Rapid Determination of Blood Coagulation Factor XIII Activity Using

Feb 16, 2011 - ABSTRACT: We developed a novel on-chip assay using protein arrays for quantitative and rapid analysis of blood coagulation factor XIII ...
0 downloads 0 Views 2MB Size
ARTICLE pubs.acs.org/ac

Rapid Determination of Blood Coagulation Factor XIII Activity Using Protein Arrays for Serodiagnosis of Human Plasma Mi-Hye Kwon,† Deok-Hoon Kong,† Se-Hui Jung,† In-Bum Suh,‡ Young-Myeong Kim,† and Kwon-Soo Ha*,† †

Department of Molecular and Cellular Biochemistry and Institute of Medical Science, and ‡Department of Laboratory Medicine, Kangwon National University School of Medicine, Chuncheon, Kangwon-do 200-701, South Korea ABSTRACT: We developed a novel on-chip assay using protein arrays for quantitative and rapid analysis of blood coagulation factor XIII (FXIII) activity in human plasma. FXIII is activated by concerted action of thrombin and Ca2þ and plays essential roles in hemostasis, angiogenesis, and wound healing. We fabricated protein arrays by immobilizing fibrinogen onto the 3-aminopropyltrimethoxysilane layer of well-type arrays and determined FXIII activity by analyzing biotinylated fibrinogen with Cy3-conjugated streptavidin. We determined optimal concentrations of Ca2þ, thrombin, and 5-(biotinamido)pentylamine (BAPA) for the on-chip activity assay, and the detection limit was 0.01 Lowey U/mL (9.9 pM). Using the on-chip activity assay, hepatocellular carcinoma patients (n = 24), but not hepatitis (n = 24) or liver cirrhosis patients (n = 41), had significantly lower FXIII activities (p < 0.001) than normal individuals (n = 41), indicating that FXIII activity is a possible diagnostic marker for hepatocellular carcinoma. In addition, we have successfully used this activity assay to reveal individual variations (37-57%, n = 65) in the inhibition rate of FXIII activity by isoniazid, the first-line antituberculosis agent. Thus, our optimized on-chip FXIII activity assay provides a quantitative and high-throughput approach to investigating the role(s) of FXIII in human diseases. Moreover, it has a strong potential to be applied toward FXIII-related personalized medicines.

B

lood coagulation factor XIII (FXIII) is a member of the transglutaminase (TG) family, which catalyzes calcium-dependent acyl transfer reactions between the γ-carboxamide group of a peptide-bound glutamine and the ε-group of a peptide-bound lysine or polyamine, leading to formation of an ε-(γ-glutamyl)lysine bond or (γ-glutamyl)polyamine bond.1-3 FXIII is present in the plasma as a heterotetramer composed of two potentially active A subunits (FXIII-A) and two regulatory B subunits (FXIII-B). FXIII-A is synthesized in monocytes/macrophages, megakaryocytes, and possibly hepatocytes. FXIII-B is synthesized in the liver and serves as a carrier for FXIII-A in plasma.4-6 FXIII is converted into an active TG (FXIII-A) through the concerted action of thrombin and Ca2þ in the terminal phase of the blood clotting cascade. Thrombin cleaves the Arg37-Gly38 peptide bond of FXIII and then, in the presence of Ca2þ, dissociates the B-subunits from the heterotetramers (A2B2). Additional Ca2þ activates the A-subunit, which contains the active site of the enzyme.5,6 In the plasma, activated FXIII catalyzes the formation of covalent cross-links between γ-glutamyl and ε-lysyl residues of adjacent fibrin monomers, a reaction that stabilizes fibrin clots.7 FXIII is known to participate in hemostasis, angiogenesis, wound healing, and maintaining pregnancy.5,6 Inherited FXIII deficiency is a bleeding disorder and is associated with poor wound healing and miscarriages.8,9 Acquired FXIII deficiency has been described in a number of diseases including inflammatory bowel disease, acute leukemia, liver disease, and ulcerative colitis.8,9 Elevated FXIII activity in plasma has been observed in patients with obliterative atherosclerosis, diabetic angiopathy, peripheral artery disease, and myocardial infarction.7,10,11 It is thought that patients with chronic liver disease have a bleeding tendency, which is considered a consequence of alterations in platelet number, hemostasis, and the fibrinolytic system.12 r 2011 American Chemical Society

