A rapid test for the diagnosis of thrombotic thrombocytopenic purpura using surface enhanced laser desorption/ionization time-of-flight (SELDI-TOF)-mass spectrometry


Haifeng M. Wu, Clinical Coagulation Laboratory, Department of Pathology, Ohio State University College of Medicine and Public Health, 288 Medical Research Facility, 420 West 12th Avenue, Columbus, OH 43210, USA.
Tel.: +1 614 292 1798; fax: +1 614 292 3144; e-mail: wu-6@medctr.osu.edu


Summary. Background: Thrombotic thrombocytopenic purpura (TTP), a life-threatening thrombotic microangiopathy, requires immediate diagnosis and plasma exchange therapy. Development of TTP is related to functional deficiency of ADAMTS-13 protease that leads to the accumulation of ultra large von Willebrand factor (VWF) and subsequent platelet thrombosis. Currently no clinical test is available for the rapid detection of ADAMTS-13 activity. Objectives: The goal is to devise a novel method to rapidly detect functional activity of ADAMTS-13 and improve clinical outcome. Methods and results: A recombinant VWF substrate containing the ADAMTS-13 cleavage site and a 6X Histidine tag was cleaved by ADAMTS-13 in a dose-dependent manner, generating approximately 7739 Da peptide containing a 6X Histidine tag. This cleaved peptide, bound to an IMAC/Nickel ProteinChip, was quantified using Surface Enhanced Laser Desorption/Ionization Time-of-flight Mass Spectrometry (SELDI-TOF-MS). The assay is capable of quantifying ADAMTS-13 activity as low as 2.5% in plasma within 4 h. When the cleaved peptide was quantified as a ratio of an internal control peptide, the test displayed good reproducibility, with an average inter-assay coefficient of variation (CV) of < 33%. Further validation revealed a mean ADAMTS-13 activity of 92.5% ± 16.6% in 39 healthy donors. Sixteen patients with idiopathic TTP displayed mean ADAMTS-13 activity of 1.73% ± 3.62%. Further utility of this novel method includes determining the inhibitory titer of ADAMTS-13 antibody in cases of acquired TTP. Conclusions: We have devised a novel SELDI-TOF-MS assay that offers a rapid, cost-effective, and functionally relevant test for timely diagnosis and management of TTP.


Thrombotic thrombocytopenic purpura (TTP) is a devastating thrombotic microangiopathy characterized by hemolytic anemia, consumptive thrombocytopenia, and ischemic injury [1–4]. The pathogenesis of TTP is attributed to the presence of ultra large von Willebrand Factor (ulVWF) multimers that lead to platelet clumping and subsequent thrombosis [5]. VWF is primarily synthesized by vascular endothelial cells and secreted as a polymer with a Mr of > 500 000 kDa [6]. These ultra-large multimers are highly active in promoting platelet thrombi [5]. Under normal physiological conditions, a metalloprotease enzyme, ADAMTS-13, cleaves VWF multimers into smaller protein units ranging from 500 to 20 000 kDa [7]. Impairment of ADAMTS-13 activity, caused either by hereditary deficiency or by acquired autoantibodies that specifically inhibit ADAMTS-13 function, leads to excessive accumulation of ulVWF and subsequent onset of TTP [8,9]. The majority of clinically observed TTP cases in adults are acquired [1–4,10], with patients showing detectable levels of autoantibodies to ADAMTS-13.

Clinical management of TTP requires rapid diagnosis followed by prompt initiation of treatment [1–4]. Plasma exchange therapy has been demonstrated to be the most successful therapy. It effectively replaces ADAMTS-13 and/or removes the ADAMTS-13 autoantibodies in the case of acquired TTP. Rapid recognition of the disease along with timely treatment has greatly improved the clinical outcome by inducing remission in more than 80% of patients [1–4]. Therefore, a clinical test that can rapidly detect deficiencies of ADAMTS-13 activity associated with acute TTP would be invaluable to confirm the clinical suspicion of TTP and support the rapid initiation of plasma exchange. On the contrary, caution must be exercised as daily plasma exchange therapies are frequently associated with adverse transfusion reactions [11]. The rapid evaluation of ADAMTS-13 activity and inhibitor titers could also prevent unnecessary exposure to the risks of plasma exchange in the patients without TTP.

