Fibrinolysis
30 June 2006
Oslo Kongressenter, Norway
Chair: O. Matsuo, Japan
Co-chairs: C. Dempfle, Germany; D. Hendriks, Belgium; C. Longstaff, UK; M. Nesheim, Canada
Section I TAFI /CPU
Dr Nesheim reviewed the structure and function of TAFI and discussed the mechanism of the prolongation of lysis time resulting from TAFI activity. Functional methods for measuring active TAFI (TAFIa) were reviewed. Three assays were described: 1. A colorimetric assay was described using the substrate anisoylazoformyl arginine (AAFR); 2. A functional assay based on FDP cofactor activity with a fluorescent plasminogen substrate and bat t-PA (DSPA); and 3. A functional assay based on the direct binding of fluorescent plasminogen to FDPs. The colorimetric assay involving AAFR is simple and shows a good dose response but is not so sensitive. In the functional assay with DSPA the normal background level of TAFIa from 6 subjects gave a mean of 11.7 ± 3.6 pM, which is only 0.2% of the circulating zymogen concentration. Thus the method is sensitive, specific and accurate, however it is technically demanding and requires specialized reagents not commercially available. Dr Nesheim described the functional assay based on the binding of fluoresecent plasminogen to FDP which was also sensitive, accurate and precise and is a one step, quick and easy method. This method should be suitable to use on plasma samples and should be free of interference from t-PA and fibrinolytic inhibitors. Dr Nesheim expressed a wish to collaborate with other workers on these assays using clinical and experimental specimens and invited them to contact him.
Drs Willemse and Hendrick introduced activity based proCPU/TAFI assays with regard to the importance of polymorphism and substrate specificity. The principle of this assay is the cleavage of hippuryl-L-rginine to hippuric acid and arginine. It was shown that the threonine 325 isoleucine polymorphism has a significant effect on CPU stability such that the thr325thr variant has a half-life of 8 minutes versus 15 min for the ile325ile variant, although there was no difference in the activation kinetics. Intrinsic activity of proCPU was analysed using Bz-AA-Arg where AA is substituted by a range of amino acids. Ala and Met were the residues showing highest intrinsic activity of proCPU. Conditions to optimize proCPU activation kinetics were explored that minimized genotype dependent artifacts. Recommendations included activation at room temperature rather than 37ºC and the use of high substrate concentrations.
Drs Gils and Declerk demonstrated ELISA methods using monoclonal antibodies for the detection of TAFI and TAFIa and isoforms. An ELISA with MA-T32F6/MAT9G12-HRP was used to measure TAFI and showed no difference between a control group and patient groups. Different combinations of antibodies were sensitive to the presence of different isoforms of TAFI. Isoform of The325 Ile was not recognized by some antibodies leading to underestimates of TAFI levels in plasma. Surprisingly some polyclonal antibodies were also sensitive to different TAFI isoforms. Monoclonal antibodies in 144 combinations were screened to detect different forms of TAFI including intact zymogen, activation peptides and active TAFIa. Thus selecting different combinations of antibodies could be used to monitor the process of TAFI activation and decay of activity. The clinical application of these assays can be used to shed light on the role of TAFI in vivo.
Section II Standardization of fibrinolytic factors
Dr Longstaff presented a summary of a recent study to measure t-PA antigen in plasma. This was a follow-up of a smaller study presented last year on t-PA antigen and PAI-1 antigen and activity. The aims of the present study were to assign an agreed value of tPA antigen to in the SSC plasma lot 3 and in a spiked plasma preparation 94/730. The previous 2 nd IS for t-PA activity was also included as this was apparently used in the past by kit manufacturers to calibrate their standards. Results were presented from 14 groups comprising 8 different methods (in-house and commercial kits). As in the previous study there was significant variability in values for SSC lot 2 and lot 3 with a mean value close to 3 ng/ml (normal value less than 10 ng/ml) and %gcv around 70%. A mean value of 25.26 ng/ml, in line with expectations, was calculated for 94/730 (%gcv 21.0 for 12 labs after removal of 2 statistical outliers). A mean value of 1.5 µg/ ml was derived for 86/670 also in line with the expectations. Recalculation using 94/730 as a common standard of t-PA antigen in SSC plasma lot 2 and 3 resulted in no change in t-PA antigen concentration and improved %gcv marginally by around 10%. There were clear differences between methods for all samples, however, results could be harmonized using a method-specific correction factor. In this way lot 2 could be used to determine the correction factor to apply to the results for lot 3 and vice versa. After this procedure the value of t-PA antigen in lot 3 was still very close to 3 ng/ml but the %gcv fell to 18.2. Results from earlier studies were also presented which showed t-PA antigen in plasma were close to 3ng/ml and in 94/730 was close to 25 ng/ml. Dr Longstaff asked if there were objections to proceeding with proposals that SSC plasma lot 3 could be assigned a t-PA antigen level of 3 ng/ml and 94/730 could proceed as a proposed international standard with a t-PA antigen of 25 ng/ml. No objections were raised.
In a second talk, Dr Longtaff summarized the results from the earlier studies including PAI-1 and highlighted the difficulties found. Recent results from NIBSC on the 1 st IS for PAI-1 suggested that this preparation was stable, after low values for antigen were determined in the earlier studies. Proposals were presented on future studies aimed at calibrating SSC plasmas for PAI-1 which could be performed using a range of plasma samples with different PAI-1 activities as a means of harmonizing different methods. These studies were planned for 2006/7.
