Full-length ADAMTS13 Research Articles

ADAMTS13 promotes angiogenesis and modulates VEGF-induced angiogenesis

Plasma ADAMTS13 deficiency results in the clinical disorder thrombotic thrombocytopenic purpura. However, other potential pathophysiological roles of ADAMTS13 in endothelial cell biology remain unexplored. To assess the possible role of ADAMTS13 and its interactions with VEGF-mediated angiogenesis, the effects of ADAMTS13 on human umbilical vein endothelial cell (HUVEC) were studied in Matrigel tube formation, proliferation, cell migration, and scratch wound assays.Treatment of endothelial cells with exogenous recombinant full-length ADAMTS13 alone promoted angiogenesis in a dose-dependent manner. HUVEC incubated with 200ng/mL ADAMTS13 (1.4nM) resulted in a 65% increase in cell tube formation when compared to the EBM-2 control. HUVEC treated with 30ng/mL ADAMTS13 (204.1pM) resulted in an 83% increase in proliferation in a visual counting assay, whereas HUVEC treated with 10ng/mL ADAMTS13 (68.0pM) yielded a 295% increase in EC migration in a Boyden chamber assay.In contrast, ADAMTS13 inhibited VEGF-induced angiogenesis in a dose-dependent manner, with 200ng/mL inhibiting tube formation by 35%. HUVEC co-incubated with ADAMTS13 and an antibody to the ADAMTS13 thrombospondin domains 5–7 reversed the inhibition of tube formation. HUVEC treated with 30ng/mL ADAMTS13 and 6.2ng/mL (323.0pM) VEGF proliferated 40% slower than the VEGF control after 24h of incubation as measured by visual counting assay. Treatment of HUVEC with 6.2ng/mL VEGF and 10ng/mL ADAMTS13 inhibited cell migration by 48%, compared to the VEGF control.Substitution of ADAMTS13 with truncated ADAMTS13 (deletion of C-terminal TSP1 domain) did not significantly increase angiogenesis or suppress VEGF-induced angiogenesis, suggesting that the TSP1 domain is involved in ADAMTS13 angiogenic activities. Co-immunoprecipitation experiments provided further evidence that ADAMTS13 binds to VEGF via its TSP1 domain.

Infusion of Recombinant Full-Length ADAMTS13 and Carboxyl-Terminal Truncated Variant Alters Composition of Ferric Chloride-Induced Arterial Thrombi in Adamts13−/− mice

Abstract 2313Proteolytic cleavage of ultra large von Willebrand factor (ULVWF) released from endothelial cells by ADAMTS13 metalloprotease is critical for maintaining normal hemostasis. However, the effect of infusing ADAMTS13 on thrombus composition remains poorly understood. In this study, we determined the morphology and composition of thrombi formed in carotid arteries after topical application of FeCl3 in Adamts13−/− mice receiving PBS, recombinant human full-length ADAMTS13 (FL) and carboxyl-terminal truncated variant after spacer domain (S), using scanning electron microscopy and quantitative image analysis. We showed that in Adamts13−/− mice 5 min after FeCl3 injury, formed arterial thrombi were comprised ∼39% platelets, ∼26% red blood cells, and ∼35% fibrin. The arterial thrombi in these mice were structurally deformed. An infusion of recombinant FL (final concentration of 10 nM) significantly reduced the accumulation of platelets (∼18%) but increased the fibrin network (57%) without affecting the composition of red blood cells (∼25%) in the arterial thrombi formed at the same time point after FeCl3 injury. Similar effects on the morphology and composition of FeCl3-induced arterial thrombi were observed after infusion of recombinant S (10 nM) into Adamts13−/− mice. Kinetic analysis showed that there was a decrease in platelet accumulation over the time of 30 min during thrombus formation with a slight increase in accumulation of red blood cells and formation of fibrin in Adamts13−/− mice. But, the infusion of recombinant FL and S into Adamts13−/− mice restored the kinetics of platelet/red blood cell accumulation and fibrin formation to those observed in wild-type mice. Our findings, revealing the apparent difference in thrombus composition, provide novel insight into the mechanism of ADAMTS13 function in vivo, which may shed more light on the pathogenesis of thrombotic thrombocytopenic purpura and other arterial thrombotic disorders associated with deficiency of plasma ADAMTS13 activity. Disclosures:No relevant conflicts of interest to declare.

Inhibitory Spleen-Derived Anti-ADAMTS13 Antibodies Are Characterized by a Limited Number of Variable Heavy Chain CDR3 Signatures in Patients with Relapsing Acquired Thrombotic Thromobocytopenic Purpura

Abstract 194 Background.Acquired Thrombotic thromobcytopenic purpura (TTP) is the result of autoantibodies neutralizing and-or accelerating clearance of ADAMTS13. We have previously isolated mononuclear cells of two spleens donated to our laboratory by patients (A and B) splenectomized for frequently relapsing TTP to generate human monoclonal anti-ADAMTS13 antibodies. Preliminary characterization of the patient’s entire IgG4 repertoire using phage display technology and Epstein Barr virus (EBV) transformation of switched memory B cells (CD19+, CD27+, IgG+) showed that the memory repertoire of Rituximab-resistant (B) and non-resistant (A) anti-ADAMTS13 IgG used the same restricted VH germline genes, namely VH1-3 (55%), VH1-69 (17%), VH3-30 (7%) and 4–28 (21%). In contrast, peripheral blood-derived IgG1 anti-ADAMTS13 antibody repertoire from 2 acquired TTP patients, inhibiting ADAMTS13 activity by maximal 15–40% used VH1-69 germline gene segment only (Luken et al, JTH 2006 and Pos et al, JTH 2009). To evaluate if the generated spleen-derived monoclonal anti-ADAMTS13 antibodies are of pathological relevance we started with their functional characterization. Methods.Both patients (A and B) were suffering from acute TTP with high titers of anti-ADAMTS13-Ab and therefore needed to be splenectomized because of frequent relapses, which occurred even after several courses of Rituximab (B). Both patients have remained in remission since splenectomy. Previous screening of the IgGk-IgGλ IgG4-Fab phage-display libraries and EBV-transformed clones on recombinant ADAMTS13 by ELISA revealed high ADAMTS13 specificity for 16/34 and 13/234 antibodies, respectively. Functional characterization of the phage-display clones included (a) determination of their inhbitiory capacity by functional FRETS inhibitor assay, (b) assessment of their affinity performed by either fluid-phase competition ELISA and/or plasma surface resonance (Biacore), and (c) epitope mapping by dot blot analysis testing their recognition of recombinant full-length wildtype ADAMTS13 and a truncated ADAMTS13 variant. Results.Detailed sequence analysis of all human monoclonal anti-ADAMTS13 analyzed showed seven unique CDR3 signatures in the variable heavy region of all clones generated using both techniques (phage display and EBV-transformation), three of which were seen in both patients. These CDR3 signatures are different from signatures found in peripheral blood-derived antibodies (Luken et al, JTH 2006 and Pos et al, Blood 2010). Strikingly, the anti-ADAMTS13 antibodies of EBV-transformed memory B cells used λ light chain exclusively, whereas the phage-display-derived clones showed usage of either κ- (A: 6/11 and B: 3/5) or λ-light chain (A: 5/11 and B: 2/5). Anti-ADAMTS13 IgG in supernatants of EBV-immortalized switched memory B-cell clones were mainly of subclass IgG4 (A: 60%; B: 25%) and IgG1 (A: 40%; B: 75%) thus IgG1 being the prevailing subclass in patient B, who relapsed despite several courses of Rituximab. Functional characterization of the anti-ADAMTS13 IgG4 antibodies revealed that 15/16 phage display clones are strongly inhibitory, showing 0–100% (A) and 60–100% (B) inhibition at 200 nM and even 0–92% (A) and 0–16% (B) inhibition at 20 nM. This represents a 5–6 times increase of inhibitory capacity when compared to previouse findingy by others, in the range of 2×10−7-7×10−10 M (A) and 1×10−9-6×10−10M (B) as measured by Biacore or/and competition ELISA. Epitope mapping revealed that 3/16 clones recognized a novel epitope at the C-terminus of ADAMTS13, in contrast to the previously reported N-terminally located spacer domain. Conclusions.The identification of several spleen-derived anti-ADAMTS13 clones displaying highly similar CDR3 variable heavy chains, in 2 unrelated acquired TTP patients is striking and novel. Our data points at the discovery of unique CDR3 signatures that characterize the pathological relevant, inhibitory antibodies paving thus the road to define their triggering ADAMTS13 epitope(s), while providing insight into the mode of success of splenectomy in relapsing TTP patients. Further functional characterization of the anti-ADAMTS13 EBV-derived clones and confirmation of our data by analysis of the immunological memory from two additional spleens are underway. Disclosures:No relevant conflicts of interest to declare.

