A phenotype–genotype correlation of ADAMTS13 mutations in congenital thrombotic thrombocytopenic purpura patients treated in the United Kingdom


Raymond S. Camilleri, Department of Biomedical Sciences, School of Life Sciences, University of Westminster, 115 New Cavendish Street, London W1W 6UW, UK.
Tel.: +44 20 7911 5000 ext 64121; fax: +44 20 7911 5087.
E-mail: r.camilleri@westminster.ac.uk


Summary.  Background:  ADAMTS13 mutations play a role in thrombotic thrombocytopenic purpura (TTP) pathogenesis.

Objectives:  To establish a phenotype–genotype correlation in a cohort of congenital TTP patients.

Patients/Methods:  Clinical history and ADAMTS13 activity, antigen and anti-ADAMTS13 antibody assays were used to diagnose congenital TTP, and DNA sequencing and in vitro expression were performed to identify the functional effects of the ADAMTS13 mutations responsible.

Results:  Seventeen (11 novel) ADAMTS13 mutations were identified in 17 congenital TTP patients. All had severely reduced ADAMTS13 activity and antigen levels at presentation. Six patients with pregnancy-associated TTP and six patients with childhood TTP were homozygous or compound heterozygous for ADAMTS13 mutations located in the metalloprotease (MP), cysteine-rich, spacer and/or distal thrombospondin type 1 domains. The adults had TTP precipitated by pregnancy, and had overall higher antigen levels (median, 30 ng mL−1; range, < 10–57 ng mL−1) than the children (median, 14 ng mL−1; range, < 10–40 ng mL−1). Presentation in the neonatal period was associated with more intensive treatment requirements. The two neonates with the most severe phenotype had mutations in the first thrombospondin type 1 motif of ADAMTS13 (p.R398C, p.R409W, and p.Q436H). Using transfected HEK293T cells, we have shown that p.R398C and p.R409W block ADAMTS13 secretion, whereas p.Q436H allows secretion at reduced levels.

Conclusions:  This study confirms the heterogeneity of ADAMTS13 defects and an association between ADAMTS13 genotypes and TTP phenotype.


Thrombotic thrombocytopenic purpura (TTP) is a rare disorder characterized by widespread microvascular platelet-rich thrombi resulting in profound thrombocytopenia and microangiopathic hemolytic anemia (MAHA) [1]. It is associated with a severe reduction in von Willebrand factor (VWF)-cleaving protease (ADAMTS13) activity and the presence of unusually large VWF multimers [2]. TTP is fatal without rapid and aggressive treatment.

ADAMTS13 (OMIM accession number 604134) contains 29 exons and spans approximately 37 kb on chromosome 9q34 [3–5]. Predominantly expressed in the liver, ADAMTS13 is 1427 amino acids long, is ∼ 180 kDa in size, and cleaves VWF at Tyr1605-Met1606 within the central A2 domain [6,7]. It consists of a signal peptide, a propeptide, a metalloprotease (MP) domain, a disintegrin-like domain, a thrombospondin type 1 (TSP1) motif, a cysteine-rich domain, a spacer domain, seven TSP1 repeats, and two CUB domains [3–5,8].

Over 80 mutations distributed throughout ADAMTS13 have been reported in congenital TTP [3,9–20]. However, several studies have also reported a number of cases of adult-onset congenital TTP [17,19,21–25], who appear to require a significant challenge, such as pregnancy or acute infection, to drive the pathophysiology of the disorder [17,24,25].

Our group has reported the ADAMTS13 mutation c.3178C>T (p.R1060W) in six Caucasian adult-onset TTP patients with a UK-wide distribution [25], and four additional adult-onset TTP patients have been reported elsewhere [17,19,24]. R1060W has not been identified in any published childhood congenital TTP cases to date [3,9–20]. This suggests that certain ADAMTS13 mutations may play a part in defining the time to presentation of TTP.

Although the ADAMTS13 mutations identified to date are distributed evenly throughout the gene, it has yet to be ascertained whether certain functional amino acid changes, and as a result the conformational changes in certain ADAMTS13 domains, are more likely to cause episodes of acute TTP with varying degrees of severity.