Dr Longstaff presented an update on work involving recombinant streptokinase which generates discrepant results in different assay formats relative to the 3 rd IS for Streptokinase. These observations have significant implications for the assignment of potency values to streptokinase therapeutic products and correct dosing for treatment of myocardial infarction. Data were presented from collaborative work with an Indian Biotchnology company (Biocon) on recombinant streptokinase suggesting that an N-terminal methionine, in place of the expected isoleucine, present as a result incomplete processing of the protein in E. coli . was the source of the problems. A modified recombinant variant without the N-terminal methionine was demonstrated to show no evidence of discrepancies in assays with or without fibrin present. In light of other results where the N-terminal sequence has been modified in recombinant streptokinase it is not unexpected that changes in activity might be seen in the presence of fibrin. It was concluded that manufacturers of recombinant streptokinase should ensure that the N-terminal sequence of their protein is correct if they are to correctly assign a potency using the 3 rd IS for Streptokinase
Section III D-dimer assays
Dr. Carl-Erik Dempfle on behalf of the FACT study group presented first the review and then possible proposal. Based on the work of Patrick Gaffney on the structure of fibrin degradation products, assays based on monoclonal antibodies generated by immunization with fibrin fragment D-dimer have been available since 1983. A variety of monoclonal antibodies have been developed, and assay technology has shifted from manual latex agglutination assays and microtiter plate ELISAs to quantitative latex particle assays with photometric detection and rapid fluorometric immunoassays.
In clinical plasma samples, D-dimer assays mainly detect high molecular weight crosslinked fibrin complexes in addition to fibrin degradation products. The consensus statement from the preceding ISTH SSC meeting concerning the definition of D-dimer antigen is as follows: ‘D-dimer antigen indicates antigenic material detected by use of monoclonal antibodies generated by immunization with fibrin fragment D-dimer and related compounds. The minimal structure detected is fibrin fragment D-dimer, but larger compounds containing dimerized D-domains are detected as well.’
Various calibrators are being used for D-dimer assays, including fibrin fragment D-dimer, terminal plasmin digests of crosslinked fibrin clots, plasmin digests of crosslinked fibrin clots with digestion stopped before the terminal stange, and plasma pools from patients with high levels of D-dimer antigen. Although fibrin fragment D-dimer and terminal digests of fibrin clots work well with some assays, others display a totally different reactivity with this material than with clinical plasma samples, resulting in over- as well as underestimation of D-dimer antigen levels in the clinical plasma samples. Therefore, the calibrator should contain a physiological array of crosslinked fibrin derivatives. This can be achieved either by using pooled plasma from patients with high levels of D-dimer antigen, or in vitro-preparations containing fibrin derivatives with similar composition of fibrin derivatives. According to the results of the Fibrin Assay Comparison Trial (FACT) part 4, a pooled plasma from patients with disseminated intravascular coagulation (DIC) shows identical performance with all 28 D-dimer assays tested both using serial dilutions with plasma from healthy blood donors, and buffer. Therefore, only a single calibrator pool plasma is needed, which may be diluted with assay-specific diluents.
Since patient plasma pools, as well as in vitro fibrin preparations may be heterogeneous, a reliable procedure for assigning D-dimer concentration values is needed. The FACT working group suggests the following approach: Aliquots of the plasma pool or fibrin preparation are incubated with a high concentration of plasmin in presence of calcium and a thrombin inhibitor. Both fibrin, and fibrinogen in the sample are degraded, resulting fibrin fragment D-dimer/E complex as main terminal breakdown product of the crosslinked fibrin. Proteolysis of the fibrinogen, in contrast, yields fibrinogen fragments D and E. After ensuring that proteolysis is complete by SDS-polyacrylamide gel electrophoresis and immunoblotting, using polyclonal anti-fibrinogen antiserum for detection, concentration of fibrin fragment D-dimer is measured. For measurement of fibrin fragment D-dimer, several D-dimer assays are used which are not influenced by presence of fibrinogen degradation product D (FDP-D) and show good reactivity with fibrin fragment D-dimer. These assays are calibrated with purified fibrin fragment D-dimer. The resulting concentration levels of fibrin fragment D-dimer reflect the total concentration of D-dimer antigen (dimerized D-domains) in the original plasma sample.
By means of plasmin proteolysis, the D-dimer antigen is ‘homogenized’ and concentration measurment is made independent of the molecular size and composition of the fibrin compounds containing the dimerized D-domains. This allows calibration with fibrin fragment D-dimer/E complex as a well-defined primary reference material.
The measured D-dimer value is then assigned to the pooled plasma, which is then used for calibration of the D-dimer assays.
The following round of FACT (FACT-5) will require calibration of the participating assays with the common calibrator pool plasma and will include a set of clinical plasma samples from patients with DIC, DVT and pulmonary embolism. The study will evaluate the effect of common calibration of D-dimer assays on the conformity of the D-dimer assay results.
Section IV General discussion
For the activity in the following term, the participants are requested to contact with speakers to collaborate in each item. The number of the participants is estimated about 150 at the end of the subcommittee.
D-dimer session with Fibrinolysis, DIC and Hemostasis and Malignancy was held after Fibrinolysis subcommittee, and this was successful as the first trail.