Carboxyl-Terminal Thrombospondin-Type 1 (TSP-1) Repeats of ADAMTS13 Inhibits Platelet Adhesion to Collagen-Coated Surface Under Shear Stress, Independent of Proteolytic Activity

Abstract 195ADAMTS13 contains multiple free thiols on its surface, which may form disulfide bonds with surface-exposed free thiols on plasma-derived von Willebrand factor (VWF). This interaction may prevent lateral association of apposed VWF under arterial shear stress. However, the functional consequence of ADAMTS13-VWF interaction without proteolysis is not known. We hypothesize that the interaction between the C-terminus of ADAMTS13 and the C-terminus of VWF inhibits thrombus formation under shear stress. Using a BioFlux microfluidic system, we showed that under arterial shear stress, 10 dyn/cm2, fluorescein-labeled platelets from PPACK (thrombin inhibitor) anti-coagulated human whole blood adhered to collagen (type I)-coated surface in a time-dependent manner. Addition of human recombinant full-length ADAMTS13 (10 nM) into whole blood dramatically reduced the surface coverage of fluorescein-labeled platelets. Conversely, addition of an inhibitory polyclonal anti-ADAMTS13 IgGs (150 ug/ml) to whole blood dramatically accelerated the accumulation of fluorescein-labeled platelets. These results suggest that this microfluidic system is highly sensitive for the assessment of anti-thrombotic function of ADAMTS13. Under the same conditions, we were able to further show that addition of recombinant C-terminal fragment of ADAMTS13 comprising of the 5th to 8th thrombospondin type 1 (TSP1) repeats and two CUB domains (T5C) or the 2nd to 8th TSP1 repeats and two CUB domains (T2C) into whole blood also inhibited the surface coverage of fluorescein-labeled platelets on collagen-coated surface in a concentration-dependent manner. In the presence of 0.1 μM and 0.5 μM of recombinant T2C or T5C, the surface coverage of fluorescein-labeled platelets was reduced by ∼40% and ∼60%, respectively. The inhibitory activity of these recombinant C-terminal fragments was nearly abolished if pre-treated with 40 mM of N-ethylmaleimide which blocked surface-exposed free thiols. Moreover, recombinant CUB domains at the highest concentration tested (1.0 μM) did not appear to alter the surface coverage of fluorecein-labeled platelets under the same conditions. These results suggest that the C-terminal TSP1 repeats of ADAMTS13 inhibit platelet adhesion and aggretion or thrombus formation through thiol-thiol interactions between ADAMTS13 and VWF (or other proteins). We conclude that the C-terminal TSP1 repeats may modulate thrombus formation independent of proteolytic activity. Disclosures:No relevant conflicts of interest to declare.

Novel recombinant glycosylphosphatidylinositol (GPI)-anchored ADAMTS13 and variants for assessment of anti-ADAMTS13 autoantibodies in patients with thrombotic thrombocytopenic purpura

Immunoglobulin Gs (IgGs) against ADAMTS13 are major causes of acquired (idiopathic) thrombotic thrombocytopenic purpura (TTP). We report here a novel cell-based assay using glycosylphosphatidylinositol (GPI)-anchored ADAMTS13 or variants expressed on cell membrane for assessment of autoantibodies in patients with TTP. We showed that IgGs from all 26 patients with acquired TTP bound to cells expressing a GPI anchored full-length ADAMTS13 (gFL) and a variant truncated after the spacer domain (gS). Also, IgGs from 25/26 (96.7%) of these TTP patients bound to cells expressing a GPI-anchored C-terminal fragment, TSP1 2-8 plus CUB (gT2C). In contrast, none of the 20 healthy blood donors showed detectable binding of their IgGs to the cells expressing gFL, gS, and gT2C. A moderate, but statistically significant correlation was observed between plasma concentrations of anti-ADAMTS13 IgG and positive cells expressing gFL (r=0.65), gS (r=0.67), and gT2C (r=0.42). These results suggest that the microtiter-plate assay and the cell-based assay may detect differential antigenic epitopes. Moreover, antigens clustered on cell membranes may enhance antibody binding affinity, thereby increasing analytical sensitivity. Finally, our assay was able to determine kinetic changes of plasma levels of anti-ADAMTS13 IgGs in TTP patients during plasma therapy. Together, our findings suggest that the novel cell-based assay may be applicable for rapid identification and mapping of anti-ADAMTS13 autoantibodies in patients with acquired TTP.