Since discovering the deficiency of ADAMTS-13, originally named VWF Cleaving Protease, as a key causal factor in TTP pathogenesis [7–9], many attempts have been made to measure plasma ADAMTS-13 activity in patients with TTP. Furlan and Tsai independently reported the original VWF protease assays [8,9]. In both of their assays human VWF, purified from plasma, was used as a substrate for VWF cleaving protease under denatured conditions. VWF cleaving protease activity was then estimated using an electrophoresis/Western blot that detects either VWF substrate disappearance or generation of VWF cleavage products. Since then, other methods have been reported including: collagen binding assays [12], immunoradiometric assays [13], and ristocetin cofactor activity assays [14]. These methods, however, do not directly measure functional activity of the ADAMTS-13 protease. One assay evaluates the cleavage of ulVWF multimeric strings from stimulated human endothelial cells [15]. Recently, a recombinant VWF protein containing a GST fusion protein and a histidine tag (rVWF73) was developed [16–18]. This substrate contains a specific cleavage site for ADAMTS-13 (Tyr1605-Met1606) and has been used to measure ADAMTS-13 activity in TTP patients by Western blot analysis [18]. Although the assay demonstrated excellent reproducibility, western blot analysis has proven difficult for clinical application because it is time-consuming and difficult to standardize. Newer studies have been reported using either enzyme-linked immunosorbent assay (ELISA) or a fluorescent substrate to measure ADAMTS-13 activity in plasma [16,19,20]. The clinical utilities of these methods remain to be further evaluated. Overall, there still remains great clinical demand for a reliable, rapid, and functionally relevant ADAMTS-13 assay to provide essential data for effective management of this devastating disorder.

Surface enhanced laser desorption/ionization time-of-flight mass spectrometry (SELDI-TOF-MS) has the ability to provide rapid protein/peptide analysis [21]. One distinguishable feature of SELDI-TOF-MS involves the surface chemistry of ProteinChips that allows for selective, rapid purification of protein/peptide candidates prior to analysis by mass spectrometry. In this application (Fig. 1), a recombinant VWF peptide containing the ADAMTS-13 cleavage site and a 6X Histidine tag at the N-terminus (rVWF73) was cleaved specifically by ADAMTS-13 in the patient's plasma that generated a peptide (Mr = 7739 Da) product containing a histidine tag. This product bound to an IMAC (Immobilized Metal Affinity Capture) ProteinChip specifically and was then effectively quantified by SELDI-TOF-MS. By this approach, functional activity of ADAMTS-13 in patient plasma can be evaluated in a very rapid fashion.

Figure 1.

Schematic diagram for detecting ADAMTS-13 cleavage product of VWF73 by SELDI TOF mass spectrometry.


Expression and purification of recombinant protein of VWF73

Human VWF73 (D1596-R1668), constructed in a pGEX-6P-1 expression vector as a GST-6xHis fusion, was a gift from Dr Toshiyuki Miyata at the National Cardiovascular Center Research Institute in Japan [18]. The VWF73 vector was used to infect Escherichia coli, BL21 (Stratagen, La Jolla, CA, USA). Protein expression was induced by 1 mm isopropyl-b-D-thiogalactoside (IPTG). The bacterial cells were then centrifuged and lysed in a CelLytic B (Sigma, St Louis, MO, USA) for 15 min, followed by sonication (4 W, five times). The recombinant VWF73 in the bacterial lysate was then purified by sequential chromatographies using glutathione affinity columns (Amersham Biosciences, Uppsala, Sweden) followed by nickel affinity columns (Qiagen, Valencia, CA, USA). Finally, the purified VWF protein was dialyzed against 5 mm Tris-HCl, 5 mm NaCl, pH 7.5. Protein concentration was determined using bicinchoninic acid (BCA) protein assay [22]. Purity of VWF73 was > 95% based on analysis by sodium dodecyl sulfate-poly acrylamide gel electrophoresis (SDS-PAGE).