Hepatocellular carcinoma, one of liver diseases, is the third most common cause of cancer-related death in the world.13 In many cases, hepatocellular carcinoma arises as a consequence of underlying liver diseases, such as viral hepatitis and liver cirrhosis.14 Recent proteomic studies have been aimed at the discovery of new biomarkers. The most widely used biomarkers for the diagnosis of hepatocellular carcinoma are R-fetoprotein, R-fetoprotein-L3, and des-γ-carboxy prothrombin.14 However, systemic studies of FXIII activity in patients with liver diseases, including cirrhosis and hepatocellular carcinoma, are limited.15-17 FXIII activity has been measured on the basis of two different properties: amine incorporation and ammonia released during the TG reaction.7,8,18-21 Photometric assays, based on the continuous monitoring of ammonia-release during the TG reaction, are used for routine determination of FXIII concentrations in plasma samples7-9 but frequently overestimate FXIII activity in the plasma of FXIII-deficient patients.8,18,21 Various assays based on incorporation of exogenous amines into peptides and proteins have been developed.7,18-21 In amine incorporation assays, radiolabeled, biotinylated, or fluorescent amine substrates are covalently attached to an amine acceptor protein by FXIII, and the amine-modified products are quantified using various approaches, including a colorimetric assay.7,9,21 Recently, a fluorometric assay using microtiter plates was developed on the basis of the isopeptidase activity of FXIIIa.21 However, these assays require a large sample volume and are not suitable for highthroughput analysis.7,21 Received: December 9, 2010 Accepted: January 27, 2011 Published: February 16, 2011 2317

dx.doi.org/10.1021/ac1032275 | Anal. Chem. 2011, 83, 2317–2323

Analytical Chemistry

ARTICLE

Figure 1. Schematic diagram of the on-chip FXIII activity assay using fibrinogen arrays. Protein arrays were fabricated by immobilizing fibrinogen on well-type amine arrays, and the FXIII activity was determined as described in the Experimental Section.

In this study, we developed a novel on-chip activity assay to determine FXIII activity using protein arrays in a high-throughput manner. Protein arrays were fabricated by immobilizing fibrinogen onto well-type amine arrays, and FXIII activity was measured by probing BAPA, incorporated into the fibrinogen, with Cy3-streptavidin as shown in Figure 1. We used this on-chip FXIII activity assay to determine FXIII activities in human plasma samples collected in heparin, sodium citrate, and ethylenediaminetetraacetic acid (EDTA) tubes. We also applied this assay system to study FXIII activity in plasma (n = 127) from normal individuals and from hepatitis, liver cirrhosis, and hepatocellular carcinoma patients and demonstrated FXIII activity as a possible candidate biomarker for hepatocellular carcinoma. Furthermore, this assay system was successfully used to analyze the differential inhibition of FXIII activity by isoniazid, an antituberculosis drug. Thus, this new assay is sensitive, suitable for high-throughput analysis, and has a strong potential for monitoring various FXIII-related human diseases.

’ EXPERIMENTAL SECTION Chemicals and Reagents. 3-Aminopropyltrimethoxysilane, ammonium hydroxide, hydrogen peroxide, bovine serum albumin, fibrinogen, Gly-Pro-Arg-Pro amide, thrombin, Cy3-conjugated streptavidin, and isoniazid were obtained from Sigma-Aldrich (St. Louis, MO). BAPA was purchased from Pierce (Rockford, IL). Purified human plasma FXIII (2226 Loewy U/mg, one plasma equivalent unit is equal to 108 Loewy U) and FXIII-deficient plasma were obtained from Innovative Research (Novi, MI) and Hyphen BioMed (Neuville-sur-oise, France), respectively. Preparation of Well-Type Amine Arrays. Amine arrays were prepared according to the procedures of Jung et al.22 In brief, glass slides (75  25 mm) were cleaned with H2O2/NH4OH/H2O (1:1:5, v/v) at 70 °C for 10 min and sequentially incubated with 1.5% (v/v) 3-aminopropyltrimethoxysilane solution in 95% ethanol for 2 h. Well-type amine arrays were made by mounting Teflon

