of the t-PA demonstrated a single band CLINICALCHEMISTRY, Vol. 35, No. 5, 1989 787
(Mr 70 000) for the one-chain t-PA and two
bands
(Mr
-30 000 and 40 000) for the two-chain t-PA. The specific activity of the t-PA preparations was determined by comparison with human melanoma t-PA reference standard, lot 83/5 17 from the National Institute for Biologic Standards and Control, London, U.K. (7). Plasminogen activator inhibitor. Human fibrosarcoma derived PM-i (product no. 109, lot no. 3, 110 mgfL) and goat polyclonal anti-PAl-i IgG (product no. 395G, lot no. 181) were from American Diagnostica Inc. Serum-free 24-h conditioned media from human umbilical vein endothelium (HuvE) stimulated with 10 mgfL endotoxin reagent was supplied by Dr. John Harlan (Seattle, WA). PM-i activity in the HUVE-conditioned media was reactivated with guanidine HC1 as described previously (8). Briefly, the procedure was as follows. We dialyzed HuvE-conditioned media against 100 volumes of 5 mol/L guanidine HC1 reagent in PBS-Triton-X buffer [per liter: 10 mmol of sodium phosphate (pH 7.4 at 25 #{176}C), 120 mmol of NaCl, 2.7 mmol of KC1, 1 g of Triton X-100J for 4.5 h at 37 #{176}C, followed by dialysis against 4000 volumes of PBS-Triton-X buffer for 18 h at 4#{176}C. After reactivation, the HuvE-conditioned media contained 1050 arb. units of PM-i activity per milliliter. Blood Sampling and Sample Preparation
acid reagent to the reaction solution, then measure the absorbance at 405 rim [we used a Cobas Bio centrifugal analyzer (Roche Analytical Instruments, Nutley, NJ)]. Determine the residual t-PA activity by comparing the change in absorbance of the plasma-containing samples with that of the t-PA standards (no plasma added). Comparison of different incubation solutions. We used two different
incubation
buffers
to dilute
the plasma
samples
and t-PA standards: PBS-Triton-X buffer containing i g of bovine albumin and 0.6 mmol of sodium aside per liter or 75 mmoliL Ti-is acetate buffer (pH 8.2 at 37#{176}C), containing 1 g of Triton X-iOO, 1 g of bovine albumin, and 0.6 mmol of sodium aside per liter. Final dilutions of plasma after t-PA was added ranged from 1:5 to 1:1280. To determine whether plasmin was being generated during the initial incubation step from plasminogen in the patient’s plasma, we added a plasmin-sensitive chromogenic substrate, n-valyl-L-phenylalanyl-L-lysyl-p-nitroaniline (0.3 mmol/L final concentration), to the initial incubation solutions described above. The plasma and t-PA were then diluted with these modified buffers, combined, and incubat-. ed for 60 mm at 37 “C. Plasmun formation during the incubation was monitored by measuring the absorbance of the solution at 405 mm. Immunoinhibition of PM-i activity. PAl-i-specific activi-
ty in samples of diluted plasma and in a preparation of Venous plasma used in developing the standardized assay purified PM-i was immunoinhibited by incubating the of PM-i activity was obtained from healthy individuals. sample with anti-PM-i IgG, 20 mg/L, for 10 mm at 37 #{176}C Pregnancy plasma samples were obtained during the third trimester, usually within one week of delivery. Blood sambefore adding the t-PA. This inhibits >98% of the PM-i ples were anticoagulated by adding 4.5 mL of whole blood to activity in a sample. We estimated PM-i specific activity as the difference between total t-PA inhibition and t-PA inhibi0.5 mL of 130 mmoIIL sodium citrate solution or by adding 5 tion with a saturating concentration (20 mg/L) of anti-PM-i mL of whole blood to 50 tL of 150 g/L K3EDTA solution. All samples were centrifuged for 10 mm at 1000 x g at room antibody. temperature. Plasma was removed and frozen at -60 #{176}C Incubation time. “Optimum incubation time” was defined as the minimum time necessary for the PM-i activity in the until analyzed. sample and the added t-PA to react to completion. To Assay Procedures determine this, we incubated diluted plasma and t-PA as Here we describe the