We hypothesized that functional alterations in the amino acid sequence in a particular domain(s) of ADAMTS13 may confer a more severe TTP phenotype than changes in other areas of ADAMTS13. We identified a number of ADAMTS13 mutations in a cohort of adult-onset and childhood congenital TTP patients referred to our unit. These heterogeneous mutations were then associated with clinical history in order to investigate whether there are genotype–phenotype correlations in TTP.

Materials and methods

Patients, family, and controls

Diagnostic criteria for TTP were based on UK national guidelines [26], and included thrombocytopenia, anemia, and morphologic evidence of red cell fragmentation. In patients with repeated ADAMTS13 activity of < 5% after resolution of the acute episode and no evidence of an inhibitor or anti-ADAMTS13 IgG antibodies, mutational analyses were routinely undertaken to confirm a congenital phenotype. Family members, when available, were also investigated.

Fifty healthy, normal adult controls (median age, 37 years; range, 23–61 years) with no history of excessive bleeding, thrombosis or hemolytic anemia were also studied. They comprised 28 non-pregnant females and 22 males. Local Ethics Committee approval as part of the UK TTP Registry (REC reference number: 08/H0810/54) since 1 January 2009 and patient consent were obtained. Blood was collected by clean venepuncture into Vacutainers (Becton Dickinson, Oxford, UK). Citrated plasma was prepared by double centrifugation, aliquoted, and stored at − 80 °C. Following confirmation of ADAMTS13 results on more than two occasions, an EDTA sample (4 mL) was sent for mutational analysis.

ADAMTS13 assays

ADAMTS13 activity was determined in all subjects by measuring the residual collagen-binding activity (CBA) of degraded exogenous VWF [27] (modified from Gerritsen et al. [28]). The results were expressed as a percentage of pooled normal plasma (PNP) (normal range: 50–160%). Inhibitors of ADAMTS13 activity were determined by incubating equal volumes of patient plasma with PNP, and reported if the ADAMTS13 activity of PNP was reduced by > 50% by residual CBA assay.

Anti-ADAMTS13 IgG antibodies were measured with an in-house ELISA [29] (modified from Scheiflinger et al. [30]). Microtiter plates were coated with recombinant ADAMTS13 (rADAMTS13) (Baxter Bioscience, Vienna, Austria). Samples were diluted 1 : 100 with phosphate-buffered saline (PBS). Bound IgG was detected with rabbit anti-human IgG conjugated with horseradish peroxidase (HRP). Results were expressed relative to an index plasma with high anti-ADAMTS13 IgG antibody titers (normal: < 4.2%).

ADAMTS13 antigen was measured with the commercial Imubind ADAMTS13 ELISA kits (Sekisui Diagnostics, Stamford, CT, USA). This employs a polyclonal anti-ADAMTS13 capture antibody and a biotinylated polyclonal antibody against ADAMTS13 TSP1 domains 5–7 (normal range: 485–1242 ng mL−1).

Genetic screening of ADAMTS13

Genomic DNA was extracted from whole blood with Nucleon BACC1 Purification kits (Tepnel Life Sciences, Manchester, UK). The 29 exons (and intronic boundaries) of ADAMTS13 were amplified by PCR. The oligonucleotide primers (Invitrogen, Paisley, UK) and PCR conditions used are available on request. NG_011934.1 was used as the genomic DNA reference sequence; NM_139025.3 was used as the cDNA reference sequence. Nucleotide numbering reflects cDNA numbering, with +1 corresponding to the A of the ATG translation initiation codon in the reference sequence, according to journal guidelines (http://www.hgvs.org/mutnomen). The initiation codon is codon 1.

Sequencing PCRs were set up with GenomeLab DTCS Quick Start kits (Beckman Coulter, High Wycombe, UK) and the following PCR regime: 96 °C for 30 s, 50 °C for 30 s, and 60 °C for 4–5 min (30 cycles). Amplification products were analyzed with an eight-channel DNA sequencer (CEQ 8000; Beckman Coulter).

The 50 healthy, normal controls were all screened for the mutations identified in our patient cohort with a variety of restriction enzyme digest screens, summarized in Table S1. The oligonucleotide primers and PCR conditions used are available on request.

