A highly sensitive fluorometric assay for determination of human coagulation factor XIII in plasma

ANALYTICAL BIOCHEMISTRY Analytical Biochemistry 367 (2007) 152–158 www.elsevier.com/locate/yabio

A highly sensitive fluorometric assay for determination of human coagulation factor XIII in plasma Kai Oertel a, Andreas Hunfeld b, Edgar Specker a, Christine Reiff a, Rainer Seitz b, Ralf Pasternack a, Johannes Dodt b,* a b

N-Zyme BioTec, D-64295 Darmstadt, Germany Paul-Ehrlich-Institut, D-63225 Langen, Germany Received 18 December 2006 Available online 17 May 2007

Abstract Based on the isopeptidase activity of human plasma FXIII, a novel fluorometric assay that determines FXIII concentrations in human plasma below 0.05 IU/ml is introduced. We considered a peptide sequence derived from a2-antiplasmin (n = 12) to yield high sensitivity. Peptide Abz-NE(Cad-Dnp)EQVSPLTLLK exhibits a Km value of 19.8 ± 2.8 lM and is used in a concentration of 50 lM. The assay design is suitable for measurements in cuvettes (1 ml volume) as well as for the microtiter plate (MTP) format (0.2 ml volume). It provides linear dose–response curves over a wide range of FXIII concentrations (0.05–8.8 IU/ml). The assay was validated with respect to precision, detection and quantitation limits, accuracy/specificity, linearity, and range. A comparison of the fluorometric assay with the photometric assay for FXIII determinations in plasma pools as well as single donor plasma revealed suitability of the fluorometric assay for FXIII determination in plasma of healthy individuals. FXIII concentrations in plasma samples of patients with severe FXIII deficiency are discussed in the context of FXIII antigen levels. These assays correlate well in the critical range below 0.1 IU/ml, whereas the photometric assay may overestimate residual FXIII activity in severe FXIII-deficient patients. Ó 2007 Elsevier Inc. All rights reserved. Keywords: Transglutaminase; Human FXIIIa; Human FXIII; Human plasma; Fluorometric assay; Isopeptidase activity

Transglutaminases (TGases)1 catalyze an acyl transfer reaction in which the carboxamide group of a glutamine residue is the acyl donor and a primary amine is the acyl acceptor. Under physiological conditions, the glutaminecontaining protein or peptide (glutamine substrate) binds highly specifically to its catalyzing TGases, whereas various amine substrates can serve as less specific acyl acceptors, leading to isopeptide bond formation. In the first step, *

Corresponding author. Fax: +49 6103 771250. E-mail address: [email protected] (J. Dodt). 1 Abbreviations used: TGase, transglutaminase; Cad-Dnp, N-(2,4-dinitrophenyl)cadaverine; FRET, fluorescence resonance energy transfer; Abz, 2-aminobenzoyl; Dns, dansyl (or 5-(diemthylamino)-1-naphthalenesulfonyl); Dnp, 2,4-dinitrophenyl; ELISA, enzyme-linked immunosorbent assay; MTP, microtiter plate; UFH, unfractionated heparin; LMWH, low-molecular weight heparin; PEG, polyethylene glycol; Ahx, e-aminocaproyl (or 6-aminohexanoyl); Cbz, benzyloxycarbonyl. 0003-2697/$ - see front matter Ó 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.ab.2007.05.011

the active site cysteine of TGases attacks as a nucleophile the carboxamide function of the glutamine substrate, resulting in the formation of a thioester and the release of ammonia as the driving force of the reaction. In the second step, the acyl group is transferred to the e-amino group of a lysine that serves as the acyl acceptor [1]. If no suitable amine substrate is available, a deamidation with water to glutamic acid occurs [2]. The covalent e(c-glutamyl)lysyl isopeptide bond is stable against further enzymatic degradation and plays an important role in physiological processes where rigidity between polypeptide chains is required. Human blood coagulation FXIII belongs to the TGase family. One of the main physiological functions of this enzyme is to catalyze the crosslinking of fibrin and the incorporation of protein substrates into the fibrin network in the last step of blood coagulation [3]. For many years,