tapes (75  25 mm), each perforated with 200 holes (25  8) of 1.5 mm diameter, onto the amine-modified slide, and were stored under vacuum until use. Plasma Samples. Human plasma samples were sequentially collected from individual donors in heparin and sodium citrate tubes (Becton Dickinson, NJ) and in EDTA tubes (Greiner BioOne, Austria). The plasma samples were obtained from the Kangwon National University Hospital (Korea), divided into aliquots, and then stored at -80 °C until use. We obtained plasma from normal, hepatitis, liver cirrhosis, and hepatocellular carcinoma patients. Experiments using human samples were performed under the approval of the Institutional Review Board of Kangwon National University Hospital for human subject research. On-Chip FXIII Activity Assay. The on-chip FXIII activity assay was performed using protein arrays, as described in Figure 1. To fabricate the fibrinogen arrays, the well-type amine arrays were incubated with 50 μg/mL fibrinogen for 1 h at 37 °C and blocked at 37 °C with 3% bovine serum albumin (BSA) containing 0.1% Tween-20 in PBS (8.1 mM Na2HPO4, 1.2 mM KH2PO4, pH 7.4, 2.7 mM KCl, and 138 mM NaCl) for 30 min. The resulting protein arrays were sequentially washed with 0.1% Tween-20 in PBS and milli-Q water. Reaction mixtures were prepared in 30 μL of a solution containing 40 mM Tris-HCl, pH 8.3, 140 mM NaCl, 2 mM CaCl2, 0.2 units thrombin, 0.5 mM dithiothreitol, 5 mM BAPA, 0.3 mM Gly-Pro-Arg-Pro amide (the fibrin polymerization inhibitor peptide), and plasma (1:400, final) or purified human FXIII. Then, a 1-μL aliquot of the reaction mixture was applied to each of the array wells at 37 °C for 1 h. After washing with 1% Tween-20 in PBS and with milli-Q water, FXIII-catalyzed cross-linking of BAPA and fibrin was probed with 10 μg/mL Cy3-conjugated streptavidin at 37 °C for 30 min. Following washing with 1% Tween-20 in PBS and with milli-Q water, the arrays were dried under compressed air and scanned with a fluorescence scanner using a 543 nm laser (ScanArray Express, PerkinElmer Life and Analytical Sciences). The fluorescence 2318

dx.doi.org/10.1021/ac1032275 |Anal. Chem. 2011, 83, 2317–2323

Analytical Chemistry

ARTICLE

Figure 2. Optimization of Ca2þ, thrombin, BAPA, and incubation time for on-chip FXIII activity assay. (A,B,C) One microliter aliquots of reaction mixtures (30 μL) containing the three types of plasma (heparin, sodium citrate, and EDTA) and indicated concentrations of Ca2þ (A), BAPA (B), thrombin (units/30 μL) (C) were loaded onto well-type fibrinogen arrays. (D) One microliter aliquots of reaction mixtures containing 2 mM Ca2þ, 0.2 U thrombin, and 5 mM BAPA were applied to fibrinogen arrays and incubated for the indicated time. FXIII activity was determined as described in the Experimental Section. The results are expressed as means of fluorescence intensities (%) ( SD from three separate experiments.

intensities of array spots were measured with the ScanArray Express program. For characterization of the chip based FXIII activity assay, data were normalized using Quantile normalization (R 2.8.1 program). The limit of detection was calculated on the basis of the mean and standard deviation of the blank, according to a previous report.23 Determination of FXIII Activity. The fluorescence intensity of array spots measured with the ScanArray Express program was representative of the FXIII activity in each sample. The activity was determined using a standard curve, plotted by the sigmoidal fit of the Origin program: x ¼ x0 ½ðA1 - A2 Þ=ðy-A2 Þ - 11=p where x is the FXIII activity of the samples, x0 is the FXIII activity at half-maximal fluorescence intensity, A1 is the minimal fluorescence intensity, A2 is the maximal fluorescence intensity, y is the fluorescence intensity of the samples, and p is the power. The FXIII activity was expressed as Loewy unit per milliliter. Inhibition Assay. To study individual variation in the inhibitory efficacy of FXIII inhibitors in human plasma, FXIII activity was measured in human plasma (n = 65) containing 2 mM isoniazid (the half-maximal inhibition concentration for human FXIII) and the percent inhibition in each plasma sample was measured. The half-maximal inhibition concentration of isoniazid was determined by performing on-chip activity assays with

reaction mixtures containing various concentrations of isoniazid and 1.5 Loewy U/mL of human FXIII. Statistical Analysis. The correlation coefficient (r2) was analyzed using the fit linear function of the Origin program. The t test of two populations and one-way analysis of variance (ANOVA) were performed to evaluate the statistical significance of differences between groups. The receiver operating characteristics (ROC) curve was generated to estimate the sensitivity and specificity using MedCalc statistical software 11.4.4.0 (Mariakerke, Belgium)