Construction, transfection and expression of ADAMTS13 variants

Site-directed mutagenesis, with QuikChange II XL kits (Stratagene Europe, Amsterdam, the Netherlands), was performed on a wild-type ADAMTS13 cDNA/pcDNA3.1 vector construct to create all 17 ADAMTS13 mutants. Two complementary oligonucleotides containing each mutation were synthesized and purified by HPLC (Invitrogen). Sample reactions were then amplified: 95 °C for 1 min (one cycle); 95 °C for 50 s, 60 °C for 50 s, 68 °C for 10 min (18 cycles); and 68 °C for 7 min (one cycle). After treatment with DpnI, PCR products were transformed in ultracompetent Escherichia coli cells and miniprepped with QIAprep spin kits (Qiagen, Crawley, UK). The presence of the expected mutations and the absence of unwanted mutations were verified by DNA sequencing.

Wild-type and mutant ADAMTS13 plasmids were prepared with HiSpeed Plasmid Maxi kits (Qiagen) prior to transfection into human embryonic kidney (HEK293T) cells. Transfection was performed with 10 mm polyethylenimine (Sigma-Aldrich, Poole, UK) when the cells were 70–80% confluent, and they were then incubated for 3 days (37 °C, 5% CO2, humidified). Conditioned media were collected and dialyzed in 20 mm Tris-HCl (pH 7.8); adherent cells were collected and centrifuged. Western blot analysis was then carried out to ascertain the levels of wild-type/mutant ADAMTS13 expression, secretion, and intracellular retention.

Western blot analyses

Dialyzed conditioned media (40 μL) and cell lysate (10 μL) samples were loaded onto separate 4–12% Bis-Tris polyacrylamide gels (Bio-Rad Laboratories, Hemel Hempstead, UK) with reducing sample loading buffer (Fig. 1A). Cell lysates were included to allow comparison of the relative amounts of secreted and total ADAMTS13. Protein samples were transferred onto separate 0.45-μm nitrocellulose membranes (Bio-Rad Laboratories), blocked with 10% milk powder (Marvel), washed with 0.02 m Borate Buffered Saline Tween, pH 7.2, and incubated for 16 h at 4 °C with goat anti-ADAMTS13 affinity-purified antibody (Universal Biologicals [Cambridge], Cambridge, UK). After further washing, the membranes were incubated with a polyclonal rabbit anti-goat IgG secondary antibody conjugated with HRP (Dako, Ely, UK) for 1 h at room temperature. ECL Western Blotting Detection Reagents (GE Healthcare UK, Little Chalfont, UK) were then used to detect expressed ADAMTS13 via autoradiographic Amersham Hyperfilm ECL (GE Healthcare UK).

Figure 1.

 Western blot analyses of wild-type and mutant ADAMTS13 clones expressed in HEK293T cells. (A) Lane 1: p.R102H. Lane 2: non-transfected negative control. Lane 3: p.A690T. Lane 4: p.R409W. Lane 5: p.Q436H. Lane 6: p.R910X. Lane 7: p.C754R. Lane 8: wild-type. (B) Lane 1: non-transfected negative control. Lane 2: wild type. Lane 3: p.C977F. Lane 4: p.Q456H. Lane 5: p.R398C. Lane 6: p.D217H. Arrows point to ADAMTS13 expressed and detected in (i) cell lysates and (ii) conditioned media.

An alternative western blot system was used in which cell lysate (15 μL) and five-fold concentrated conditioned media (40 μL) samples were loaded onto separate 4–12% Bis-Tris polyacrylamide gels (Bio-Rad Laboratories) with reducing sample loading buffer (Fig. 1B). Once run, the protein samples were transferred onto separate poly(vinylidene difluoride) membranes (Immobilon-P; Millipore (UK) Ltd, Watford, UK). These membranes were then blocked with 5% milk powder (Marvel), washed with PBS Tween (pH 7.2), and incubated for 16 h at 4 °C with an anti-Myc antibody conjugated with HRP (Invitrogen). After further washing, ECL Western Blotting Detection Reagents (GE Healthcare UK) were used to detect expressed ADAMTS13 via autoradiographic Amersham Hyperfilm ECL (GE Healthcare UK).

Levels of secretion of mutant ADAMTS13 proteins as compared with wild-type ADAMTS13 were quantified in triplicate by measuring the surface area of the relevant protein bands from both the cell lysate and conditioned media blots. When the the level of secretion for each mutant ADAMTS13 protein had been determined, this was compared with the level of secretion of wild-type ADAMTS13.

