Recombinant ADAMTS-13: first-in-human pharmacokinetics and safety in congenital thrombotic thrombocytopenic purpura

The plasma metalloprotease ADAMTS-13 (a disintegrin and metalloproteinase with a thrombospondin type 1 motif, member 13) is a constitutively active enzyme that cleaves von Willebrand factor (VWF) at the Tyr1605–Met1606 bond in the A2 domain, an otherwise cryptic site rendered susceptible, by the application of shear stress, to regulate the size of VWF multimers.¹  VWF is a multimeric glycoprotein synthesized principally by vascular endothelial cells.²  The inability to cleave ultralarge (UL) VWF multimers into smaller forms as a result of congenital or acquired ADAMTS-13 deficiency results in thrombotic thrombocytopenic purpura (TTP), a rare potentially fatal disorder of the microcirculation caused by increased binding of platelets to UL VWF.³  The congenital form of TTP (cTTP, previously termed hereditary TTP or Upshaw-Schulman syndrome) is an ultrarare, although likely underestimated, condition with a prevalence of approximately 1 case per million.^4,5  It exhibits an autosomal recessive mode of inheritance caused by homozygous or compound heterozygous mutations in both ADAMTS-13 alleles on chromosome 9.⁶  The nature of the mutations is diverse and includes single amino acid missense substitution (approximately 60%), as well as nonsense, frame-shift, splice site mutations and deletions and insertions (collectively, approximately 40%).^1,7-9 

The principal pathophysiology arises from platelet aggregates in the microcirculation affecting critical organs, including the brain, heart, and kidneys. TTP crises are associated with cerebral vascular incidents in at least 30% of patients, with a risk for neurologic sequelae in approximately 20% of patients.¹⁰  Acute renal failure has been reported in 11% of patients with severe cTTP,¹¹  and chronic, possibly progressive, renal involvement is often seen. Sudden cardiac death resulting from myocardial infarction, heart failure, and arrhythmia has also been reported.¹² 

Although considered a monogenetic disorder, the clinical presentation of cTTP is variable. Symptoms develop soon after birth in some patients, whereas others remain asymptomatic until the second or third decade of life. This phenotypic variability is thought to be related to the causative mutations and the level of plasma ADAMTS-13 activity.^13,14  In the newborn, cTTP typically presents as neonatal jaundice and thrombocytopenia,³  whereas in early childhood, symptomatic episodes are often associated with intercurrent infections or vaccinations. Among patients with cTTP presenting with a first TTP episode later in life, pregnancy is often the inciting event.¹⁵  Intrauterine fetal death is common in patients with cTTP who do not receive regular plasma therapy throughout pregnancy.^16,17  Other precipitants associated with increased VWF levels, such as infection, surgery, and alcohol binge drinking, provide additional triggers for acute TTP events.^18-22  Despite the wide range of ages of the first TTP event, most patients with cTTP subsequently demonstrate a chronic, relapsing course and require prophylactic treatment to prevent long-term neurological, renal, and other sequelae.

There are no drugs currently approved for the specific treatment of cTTP. Acute TTP episodes are generally treated with infusions of fresh frozen plasma (FFP) or solvent/detergent-treated plasma, typically 10 to 20 mL/kg body weight (BW). Some intermediate-purity FVIII:VWF concentrates have been shown to contain low levels of ADAMTS-13 and have been used as an alternative treatment in select patient populations.²³  Although the half-life of infused plasma ADAMTS-13 was found to be 2 to 4 days,²⁴  prophylaxis with 10 to 15 mL plasma /kg BW every 2 to 3 weeks has been shown to be effective in preventing acute episodes in the majority of patients with TTP.^6,25,26 

Although the treatment of TTP via plasma infusions is generally effective, the therapy is frequently complicated by allergic and anaphylactic reactions or volume overload.²⁷  Plasma infusions also carry risks for infection as a result of bloodborne pathogens. In addition, plasma infusions carried out in the hospital or outpatient settings are burdensome and time-consuming, and can be stressful for younger patients. As such, the development of a recombinant ADAMTS-13 (BAX 930; SHP 655; rADAMTS-13) represents a potential new therapeutic option to improve the current standard of care.

In this phase 1, first-in-human study, we investigated the safety, including immunogenicity, tolerability, and pharmacokinetics (PKs), of rADAMTS-13 in patients diagnosed with severe congenital ADAMTS-13 deficiency to establish dosing for future studies.