Thrombospondin-1 and ADAMTS13 competitively bind to VWF A2 and A3 domains in vitro

Introduction ADAMTS13 (a disintegrin-like and metalloprotease with thrombospondin type 1 repeat motif. 13) is the major metalloprotease for VWF degradation. ADAMTS13 deficiency causes the accumulation of uncleaved VWF and might lead to a lethal thrombotic thrombocytopenic purpura (TTP). Thrombospondin-1 (TSP1) is considered as a reductase of VWF (von Willebrand factor) which can mildly downregulate the size of VWF by targeting on disulfide bond between VWF dimers. It was reported that TSP1 might protected VWF from cleaving by ADAMTS13, yet the underlying mechanism of this VWF protection has remained unknown. Materials and Methods Full-length ADAMTS13 and different domains (A1,A2,A3) of human VWF were constructed and expressed respectively. The binding ability of TSP1 or ADAMTS13 with each VWF domain or full-length VWF was investigated by using enzyme linked immunosorbent assay. The inhibition of ADAMTS13 activities by the different concentrations of TSP1 were observed by western blot and residual-collagen binding assay (R-CBA) under the denaturing condition. Results We found that ADAMTS13 interacted with the rVWF A1, A2, A3 domains and full-length VWF, while TSP1 also bound to three A domains, especially to A2 and A3 domains. We observed that TSP1 partially blocked ADAMTS13 binding to A2 domain, A3 domain and full length VWF. The results of our assays showed that TSP1 could restrain ADAMTS13 activity up to 70%. Conclusions Our study suggested that TSP1 played competitively inhibitory role in ADAMTS13 binding and cleaving of VWF, and the potential competition might happen within A2 and A3 domains.

Multiple domains of ADAMTS13 are targeted by autoantibodies against ADAMTS13 in patients with acquired idiopathic thrombotic thrombocytopenic purpura

Type G immunoglobulins against ADAMTS13 are the primary cause of acquired (idiopathic) thrombotic thrombocytopenic purpura. However, the domains of ADAMTS13 which the type G anti-ADAMT13 immunoglobulins target have not been investigated in a large cohort of patients with thrombotic thrombocytopenic purpura. Sixty-seven patients with acquired idiopathic thrombotic thrombocytopenic purpura were prospectively collected from three major U.S. centers. An enzyme-linked immunosorbent assay determined plasma concentrations of anti-ADAMTS13 type G immunoglobulins, whereas immunoprecipitation plus western blotting determined the binding domains of these type G immunoglobulins. Plasma anti-ADAMTS13 type G immunoglobulins from 67 patients all bound full-length ADAMTS13 and a variant truncated after the eighth TSP1 repeat (delCUB). Approximately 97% (65/67) of patients harbored type G immunoglobulins targeted against a variant truncated after the spacer domain (MDTCS). However, only 12% of patients' samples reacted with a variant lacking the Cys-rich and spacer domains (MDT). In addition, approximately 37%, 31%, and 46% of patients' type G immunoglobulins interacted with the ADAMTS13 fragment containing TSP1 2-8 repeats (T2-8), CUB domains, and TSP1 5-8 repeats plus CUB domains (T5-8CUB), respectively. The presence of type G immunoglobulins targeted against the T2-8 and/or CUB domains was inversely correlated with the patients' platelet counts on admission. This multicenter study further demonstrated that the multiple domains of ADAMTS13, particularly the Cys-rich and spacer domains, are frequently targeted by anti-ADAMTS13 type G immunoglobulins in patients with acquired (idiopathic) thrombotic thrombocytopenic purpura. Our data shed more light on the pathogenesis of acquired thrombotic thrombocytopenic purpura and provide further rationales for adjunctive immunotherapy.

In Vitro and In Vivo Mouse Models of the Type 1 Von Willebrand Disease Mutations Y1584C and R1205H.

Abstract 26 Introduction:Type 1 von Willebrand disease (VWD) is caused by mutations that result in moderate decreases in VWF (5-50% of normal levels) and a mild bleeding phenotype. The VWF missense mutation Y1584C is associated with mildly decreased VWF:Ag levels, increased ADAMTS13 cleavage, as well as a possible increase in clearance. The Vicenza mutation, R1205H, exhibits a more severe phenotype (VWF:Ag ∼10%), as well as accelerated clearance. However, extensive controlled in vitro and in vivo investigation of these mutations has yet to be described. In this study, we examine both Y1584C and R1205H in comparison to wild type VWF using in vitro and in vivo strategies employing both human and mouse VWF. Methods:Recombinant murine and human VWF and ADAMTS13 were produced via transient transfection in HEK293T cells in serum free OPTIMEM for 72 hours. Full length ADAMTS13 digests were performed in a Tris-Urea system and analyzed via multimer pattern. VWF115 digests were performed in an ELISA based assay. Hydrodynamic injections were performed using 100 μg wild type (WT) or mutant ET-mVWF plasmid DNA in Ringer's solution in 7-9 week old C57Bl6 VWF knockout mice. Mice were sampled at days 2, 5, 8, and then weekly. Mouse plasma was analyzed for CBCs, VWF:Ag, and VWF multimer structure. Results:In the HEK293T transient transfection system, secreted mutant protein was similar to that of wild type recombinant protein, with high molecular weight material present. ADAMTS13 digestion of full length recombinant Y1584C versus wild type showed no statistical difference: 50% cleavage for hVWF WT 1.78U hADAMTS13, hVWF Y1584C 1.67 U, P=0.58; mVWF WT 0.32 U mADAMTS13; Y1584C 0.42 U, P=0.11. In contrast, the Y1584C substitution in the hVWF115 construct required 40% less hADAMTS13 to effect equivalent cleavage (WT 0.087±0.014, 4; Y1584C 0.052±0.005, 8, mean U/ml ADAMTS13±SEM, N, P=0.013). Mouse ADAMTS13 cleavage of mVWF115 was also increased 20% for Y1584C (WT: 0.20±0.06, 4; Y1584C: 0.16±0.01, 4, P=0.014). Hydrodynamic injection caused no adverse events in any animals. CBC values were not statistically significantly different between wild type and mutants. Initial high VWF:Ag values were similar for wild type VWF (25.6±2.9, 15, mean U/ml±SEM, n) and Y1584C (27.2±5.0, 11), but R1205H levels were 36% lower (16.3±2.1, 10). At 14 days, WT VWF:Ag was 5.33±1.13, 15, with R1205H (1.73±0.44, 12) and Y1584C (1.84±0.59, 11) VWF:Ag levels being 68% and 65% lower, respectively. R1205H continued to remain approximately 40% of WT values for the next three weeks, while Y1584C continued to decrease, dropping to 15% on day 21 and 8% on day 28, compared to the wild type values at these time points. The R1205H mutation showed no significant difference in multimer structure defined by observed number of bands (days 2 to 42, mean difference -0.49, P>0.05) to wild type VWF. In contrast, Y1584C had a significant decrease in band number (3.38, P<0.001). Conclusions:This study demonstrates that these two type 1 VWD mutations have a strong observable effect in the VWF knockout mouse model. R1205H exhibits a large decrease in VWF:Ag levels at all measured time points, but no alteration in multimer structure. Y1584C, in contrast, shows a loss of high molecular weight material, and an initially high VWF:Ag level that rapidly decays from day 14 onward, suggesting increased ADAMTS13 cleavage and increased clearance. In addition, in vitro ADAMTS13 testing shows that Y1584C responds differently in the two assay systems with only the VWF115 assay showing significant increases in ADAMTS13-mediated cleavage. Disclosures:No relevant conflicts of interest to declare.