Cleavage of rVWF73 substrate by plasma ADAMTS-13

The cleavage reaction, containing 2.2 μL of plasma and 1.8 μg VWF73 in a 30 μL of buffer (5 mm Tris HCl, 5 mm NaCl, 1 mm BaCl2, pH 7.5), was performed for 60 min at 37 °C and terminated by boiling at 95 °C for 2 min. Each experiment included a standard curve performed under identical conditions except that the plasma sample was replaced by pooled normal plasma (PNP) dilutes at 50%, 25%, 10%, 5%, 2.5% in 100 mm NaCl containing 0.1% bovine serum albumin (BSA).

Preparation of an internal peptide standard containing 6XHis

Recombinant VWF73 was cleaved by PreScissionTM Protease (Amersham Biosciences) to generate a 9340 Da peptide containing a 6X Histidine tag. Briefly, the VWF73, at 1.4 mg mL−1, was incubated with PreScissionTM at 60 units mL−1 in 150 mm NaCl, 50 mm Tris-HCl, pH 7.0 containing 1 mm DTT overnight at 4 °C. Boiling the mixture for 2 min at 95 °C terminated the reaction.

Analysis of the cleaved peptide by IMAC ProteinChip on SELDI –TOF-MS

This experiment was performed using a SELDI Bioprocessor. Each spot on the IMAC ProteinChip was charged with 50 mm nickel sulfate for 15 min. The spot was then washed extensively with washing buffer (50 mm sodium phosphate, pH 7.2, containing 0.8 m sodium chloride and 0.1% Triton X-100) followed with binding buffer (50 mm sodium phosphate, pH 7.2). Twenty μL of VWF73 cleavage product (described above) and 40 μL of internal standard (1 : 20 diluted in binding buffer) were mixed and added to each spot for 30 min incubation at room temperature with constant shaking. Afterwards, each spot was washed five times with 200 μL of washing buffer with three washes for 5 min and then two quick washes. This was followed by one quick wash with 1 mm4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), pH 7.0. Finally, 1 μL of energy absorbing molecule (EAM) solution (100% saturated Sinapinic acid in 50% acetonitrile and 0.5% trifluoroacetic acid) was added to each spot, the chips were allowed to air dry, and then the step was repeated. IMAC ProteinChips were analyzed by SEDLI-TOF-MS instrument (Protein Biology System II, Ciphergen, Fremont, CA, USA). The data collection was acquired using the following settings: data acquisition using SELDI quantitation mode; mass settings with optimized range of mass to charge ratio (m/z) from 3000 to 20 000; laser intensity at 165; detector sensitivity at 8. After baseline subtraction, data was analyzed with ProteinChip Software (version 3.1). Peak labeling was performed manually for cleaved peptide and internal control based on calculated molecular weight. The peak area was used to quantify the analytes. Because of the fact that virtually all analytes displayed a signal to noise ratio > 10 we defined the peak areas by slope-based option using the peak boundary: 25 times the noise per 11 times expected peak width. Data was then exported to Excel for further analysis.

Determining the inhibitory activity of ADAMTS-13 autoantibody

To inactivate ADAMTS-13 protease activity, patient's plasma was heated for 30 min at 56 °C followed by centrifugation for 15 min at 15 000 g to remove insoluble proteins. Twenty microliters of supernatant were mixed with 20 μL of PNP and incubated at 37 °C for 1 h. An aliquot of the mix was then assayed for residual ADAMTS-13 activity, as described above. Inhibition was indicated by a decrease in the amount of cleaved peptide by PNP after incubation with patient's plasma. Results were expressed as percentage of PNP ADAMTS-13 activity neutralized by an equal amount of patient's plasma.