This was a prospective, open-label, multicenter, first-in-human dose escalation study on safety, immunogenicity, tolerability and PKs of BAX 930 in patients diagnosed with cTTP assigned to 1 of 3 dose cohorts, receiving a single dose of BAX 930 at 5, 20, or 40 U/kg BW, where 1 unit is approximately equivalent to the ADAMTS-13 activity in 1 mL plasma. The study comprised 2 dose-escalation steps. Fifteen patients were enrolled (of 16 screened; 1 was unable to participate because of a scheduling conflict) and dosed sequentially. Participants were recruited to the next higher dose level only after short-term safety had been demonstrated and reviewed by an independent Data Monitoring Committee at the preceding dose level. The study was approved by the responsible ethics committees or institutional review boards and regulatory authorities. The study was registered at www.clinicaltrials.gov as #NCT02216084. All participants gave written informed consent. Data were analyzed by biostatisticians at Shire and Quintiles; all authors had access to primary clinical trial data.

The principal inclusion criteria consisted of a documented diagnosis of severe congenital ADAMTS-13 deficiency (baseline plasma ADAMTS-13 activity < 6%) confirmed by genetic testing, ages 12 to 65 years, and the absence of severe TTP symptoms at the time of screening. This included stable laboratory parameters (platelet count, >100 × 10⁹/L; lactate dehydrogenase [LDH] <3 times the upper limit of normal). Evidence of end organ damage such as stroke, heart, renal disease, or abnormal liver function tests were further exclusions to study enrollment. Patients with a medical history or presence of functional neutralizing ADAMTS-13 inhibitor at screening, a medical history of immunological or autoimmune disorders, or significant neurological events were excluded. A sample size of 14 was selected on the basis of feasibility.

BAX 930 is a fully glycosylated recombinant human ADAMTS-13 protein produced in a Chinese hamster ovary (CHO) mammalian expression system in a plasma protein-free milieu. The cell culture and purification processes used in the manufacture of BAX 930 do not employ additives of human or animal origin. The purification process of BAX 930 is based on chromatography using conventionally available resins, including anion exchange, hydroxyapatite, mixed mode (hydrophobic interaction), and cation exchange. In addition, BAX 930 undergoes 2 dedicated virus reduction steps: solvent/detergent treatment and nanofiltration. The final product is lyophilized for intravenous injection. The sterile, lyophilized BAX 930 was stored in 10-mL vials, which were reconstituted with 5 mL sterile water, giving the solution a nominal strength of 300 U/mL that is stable for 3 hours at room temperature after reconstitution. BAX 930 was infused at a rate of 2 to 4 mL/minute. In patients receiving regular prophylaxis, there was at least a 10-day window between their last prophylactic treatment and their BAX 930 dose.

Patients were hospitalized for at least 96 hours after investigational product infusion for safety observation and PK assessments. Follow-up visits were scheduled 12 days and 30 days (or 6, 9, and 12 days for the 2 participating adolescents) postinfusion. Safety was evaluated through clinical assessment, hematology, serum chemistry, urinalysis, coagulation tests, antibodies against ADAMTS-13 or CHO cell protein, viral serology, cardiac troponin, d-dimers, and seromarkers for brain or renal impairment. Adverse events (AEs) were recorded throughout the study, but specifically by the patient in a diary, beginning 96 hours postinfusion until the study completion visit.

ADAMTS-13 activity was calculated by analysis of fluorescence over time, using the synthetic fluorogenic VWF73 peptide substrate (FRETS-VWF73) and a commercially available, chromogenic GST-VWF73-based ADAMTS-13-activity ELISA (Technozym® ADAMTS-13 Activity ELISA).^28,29  ADAMTS-13 antigen was measured using a validated sandwich ELISA employing anti-human-ADAMTS-13 antibodies.^30-32  VWF:RCo and VWF:Ag were determined as previously described.^33,34 

Analysis of VWF multimers was performed using sodium dodecyl sulfate-agarose gel electrophoresis followed by western blotting and sensitive luminescence detection.^4,35  VWF degradation products generated by ADAMTS-13-mediated cleavage were assessed by sodium dodecyl sulfate‐polyacrylamide gel electrophoresis under reducing conditions followed by western blotting and immunostaining with a horseradish peroxidase‐labeled polyclonal rabbit antihuman VWF antibody with enhanced chemiluminescence detection.^4,35,36 