Assay for determination of human coagulation factor XIII / K. Oertel et al. / Anal. Biochem. 367 (2007) 152–158

there has been considerable clinical interest in the development of novel methods for measuring FXIII activity in plasma samples. The benefits for patients with inherited or acquired FXIII deficiency are obvious. Estimating FXIII activity, however, is a challenging task for monitoring hemostasis and wound healing in such patients and, if necessary, for providing these patients with therapeutic FXIII preparations to avoid severe bleeding problems. A number of assays have been developed for measuring FXIII activity [4–10]. These assays can be divided into two major categories using either the endogenous TGase activity of FXIII or the isopeptidase activity of FXIII. A coupled optical assay based on the TGase activity of FXIII was introduced by Fickenscher and coworkers in 1991 [8]. The principle of this method is based on the measurement of ammonia release when a specific synthetic peptide is converted through catalysis by FXIII to the deamidated product or crosslinked with glycine ethyl ester. In clinical and analytical laboratories, the photometric FXIII assay is one of the most widely used and accepted tests for routine determination of FXIII concentration in plasma samples. It is also the standard test for FXIII in fibrinogen components of fibrin sealant kits described by the European Pharmacopoeiea [11]. This assay was modified and improved by Ka´rpa´ti and coworkers [9], who used an a2-antiplasmin-derived substrate peptide. In 1997, Parameswaran and coworkers discovered a novel and unexpected feature of FXIII activity, the so-called ‘‘isopeptidase’’ activity [10], which is characterized by a release of a primary amine followed by incorporation of water or primary amines (R2NH2) into the glutamine substrate (Fig. 1). Parameswaran and coworkers demonstrated the exchange of primary amines such as N-(2,4-dinitrophenyl)cadaverine (Cad-Dnp) in their synthesized peptides [10]. Their detection method is based on a fluorescence resonance energy transfer (FRET) system [12]. The increase of fluorescence results from N-terminal attached fluorophores, such as 2-aminobenzoyl (Abz) and 5-(dimethylamino)-1-naphthalenesulfonyl (Dns), which are internally quenched by the 2,4-dinitrophenyl (Dnp) moiety before hydrolysis of the peptide substrate. Although the cleavage

153

of 1,5-diaminopentane as N-alkyl spacer between the peptide backbone and the quenching moiety Dnp is catalyzed by FXIII, the concentration of FXIII (0.5–8.0 lM) used by Parameswaran and coworkers is beyond the physiological concentration (1 IU/ml corresponds to 50 nM plasma FXIII) for the tetrameric complex in normal human blood [13]. In comparison with the photometric assay, the FXIII isopeptidase method does not rely on a subsequent enzymatic step and is simple and direct. Our aim was to develop a sensitive assay for plasma FXIII based on the isopeptidase activity that allows determination of plasma FXIII in the pathophysiological range (<0.1 IU/ml). Materials and methods Starting materials for organic synthesis were purchased by Sigma (Munich, Germany) and used without further purification. Salts and buffers were purchased from Merck (Darmstadt, Germany) or Sigma. Protease-free bovine serum albumin was supplied by Roth (Karlsruhe, Germany), glycine methylester was supplied by Bachem (Weil am Rhein, Germany), FXIII antigen enzyme-linked immunosorbent assay (ELISA) was supplied by Haemochrom (Essen, Germany), and the photometric FXIII assay ‘‘Berichrom FXIII’’ was supplied by DADE Behring (Schwalbach, Germany). Microtiter plates (MTPs, 96 wells) for the fluorometric assay were Nunc flat (Wiesbaden, Germany). ELISA tests were performed with Immulon 4HBX (Thermo Electron, Langenselbold, Germany), and the photometric assay was performed with PS Microplate F (Greiner, Solingen, Germany). Human a-thrombin, human fibrinogen, and human FXIII were purchased from ZLB Behring (Marburg, Germany). The fifth international standard for unfractionated heparin (UFH, code 97/578), the first international standard for low-molecular weight heparin (LMWH, code 85/600), and the first international standard for factor XIII, plasma (code 02/206) with assigned potencies of 0.91 IU/ml for FXIII and 0.93 IU/ml for FXIII antigen (A2B2 complex), were obtained from the National Institute for Biological Standards and Control (Potters Bar, UK).