’ RESULTS AND DICUSSION Characterization of the On-Chip FXIII Activity Assay. FXIII is essential for maintaining hemostasis; its deficiency leads to severe bleeding disorders, while elevated FXIII activity is associated with various diseases.9-11 The ability to determine FXIII activity is crucial for adequate diagnosis and classification of FXIII deficiencies. We have developed a novel assay that uses protein arrays for sensitive and high-throughput analysis of FXIII activity, as illustrated in Figure 1. We fabricated the protein arrays by immobilizing fibrinogen, a FXIII substrate, on well-type amine arrays, and directly applied reaction mixtures containing BAPA, another FXIII substrate, and human plasma samples collected in heparin, sodium citrate, or EDTA tubes, onto the fibrinogen 2319

dx.doi.org/10.1021/ac1032275 |Anal. Chem. 2011, 83, 2317–2323

Analytical Chemistry

Figure 3. Comparison of FXIII activity in three types of plasma. Three types of plasma (n = 40) were collected in heparin, sodium citrate, and EDTA tubes, and then, the FXIII activity was determined as described in the Experimental Section. The results are expressed as means of fluorescence intensities (a.u.) ( SD from three separate experiments (/, p < 0.001 after ANOVA). a.u., arbitrary unit.

surface of the protein arrays. We then probed the resulting biotinylated fibrinogen with Cy3-conjugated streptavidin and determined the level of FXIII activity by measuring the fluorescence intensities of each spot on the array using a fluorescence scanner. Since FXIII is known to be activated by Ca2þ and thrombin,6 we initially determined an optimal concentration of Ca2þ using human plasma in the on-chip FXIII activity assay. We used three types of human plasma collected in heparin, sodium citrate, or EDTA tubes from a single donor. As shown in Figure 2A, the results of the FXIII activity assays from all three types of plasma samples showed a Ca2þ concentration-dependent increase, with evident activation at 0.2 mM and saturation at 2 mM. Thrombin also induced the activation of FXIII activity, in a dose-dependent manner, with a maximal effect at 0.2 units (Figure 2B). We then determined the optimal concentration of BAPA, an amine donor in the cross-linking reaction used in the FXIII activity assay. As shown in Figure 2C, BAPA elevated FXIII activity in a dosedependent manner, with a maximal stimulation at 5 mM. Next, we investigated the change in FXIII activity with respect to incubation time, and the FXIII activity showed a time-dependent increase, which reached saturation at 60 min (Figure 2D). In the assays to determine the optimal concentrations of Ca2þ, thrombin, and BAPA and incubation time, we found that the relative fluorescence intensities were higher in the EDTA plasma samples than in the heparin or sodium citrate plasma samples (Figure 2), indicating that EDTA plasma provides higher sensitivity than the others. To confirm the finding, we collected three types of plasma samples in heparin, sodium citrate, and EDTA tubes from 40 donors and determined their FXIII activities. As shown in Figure 3, the average FXIII activity (fluorescence intensity) of EDTA plasma samples (33 205 ( 3822) was higher (p < 0.001, n = 40) than that of heparin (29 325 ( 3812) and sodium citrate (27 507 ( 3465) samples. We concluded that the FXIII activity assay was more sensitive when used with EDTA plasma than with heparin or sodium citrate plasma. Anticoagulants such as heparin, sodium citrate, and EDTA are commonly used in the collection of blood samples and inhibit coagulation by chelating calcium ions (EDTA and sodium citrate) or by inhibiting thrombin (heparin).24-28 It is desirable

ARTICLE

Figure 4. Limit of detection of on-chip FXIII activity assay. Reaction mixtures containing the indicated concentrations of the purified human plasma FXIII were applied in 1 μL aliquots to the fibrinogen arrays, and the FXIII activities were determined as described in the Experimental Section. The results are expressed as means of fluorescence intensities (a.u.) ( SD from three separate experiments.