final assay conditions used to described above for 5, 10, 15, and 20 mm, then acidified the measure PM-i activity, then discuss the experiments persolution and measured its residual t-PA activity. The reacformed to determine these standard conditions. tion was defined as being complete when there was 50% or 80% of the original t-PA activity is inhibited, dilute the sample with sodium phosphate buffer (20 mmol/L, pH 7.3) containing 0.1 mol of NaCl per liter, and reassay. Method C (4). Add 10 zL of plasma to 240 ,UL of plasminogen-chromogenic substrate reagent (described under the standardized assay) containing from 0 to 0.6 int. units oft-PA activity per milliliter. Determine residual t-PA activity from the change in absorbance at 405 mm per hoursquared at room temperature in comparison with the change for t-PA containing no plasma. Estimate the total t-PA inhibition by determining the linear-regression line between added t-PA activity vs residual t-PA activity. To estimate PM-i specific activity, add 20 mg of anti-PAl-i IgG per liter to the plasma before adding the plasma to the plasminogen-chromogenic substrate reagent. Calculate PM-i specific activity as the difference between total t-PA inhibited with and without anti-PM-i IgG. Statistics. We tested group distributions for normality by using the Wilk-Shapiro test and by determining the coeffi-
cients of skew and kurtosis. Normal distributions are reported as mean ± SD. Differences between groups were tested
for significance by using two-tailed Student’s t-tests (paired or unpaired, depending on the specific groups). Linear regressions and correlation coefficients were calculated by conventional methods. Besides the slopeand SD of the slope, and intercept and SD of the intercept, we also calculated the SD of the regression (SDR).
Results Analytical recovery of t-PA activity during the incubation step. During the initial incubation step, the PM-i in the sample reacts with the added t-PA. This assay is based on the assumption that, if the sample contains no PM-i activity, all the original t-PA activity added will be accounted for. We tested this assumption for both one-chain t-PA and two-chain t-PA by using two different solutions to dilute the plasma sample and t-PA: PBS buffer (pH 7.4, ionic strength = 0.15) and Ti-is buffer (pH 8.2 at 37 “C, ionic strength = 0.03). Plasma samples with 50% of the original t-PA was inhibited, the assay became progressively nonlinear and began to underestimate the concentration of PM-i activity in the sample. Added t-PA activity was never completely inhibited. Even when we added a twofold excess of PM-i activity (as compared with t-PA activity) to the assay, a small but measurable t-PA activity was still found. We confirmed by immunoinhibition, using anti-t-PA antibodies,that this small residual activity was caused by t-PA. Figure 5 shows plots of original vs residual t-PA activity for a range of plasma dilutions and initial added t-PA concentrations. Use of final plasma dilutions from fivefold to 20-fold yielded parallel curves, indicating that varying the final concentration of t-PA activity in the assay from 4 to 7 nit. units/mL did not change the estimated PM-i activity in the sample. When twofold-diluted plasma was used, >50% of the original t-PA was inhibited and the curve became nonparallel with the other dilution curves, again indicating that the assay is inaccurate when >50% of the original t-PA activity is inhibited.
(‘3 0
4-i
A.
Q.E
0.05
‘S.’
4
I.-
0.00
0
10
20
30
TIME
40
50
60
0 H’
(1)1-
wz
(MINI
.
Fig. 2. Effectof incubationsolutionand pH on theformationof plasmin activity during the incubation step, as measured with the plasminsensitive chromogenic substrate D-valyl-L-phenylalanyl-L-lysyl-frni-
troaniline (0) one-chain t-PA;(0) two-chaint-PA; ( = 0.15); (-) Tris buffer of the buffersas in Fig. 1
PURIFIED
) PBSbuffer(pH7.4,ionic strength (pH 8.2, ionicstrength = 0.03).Completecomposition
B.
4-i 0_C I-...0W
O..E IS.,.