Antigen assays were also performed in triplicate on all relevant conditioned media, and CBA assays were performed on selected mutants with adequate amounts of protein secreted in vitro, standardized for the amount of antigen in the sample. ADAMTS13 activity of in vitro expressed protein was also determined in a chromogenic assay with a 73 amino acid peptide substrate based on the cleavage site of the VWF A2 domain (Technozym ADAMTS13 activity; Technoclone, Vienna, Austria) (normal range: 40–130%).


Seventeen patients, seven males and 10 females, from 14 unrelated families, who presented with congenital TTP in the UK met the study eligibility criteria. Six patients presented with de novo pregnancy-associated TTP, five presented as neonates (three were siblings), and six presented in childhood (two were siblings). The median age for adult-onset TTP was 24.5 years (range: 18–33 years) and that for children was 27 months (birth to 120 months). Consanguinity was confirmed in cases 5, 7, and 12. All patients had profoundly decreased ADAMTS13 activity (< 5%; normal range, 50–160%) and persistently low/undetectable ADAMTS13 antigen levels (< 60 ng mL−1; normal range, 485–1242 ng mL−1), confirmed at presentation and repeated on more than one occasion in remission or prior to further therapy. None had detectable ADAMTS13 inhibitor/IgG antibodies. All 50 healthy untransfused controls had ADAMTS13 activity within the normal range (median, 109%; interquartile range, 44%) and no detectable anti-ADAMTS13 IgG antibodies.

The clinical history, age at onset and ADAMTS13 assay results of all patients are summarized in Table 1. ADAMTS13 genotype and single-nucleotide polymorphisms (SNPs) are shown in Table 2. Eleven novel mutations and six previously described mutations [9,11,12,15,20,24,25] were detected in the 17 patients. Twelve cases were found to have a variety of SNPs present. None of the 100 alleles of the healthy, normal control group studied were found to have any of the ADAMTS13 mutations described following a program of restriction enzyme digest screening (Table S1).

Table 1.   Clinical history and plasma ADAMTS13 antigen levels of 17 congenital thrombotic thrombocytopenic purpura (TTP) patients
PatientSexAge at onsetEthnicityCongenital TTP historyClinical detailsADAMTS13 antigen (ng mL–1)
  1. F, female; M, male; NA, not available; Rx, treatment. *Siblings. †Siblings.

1F33 yearsCaucasianPregnancy-associatedOne loss, two live pregnanciesNA
2F23 yearsCaucasianPregnancy-associatedOne loss, two live pregnancies< 10
3F29 yearsAsianPregnancy-associatedNo live pregnancies< 10
4F18 yearsCaucasianPregnancy-associatedThree live pregnancies57
5F24 yearsAsianPregnancy-associatedOne live pregnancy38
6F20 yearsCaucasianPregnancy-associatedFour losses, one live pregnancy35
7MBirthAsianNeonatalRx every 1–2 weeks< 10
8FBirthAsianNeonatalRx every 1–2 weeks< 10
9*MBirthCaucasianNeonatalRx every 1–2 weeks40
10*MBirthCaucasianNeonatalRx every 1–2 weeks19
11*MBirthCaucasianNeonatalRx every 1–2 weeks10
12M18 monthsAsianChildhoodRx every 2 weeks11
13F3 yearsAsianChildhoodRx as required< 10
14†F5 yearsCaucasianChildhoodRx as required< 10
15†M3 yearsCaucasianChildhoodRx as required< 10
16F18 monthsCaucasianChildhoodInfrequent Rx: only severe infections17
17M10 yearsAsianChildhoodInfrequent Rx: only severe infections< 10
Table 2. ADAMTS13 genotype and single nucleotide polymorphisms (SNPs) of 17 congenital thrombotic thrombocytopenic purpura (TTP) patients
Patient ADAMTS13 genotype
Mutation(s) identifiedSNPs present (not haplotype-specific)
Nucleotide change*ExonAmino acid changeDomain
  1. del, deletion; MP, metalloprotease; TSP1, thrombospondin type 1; –, no SNPs detected. *Compound heterozygous or homozygous unless otherwise stated. †Indicates pregnancy-associated TTP; ‡ Indicates neonatal TTP; §indicates childhood TTP. NG_011934.1 was used as the genomic DNA reference sequence; NM_139025.3 was used as the cDNA reference sequence. Nucleotide numbering reflects cDNA numbering, with +1 corresponding to the A of the ATG translation initiation codon in the reference sequence, according to journal guidelines (http://www.hgvs.org/mutnomen). The initiation codon is codon 1.