to use a universal anticoagulant for the evaluation of several laboratory parameters, such as hematological testing, coagulation tests, and clinical chemistry analysis.24 The use of a universal anticoagulant may have several advantages, including the possible elimination of preanalytical bias in serum and whole blood, a reduction in the number of error-prone steps, such as collection, handling, and sorting of separate blood tubes, and ultimately an improvement in the time and cost effectiveness of the total testing process.24,26 However, no universal anticoagulant is available so far. Heparin is used for clinical chemistry. Sodium citrate is the anticoagulant recommended for coagulation tests, such as prothrombin time, activated partial thromboplastin time, and antithrombin activity, and for measurement of coagulation factors.25 EDTA has been used in hematological tests because it preserves well components and the morphology of blood cells.24 EDTA is also recommended for cytokine measurement, as it stabilizes the blood until centrifugation.24 In addition, EDTA has been recommended for the measurement of prothrombin time26,27 and hemostatic factors, including the von Willebrand factor,28 and for proteomic analysis.24 Furthermore, in this report, we show that EDTA plasma was more suitable for the on-chip FXIII activity assay than heparin and sodium citrate plasma. Thus, it is likely that EDTA is the optimal anticoagulant for multiple laboratory analyses. Limit of Detection and Reproducibility. To determine the limit of detection of the FXIII activity assay, we applied reaction mixtures containing various concentrations of purified human plasma FXIII to fibrinogen arrays, and the resulting fluorescence intensities were plotted as a function of human FXIII concentration (Figure 4A). The fluorescence intensity of the blank sample without the human plasma FXIII was negligible (data not shown), indicating that fibrinogen immobilized on amine arrays has no effect on the FXIII activity assay. Using the mean and the standard deviation of the blank sample from Figure 4A, we estimated the limit of detection of the on-chip FXIII activity assay to be 0.01 Loewy U/mL (9.9 pM or 9.26  10-5 PEU; p < 0.05), which is at least three times more sensitive than enzymelinked immunosorbent assay, based on amine incorporation, using BAPA and fibribogen as substrates and avidin-conjugated horseradish peroxidase for probing the reaction product. 2320

dx.doi.org/10.1021/ac1032275 |Anal. Chem. 2011, 83, 2317–2323

Analytical Chemistry

ARTICLE

Figure 5. Determination of FXIII activity in the patients of liver diseases. (A) Human plasma samples (n = 127) were collected in EDTA tubes from normal individuals (n = 41) and from patients with hepatitis (n = 21), liver cirrhosis (LC, n = 41), or hepatocellular carcinoma (HCC, n = 24). Next, reaction mixtures containing plasma (400-fold dilution) and the indicated concentrations of human plasma FXIII were applied to the fibrinogen arrays, and the arrays were analyzed with a fluorescence scanner. (B) The FXIII activities of 127 plasma samples were determined using a standard curve (r2 = 0.99), as described in the Experimental Section, and the distribution is expressed in box plots (/, p < 0.001). (C) ROC curve for HCC plasma.

These results indicate that the on-chip activity assay is highly sensitive. Next, we evaluated the reproducibility of the activity assay by analyzing interarray and inter-reaction/-array reproducibility using a human EDTA plasma sample. Interarray reproducibility was determined by analyzing the same batch of reaction mixtures on different fibrinogen arrays. We found the interarray reproducibility to be high, with average correlation coefficients (r2) of 0.99 ( 0.01 (n = 3). Inter-reaction/array was determined by analyzing different batches of reaction mixtures on different arrays, and the average correlation coefficient was found to be 0.94 ( 0.04 (n = 3). In addition, interspot reproducibility was determined by analyzing ten replicate spots, and the average coefficient of variation (%) was found to be 2.7 ( 0.5 (n = 3). The reproducibility of purified human plasma FXIII was also determined, and we obtained a degree of reproducibility similar to that of the human EDTA plasma (data not shown). Thus, the on-chip FXIII activity assay is highly sensitive and reproducible. In the past, FXIII activity has been determined in photometric assays that measure the ammonia released during the TG reaction or colorimetric assays of amine incorporation.7-9,18 It has been reported that approaches that measure amine incorporation are sensitive but cumbersome, time-consuming, and difficult to standardize.8 A fluorometric assay using microtiter plates was recently developed, but this assay is not suitable for the high-throughput determination of FXIII activities.21 However, our FXIII activity assay uses protein arrays rather than microtiter plates and is performed in a high-throughput manner in the following two step process, without separation of the biotinylated products: (1) incubation of reaction mixtures on fibrinogen arrays and (2) probing with Cy3-conjugated streptavidin. To our knowledge, this is the first report of a FXIII activity assay that