.JH
.00) I.JH
(ng)
PAl-I
I-
0 H
0 I-I
(‘31-
0)1-
wz
-
wz
10
#{149}0 4
8
12
16
PLASMA
INCUBATION
TIME
(MINI
Fig.3. Residualt-PAactivityvsincubationtimeat 37#{176}C forthe reaction between t-PA and dilutedplasma (S) 200 4. of 10-fold-dilutednonpregnancyplasmaplus 2004. of 10mt.unitsof one-chaint-PA per milliliter; (#{149}) 20 mg of anti-PAl-i tgG added per liter to the dilutedplasma before t-PA activitywas added
790
CLINICALCHEMISTRY, Vol. 35, No. 5, 1989
20
30
20 VOLUME
40
50
(uL)
Fig. 4. Linearity of t-PA inhibitionvs the amount of purified PAl-i or volumeof plasmaaddedto the assay The solidlines show the linear least-squares fit to the linearportions of thecurves. A:t-PAactivity (mt. units/rnL) = -0.224 ± 0.012(ng PAl-i) + 4.82 ± 0.08(SDR = 0.13); & t-PA activity (mt. units/mL) = -0.130 ± 0.005(4. plasma) + 4.92 ± 0.06 (SDR = 0.09)
WA
OP&V
PLASMA
4._I
QE IN
430)
1: 20 10
PLASMA 1: PLASMA
1: 5
PLASMA
1: 2
I-
0 H
-
(‘31I1Jz EH
0
2
4
6
8
ORIGINAL t-PA (INT. UNITS/mL) Fig. 5. Relationship between originalt-PA activity, plasma dilution, and residual t-PA activity
Plasma dilutions that inhibited less than 8% of the original t-PA resulted in imprecise estimates of PM-i activity. Thus, the useful range of t-PA inhibition in the assay was 8% to 50%. Plasma dilutions resulting in inhibition of t-PA activity outside these ranges were not used in the estimation of PM-i activity. We selected 10 mt. units of t-PA activity per milliliter as the standard t-PA concentration for use in the assay. When mixed 1:1 with diluted plasma this results in a t-PA activity of 5 int. units/mL in the incubation step. Assuming that plasma is diluted at least fourfold, the minimum ratio oft-PA activity to plasma volume in the assay is 20 mt. units of t-PA per milliliter of plasma added. For most plasma samples, final plasma dilutions of four- to 40-fold were used; with this range of dilutions, samples with PM-i activities from i.6 to 100 arb. units/mL could be measured. Comparison of EDTA vs citrate anticoagulants. Average PM-i activity in plasma from iO healthy individuals was 23.9 (SD 15.4) arb. units/mL for citrate-anticoagulated plasma (corrected for citrate dilution of plasma) and 25.0 (SD 14.2) arb. units/mL for EDTA-anticoagulated plasma. The difference between PM-i activity measured in citrated vs EDTA plasma was of borderline significance (P = 0.049, paired
t-test).
Comparison of results with one-chain vs two-chain t-PA. Average PM activity determined in 12 samples of nonpregnancy plasma was not significantly different (P = 0.95, paired t-test) when measured with one-chain t-PA, 17.5 (SD 9.5) arb. units/mL, as compared with two-chain t-PA, 17.6 (SD 9.5) arb. units/mL. In contrast, average PM activity determined in 18 samples of pregnancy plasma was lower when measured with one-chain t-PA, 32.1 (SD 20.4) arb. units/mL, as compared with two-chain t-PA, 38.0 (SD 21.9) arb. units/mL (P = 0.0001, paired t-test). The ratio of PM-i activity determined by using one-chain vs two-chain t-PA was 1.00 (SD 0.07) for the nonpregnancy samples and 0.81 (SD 0.10) for the pregnancy samples (P = 0.0001, unpaired t-test). Thus, on average, pregnancy plasma inhibited significantly more two-chain t-PA than one-chain t-PA as compared with nonpregnancy plasma. When measured with either one-chain t-PA or two-chain t-PA, anti-PM-i IgG inhibited >98% of the total PAl activity in nonpregnancy plasma. When we used one-chain t-PA, anti-PM-i IgG inhibited more than 98% of the total
PAl activity in pregnancy plasma. In contrast, when twochain t-PA was used in the assay, pregnancy plasma contained up to i2 arb. units of PM activity per milliliter that could not be inhibited with anti-PM-i IgG. This suggests that other t-PA inhibitors, such as PM-2, are interfering in the assay when PM activity is measured in pregnancy plasma with two-chain t-PA, and that the assay is specific for PM-i activity when one-chain t-PA is used. The current assay is most specific for PM-i activity when the plasma contains >5 arb. units of PM-i activity per milliliter. Samples with