1†c.3178C>T24p.R1060WTSP1–7p.[R7W]+[R7W], p.[A1033T]+[A1033T]
2†c.[794G>C(+)796A>T] (novel)

p.[C265S(+)S266C] (novel)

p.R7W, p.A1033T
c.3284G>A (novel)
p.R1095Q (novel)
4†c.305G>A (novel)3p.R102H (novel)MPp.[Q448E]+[Q448E], p.[P618A]+[P618A], p.[A900V]+[A900V]
5†c.2930G>T (novel)23p.C977F (novel)TSP1-5
c.2068G>A (novel)
p.A690T (novel)
p.R7W, p.Q448E, p.P618A, p.A732V, p.A900V
7‡c.1225C>T (novel)10p.R409W (novel)TSP1-1
8‡c.1192C>T (novel)
c.1308G>C (novel)
p.R398C (novel)
p.Q436H (novel)
14§c.2260T>C (novel)19p.C754R (novel)TSP1-3p.[Q448E]+[Q448E]
15§c.2260T>C (novel)19p.C754R (novel)TSP1-3p.[Q448E]+[Q448E]
16§c.649G>C (novel)
p.D217H (novel)
p.R7W, p.Q448E, p.A1033T
c.1368G>T (novel)
p.Q456H (novel)

Pregnancy-associated TTP

Patients 1–6 had adult-onset TTP precipitated by pregnancy. They had higher antigen levels than the children (patients 7–17); median of 30 ng mL−1 and range of < 10–57 ng mL−1 as compared with median of 14 ng mL−1 and range of < 10–40 ng mL−1. The clinical course, ADAMTS13 measurements and p.R1060W genotype of patient 1 were previously reported by our group, as was the effect that p.R1060W had on ADAMTS13 functionality [25]. Patient 2, who had experienced a previous pregnancy loss, but received plasma therapy in subsequent pregnancies and had two live births, was compound heterozygous for c.[794G>C(+)796A>T] (p.[C265S(+)S266C]) and p.R1060W. Patient 3, who presented in the first trimester and had a termination of pregnancy, also had non-pregnancy-associated TTP episodes and was compound heterozygous for the previously reported c.1787C>T (p.A596V) [20] and c.3284G>A (p.R1095Q). Both had ADAMTS13 antigen levels below the detectable limits of the assay.

Patient 4 was homozygous for c.305G>A (p.R102H), developed TTP during the third trimester of her first pregnancy, and had an antigen level of 57 ng mL−1. She has had three live births and one episode unrelated to pregnancy. Patient 5, who presented in the first trimester of pregnancy and was treated with plasma exchange, resulting in a live birth, was homozygous for c.2930G>T (p.C977F) and had an antigen level of 38 ng mL−1.

Patient 6, a 60-year-old woman who developed TTP during her first pregnancy and has had no other TTP episodes other than in pregnancy, was found to be compound heterozygous for the previously reported deletion mutations c.719_724del (p.G241_C242del) [11] and c.2068G>A (p.A690T), and had an antigen level of 35 ng mL−1. Interestingly, the sister of patient 6, who developed thrombocytopenia at the end of her third pregnancy, was only heterozygous for p.G241_C242del, and had an antigen level of 256 ng mL−1. The family trees of patients 5 and 6 are shown in Fig. 2A,B.

Figure 2.