Total binding antibodies to ADAMTS-13 were measured by ELISA-based assays, detecting total immunoglobin (IgG, IgA, IgM). Neutralizing antibodies to ADAMTS-13 were measured using a Bethesda-like approach with the ADAMTS-13 activity assay.³⁷ 

PK parameters, including area under the concentration-time curve (AUC [h∙U/mL]), plasma half-life (t_1/2 [hours]), clearance (Cl [mL/kg/hours]), mean residence time (MRT [hours]), volume of distribution at steady state (V_ss [mL/kg]), maximum plasma concentration (C_max [U/mL]), and incremental recovery (IR [U/mL]) were assessed for ADAMTS-13 activity and ADAMTS-13 antigen (ADAMTS-13:Ag). Concentrations were measured at standardized time points: within 60 minutes preinfusion and 15, 30, 60 minutes and 3, 6, 9, 12, 24, 48, 96, 144, 168, 192, 216, 240, 264, and 288 hours postinfusion in adults. The 2 adolescent subjects (16 and 17 years of age) had sparse sampling preinfusion and at 6, 48, 144, 216, and 288 hours postinfusion. The hematology and clinical chemistry assessments, conducted as part of the safety assessments (see earlier), were performed on EDTA anticoagulated whole blood and serum, respectively, at the screening visit, just before investigational product infusion, and at regular increments up to 288 hours postinfusion. The hematology panel included platelet counts, and the clinical chemistry panel included serum LDH, both of which were used for assessing the pharmacodynamic effects of BAX 930.

ADAMTS-13 concentration data were analyzed using a noncompartmental approach, and individual PK parameter estimates were derived. Total AUC_0‐∞ was calculated by the linear trapezoidal rule up to the last quantifiable concentration with a tail area correction. Terminal plasma half-life was determined from the terminal rate constant obtained by log-linear fitting of a regression line by the least squares deviation method to the last 5 quantifiable concentrations that were above preinfusion levels. MRT was calculated as total area under the moment curve divided by the total area under the curve corrected for the duration of the infusion. Systemic clearance (Cl) was calculated as dose per body mass (kg) divided by AUC_0‐∞. IR was calculated as C_max divided by dose per body mass.

To better address the variability in baseline ADAMTS-13 levels, a population PK model was also developed. A 2-compartment model with first-order elimination and constant rate infusion was selected iteratively among different structural and stochastic candidates and provided population PK parameter estimates. Individual PK parameters from a noncompartmental approach and from the population PK model were similar.

Descriptive statistics were used to assess safety in terms of product-related AEs.

Fifteen patients received study drug: 3 each in the 5 and 20 U/kg dose cohorts, and 9 in the 40 U/kg dose cohort (Figure 1; Table 1); 2 patients in the 40 U/kg cohort were adolescents. All patients completed the study. Doses administered were within 0.3 U/kg of the planned doses, with the exception of 1 patient who received 47.5 U/kg instead of the planned 40 U/kg.

After the single infusion of BAX 930, dose-related increases in ADAMTS-13 FRETS-VWF73 concentrations were observed (data from the 40 U/kg cohort shown in Figure 2). ADAMTS-13 activity remained quantifiable for 24 hours in the 5 U/kg cohort, for 240 hours in the 20 U/kg cohort, and for the full 288-hour study period in the 40 U/kg cohort. ADAMTS-13 activities measured by FRETS-VWF73 and chromogenic assay methods were highly comparable (Table 2).

In the 40 U/kg cohort, the geometric mean of IR was 0.0232 (U/mL∙kg/U), mean terminal half-life was 59.2 hours, and initial half-life was 17.0 hours for ADAMTS-13 activity (FRETS-VWF73). ADAMTS-13 activity, C_max, and total exposures increased approximately in proportion to the dose escalations. The geometric mean C_max was 0.398 U/mL after 20 U/kg infusion and 0.948 U/mL after 40 U/kg infusion; the geometric mean AUC_(0-inf) was 19.1 U∙h/mL after 20 U/kg infusion and 53.1 U∙h/mL after 40 U/kg infusion.