O2N H N

O

NO2

NH

OH or NHR2

O

n = 1,2

n = 1,2

(Xaa)n fluorophore-NH

peptide fluorophore-NH O

O

(Xaa)n

N H

O

O

FXIIIa

+

peptide N H

+

O2N

H2O or R2- NH2 H2N

NH

NO2

Fig. 1. Molecular structure of FXIII isopeptidase substrates. The peptide sequence is modified with a fluorophore/quencher pair. During the isopeptidase reaction, the quencher is released and water or primary amines (R2-NH2, e.g., glycine methylester) are incorporated into the glutamine substrate. The release of the quencher (Cad-Dnp) is accompanied by an increase of fluorescence.

154

Assay for determination of human coagulation factor XIII / K. Oertel et al. / Anal. Biochem. 367 (2007) 152–158

Fluorometric FXIII assay in cuvettes The synthesized peptides were dissolved in dimethyl sulfoxide to a final concentration of 56.8 mM (stock solution), which subsequently was diluted 1:1000 with reagent solution containing 56.8 mM Tris, 11.36 mM CaCl2, 113.6 mM NaCl, 0.113% (w/v) polyethylene glycol (PEG) 8000, 0.2% sodium azide, 5.68 mM H-Gly-OMe Æ HCl, 5.68 mg/L hexadimethrinbromide, and 1.1 mg GPRPNH2 Æ 2HCl. Reagent solution (pH 7.5, 880 ll) was prewarmed to 37 °C, and the enzymatic reaction was started by the addition of 20 ll thrombin solution (10 IU/20 ll in buffer) and 100 ll plasma sample. Final concentrations of each component in the assay are calculated to Tris–HCl (50 mM), CaCl2 (10 mM), NaCl (100 mM), 0.1% (w/v) PEG 8000, 0.2% sodium azide, H-Gly-OMe Æ HCl (5 mM), hexadimethrinbromide (5 mg/L), synthetic peptides (50 lM), GPRP-NH2 Æ 2HCl (2 mM), thrombin (10 IU), and FXIII activity in the range of 0.1 IU/ml (0.005 lM). Measurements were carried out using a Cary Eclipse fluorescence spectrophotometer (Varian, Darmstadt, Germany) with kex = 313 nm and kem = 418 nm. Fluorometric FXIII assay in MTPs The FXIII assay was adapted to the MTP format by reducing the total volume from 1.0 to 0.2 ml but maintaining the ratio of sample volume to reagent volume and the final concentration of reagents and buffer substances in the assay. The tests are performed at 37 °C in an MTP filter fluorometer Fluoroscan Ascent (Thermo Electron, Dreieich, Germany) or Infinite F200 (Tecan, Crailsheim, Germany) with filters for excitation at 320 nm and emission at 430 nm. In all optimization and validation experiments, pooled plasma (>100 donors) that was calibrated against the first international standard for FXIII (plasma, 02/206) was used. Three independent series of experiments were performed, and results within a series represent the means of duplicates.

national standard of factor XIII (plasma, 02/206) was used as a standard that has an assigned antigen content of 0.93 IU/ml (A2B2 complex). Results and discussion The initiative for our development of a new, more sensitive FXIII assay was fueled by the discovery of the isopepditase activity [10]. We envisaged that optimization of this method would provide a method that permitted the measurement of FXIII activity down to the low range found in congenital FXIII deficiency. First, we searched for a peptide substrate that displays a higher affinity for FXIII but with the same FRET fluorophore/quencher pair introduced by Parameswaran and coworkers [10]. In our view an a2-antiplasmin-derived peptide sequence is an optimal choice because the dodecapeptide NQEQVSPLTLLK has already been evaluated and recommended as an excellent substrate for FXIII [4]. The glutamine at position 2 serves as primary acyl donor in the FXIII-catalyzed reaction [14]. The synthesized substrates represent c-glutamyl branched peptides. Compounds 1 and 2 were synthesized according to the literature [10] and served as internal controls in comparison with a2-antiplasmin-derived peptides 3 to 7 (Fig. 2). Each peptide was evaluated under the same assay conditions at a concentration of 50 lM. For peptides 1 and 2, an increase of fluorescence was not detectable at low concentrations of FXIII (1.08 IU/ml standard plasma). Peptides 3 and 4 were labeled with the fluorophore Dns, and the distances to the quencher Dnp were varied by insertion of an asparagine between fluorophore and glutamine. Neither of the two peptides generated an increase of fluorescence. Changing the fluorophore to Abz in 5 and 6 was judged as not sufficient to gain reliable sensitivity in plasma samples with low FXIII content, although for 5 a small increase of fluorescence was observed. Because the 40 35