makes use of protein arrays. Thus, this novel on-chip activity assay is a very efficient and high-throughput approach and has strong potential to be used as an alternative to previous FXIII activity assays. Analysis of FXIII Activity in Patients with Liver Diseases. Our previous results showed that the FXIII activity assay was more sensitive when used with EDTA plasma than with heparin or sodium citrate plasma (Figures 2 and 3), and EDTA is the recommended anticoagulant for hematological testing and proteomics.24 Patients with chronic liver diseases tend to bleed, which is thought to be caused by alterations in blood coagulation.17 However, systemic study of FXIII activity in patients with liver diseases, including hepatocellular carcinoma, is limited. Thus, we investigated the differences in the FXIII activity between normal individuals and patients with liver diseases. We collected plasma samples in EDTA tubes from a normal group (n = 41) and from patients with hepatitis (n = 21), liver cirrhosis (n = 41), or hepatocellular carcinoma (n = 24). We then determined the FXIII activity using the standard curve and compared the FXIII activities of the normal group with those of liver disease patients (Figure 5A,B). The mean FXIII activity of the normal group was 550.5 ( 152.8 Loewy U/mL (n = 3). The FXIII activities of patients with hepatitis and liver cirrhosis (599.3 ( 212.0 and 524.7 ( 176.2 Loewy U/mL, n = 3, respectively) were not significantly different from that of the normal group. However, FXIII activity of the hepatocellular carcinoma patients (403.3 ( 158.9 Loewy U/mL, n = 3) was significantly lower than the control level (p < 0.001). Then, a ROC curve was generated to evaluate FXIII activity as a possible biomarker for hepatocellular carcinoma (Figure 5C). Sensitivity and specificity of FXIII activity were 91.7% and 61.0% at the cutoff point of 575.52 Loewy U/mL, respectively, with an area under ROC curve 2321

dx.doi.org/10.1021/ac1032275 |Anal. Chem. 2011, 83, 2317–2323

Analytical Chemistry

Figure 6. Differential inhibition of FXIII activity by isoniazid in human plasma. (A) Reaction mixtures containing 1.5 Loewy U/mL of the human plasma FXIII were loaded onto the fibrinogen arrays in the presence of the indicated concentrations of isoniazid for 15 min. The half-maximal inhibition concentration was determined as described in the Experimental Section. Results are means ( SD from three separate experiments. (B) Distribution of inhibition rate of human plasma of normal (n = 41) and hepatocellular carcinoma patient (HCC, n = 24) groups. The inhibition rates (average of three separate experiments) were determined as described in the Experimental Section. (C) A histogram of inhibition rates of FXIII by isoniazid in human plasma (n = 65). Inhibition rates (average of three separate experiments) were determined as described in the Experimental Section.

(AUC) of 0.80 ( 0.06 (range 0.68-0.89 at the 95% confidence interval). These results indicate that FXIII activity has a potential to be a biomarker for hepatocellular carcinoma. Thus, chip based analysis of FXIII activity and biomarker proteins may be an effective approach for the diagnosis of hepatocellular carcinoma. Differential Inhibition of FXIII Activity with Isoniazid in Human Plasma. To investigate personal variation in the efficacy of the FXIII inhibitor, isoniazid, FXIII activity was measured in human plasma samples collected in EDTA tubes in the presence of the inhibitor. Isoniazid is the first-line antituberculosis medicine and can be used as a substrate of FXIII and incorporated into fibrin, in vitro.29 Initially, we determined the half maximal inhibitory concentration of isoniazid using a human plasma