With the more sensitive chromogenic ELISA assay (lower limit of quantitation [LLOQ]: 0.0073 U/mL compared with 0.05 U/mL for FRETS-VWF73), low levels of ADAMTS-13 activity were measurable at the preinfusion time in all patients except 1 adult and both adolescents.

Despite the differences in the LLOQ, activity measurements by FRETS-VWF73 and chromogenic assay methods were otherwise similar (Table 2).

ADAMTS-13 antigen (ADAMTS-13:Ag) concentrations at predose (baseline) were low, ranging from 0.003 to 0.031 U/mL, but quantifiable in all but 2 patients (1 patient each in the 5 and 40 U/kg dose cohorts). Individual postdose ADAMTS-13:Ag levels remained quantifiable in all samples up to the last collection at 288 hours postdose and were higher than baseline (predose) values.

The geometric mean of ADAMTS-13:Ag Cl was 61.5 mL/h, and of V_ss was 5300 mL (at 40 U/kg dose level), suggesting the protein distributed primarily within the intravascular compartment.³⁸  There was approximate dose proportionality with respect to C_max (geometric means of 0.323 µg/mL after 20 U/kg infusion and 0.672 µg/mL after 40 U/kg infusion) and AUC_(0-inf) (geometric means of 18.3 µg∙h/mL after 20 U/kg infusion and 36.0 µg∙h/mL after 40 U/kg infusion).

The PK parameter estimates for ADAMTS-13:Ag were comparable to ADAMTS-13 activity. Above the 5 to 40 U/kg dose range, ADAMTS-13:Ag geometric mean C_max increased 5-fold between 5 and 20 U/kg, and approximately proportional to the dose between 20 and 40 U/kg (2.1-fold) (Table 2). The fold increases in geometric mean total exposures (AUC_(0-t) and AUC_(0-inf)) were approximately proportional to the dose increase for both dose level comparisons.

The available concentration data for both ADAMTS-13 activity and ADAMTS-13:Ag, did not suggest any major differences between adolescents and adults. Despite the low number of adolescents in the study (n = 2) and reduced sampling per protocol, ADAMTS-13 activity, IR, t_1/2, and AUC were generally within the data ranges for adults treated at the same dose level (40 U/kg). The estimated clearances for the 2 adolescent subjects were 64.9 and 54.5 mL/h, respectively, which were within the range observed for adult subjects (n = 7; range, 44.4-115 mL/h).

In addition to the small, intermediate, and large multimeric sizes typically present in normal plasma, UL VWF multimers were observed in the samples collected before dosing, at screening or predose, in all patients, and at most points postdose. The trend for decreasing large multimers, including UL multimers, and increasing levels of the intermediate fraction was observed during the first 12 to 24 hours postinfusion in individual profiles at BAX 930 doses of 20 or 40 U/kg before slowly returning to preinfusion levels during a 288-hour period (Figure 3B).

Likewise, an apparent dose-dependent effect was seen with the detection of ADAMTS-13-mediated VWF cleavage products. In the 5 U/kg dose cohort, ADAMTS-13-mediated VWF cleavage products were present up to 3 hours postdose in all patients (100%) and up to 6 hours postdose in 1 patient (33.3%), and no longer detectable thereafter (Figure 3A). ADAMTS-13-mediated VWF cleavage products were observed during a longer period of time at higher dose levels: for 20 U/kg, up to 24 hours postdose in all patients and up to 48 hours in 2 patients (66.7%); for 40 U/kg, up to 48 hours postdose in all patients and as long as 264 hours in 1 patient (14.3%) (Figure 3C).

In addition, the function of VWF to bind platelet GPIb, as measured by the VWF:RCo assay, was lower than mean baseline levels at most points across the 3 dose levels and decreased by approximately 30% within the first 9 hours. A modest decrease in LDH of 5% to 10% relative to baseline was also observed in the first 48 hours postinfusion. In addition, although subjects in this study were not experiencing acute manifestations of cTTP, there was an increasing trend in the platelet count up to 6 days after infusion in all dosing cohorts (Figure 4).