FXIII photometric activity and antigen assays Relative fluorescence

The photometric FXIII activity assay (Berichrom FXIII) was performed according to the manufacturer’s instructions but adapted to the MTP format at 37 °C. Using a SpectraMax (Molecular Devices, Ismaning, Germany) MTP reader at 340 nm in the kinetic mode, absorbance was read every 20 s for 30 min and the change of absorbance between 10 and 15 min was used for calculations. Dilutions of standards and samples were performed in 50 mM Tris–HCl buffer (pH 7.5) containing 2.5 mg fibrinogen. The FXIII antigen ELISA kit provided capture and detection antibodies. All steps were performed according to the manufacturer’s instructions. Readings were performed with a Sunrise MTP reader (Tecan) at 450 nm using 3,3’,5,5’-tetramethylbenzidine as substrate. The first inter-

30 25

Compound

Sequence

1

Dns-Ahx-AE(Cad-Dnp)QIV

2

Abz-Ahx-E(Cad-Dnp)QIV

3

Dns-Ahx-E(Cad-Dnp)EQVSPLTLLK

4

Dns-Ahx-NE(Cad-Dnp)EQVSPLTLLK

5

Abz-Ahx-E(Cad-Dnp)EQVSPLTLLK

6

Abz-Ahx-NE(Cad-Dnp)EQVSPLTLLK

7

Abz-NE(Cad-Dnp)EQVSPLTLLK

7

20 15

1

10

2 3, 4 5

5

6

0 0

2

4

6

8

10

12

14

Time [min]

Fig. 2. Evaluation of FXIII isopeptidase activity with substrates 1 to 7. Reactivity of 1.08 IU/ml FXIII with peptides 1 to 7 (50 lM) was evaluated. Internal standards 1 (red) and 2 (blue) and peptides 3 (yellow), 4 (magenta), 5 (brown), and 6 (green) did not show any appreciable increase of fluorescence. Only with peptide 7 (dark blue) an increase of fluorescence was detected.

Assay for determination of human coagulation factor XIII / K. Oertel et al. / Anal. Biochem. 367 (2007) 152–158

switch to fluorophore 6-aminohexanoyl (Ahx) peptides 3 to 6 with a2-antiplasmin-derived sequences did not provide a satisfying sensitivity for FXIII catalysis, we considered whether it might be useful for catalysis to increase the hydrophobicity with aromatic groups in the immediate neighborhood to the substrate glutamine. Chica and coworkers [15] emphasized in a modeling approach the importance of an aromatic benzyloxycarbonyl moiety (Cbz) in close proximity to glutamine in the peptide substrate Cbz-QG-OH for recognition and affinity to Red Sea bream TGase. Hence, peptide 7 was synthesized to evaluate the impact of a direct link of aromatic fluorophore Abz to peptide sequence NE(Cad-Dnp)EQVSPLTLLK lacking the Ahx spacer. An unexpected increase of fluorescence was obtained when peptide 7 was tested. Presumably, the catalysis is induced specifically by optimal orientation and interaction of 7 on binding to FXIII, resulting in a cleavage of the quencher Dnp. Although fluorophore and quencher are in such close proximity compared with peptides 3 to 6, only for 7 is an optimal balance of recognition and catalysis of FXIII obtained, leading to the observed increase of fluorescence. Fig. 2 summarizes the FXIII assay results for peptides 1 to 7 in a graphical form. After identification of the optimized peptide sequence, we proceeded to evaluate the suitability of the assay for clinical purposes by determining FXIII in human plasma samples from healthy individuals and patients with inherited FXIII deficiency.