ARTICLE

sample that had the median activity of FXIII in the normal group. Reaction mixtures containing various concentrations of isoniazid and a 1:400 dilution of plasma sample (final concentration, 1.5 Loewy U/mL) were loaded onto fibrinogen arrays, and the half maximal inhibition concentration was found to be 2.0 mM (Figure 6A). Patients with acute and chronic liver disease have a bleeding tendency, which is considered to be caused by thrombocytopenia, decreased hepatic synthesis of coagulation factors, and hyperfibrinolysis.12,17 Our results demonstrated that the FXIII activity of hepatocellular carcinoma patients was lower than the normal group. Thus, we compared the inhibition rate of FXIII activity in the normal group and hepatocellular carcinoma groups using 2.0 mM isoniazid. We used isoniazid in the inhibition study because it is a known FXIII inhibitor.29 As shown in Figure 6B, the average inhibition rate of the normal group plasma samples was 47.6 ( 5.0 (n = 41) and that of the patient group plasma samples was 47.2 ( 3.5 (n = 24), demonstrating that there was no difference in the inhibition rate of FXIII activity between the two groups. We then analyzed individual variation in isoniazid efficacy among the plasma samples (n = 65) of the normal and the patient groups, and the distribution of the inhibition rate was shown in Figure 6C. The FXIII inhibition range of the human plasma samples was 37-57%, indicating that the inhibition of FXIII activity by isoniazid varies among individuals. Thus, isoniazid differentially inhibited FXIII activity in human plasma; however, there was no difference in the inhibition rate between the normal and the hepatocellular carcinoma groups. It is not understood how isoniazid differentially inhibits FXIII activity in human plasma. It is reported that bleeding of patients administered with isoniazid is associated with an acquired factor XIII inhibitor, immunoglobulin G, but is rare.29 In this report, we show differential inhibition of FXIII activity in human plasma by isoniazid, indicating that other plasma factors rather than immunoglobulin G could be involved in the inhibition of FXIII activity by isoniazid. Thus, isoniazid may regulate FXIII activity by various factors including the acquired factor XIII inhibitor, although it is necessary to elucidate regulation factors of FXIII activity in vitro and In Vivo. Furthermore, the individual variation of FXIII inhibition by isoniazid highlights the importance of personalized medicine in the treatment of human diseases. It was recently reported that protein arrays have enormous possibility in the personalization of treatment for many diseases.30,31 Thus, the on-chip based FXIII activity assay has a strong potential for use in the personalized treatment of FXIII-related diseases.

’ CONCLUSION In this study, we present a novel on-chip activity assay that uses fibrinogen arrays for the rapid determination of FXIII activity in human plasma. We fabricated fibrinogen arrays by immobilizing fibrinogen onto well-type amine arrays and determined FXIII activity using a fluorescence scanner. This on-chip FXIII activity assay is highly sensitive and reproducible. Using this assay, we indicated FXIII activity as a potential diagnostic biomarker for hepatocellular carcinoma. In addition, we successfully applied this assay to demonstrate individual variations in the inhibition rate of FXIII activity by isoniazid. Thus, the on-chip based FXIII activity assay is a highly effective system for investigating FXIII activity in human plasma and has a strong potential for monitoring FXIII-related human diseases and screening FXIII inhibitors. 2322

dx.doi.org/10.1021/ac1032275 |Anal. Chem. 2011, 83, 2317–2323

Analytical Chemistry

’ AUTHOR INFORMATION Corresponding Author

*Tel.: þ82-33-250-8833. Fax: þ82-33-250-8807. E-mail: ksha@ kangwon.ac.kr.

’ ACKNOWLEDGMENT This work was supported in part by the Ministry of Health and Welfare through the grant of the Korea Health 21 R&D project (A030003) and the National R&D Program for Cancer Control (1020420) and the Korea Research Foundation through the Basic Research Program (2008-05943). We would like to thank the Korea Basic Science Institute for help with array preparation.

ARTICLE

(28) Green, D.; McMahon, B.; Foiles, N.; Tian, L. Am. J. Clin. Pathol. 2008, 130, 811–815. (29) Krumdieck, R.; Shaw, D. R.; Huang, S. T.; Poon, M. C.; Rustagi, P. K. Am. J. Med. 1991, 90, 639–645. (30) Jain, K. K. Curr. Opin. Drug Discovery Dev. 2004, 7, 285–289. (31) Yu, X.; Schneiderhan-Marra, N.; Joos, T. O. Clin. Chem. 2010, 56 (3), 376–387.