Correction of murine ADAMTS13 deficiency by hematopoietic progenitor cell–mediated gene therapy

AbstractADAMTS13, a metalloprotease primarily synthesized in liver and endothelial cells, cleaves von Willebrand factor (VWF) at the central A2 domain, thereby reducing the sizes of circulating VWF multimers. Genetic or acquired deficiency of plasma ADAMTS13 activity leads to a potentially fatal syndrome, thrombotic thrombocytopenic purpura (TTP). To date, plasma infusion or exchange is the only proven effective therapy for TTP. In search for a better therapy, an autologous transplantation of hematopoietic progenitor cells transduced ex vivo with a self-inactivating lentiviral vector encoding a full-length murine Adamts13 and an enhanced green fluorescent protein (GFP) reporter gene was performed in Adamts13−/− mice after irradiation. All recipient mice showed detectable ADAMTS13 antigen and proteolytic activity in plasma despite only low levels of bone marrow chimerism. The levels of plasma ADAMTS13 were sufficient to eliminate the ultralarge VWF multimers and offered systemic protection against ferric chloride–induced arterial thrombosis. The data suggest that hematopoietic progenitor cells can be genetically modified ex vivo and transplanted in an autologous model to provide adequate levels of functional ADAMTS13 metalloprotease. This success may provide the basis for development of a novel therapeutic strategy to cure hereditary TTP in humans.

Correction of ADAMTS13 Deficiency by In Utero Gene Transfer of Lentiviral Vector encoding ADAMTS13 Genes

Deficiency of A Disintegrin And Metalloprotease with ThromboSpondin (ADAMTS13) results in thrombotic thrombocytopenic purpura (TTP). Plasma infusion or exchange is the only effective treatment to date. We show in this study that an administration of a self-inactivating lentiviral vector encoding human full-length ADAMTS13 and a variant truncated after the spacer domain (MDTCS) in mice by in utero injection at embryonic days 8 and 14 resulted in detectable plasma proteolytic activity (approximately 5-70%), which persisted for the length of the study (up to 24 weeks). Intravascular injection via a vitelline vein at E14 was associated with significantly lower rate of fetal loss than intra-amniotic injection, suggesting that the administration of vector at E14 may be a preferred gestational age for vector delivery. The mice expressing ADAMTS13 and MDTCS exhibited reduced sizes of von Willebrand factor (vWF) compared to the Adamts13(-/-) mice expressing enhanced green fluorescent protein (eGFP). Moreover, the mice expressing both ADAMTS13 and MDTCS showed a significant prolongation of ferric chloride-induced carotid arterial occlusion time as compared to the Adamts13(-/-) expressing eGFP. The data demonstrate the successful correction of the prothrombotic phenotypes in Adamts13(-/-) mice by a single in utero injection of lentiviral vectors encoding human ADAMTS13 genes, providing the basis for developing a gene therapy for hereditary TTP in humans.

Multiple Non-Linear Epitopes of ADAMTS13 Were Targeted by Anti- ADAMTS13 IgG in Patients with Acquired Idiopathic Thrombotic Thrombocytopenic Purpura

Most patients with idiopathic thrombotic thrombocytopenic purpura (TTP) are caused by acquired autoantibodies that bind and inhibit proteolytic activity of ADAMTS13, a metalloprotease that cleaves von Willebrand factor (vWF). However, the binding epitopes of these anti-ADAMTS13 autoantibodies in many cases have not been fully defined. The limited data in the literature were largely from studies of small sample size, using less desirable recombinant ADAMTS13 fragments expressed in E. coli bacteria or insect cells or employing a non-physiological technique for detection. The results were therefore either inconclusive or controversial. To overcome these pitfalls, the full-length ADAMTS13 and various fragments of ADAMTS13 were expressed in human embryonic kidney (HEK293) cells. The recombinant proteins were purified by a combination of anion exchange chromatography and Ni-NTA affinity column. The antibody and antigen reactions were all performed in phosphate-buffered saline at pH 7.4. The immune complexes were collected by incubation with protein A/G Sepharose. After electrophoresis under denatured and reduced conditions, the antigens (ADAMTS13 and fragments) were then detected by Western blot with anti-V5 IgG, which recognizes the carboxyl terminal V5- His epitopes of the recombinant proteins. We showed that IgGs from 67 patients with acquired idiopathic TTP all interacted with full-length ADAMTS13 (construct FL-AD13) and the CUB domainless variant (construct delCUB). Approximately 97% (65/67) of IgGs in TTP patients bound the variant truncated after the spacer domain (construct MDTCS). However, a removal of the Cys-rich and spacer domains (construct MDT) reduced its reactivity with IgGs from TTP patients by 88%, suggesting the important contribution of these domains to the antibody binding epitopes. Unexpectedly, a fragment consisting of a disintegrin domain, the first TSP1 repeat, the Cys-rich and spacer domains (construct DTCS) reacted with IgGs from only 52% of TTP patients, indicating the additional contribution of the metalloprotease domain to these non-linear binding epitopes. Moreover, approximately 31%, 37%, and 46% of patients' IgGs targeted against the carboxyl terminal CUB domains, TSP1 2–8, and TSP1 5-8CUB, respectively. Low platelet count was observed at the time of admission in patients with the IgGs that reacted with the TSP1 2–8 and CUB domains of ADAMTS13, suggesting that the autoantibodies with the specificity toward more distal carboxyl-terminal domains of ADAMTS13 may impair ADAMTS13 function in vivo. We conclude that while the Cys-rich and spacer domains of ADAMTS13 contain the core epitopes, other adjacent functional domains may also contribute to the non-linear binding epitopes that are recognized by anti-ADAMTS13 IgGs. The data further support the role of autoantibodies against the middle and distal carboxyl terminal domains of ADAMTS13 in pathogenesis of TTP. Our findings provide novel insight into the mechanism of autoimmune-mediated TTP.