Results and discussion

When a fixed amount of rVWF73 was incubated with PNP dilutions containing a known percentage of ADAMTS-13, the cleaved peptide (m/z = 7739) of rVWF73 was produced in a dose-dependent manner (Fig. 2). Initial attempts were made to quantify the analyte according to ion peak area. As shown in Table 1, when assays for a standard curve, using various known percentages of PNP, were performed multiple times independently (n = 9) inter-assay variation of the measurements, expressed as coefficient of variation (CV), ranged from 44% to 77%. As expected, more inter-assay variation was seen when the smaller peaks, corresponding to low ADAMTS-13 activity, were measured by mass spectrometry. The overall inter-assay variation probably resulted from the cumulative effect of experimental variations and detection variations intrinsic to mass spectrometry. In fact, numerous studies have reported broad ranges of variation when the same sample is quantified multiple times by mass spectrometry [21,23]. Variations associated with SEDLI mass spectrometry may include: differences of ionization from sample to sample, heterogeneity of surface chemistry on the SELDI ProteinChip, and others. To compensate for these known variations, we included a fixed amount of internal control peptide in all samples prior to quantifying the analyte by SELDI. This internal control was prepared from the same rVWF73 but cleaved with PreScission protease that cleaves and produces a peptide at m/z = 9340 with a 6X Histidine tag. For all subsequent calculations, we used the ratio of analyte to the internal standard rather than the ion peak area. As shown in Table 1, quantification as a ratio of an internal control significantly enhanced assay reproducibility with an average CV of 33% when compared with an average of CV at 59% without the internal control. Figure 3 was plotted by the mean peak areas (A) and peak ratios (B) as a function of percentage of ADAMTS-13. Again, the standard curve using peak ratios for quantification displays a better data fit at R2 = 0.995 as compared with R2 = 0.98 when peak area was used for calculation. Ideally, the assay would display a linear relationship between the percentage of ADAMTS-13 activity in the plasma and peak ratios of cleaved products. However, in this case, the SELDI mass spectrometer indirectly quantified plasma ADAMTS-13 activity by measuring the cleaved protein product. The rate of the cleavage reaction is likely influenced by multiple factors including the catalytic property of the enzyme, concentration of reactant, and other conditions. Similar to other methods that evaluate in vitro cleavage of recombinant VWF peptide, this approach will not be able to detect the defect(s) that causes TTP by impairing effective interaction of ADAMTS-13 with A1 or A3 domain of VWF substrate, as reported recently [24].

Figure 2.

Quantification of VWF73 cleavage by ADAMTS-13 using SELDI-TOF mass spectrometry. Different percentages of PNP were incubated with a fixed amount of recombinant VWF for 60 min at 37 °C. The cleaved product, 7739 Da peptide, was then detected by SELDI-TOF. An internal standard (9340 Da peptide) was added before ionization reaction during SELDI TOF analysis.

Table 1.  Comparison of inter-assay variations
% of PNP Peak area (n = 9)Ratio of peak area (n = 9)
MeanSD% CVMeanSD% CV
Figure 3.

A dose-dependent cleavage of VWF73 by ADAMTS-13. Assays were performed using a fixed amount of VWF73 and various dilutions of PNP. VWF cleavage product, 7739 Da peptide, was quantified by SELDI TOF. The standard curves were plotted using either mean peak areas (A) or mean peak ratios (B), shown in Table 1, as a function of PNP dilutions in the reaction.

Some key issues were considered during development and validation of this method. Test turnaround time is a critical factor that helps for timely clinical decision-making and influences clinical outcomes. Test detection sensitivity is also essential for adequate monitoring of patient's ADAMTS-13 levels during the course of therapy. To address these issues, various assay conditions were tested extensively. By meticulous experimentation and optimization of assay conditions, we finally chose an assay condition that achieved a turnaround time of < 4 h, with excellent detection sensitivity, and with acceptable reportable ranges. To summarize, the assay included a 60 min reaction time for plasma ADAMTS-13 to cleave rVWF73, binding of the analyte to the IMAC ProteinChip, washing steps, peptide ionization, and calculation of ADAMTS-13 activity according to the standard curve. The assay can adequately quantify ADAMTS-13 activity as low as 2.5% in plasma. This is important as a severe deficiency of ADAMTS-13 activity associated with idiopathic TTP is typically defined as being < 5% activity. This detection sensitivity was achieved, however, at the expense of using a standard curve that only covered ADAMTS-13 levels between 2.5% and 50% (Fig. 3). Under the final chosen condition, undiluted PNP (100% ADAMTS-13) appears to exhibit much greater inter-assay variation with CV at 98%. This less reproducible result may be because of the limitation in the overall range of ion peak detection by mass spectrometry and the complex relationship between amount of enzyme, cleavage of substrate, binding of analyte to the IMAC ProteinChip, and ionization efficiency of the analyte. However, this situation would not impact clinical practice as nearly all idiopathic TTP patients display an ADAMTS-13 activity at < 10% at the onset of disease. When a specimen exhibits an activity > 50%, it can be reported either as > 50% or repeated after obtaining the result using a diluted sample.