 Family trees of patients 5–8, 12, and 13. Genotypes are given. ADAMTS13 activities are given as a percentage of that in pooled normal plasma. ADAMTS13 antigen levels are in ng.mL−1. Circles indicate females, and squares indicate males. (A) The consanguineous parents of patient 5 were both heterozygous for p.C977F, and both had subnormal ADAMTS13 activity and antigen levels. One brother was also heterozygous for p.C977F and had low ADAMTS13 activity and antigen levels, but her youngest brother was devoid of this mutation and possessed normal ADAMTS13 activity and antigen levels. (B) The sister of patient 6 was only heterozygous for p.G241_C242del and the single-nucleotide polymorphisms (SNPs) p.Q448E and p.A900V, but had subnormal ADAMTS13 activity and antigen levels. (C) The consanguineous parents of patient 7 were heterozygous for p.R409W, and had normal/relatively normal ADAMTS13 activities and subnormal antigen levels. The father was also heterozygous for the SNP p.Q448E, even though no SNPs were found in the proband. (D) The mother of patient 8 was heterozygous for p.R398C and the SNP p.Q448E, whereas her father was heterozygous for p.Q436H and p.Q448E. Her sister was also heterozygous for p.R398C, and all had ADAMTS13 activities and antigen levels indicative of heterozygosity. Her elder brother had neither mutation (he was heterozygous for p.Q448E) and had normal ADAMTS13 activity and antigen levels, whereas her younger brother had normal ADAMTS13 activity and near-normal antigen levels. (E) The consanguineous parents of patient 12 were both heterozygous for p.A596V, and had normal ADAMTS13 activity and subnormal antigen levels. His stepfather (brother of his father) was also heterozygous for p.A596V, and had decreased ADAMTS13 activity and antigen levels, whereas his half-sister had normal ADAMTS13 activity and antigen levels. No common SNPs were found in this family. (F) The parents of patient 13 were both heterozygous for p.T196I and the SNP p.Q448E, but their ADAMTS13 activity and antigen levels were unavailable. NA, not available; WT, wild type.

Neonatal TTP

Patients 7–11 first presented with TTP as neonates, and their clinical course, prophylaxis and treatment have been described previously [31]. Patients 7 and 8 had a classic, chronic MAHA picture, and patients 9–11 also required regular ADAMTS13 replacement therapy. All had exacerbations with infections/immunizations, and received treatment with intermediate-purity human factor VIII concentrate BPL 8Y, which has some ADAMTS13 activity [31]. None have long-term organ damage.

Patient 7 was found to be homozygous for p.R409W, whereas patient 8 was compound heterozygous for p.R398C and p.Q436H, and both had undetectable antigen levels. The family trees of patients 7 and 8 are shown in Fig. 2C,D. Patients 9–11 were brothers who had antigen levels of 40, 19 and 10 ng mL−1, respectively. The eldest two (patients 9 and 10) have previously been reported [11], and their younger brother (patient 11) was also found to be compound heterozygous for p.G241_C242del and the nonsense mutation p.R910X.

Childhood TTP

Patients 12–17 first presented with TTP during childhood. The clinical course, prophylaxis and treatment of patients 12 and 16 have been reported previously [31]. Patient 12 was found to be homozygous for the previously described p.A596V [20], and had an antigen level of 11 ng mL−1. Patient 13, a 21-year-old female who first developed TTP at the age of 3 years, was found to be homozygous for c.587C>T (p.T196I) [12], and had undetectable antigen levels. The family trees of patients 12 and 13 are shown in Fig. 2E,F. Patients 14 and 15 were a 5-year-old and 3-year-old sister and brother who were both found to be homozygous for c.2260T>C (p.C754R), with undetectable antigen levels. Patient 16, who only presented on three occasions associated with severe infections, was compound heterozygous for c.649G>C (p.D217H) and p.R1060W, and had an antigen level of 17 ng mL−1. Patient 17 first presented with TTP at 10 years, associated with severe infection, has only had one subsequent episode, and was found to be compound heterozygous for the previously reported c.803G>C (p.R268P) [9,15,20] and c.1368G>T (p.Q456H), and also had undetectable antigen levels.

In vitro expression studies

The western blot results for expression of the ADAMTS13 mutants are shown in Fig. 1. Secretion levels as compared with wild-type ADAMTS13 by western blotting and antigen assay are shown in Table 3.

Table 3.   Secretion of ADAMTS13 mutants expressed in HEK293T cells
MutationNovel% Secretion of wild type (western blot)% Secretion of wild type (antigen assay)
  1. NA, not assessed.

p.R398CYes< 5< 3
p.R409WYes< 5< 3
p.C977FYes10< 3
p.G241_C242delNoNA< 3
p.A596VNoNA< 3
p.R910XNo7Not detectable with assay antibody
p.R1060WNo< 5 [25]NA

Novel mutations

p.R398C and p.R409W abolished ADAMTS13 secretion, whereas p.[C265S(+)S266C], p.Q436H, p.A690T, p.C754R and p.C977F reduced ADAMTS13 secretion. p.R102H reduced secretion to 28–51% of that of the wild type, but the mutant protein had reduced, albeit detectable, activity (12% of that of the wild-type ADAMTS13 by fluorescence resonance energy transfer). p.D217H allowed secretion (29–112% of that of the wild type), but the mutant protein also had reduced specific activity (24%) as compared with wild-type ADAMTS13 when a VWF peptide substrate was used.

p.Q456H increased secretion of the protein (117–261% of that of the wild type) and the mutant had normal activity when a VWF peptide substrate was used. However, ADAMTS13 activity was markedly reduced (16%) when a full-length VWF substrate was used. p.R1095Q was not assessed, as site-directed mutagenesis failed to give a viable clone.