Binding of ADAMTS13 to Von Willebrand Factor Under Static Conditions and Under Fluid Shear Stress

ADAMTS13 is a plasma metalloprotease that is essential for the normal proteolytic processing of von Willebrand factor (VWF). Dysfunctional ADAMTS13 may lead to thrombotic thrombocytopenic purpura, as uncleaved and unusually large VWF multimers accumulate in the blood and cause intravascular platelet aggregation. Many studies indicate that proteolysis of multimeric VWF involves conformational changes in the VWF A2 domain that expose the Y1605-M1606 scissile bond and also allow substrate binding to multiple exosites on ADAMTS13. For example, VWF is resistant to proteolysis by ADAMTS13 unless the VWF is subjected to fluid shear stress, mild denaturation with guanidine or urea, or adsorption onto a surface. However, the functional interactions between shear stress, various ADAMTS13 binding sites and VWF cleavage are not understood. Therefore, we investigated the effect of fluid shear stress and ADAMTS13 structure on ADAMTS13-VWF binding and VWF cleavage. Upon mixing recombinant VWF (rVWF) and ADAMTS13 in a physiological buffer (50 mM HEPES, 5 mM CaCl2, 1 μM ZnCl2, 150 mM NaCl, pH 7.4), we found that immunoprecipitation with anti-VWF also pulled down substantial amounts of ADAMTS13. Although less striking, a similar result was obtained with purified plasma VWF. Therefore, ADAMTS13 can bind VWF without gaining access to the cleavage site in VWF domain A2. When fluid shear stress was applied for 2 min with a bench-top vortexer, ADAMTS13 binding increased 3-fold and VWF was also cleaved. Lowering the ionic strength markedly increased the rate of VWF cleavage but did not affect ADAMTS13 binding, which suggests that cleavage and binding depend on distinct VWF-ADAMTS13 interactions. Shear-induced binding was reversible slowly upon removal of unbound ADAMTS13 or rapidly by addition of SDS. ADAMTS13-VWF binding was stable for at least 24 h after cessation of shear stress, indicating that the structural change in VWF that promotes binding was not readily reversible. Using a catalytically inactive ADAMTS13 variant to simplify the analysis of binding assays, 30 nM ADAMTS13(E231Q) bound to 30 μg/ml rVWF (120 nM subunits) with a stoichiometry of 0.012 ± 0.004 under static conditions and 0.098 ± 0.023 after shearing (mean ± SD, n = 3, P = 0.019). With 120 nM ADAMTS13(E231Q) the stoichiometry increased to 0.086 ± 0.036 under static conditions and 0.469 ± 0.033 after shearing for 2 min. Recombinant ADAMTS13 truncated after TSP-1 repeat 8 (lacking the C-terminal CUB domains, delCUB), or truncated after the Spacer domain (consisting of domains MDTCS), did not bind rVWF under static conditions, implicating the CUB domains in binding to VWF. In contrast, full-length ADAMTS13, delCUB and MDTCS bound similarly to rVWF after shearing. In a previous study, delCUB and MDTCS did not cleave VWF subjected to fluid shear stress (Zhang et al, Blood 2007; 110: 1887–1894). However, under the conditions employed in these experiments, MDTCS and delCUB displayed significant proteolytic activity, cleaving VWF at a rate comparable to that of full length ADAMTS13 when shear stress was applied over a time course of 0–160 sec. We conclude that ADAMTS13 CUB domains contribute to binding a few sites on multimeric VWF under static conditions, whereas ADAMTS13 MDTCS domains are sufficient to bind many sites in an altered conformation of VWF that is induced by fluid shear stress. Binding of ADAMTS13 to unsheared VWF multimers may facilitate the cleavage of VWF within a growing thrombus.

Evidence for the Interaction Between VWF Carboxyl-Terminal Domains and ADAMTS13 Domains Distal to Its Spacer Domain

ADAMTS13 plasma metalloprotease regulates the multimeric size of von Willebrand factor (VWF) by cleaving the Y1605-M1606 bond in the A2 domain. Several studies have characterised high affinity binding sites in the A2 domain, probably interacting with the ADAMTS13 spacer and proximal domains. We have previously provided preliminary evidence for an additional ADAMTS13 binding site distal to the A domains, by measuring a ~7-fold increase in affinity for ADAMTS13 of plate-immobilised VWFA1CK (residues 1260-2813) (KD 11 nM) compared with VWFA1A2A3 (residues 1260-1874) (KD 78 nM) (Zanardelli et al, JTH, volume 5 Supplement I, XXI ISTH congress 2007, abstract number O-W-018). We subsequently expressed a series of VWF carboxyl-terminal fragments, VWFD4CK (residues 1874-2813), VWFA3CK (residues 1671-2813), VWFD4 (residues 1874-2299) and VWFB1CK (residues 2296-2813). Using an equilibrium binding assay, we assessed their binding to full-length ADAMTS13 and truncated ADAMTS13 proteins MDTCS (residues 34-685) and del(TSP5-CUB) (residues 34-894). KD values calculated for VWFD4CK and VWFA3CK binding to full-length ADAMTS13 were 134 and 108 nM, respectively. This confirmed the presence of an ADAMTS13 binding site outside VWF A2 domain, in the region D4CK. It also excluded a major role in the interaction for the VWF A3 domain. However, isolated VWFD4 and VWFB1CK did not bind significantly, suggesting that integrity of the VWF C-terminal region is also required for the binding. Furthermore, MDTCS bound to VWFA1CK and VWFA1A2A3 with similar affinity (KD 120 and 82 nM, respectively) and no significant interaction occurred with VWFD4CK. Similar results were found for the binding of del(TSP5-CUB) to the VWF fragments (VWFA1CK KD 85 nM, VWFA1A2A3 KD 65 nM and VWFD4CK no binding). Together these results suggested a multiple interaction between the carboxyl-terminal region of VWF (within D4CK) and the distal domains of ADAMTS13 (TSP5–8 and CUB domains). The VWF carboxyl-terminal binding site may be important for initial anchoring of ADAMTS13, allowing the subsequent binding between the spacer domain and VWF A2 domain.

Correction of Murine Adamts13 Deficiency by Hematopoietic Progenitor Cell-Mediated Gene Therapy