’ REFERENCES (1) Greenberg, C. S.; Birckbichler, P. J.; Rice, R. H. FASEB J. 1991, 5, 3071–3077. (2) Griffin, M.; Casadio, R.; Bergamini, C. M. Biochem. J. 2002, 368, 377–396. (3) Park, D.; Choi, S. S.; Ha, K. S. Amino Acids 2010, 39, 619–631. (4) Muszbek, L.; Yee, V. C.; Hevessy, Z. Thromb. Res. 1999, 94, 271– 305. (5) Dardik, R.; Loscalzo, J.; Inbal, A. J. Thromb. Haemostasis 2006, 4, 19–25. (6) Muszbek, L.; Bagoly, Z.; Bereczky, Z.; Katona, E. Cardiovasc. Hematol. Agents Med. Chem. 2008, 6, 190–205. (7) Thomas, V.; El Alaoui, S.; Massignon, D.; Clement, S.; Simonet, F.; Quash, G. Biotechnol. Appl. Biochem. 2006, 43, 171–179. (8) Karpati, L.; Penke, B.; Katona, E.; Balogh, I.; Vamosi, G.; Muszbek, L. Clin. Chem. 2000, 46, 1946–1955. (9) Hsieh, L.; Nugent, D. Haemophilia 2008, 14, 1190–1200. (10) Shemirani, A. H.; Szomjak, E.; Csiki, Z.; Katona, E.; Bereczky, Z.; Muszbek, L. Haematologica 2008, 93, 1430–1432. (11) Bereczky, Z.; Balogh, E.; Katona, E.; Czuriga, I.; Edes, I.; Muszbek, L. Haematologica 2007, 92, 287–288. (12) Lisman, T.; Leebeek, F. W. Dig. Surg. 2007, 24, 250–258. (13) Srivatanakul, P.; Sriplung, H.; Deerasamee, S. Asian Pac. J. Cancer Prev. 2004, 5, 118–125. (14) Pei, Y.; Zhang, T.; Renault, V.; Zhang, X. Acta Biochim. Biophys. Sin. (Shanghai) 2009, 41, 1–15. (15) Biland, L.; Duckert, F.; Prisender, S.; Nyman, D. Thromb. Haemostasis 1978, 39, 646–656. (16) Ballerini, G.; Guerra, S.; Rodeghiero, F.; Castaman, G. Semin. Thromb. Haemostasis 1985, 11, 357–361. (17) Tacke, F.; Fiedler, K.; von Depka, M.; Luedde, T.; Hecker, H.; Manns, M. P.; Ganser, A.; Trautwein, C. Liver Int. 2006, 26, 173–181. (18) Wilmer, M.; Rudin, K.; Kolde, H.; Poetzsch, B.; Lenz, W.; Moessmer, G.; Meili, E.; Egbring, R.; Gempeler-Messina, P.; Gempeler, M.; Bastian, S.; Kohler, H. Thromb. Res. 2001, 102, 81–91. (19) Fickenscher, K.; Aab, A.; Stuber, W. Thromb. Haemostasis 1991, 65, 535–540. (20) Lorand, L.; Campbell-Wilkes, L. K.; Cooperstein, L. Anal. Biochem. 1972, 50, 623–631. (21) Oertel, K.; Hunfeld, A.; Specker, E.; Reiff, C.; Seitz, R.; Pasternack, R.; Dodt, J. Anal. Biochem. 2007, 367, 152–158. (22) Jung, J. W.; Jung, S. H.; Yoo, J. O.; Suh, I. B.; Kim, Y. M.; Ha, K. S. Biosens. Bioelectron. 2009, 24, 1469–1473. (23) Armbruster, D. A.; Tillman, M. D.; Hubbs, L. M. Clin. Chem. 1994, 40, 1233–1238. (24) Banfi, G.; Salvagno, G. L.; Lippi, G. Clin. Chem. Lab. Med. 2007, 45, 565–576. (25) Ceron, J. J.; Carli, E.; Tasca, S.; Martinez-Subiela, S.; Caldin, M. Am. J. Vet. Res. 2008, 69, 1141–1147. (26) Horsti, J. Haematologica 2001, 86, 851–855. (27) Horsti, J. Clin. Chem. 2001, 47, 1731–1732. 2323

dx.doi.org/10.1021/ac1032275 |Anal. Chem. 2011, 83, 2317–2323