155

Validation of the fluorogenic FXIII assay The method was validated in line with the recommendations of the International Conference on Harmonization (ICH) guideline on validation of analytical procedures [16]. Accuracy and specificity of the assay were evaluated by recovery of FXIII after spiking of in-house plasma samples with purified FXIII concentrates. The mean recovery was found to be 104% (minimum 98%, maximum 107%). Using plasma samples spiked with FXIII concentrate, there is a linear signal-to-analyte concentration relationship over the range of 0.1 to 8.8 IU/ml FXIII after log–log transformation of the data. The precision was determined to be 5.1% for repeatability and 3.6% for intermediate precision by performing a total of six series on 3 different days. A detection limit of 0.02 IU/ml and a quantitation limit of 0.05 IU/ml were determined from the data of the repeatability study based on the standard deviation of the blank (n = 6) and the slope of the calibration curve [16]. It is remarkable that the assay shows a limit of quantitation of 0.05 IU/ml (2.5 nM FXIII), which is in the range of FXIII levels of patients with severe FXIII deficiency. Compared with the lowest concentration used in the literature (0.5 lM FXIII) in FXIII isopeptidase assays [10], this assay exhibits a 100- to 200-fold increase of sensitivity. Comparison of the fluorometric assay with the photometric assay and antigen assay

Evaluation of assay conditions Assay conditions for the isopeptidase assay were optimized with respect to reagents that are known to be necessary for the activation of FXIII in plasma (e.g., calcium, thrombin, hexadimethrinbromide) or those that could affect the velocity of FXIIIa catalysis (e.g., glycine methylester). As a result of these studies, we recommend the use of 10 mM calcium chloride, 10 IU/ml thrombin, and 5 lg/ml hexadimethrinbromide as heparin antagonist. The glycine methylester, which functions as amine donor (Fig. 1), increases slightly the isopeptidase activity of FXIIIa in a concentration-dependent manner. Using the ester under the proposed assay conditions of 4 mM, the increase of activity is approximately 20%. Because fibrin formation could interfere with signal detection, the clot inhibitor GPRP-NH2 was added to a concentration of 2 mM that was shown to be sufficient for clot prevention by visual inspection. For the determination of the apparent Km value of the fluorometric substrate, we analyzed the dependence of the initial reaction velocity from the substrate concentration over the range of 2.5 to 50.0 lM. Initial velocity was plotted versus the substrate concentration and fitted by nonlinear regression to the Michaelis–Menten equation. An apparent Km value of 19.8 ± 2.8 lM (n = 3) was derived for the fluorometric substrate. Based on these results, the substrate was used at a concentration of 50 lM.

The reference method for FXIII determination in FXIII concentrates currently is the photometric assay [8]. It is used for screening of FXIII in patient plasma, but it has also been applied for potency assignment of the first international standard for FXIII (plasma). Therefore, we carried out an in-depth comparison of the fluorometric assay with the reference assay for the determination of FXIII levels in plasma of individual patients as well as in pooled plasma. First, we used plasma pools because we assumed that differences of individual plasma samples should be compensated, and therefore we expected a good correlation and agreement of results. For each pool, two independent series with four dilutions were tested with both assays, and results were calculated according to the European Pharmacopoeia using the parallel line assay [17]. In Fig. 3A, the means of the two estimates obtained for both tests are compared for 19 plasma pool samples. The coefficient of correlation between both assays is 0.95 (95% confidence interval, range = 0.88–0.98). From an inspection of the individual data, one might conclude that the fluorometric assay yields slightly higher FXIII potencies than does the photometric assay (mean difference = 0.03 IU/ml). However, it can be seen that the variability of the differences between both methods is small. For individual plasma samples from 22 healthy individuals, the situation is slightly different (Fig. 3B). The coefficient of correlation is 0.77 (95% CI, range = 0.50–0.90),