ADAMTS13, a metalloprotease that cleaves von Willebrand factor (vWF), is primarily synthesized in liver, endothelial cells and megakaryocytes/platelets. It circulates in plasma as an active protease with concentrations between 0.5 and 1.0 μg per milliliter of plasma. Proteolytic cleavage of endothelial cell bound and plasma vWF by ADAMTS13 is critical for maintaining normal hemostasis. Inability to cleave newly released unusually large vWF as a result of hereditary or acquired deficiency of ADAMTS13 activity leads to a potentially fatal syndrome, thrombotic thrombocytopenic purpura (TTP). Although plasma infusion or exchange is a proven effective therapy for TTP, the life-long plasma infusion for hereditary TTP is not without complications. In search for a better therapy, we performed an autologous transplantation of hematopoietic progenitor cells (HPCs) for correcting Adamts13 deficiency in a mouse model. The bone marrows were harvested from Adamts13−/− mice (C57BL6/129Sv) at 6–8 weeks of age and HPCs (CD48−/CD150+) were purified to a greater than 90% of purity in all cases using serial immune affinity depletion and enrichment techniques. The purified HPCs were transduced in culture with a self-inactivated lentiviral vector either encoding GFP or GFP plus murine full-length Adamts13 at 100 × MOI (multiplicity of infection) for 14–16 hours. The transduced HPCs (1 × 105) mixed with freshly prepared non-transduced bone marrow mononucleated cells (2 × 105) were then infused via tail vein into Adamts13−/− mice (recipients ) after lethal irradiation (450 cGy, twice, 3 h apart, at a dose rate of 2.2 cGy per minute). Peripheral blood was collected at 1, 3, and 5 months of post-transplantation for assessing the bone marrow chimerism and Adamts13 activity. The GFP positive cells were determined by flow cytometry and Adamts13 activity was determined by FRETS-vWF73 and GST-vWF73 peptides as described previously. We showed that all ten mice transplanted with the transduced HPCs encoding GFP plus murine Adamts13 exhibited persistent levels of plasma Adamts13 proteolytic activity, ranging from 10% to 60% of pooled normal murine plasma (PNP), despite low levels of bone marrow chimerism based on the percentage of GFP-positive mononucleated cells (~10–40%) in the peripheral blood. The expressed Adamts13 in plasma was able to cleave process unusually large vWF, resulting in a reduction of vWF multimers sizes compared with those in Adamts13−/− mice transplanted with HPCs transduced with the vector encoding GFP. Moreover, the levels of plasma Adamts13 appeared to be sufficient to protect mice against ferric chloride-induced carotid arterial thrombosis. The carotid arterial occlusion times for Adamts13−/− mice which did not undergo bone marrow transplantation were 5.6 ± 1.6 min, similar to those transplanted with HPCs transduced with the vector encoding GFP (4.3 ± 0.8 min) (mean ±SD) (p>0.05). However, the carotid arterial occlusion times were significantly prolonged in Adamts13−/− mice transplanted with the vector encoding GFP plus murine Adamts13 (15.5 ± 6.8 min). The differences between the experimental group and the control groups were all statistically very significant (p<0.01). The data suggest that hematopoietic progenitor cells can be genetically modified ex vivo and transplanted in an autologous model to provide adequate levels of functional Adamts13 metalloprotease. This success may provide basis for development of a novel therapeutic strategy to cure hereditary TTP in humans.

Interaction of ADAMTS13 with the Endothelial Cell Surface.

ADAMTS13 proteolysis of von Willebrand Factor (VWF) generates smaller multimers that are less likely to promote blood clotting. Deficiency of ADAMTS13 leads to thrombotic thrombocytopenic purpura, a frequently fatal disease, characterized by microangiopathic hemolytic anemia and thrombocytopenia. ADAMTS13 has a characteristic domain structure that includes metalloprotease and disintegrin domains, a thrombospondin type 1 repeat (TSR), cysteine-rich and spacer domains, 7 additional TSRs, and 2 carboxyl-terminal CUB domains. The ADAMTS13 substrate, VWF, is synthesized in endothelial cells and forms large multimers within the cell. These large multimers are secreted and adhere to the endothelial cell surface where they can bind platelets flowing in blood leading to thrombosis. ADAMTS13 has been shown to cleave VWF on the surface of endothelial cells, but it is unclear if ADAMTS13 also interacts with the endothelial cell surface. We have used iodinated ADAMTS13, fluorescence-activated cell sorting (FACS), and biochemical analysis using flow conditions to demonstrate that ADAMTS13 does interact with the endothelial cell surface. Iodinated ADAMTS13 bound the endothelial cell surface at 4oC. This binding was specific since the binding was inhibited in the presence of 40-fold excess unlabeled ADAMTS13. Binding of ADAMTS13 to the cell surface was time-dependent with maximal binding occurring within two hours. The binding was also reversible; the half-time for dissociation was four hours. Binding was inhibited by heparin but not by dextran sulfate. The Kd of binding to endothelial cells was 75 nM (range 40–100 nM). FACS analysis also demonstrated binding of ADAMTS13 to endothelial cells. A fluorescein isothiocyanate labeled anti-epitope antibody bound to endothelial cells in the presence but not the absence of ADAMTS13. A polyclonal antibody to VWF inhibited binding of ADAMTS13 to VWF, but this antibody did not affect binding of ADAMTS13 to endothelial cells, suggesting that ADAMTS13 interacts with endothelial cells independently of VWF. Studies with C-terminal truncation constructs of ADAMTS13 indicated that the carboxyl-terminal TSRs are important for binding since constructs terminating with the metalloprotease domain, the first TSR, or the sixth TSR failed to compete with full-length ADAMTS13 for binding to endothelial cells, but constructs terminating with either the seventh or eighth TSR did compete for binding. Lastly, recombinant ADAMTS13 was found to be associated with endothelial cells in flow experiments. Endothelial cells were perfused with medium containing plasma concentrations of ADAMTS13 (1 μg/ml) at 10 dynes/cm2. After perfusion, the endothelial cells were washed and bound ADAMTS13 was identified from whole cell lysates through SDS-PAGE and immunoblotting with an anti-V5 epitope antibody. ADAMTS13 was found associated with endothelial cells after perfusion. Binding of ADAMTS13 to the endothelial cells prior to perfusion led to enhanced proteolysis of VWF as compared to addition of ADAMTS13 during perfusion only. This suggests that the interaction of ADAMTS13 with endothelial cells is important since it enhances the cleavage of VWF as compared to that of ADAMTS13 in solution.

An IAP retrotransposon in the mouse ADAMTS13 gene creates ADAMTS13 variant proteins that are less effective in cleaving von Willebrand factor multimers

AbstractSevere deficiency of ADAMTS13, a von Willebrand factor (VWF)–cleaving metalloprotease, causes thrombotic thrombocytopenic purpura. When analyzed with VWF multimers, but not with an abbreviated VWF peptide (VWF73) as the substrate, the plasma ADAMTS13 activity levels of mouse strains segregated into a high and a low group that differed by approximately 10 fold. Low ADAMTS13 activity was detected in mice containing 2 alleles of intracisternal A-type particle (IAP) retrotransposon sequence in the ADAMTS13 gene. Molecular cloning of mouse ADAMTS13 identified 2 truncated variants (IAP-a and IAP-b) in the low-activity mice. Both of the IAP variants lacked the 2 carboxyl terminus thrombospondin type 1 repeat (TSR) and CUB domains of full-length ADAMTS13. The IAP-b variant also had splicing abnormalities affecting the spacer domain sequence and had miniscule enzymatic activity. Compared with full-length ADAMTS13, the IAP-a variant was approximately one ninth as active in cleaving VWF multimers but was only slightly less active in cleaving VWF73 peptide. Recombinant human ADAMTS13 was also less effective in cleaving VWF multimers than VWF73 when the C-terminal TSR sequence was deleted. In summary, the carboxyl terminus TSR sequence is important for cleaving VWF multimers. Assay results should be interpreted with caution when peptide substrates are used for analysis of variant ADAMTS13 proteins.