Previously described mutations

Of the previously described mutations, expression investigations performed for this study showed that p.G241_C242del and p.A596V abolished ADAMTS13 secretion, whereas p.T196I, p.R268P and p.R910X markedly reduced secretion (Table 3). These results are in keeping with previously published expression studies on some of these mutants. No secretion data have been reported for p.T196I, but it exhibited markedly reduced proteolytic activity [12]. In another study, the percentage secretion levels of p.R268P and p.A596V were 38% and 17% as compared with wild-type ADAMTS13, and neither mutant exhibited activity against full-length VWF substrate [20]. The difference in secretion levels between this study and our data may reflect the cell type used for expression (COS-7 vs. HEK293T cells). No functional data are available in the literature for p.G241_C242del and p.R910X.

p.R910X clearly truncated the ADAMTS13 protein (Fig. 1-Ai, lane 6 as compared with the other ADAMTS13 variants [lanes 1, 3–5, and 7–8]), but still allowed secretion in vitro (Fig. 1Aii, lane 6). It is important to note that the antigen assay used in this study may not have been able to detect the p.R910X-truncated ADAMTS13 protein in patients 9–11, as the detection antibody present was raised against TSP1 repeats 5–7.


We have identified 17 mutations (11 novel) in 17 congenital TTP patients diagnosed and treated in the UK following complete ADAMTS13 DNA sequencing. In 12 of 17 cases, a variety of SNPs were also present. The most prevalent SNPs were p.Q448E in seven patients (two siblings), p.R7W in four patients, and p.A900V in four patients (two siblings). All patients had severely reduced ADAMTS13 activity, decreased antigen levels and no detectable ADAMTS13 inhibitor/IgG antibodies before mutational analysis.

The six pregnancy-associated TTP patients presenting in adulthood were homozygous for the novel mutations p.R102H and p.C977F, compound heterozygous for: the previously unidentified p.A690T and the 6-bp in-frame deletion p.G241_C242del [11]; p.A596V [20] and the novel p.R1095Q; the new mutations p.[C265S(+)S266C] and p.R1060W, and homozygous for p.R1060W [25]. These patients possessed the mildest subtype of congenital TTP, as they developed the disorder during adulthood and only after suffering a significant physiologic challenge. The ADAMTS13 mutations found in these patients were located in the MP domain, the spacer domain, and/or the distal TSP1 repeats.

The in vitro functional analyses in this study have shown that both p.R102H and p.A690T allow ADAMTS13 secretion to take place (albeit at a severely reduced level), although p.R102H in the MP domain resulted in reduced activity of the enzyme. In contrast, p.G241_C242del, p.[C265S(+)S266C] and p.C977F severely reduced ADAMTS13 secretion in vitro. It has been previously shown that both p.A596V [20] and p.R1060W [24,25] greatly reduce ADAMTS13 secretion in HEK293T cells, but differ in their effects on other ADAMTS13 functionalities (no activity or binding to VWF observed with p.A596V [20], and full activity observed with p.R1060W [24,25]).

The patients with the severest TTP phenotype associated with chronic MAHA, despite regular ADAMTS13 replacement in the form of intermediate-purity human FVIII concentrate, were patients 7 and 8, who first presented with TTP as neonates and were homozygous and compound heterozygous for the novel ADAMTS13 missense mutations p.R409W and p.R398C/p.Q436H respectively. All three mutations were located in the first TSP1 motif and were the only mutations identified in this study that were located in this particular region of ADAMTS13.