Assay for determination of human coagulation factor XIII / K. Oertel et al. / Anal. Biochem. 367 (2007) 152–158

1.2 1.0 0.8 0.4

0.6

0.6

Fluorometric Assay [IU/ml]

Fluorometric Assay [IU/ml] 0.8 1.0 1.2

1.4

1.6

1.4

156

0.8

1.0 1.2 Photometric Assay [IU/ml]

1.4

0.6

0.4

0.6

0.8 1.0 1.2 Photometric Assay [IU/ml]

1.4

1.6

1.4

1.6

1.2 1.0 0.4

0.4

0.6

0.8

Photometric Assay [IU/ml]

1.4

1.4 1.2 1.0 0.8 0.6

Fluorometric Assay [IU/ml]

0.4

1.6

1.6

0.6

0.4

0.6

0.8

1.0 1.2 FXIII-Ag [IU/ml]

1.4

1.6

0.8

1.0

1.2

FXIII-Ag [IU/ml]

Fig. 3. Comparison of fluorometric versus photometric FXIII activity assay for pooled plasma samples and healthy donor samples as well as comparison of fluorometric and photometric FXIII activity assay for healthy donor samples versus FXIII antigen level. (A) The FXIII activity in 19 plasma pools was determined using the FXIII fluorometric assay and the FXIII photometric assay. Two independent series with four dilutions were run, and the results were calculated using the parallel line assay approach according to the European Pharmacopoeia with the CombiStats software. The average of the two series is given in the plot. The dashed line represents the ‘‘line of identity’’ in the scatter plot. (B) Plasma samples from healthy donors were tested for FXIII activity using the fluorometric assay and the photometric assay. Each point represents the average of two independent determinations. Samples were run in duplicate. The FXIII content was calculated from the slope of the standard curve. The dashed line represents the line of identity in the scatter plot. (C,D) FXIII antigen levels were determined in plasma samples of 22 healthy donors by an enzyme immunoassay using an affinity purified polyclonal anti-FXIIIA antibody as the capture antibody and a peroxidase-conjugated polyclonal anti-FXIII antibody for detection of captured plasma FXIII. The first international standard for FXIII (plasma) was used for calibration because this standard is also labeled with the antigen content (IU/ml). The dashed lines represent the line of identity in the scatter plot.

similar to the concordance coefficient (0.76, 95% CI, range = 0.50–0.90). In contrast to the plasma pool samples, there is no tendency for the fluorometric assay to yield slightly higher estimates, but the variability of the differences between the methods is larger. Indeed, it is not known which of the two methods yields the true value, and the observed differences may be due to factors inherent in a single plasma sample that may affect both assays differently. Using an FXIII antigen ELISA (A subunit), we determined the FXIII antigen concentration in individual plasma samples (Fig. 3C and D). The coefficient of correlation between the fluorometric assay and the antigen assay is 0.77 (95% CI, range = 0.51–0.90), whereas the coefficient of

correlation between the photometric assay and the antigen assay is 0.90 (95% CI, range = 0.76–0.96). Although there appears to be a quite good correlation between the antigen level and the FXIII activity estimated by the photometric assay, and a rather poorer correlation between antigen levels and FXIII activity for the fluorometric assay, we do not know the true quantity of FXIII activity in a sample when the antigen level is known. Therefore, the results support the suitability of the fluorometric assay when it is used for plasma samples with FXIII levels that are in the reference range of 0.7 to 1.4 IU/ml [18]. According to the literature, there is controversy regarding bleeding tendencies associated with the severity of

Assay for determination of human coagulation factor XIII / K. Oertel et al. / Anal. Biochem. 367 (2007) 152–158 Table 1 Comparison of photometric and fluorometric FXIII activity assay determinations versus FXIII antigen levels in plasma samples of 10 FXIII-deficient patients Sample

Photometric assay (IU/ml)

Fluorometric assay (IU/ml)

FXIII A subunit Ag (IU/ml)