To further validate the test accuracy and to develop test reference ranges, we assayed plasma samples from healthy donors (n = 39) and from patients with known idiopathic TTP (n = 16). As shown in Fig. 4, the sampling of healthy donors included equal distribution of genders with age range from 18 to 70. These subjects exhibited ADAMTS-13 activities ranging from 61% to 122% (mean = 92.5% ± 16.6%). From our TTP cohort at the Ohio State University Medical Center, we selected a group of patients with the characteristics as shown in Table 2. They all have: (i) clinical presentation as idiopathic TTP; (ii) low ADAMTS-13 activity measured by western blot analysis using rVWF73 as a substrate [18]; (iii) strong evidence of microangiopathic findings at the time of diagnosis; and (iv) a history of TTP relapse. As expected, these clinically well-characterized TTP patients all demonstrated low ADAMTS activities (mean = 1.73% ± 3.6%). Notably, there was virtually an absence of overlap in ADAMTS-13 levels between TTP patients and healthy donors. Thus, a diagnostic test to measure functional level of ADAMTS-13 will likely be robust with regard to potential inter-assay variation.

Figure 4.

ADAMTS-13 activity in normal donors and idiopathic TTP patients.

Table 2.  Characteristics of patients with a diagnosis of idiopathic TTP
Patient no.SexAgeRaceInitial presentation
Initial ADAMTS-13 by SELDI (%)Mixing studies, residual ADAMTS-13 activity (%)Number of occurrences over duration of follow-up (months) Platelet count (150–400) LDH (100–190) Haptoglobin (20–230) Hg (13.2–17.3) Bilirubin, indirect (0–0.9)Schisto-cytesRenal failureChanges of mental status
  1. *First occurrence is recorded from patient medical history. Laboratory results in the table were from second occurrence.

*16M51AA<2.58.42/11 yrs7724<89.23.4XX 

Finally, we examined whether this method is able to functionally determine the inhibitory activity of ADAMTS-13 autoantibody. As the vast majority of adult TTP cases are acquired, we tested the antibody's neutralizing activity in the same group of idiopathic TTP patients selected for our validation study. The experiments were performed by mixing an equal amount of patient plasma with PNP. The residual ADAMTS-13 activity in the mix was determined that is inversely related to neutralizing activity of ADAMTS-13 autoantibodies in patient plasma. The residual ADAMTS-13 activity is reported in Table 2. The percentage of ADAMTS-13 activity inhibited by each TTP sample is illustrated in Fig. 5. As expected, the plasma samples from these idiopathic TTP patients all demonstrated inhibitory activity toward ADAMTS-13 protease in PNP but levels of neutralization varied from 14% to as high as 92%. Thus, this novel method offers the additional utility to quantify the neutralizing ability of the patient's plasma to inhibit in vitro ADAMTS-13 activity in a timely fashion. For example, a direct comparison of ADAMTS-13 autoantibody level to their inhibitory activity in TTP cohort may provide new insight into the pathobiology of TTP autoantibodies.

Figure 5.

Percentage of ADAMTS-13 activity in PNP neutralized by an equal amount of plasma from each TTP patient. The data are from duplicates.

In conclusion, this novel method employs SELDI-TOF-MS technology to provide the most rapid analysis of functional ADAMTS-13 activity to date. With a turnaround time of < 4 h, this method will offer great utility for making timely clinical decisions with respect to the necessity for initiating and continuing plasma exchange therapy in patients with clinical suspicion of a thrombotic microangiopathy. In addition, this method offers excellent detection sensitivity and quantification of ADAMTS-13 activity. Moreover, this method is able to determine functional activity of ADAMTS-13 autoantibodies, providing valuable information to help with the differential diagnosis and for selecting an appropriate treatment strategy.


This study is supported in part by grants from National Institutes of Health K08HL03279 and Ohio Biomedical Research and Technology Transfer Commission.