We have shown through in vitro functional analyses that the p.R409W missense mutation prevents ADAMTS13 secretion. We have also shown that p.R398C blocks ADAMTS13 secretion, whereas it has previously been reported that c.1193G>A (p.R398H) affects ADAMTS13 activity [10]. However, the third TSP1-1 mutation identified in this study, p.Q436H, allowed ADAMTS13 secretion in vitro, but at a reduced level. Therefore, the mechanism by which p.R409W confers a severe TTP subtype in patient 7 is very likely to differ from that of the two heterozygous TSP1-1 mutations p.R398C and p.Q436H in patient 8.

The remaining three neonates (patients 9–11) were brothers who presented with varying phenotypes of TTP, but required regular ADAMTS13 replacement [31]. They were compound heterozygous for an MP (p.G241_C242del) and a TSP1-5 (p.R910X) mutation, suggesting that this ADAMTS13 genotype resulted in a less severe form of TTP. This can also be observed in the six other childhood congenital TTP patients, who were homozygous or compound heterozygous for mutations located in the MP (p.T196I, p.D217H, and p.R268P), cysteine-rich (p.Q456H), spacer (p.A596V) and/or distal TSP1 (p.C754R, p.R910X, and p.R1060W) domains.

Both p.G241_C242del and p.R910X reduced ADAMTS13 secretion to < 10%. p.D217H, p.Q456H and p.C754R all allowed ADAMTS13 secretion in vitro (p.Q456H > p.D217H > p.C754R), although p.C754R ADAMTS13 was not secreted at an appreciable level. With a VWF peptide substrate, p.D217H was shown to cause reduced ADAMTS13 activity. With the same peptide substrate, p.Q456H did not affect ADAMTS13 activity. However, when full-length VWF substrate was used, p.Q456H markedly reduced ADAMTS13 activity (16%), suggesting an effect on ADAMTS13 binding to (full-length) VWF. p.T196I was previously reported to result in the loss of ADAMTS13 activity [12], whereas p.R268P has been shown to effect both ADAMTS13 secretion and activity [9,20].

Patient 16, who first presented with TTP at 18 months, is the first known case of a child possessing the previously adult-only p.R1060W ADAMTS13 mutation. The severely reduced secretion caused by p.R1060W [24,25] and the effect that the secreted novel mutation p.D217H had on ADAMTS13 activity would explain why she had profoundly decreased ADAMTS13 activity and undetectable antigen levels.

Although TTP developed during childhood in this group of patients, the relative delay in onset and lack of disease severity, as compared with those with a neonatal presentation, coupled with the striking similarity in ADAMTS13 mutation distribution in the adult-onset TTP cases, may suggest that patients possessing mutations in these regions of ADAMTS13 (namely the MP domain and distal TSP1 repeats) may have a less aggressive phenotype, although some patients still require regular therapy.

In conclusion, we describe a cohort of 17 congenital TTP patients with 17 ADAMTS13 mutations (11 novel). This study confirms that ADAMTS13 defects are heterogeneous, and suggests that certain mutations appear to cause more severe clinical phenotypes of TTP than other abnormalities in ADAMTS13. This may have long-term implications for the management of patients with congenital TTP.


R. S. Camilleri: designed and performed research, collected data, analyzed and interpreted data, and wrote the manuscript; M. Scully: contributed vital samples, collected data, analyzed and interpreted data, and wrote the manuscript; M. Thomas: performed research, analyzed and interpreted data, and wrote the manuscript; I. J. Mackie: interpreted data, and reviewed the manuscript; R. Liesner: contributed vital samples, collected data, and interpreted data; W. J. Chen: performed research, and analyzed and interpreted data; K. Manns: performed research, and analyzed and interpreted data; S. J. Machin: contributed vital samples, analyzed and interpreted data, and reviewed the manuscript.


The authors would like to thank M. Gattens (Addenbrooke’s Hospital, Cambridge), W. Lester (Queen Elizabeth Hospital, Birmingham), A. Mumford (Bristol Royal Infirmary) and S. Paneesha (Birmingham Heartlands Hospital) for providing some of the patient samples for ADAMTS13 analyses, and F. Scheiflinger (Baxter Bioscience) for providing rADAMTS13. The authors would also like to thank M. Underwood for her technical support.

R. S. Camilleri was supported by BHF Grant PG/09/017. The UK TTP registry was sponsored by the MRC. M. Thomas is funded by the British Heart Foundation (BHF Clinical Research Training fellowship no. FS/10/013/28073).

Disclosure of Conflict of Interests

The authors state that they have no conflict of interest.