1 2 3 4 5 6 7 8 9 10

0.42 0.29 0.06a 0.05a 0.11 0.78 0.06a 0.1a 0.06a 0.52

0.31 0.22 <0.02b <0.02b <0.02b 0.67 <0.02b <0.02b <0.02b 0.27

0.44 0.38 <0.005c <0.005c <0.005c 0.67 <0.005c <0.005c <0.005c 0.48

a

The linear range of the assay needs to be determined in each laboratory. According to our validation, these data are out of the linear range of the assay. b The detection limit is 0.02 IU/ml, and the limit of quantitation is 0.05 IU/ml. c The concentration of the lowest standard within the linear range of the assay is 0.005 IU/ml. Values below this limit are given as <0.005 IU/ml.

157

plasma samples. Based on the isopeptidase activity of FXIIIa discovered by Parameswaran and coworkers [10], the cleavage of a fluorometric peptide can be detected via an increase of fluorescence in a continuous, simple, and rapid manner. A comprehensive evaluation in terms of assay conditions and components, as well as concise comparison to already established FXIIIa assays, was presented. Our current data imply that in patients with severe FXIII deficiency, the fluorometric assay corresponds quite well with the FXIII antigen assay results and the photometric assay appears to overestimate the FXIII content. These findings may be an advantage in situations where overestimations may have critical clinical consequences. Based on the precision of FXIII activity determinations in the range of 0.7 to 1.4 IU/ml and the good agreement with FXIII antigen levels in the range below 0.1 IU/ml, the fluorometric assay is regarded as an excellent alternative method for measurements of FXIII concentrations in human plasma samples. Acknowledgments

FXIII deficiency. FXIII levels above 7% are considered as sufficient to protect patients against bleeding after small lesions. However, excessive bleeding after surgery has been reported in patients showing plasma levels of approximately 30% [19]. Patients with plasma levels below 7% suffer from bleeding that is similar to bleeding of haemophilia patients. Therefore, we were interested in the comparison of the different FXIII assays, especially in patients with FXIII deficiencies. In addition to the single donations from healthy donors, we analyzed 10 plasma samples from patients with an FXIII deficiency or their relatives with a heterozygote FXIII deficiency using the fluorometric assay, the photometric assay, and the antigen ELISA. The results are presented in Table 1. The correlation for the whole number of analyzed single donation samples is 0.90 (95% CI, range = 0.77–0.96, based on 10,000 bootstrap samples). The data for the FXIII-deficient patients should be discussed with caution because only a very limited number of estimates are available. The current data would suggest that in the case of heterozygous patients (samples 1, 2, 6, and 10) with FXIII levels above 0.1 IU/ml, both activity assays show reliable results for diagnosis. However, in pathological samples for patients with antigen levels below 0.05 IU/ml (samples 3, 4, 5, 7, 8, and 9), the photometric assay appears to overestimate residual FXIII activity. The antigen content of these samples is in better agreement with the results of the fluorometric assay, particularly in the critical range below 0.1 IU/ml FXIII activity.

Conclusions We have developed a highly sensitive fluorometric assay for the determination of human coagulation factor XIII in