The Carboxyl-Terminal Domains of ADAMTS13 Cause Substrate-Dependent Divergence of ADAMTS13 Activity.

ADAMTS13, a circulating metalloprotease that cleaves the Y1605-M1606 bond of VWF, is critical for preventing intravascular platelet thrombosis of thrombotic thrombocytopenic purpura. Unlike genetic deficiency of ADAMTS13 in patients, inactivation of ADAMTS13 created by homologous recombination causes either no or delayed-onset phenotypic abnormalities in mice, depending on the strains examined. In order to further understand the role of ADAMTS13 in VWF homeostasis, we investigated the structure and function of the enzyme in various mouse strains. As in human subjects, RT PCR analysis revealed that ADAMTS13 was expressed primarily in the mouse liver. However, the enzyme activity levels of mouse plasma ADAMTS13 segregated in two groups: a low group (C56BL/6J strain, 24%±10% of normal human plasma, N=21; and DBA/2J strain, 24%±7%, N=6), and a high group (FVB/NJ strain, 288%±60%, N=23; and 129×1/SvJ strain, 313%±39%, N=6). No such difference was detected when GST-His fusion or FRET form of VWF73 (D1596-R1668) peptide was used for activity measurement. Real-time analysis of RNA extracted from mouse livers showed that the ADAMTS13 transcript levels were similar between the two groups. Cloning of the mouse liver ADAMTS13 yielded a predominant 3.5kb instead of the expected 4.3kb species from the low-activity group. This truncated cDNA contained the first 23 exons of the full-length (FL) ADAMTS13 and an extraneous sequence derived from the long terminal repeat sequence of a previously described IAP-type retrotransposone located in intron 23 of affected mouse genome. Protein expression analysis of the cloned cDNA in HEK 293 and MDCK cells showed that both full-length and IAP-type ADAMTS13 proteins were secreted efficiently without evidence of polarity. Nevertheless, the IAP-type ADAMTS13 was only 10% as active as FL ADAMTS13 in cleaving VWF multimers, but was 50% – 70% as effective in cleaving human or mouse VWF73 peptides, suggesting that truncation and/or presence of IAP sequence at the C-terminus sequence of ADAMTS13 markedly impeded cleavage of VWF multimers. Crossbreeding between C57BL/6J and FVB/NJ strains of mice confirmed that the low-activity phenotype was linked to homozygous IAP genotype. In SDS agarose gel analysis, both C57BL/6J and FVB/NJ strains contained ultra large multimers similar to those observed in ADAMTS13-null mice (kindly provided by D. Motto and D. Ginsburg). Analysis of recombinant human ADAMTS13 proteins also showed that proteins truncated at the disintegrin or spacer domain were relatively more active in cleaving VWF peptide substrates than cleaving VWF multimers. In conclusion, the presence of an IAP-type retrotransposone in the ADAMTS13 gene enhances alternative splicing, resulting in predominant expression of IAP types of ADAMTS13 in the affected mouse strains. Nevertheless, ADAMTS13 does not regulate VWF multimer size, and deficiency of ADAMTS13 is phenotypically silent in the mice examined. The carboxyl-terminus sequence of ADAMTS13 may enhance interaction with VWF multimers and its subsequent cleavage. Clinically, ADAMTS13 assay results should be interpreted with caution when VWF peptide-based proteins are used as the substrate, because they yield higher activity values for ADAMTS13 proteins lacking the carboxyl terminus sequence.

A C-Terminal Exosite of von Willebrand Factor Domain A2 Interacts with ADAMTS13 Spacer Domain To Accelerate Substrate Cleavage.

ADAMTS13 is a plasma metalloproteinase that limits platelet-rich thrombi by cleaving von Willebrand factor (VWF) at the Tyr1605-Met1606 bond in the A2 domain. A minimal substrate that consists of GST linked to VWF residues Asp1596-Arg1668 with a C-terminal 6×His tag (GST-VWF73) is cleaved rapidly by plasma ADAMTS13. Further removal of Glu1660-Arg1668, which comprises a predicted C-terminal α-helix (GST-VWF64) markedly reduced the rate of cleavage, suggesting this helix comprises an exosite for substrate recognition. By amino acid sequencing, ADAMTS13 was shown to cleave the Tyr1605-Met1606 bond of GST-VWF64. Truncation of ADAMTS13 after certain structural domains had different effects on substrate cleavage. After removal of the GST moiety by proteolysis, the kinetic constants for cleavage of VWF73 and VWF64 were determined for several recombinant ADAMTS13 variants using a quantitative MALDI-MS assay (Table). Comparison of the specificity constants (kcat/Km) shows that ADAMTS13 truncated after the spacer domain (construct MDTCS) cleaved VWF73 ~20-fold faster than did a similar enzyme without the spacer domain (construct MDTC). In contrast, both MDTCS and MDTC cleaved VWF64 slowly at a rate similar to the cleavage of VWF73 by MDTC. Most of the variation in cleavage rates was explained by differences in Km, suggesting that the spacer domain recognizes an exosite at the C-terminus of VWF73 that is missing from VWF64. Cleavage of VWF73 yields a C-terminal product (cVWF63). Purified cVWF63 (7.5 μM) inhibited MDTCS activity toward VWF73 or VWF64 (1.5 μM) by »90%, but did not inhibit MDTC, suggesting that the exosite of VWF73 or cVWF63 interacts directly with the spacer domain. Moreover, cVWF63 inhibited the cleavage of multimeric VWF by full-length ADAMTS13 and MDTCS up to 80%, but did not inhibit MDTC. The selective effects of deleting the ADAMTS13 spacer domain and Glu1660-Arg1668 of the VWF domain A2 suggest that the C-terminal exosite of the VWF A2 domain accelerates substrate cleavage by binding specifically to the ADAMTS13 spacer domain.Table Kinetics studies of VWF73 and VWF64 cleavage by ADAMTS13 and its truncationsVWF73VWF64EnzymeKm (μM)kcat (s−1)kcat/Km (×105 M−1 s−1)Km (μM)kcat (s−1)kcat/Km (×105 M−1 s−1)FL-ADAMTS131.7±0.41.3±0.17.5±2.037.7±12.81.9±0.40.5±0.3MDTCS0.8±0.21.7±0.120.5±6.65.5±1.40.6±0.11.0±0.3MDTC16.0±4.51.8±0.31.1±0.517.9±6.31.6±0.30.9±0.5

The Middle and Distal Carboxyl Terminal Domains of ADAMTS13 Are Critical for Docking ADAMTS13 to von Willebrand Factor.