We thank Viola Daub and Ju¨rgen Ohsam (N-Zyme BioTec [NZBT]), as well as Dorothea Hausleithner and Andrea Schroda (Paul-Ehrlich-Institut [PEI]), for technical assistance; Peter Volkers (PEI) for statistical advice and analysis of the data; and Verena Schro¨der and Hans-Peter Ko¨hler (Inselspital, Bern, Switzerland) for plasma samples from FXIII-deficient patients and their relatives. This work was supported by ‘‘Arbeitsgemeinschaft industrieller Forschungsvereinigungen Otto von Guericke e.V.’’ Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.ab.2007. 05.011. References [1] J.E. Folk, Mechanism and basis for specificity of transglutaminase catalyzed (((-glutamyl) lysine bond formation, Adv. Enzymol. Relat. Areas Mol. Biol. 54 (1983) 1–56. [2] L. Muszbek, J. Polga´r, L. Fe´sus, Kinetic determination of blood coagulation factor XIII, Clin. Chem. 31 (1985) 35–40. [3] L. Muszbek, V.C. Lee, Z. Hevessy, Blood coagulation factor XIII: Structure and function, Thromb. Res. 94 (1999) 271–305. [4] T.F. Slaughter, K.E. Achyuthan, T.S. Lai, C.S. Greenberg, A microtiter plate transglutaminase assay utilizing 5-(biotinamido) pentylamine as substrate, Anal. Biochem. 205 (1992) 466–471. [5] Y.C. Song, D.Q. Sheng, S.M. Taubenfeld, G.R. Matsueda, A microtiter assay for factor XIII using fibrinogen and biotincadaverine as substrates, Anal. Biochem. 223 (1994) 88–92. [6] V. Thomas, S. El Alaoui, D. Massignon, S. Cle´ment, F. Simonet, G. Quash, Development and evaluation of a modified colorimetric solidphase microassay for measuring the activity of cellular and plasma (factor XIII) transglutaminases, Biotechnol. Appl. Biochem. 43 (2006) 171–179. [7] M. Kusch, C. Grundmann, S. Keitel, R. Seitz, H. Ko¨nig, A novel assay for factor XIII based on cross-linking of synthetic peptides:

158

[8] [9]

[10]

[11] [12] [13]

Assay for determination of human coagulation factor XIII / K. Oertel et al. / Anal. Biochem. 367 (2007) 152–158 Analysis of different substrates, Blood Coagul. Fibrinolysis 17 (2006) 575–580. K. Fickenscher, A. Aab, W. Stu¨ber, A photometric assay for blood coagulation factor XIII, Thromb. Haemost. 65 (1991) 535–540. ´ . Katona, I. Balogh, G. Va´mosi, L. Muszbek, L. Ka´rpa´ti, B. Penke, E A modified, optimized kinetic photometric assay for the determination of blood coagulation factor XIII activity in plasma, Clin. Chem. 46 (2000) 1946–1955. K.N. Parameswaran, X.-F. Cheng, E.C. Chen, P.T. Velasco, J.H. Wilson, L. Lorand, Hydrolysis of c:e isopeptides by cytosolic transglutaminases and by coagulation factor XIIIa, J. Biol. Chem. 272 (1997) 10311–10317. Fibrin sealant kit, in: European Pharmacopoeiea, 5th ed., Council of Europe, Strasbourg, France, 2004, pp. 1593–1594. R.M. Clegg, Fluorescence resonance energy transfer and nucleic acids, Methods Enzymol. 211 (1992) 353–388. H. Ostermann, Biochemie und funktion des faktors XIII, in: R. Egbring, R. Seitz, G. Wozniak (Eds.), Klinische Aspekte des FaktorXIII-Mangels, Karger, Basel, Freiburg, Paris, London, New York, New Delhi, Bangkok, Singapore, Tokyo, Sydney, 1999, pp. 1–6.

[14] D.B. Cleary, M.C. Maurer, Characterizing the specificity of activated factor XIII for glutamine-containing substrate peptides, Biochim. Biophys. Acta 1764 (2006) 1207–1217. [15] R.A. Chica, P. Gagnon, J.W. Keillor, J.N. Pelletier, Tissue transglutaminase acylation: Proposed role of conserved active site Tyr and Trp residues revealed by molecular modelling of peptide substrate binding, Protein Sci. 13 (2004) 979–991. [16] International Conference on Harmonization, Topic Q2: Note for guidance on validation of analytical procedures (1995). [17] Statistical analysis of results of biological assays and tests, in: European Pharmacopoeiea, 5th ed., Council of Europe, Strasbourg, France, 2004, pp. 473–504. [18] M. Barthels, M. von Depka, Faktor XIII, in: M. Barthels, M. von Depka (Eds.), Das Gerinnungskompendium, Georg Thieme Verlag, Stuttgart, Germany, 2003, pp. 477–486. [19] R. Seitz, S. Duckert, S. Lopaciuk, L. Muszbek, F. Rodeghiero, U. Seligsohn, ETRO working party on factor XIII questionnaire on congenital factor XIII deficiency in Europe: Status and perspectives, Sem. Thromb. Hemost. 22 (1996) 415–418.