ADAMTS13 cleaves von Willebrand factor (VWF) at the Tyr1605-Met1606 bond of the central A2 domain. However, the domains of ADAMTS13 required for recognition and cleavage of VWF remain controversial. On the one hand, we and others have shown that the proximal carboxyl terminal domains (i.e. Cys-rich and spacer domains) of ADAMTS13 appeared to be sufficient for recognition of VWF under the static and denatured conditions (Zheng et al, JBC, 2003; Soejima et al, Blood, 2003, Ai et al, JBC, 2005). On the other hand, the distal carboxyl terminus (i.e. the first CUB domain) of ADAMTS13 has been proposed to dock ADAMTS13 to VWF under flow (Tao et al, Blood, 2005). To determine the interaction between ADAMTS13 and native VWF in real time, we employed Biacore technology based on surface plasmon resonance. To avoid conformational change or activation of VWF during the ligand immobilization, we chose to flow purified VWF or A2 fragment (VWF73) over the carboxymethylated dextran (CM5) surface to which was covalently attached via amine coupling either purified full-length recombinant ADAMTS13 (FL-A13) (~8,200 RU), the variant truncated after the 8th thrombospondin type 1 repeat (delCUB) (~7,200 RU), or that truncated after the spacer domain (MDTCS) (~4,200 RU). The control surface was prepared similarly by activation and deactivation processes without protein being coupled. The VWF at various concentrations in 20 mM HEPES buffer, pH 7.5, 150 mM NaCl, 0.02% Tween 20 and 0.05% BSA was flown at 20 μl/min over the surfaces as prepared above. The sensorgrams were recorded simutaneouly in all four microfluid cells and the kinetic parameters were determined. We showed that VWF bound immobilized FL-A13 in a dose and time-dependent manner with an equilibrium dissociation constant (Kd) of 50.0 ± 21.8 nM (n=6). Removal of the CUB domains (construct delCUB) reduced its affinity for VWF by ~5.5 fold (Kd=274 ± 158 nM (n=6). Further removal of the TSP1 2~8 repeats (construct MDTCS) almost completely abolished its binding to VWF (n=4) (Kd not determined). The affinity between ADAMTS13 (or variants) and VWF was not significantly altered by addition of 10 mM EDTA. Pre-denaturization of VWF with 1.5 M guanidine-HCl at 37 °C for 1 h, however, modestly enhanced VWF binding to immobilized delCUB (Kd=185 ± 96 nM, n=4) and MDTCS (Kd not determined), but reduced VWF binding to FL-A13 by ~2 fold, perhaps due to the mild denaturization of FL-A13 by 0.15 M guanidine-HCl in the flowing buffer. Interestingly, VWF73 bound FL-A13, delCUB and MDTCS with the Kd (s) of 2.6 ± 0.5 nM, 4.9 ± 1.0 nM and 3.0 ± 0.5 nM, respectively, which was at least 20~55 fold higher than native VWF did, suggesting that the protein structure surrounding the cleavage site at the central A2 domain of VWF negatively regulates ADAMTS13-VWF interaction. Moreover, the Michael-Menton constant Km (s) for cleavage of the denatured VWF (or VWF73) by FL-A13, delCUB and MDTCS were 260 ± 117 nM, 276 ± 118 nM and 534 ± 657 nM (or 1.8 ± 1.5 μM, 0.3 ± 0.1 μM and 0.6 ± 0.2 μM), respectively. These data demonstrate that the interaction between ADAMTS13 and native VWF is a high affinity one and that the middle and distal carboxyl terminal domains of ADAMTS13 may be critical for initial docking of ADAMTS13 to the exosites of native VWF in solution.

Long-Term Expression of ADAMTS13 at Therapeutic Levels in Mice by In Utero Administration of Lentiviral Vector Encoding the Human ADAMTS13 Gene.

Thrombotic thrombocytopenic purpura (TTP), characterized by profound thrombocytopenia and microangiopathic hemolytic anemia, may occur in neonates, children and adults. Two types of TTP: hereditary and acquired, have been described. Hereditary TTP is caused by mutations in the ADAMTS13 gene, whereas the acquired TTP is mainly caused by autoantibodies against ADAMTS13 protease. Despite of low incidence of hereditary TTP, patients with hereditary TTP tend to recur or relapse after an initial episode, which may result in sustained neurological and/or renal complications or death. Therefore, early diagnosis and treatment of hereditary TTP are crucial to prevent catastrophic complications. Gene therapy may provide an attractive alternative therapeutic approach to plasma infusion or exchange for long-term correction of ADAMTS13 deficiency. Lentiviral vector encoding the full-length human ADAMTS13 gene tagged with V5-His at the carboxyl terminus and an enhanced green fluorescent protein (eGFP) reporter or PBS was administered to day 14 mouse fetuses (C57/BL6) via direct injection into yolk sac vessels. The blood was collected into sodium citrate at birth, 1, 4, 8, 12, 18, 24, 28 and 36 weeks following the injection. The plasma ADAMTS13 activity, VWF antigen, VWF activity and VWF multimers were determined by FRETS-VWF73, enzyme-linked immunoassay, type III collagen binding and agarose gel electrophoresis, respectively. The transduction of lentiviral vector and expression of human ADAMTS13 in vivo were determined by direct visualization of eGFP reporter by fluorescence stereomicroscopy and immunohistochemistry with monoclonal anti-V5 antibody. We showed that liver appeared to be the predominant organ expressing human ADAMTS13 and eGFP reporter. Other organs including brain, heart, kidney, lung, intestines and testes were also positive for the transgene products, particularly in the vessels. Full-length human ADAMTS13-V5-His (~195kDa) protein was detectable in mouse plasma after immunoprecipitation with mouse anti-V5 and Western blot with rabbit anti-ADAMTS13 IgG. The plasma ADAMTS13 activity in lentiviral vector-injected mice (n=8) was approximately 15~20% higher on average than that in the buffer-injected controls (n=8). The elevated levels of ADAMTS13 persisted for at least 9 months tested. Furthermore, the ratio of VWF activity to VWF antigen levels and the multimer sizes in the lentiviral vector-injected mice were significantly smaller than those in the PBS-injected control. The mice overexpressing human ADAMTS13 had normal complete blood count and did not show signs of spontaneous bleeding. To our surprise, these mice exhibited a significant prolongation of occlusion time in the carotid arteries after 10% FeCl3 injury (28.6 ± 13.0 min, n=8) compared to the control mice (8.5 ± 1.9 min, n=8) (p<0.001). These data demonstrate for the first time that long-term expression of human ADAMTS13 is achieved in mice by in utero administration of lentiviral vector and that the over expression of human ADAMTS13 may offer a systemic anti-thrombotic effect in vivo. Our study supports the feasibility of successful gene therapy for hereditary TTP and/or other arterial thrombotic disorders that are mediated by “unusually large” or “hyperactive” VWF.