BAX 930 was well tolerated in all 15 patients, and no patient exhibited anaphylaxis or other allergic manifestations. No serious AEs, AEs leading to discontinuation from the study, or breakthrough TTP events were seen in the study after treatment. Overall, 12 of 15 patients (80.0%) reported at least 1 temporally emergent AE (TEAE; supplemental Table 1, available on the Blood Web site). Three subjects, all in the 40 U/kg BAX 930 cohort (40 U/kg: 3/9 subjects, 33.3%; overall: 3/15 subjects, 20.0%) reported a total of 5 TEAEs considered by the investigators to be temporally associated with BAX 930 infusion: decrease in VWF antigen and VWF activity at 30 minutes postinfusion, returning to normal activity levels at 1 hour postinfusion (n = 1), flatulence (n = 1), or nausea on 2 separate days (n = 1). No trends were observed over time or among the 3 dose cohorts in the laboratory parameters, vital sign assessments, electrocardiogram measurements, or physical examination.

All anti-ADAMTS-13 neutralizing antibody results were less than 0.6 BU/mL across the 3 dose cohorts and at all 3 scheduled times; that is, screening, infusion predose, and study completion, day 28 ± 3 days. None of the 15 patients who received BAX 930 exhibited anti-ADAMTS-13 binding antibodies at any time in the study. Similarly, all patients tested negative for anti-CHO protein antibodies at all times.

Recombinant ADAMTS-13 potentially offers a novel option for the treatment of cTTP, eliminating many of the risks associated with products derived from human plasma and facilitating the introduction of more precise and individualized dosing regimens.^24,39-43  In our prospective, dose-escalating clinical trial, we investigated the safety, tolerability, and PK of BAX 930 infused for the first time in humans.

BAX 930 was expected to behave in the same way as endogenous ADAMTS-13, as demonstrated in several nonclinical in vitro and in vivo studies. In current clinical practice, infusions of 10 to 20 mL/kg FFP are recommended for treating acute episodes or prophylaxis of cTTP.^44-46  As such, a low dose of 5 U/kg, an intermediate dose of 20 U/kg, and a high dose of 40 U/kg BAX 930 were chosen for this study. The 40 U/kg dose has been calculated to correspond to doses of ADAMTS-13 typically administered during single-volume plasma exchange sessions.⁴⁷ 

As would be expected in this population of patients with cTTP, predose baseline levels of ADAMTS-13:Ag and ADAMTS-13 activity (assessed by both FRETS-VWF73 and chromogenic ELISA assays) were very low or below the LLOQ. Immediately after infusion of all doses, of BAX 930 there was an increase in individual ADAMTS-13:Ag and activity observed, reaching a maximum within 1 hour postinfusion. Similar ADAMTS-13:Ag, ADAMTS-13 FRETS-VWF73, and ADAMTS-13 activity ELISA profiles suggest good correlation between protein concentrations and activity. No apparent differences were observed for ADAMTS-13 activity and ADAMTS-13:Ag between adolescents (n = 2) and adults (n = 7), treated with the same BAX 930 doses.

Overall, there was evidence suggesting dose proportionality with respect to AUC_(0-inf) and C_max (Table 2; Figure 5). The change in the geometric mean C_max for 5 to 20 U/kg was approximately 5-fold, and was approximately 2.4-fold for 20 to 40 U/kg. Greater increases were observed for geometric mean AUC_(0-inf) (47-fold for 5 to 20 U/kg and 3.1-fold for 20 to 40 U/kg), which could be attributable to concentrations falling below the LLOQ at lower doses. Available data for AUC_(0-inf) showed an approximately 2.8-fold increase in exposure for dose increases from 20 to 40 U/kg. Available data at the 20 and 40 U/kg dose levels show no major changes in Cl, V_ss, t_1/2, and MRT_(0-inf) with increasing dose.

PK parameters derived from the noncompartmental and population PK approaches were consistent. In the population PK model, a shorter initial half-life was observed followed by a longer terminal half-life. Whether the initial phase is a result of VWF binding, consumption of ADAMTS-13 activity, or distribution to another physiologic compartment remains to be determined.

In addition to the PK results, evidence of pharmacodynamic activity was observed. With escalating BAX 930 doses, prolonged detectable ADAMTS-13-mediated VWF cleavage products were present, in line with dose-related increases of ADAMTS-13:Ag and activity. Specifically, detectable ADAMTS-13-mediated VWF cleavage products were present in all patients (100%) and up to 48 hours after the highest dose, 40 U/kg. Patients with cTTP typically exhibit UL molecular weight forms of VWF at baseline consistent with the severe reduction in ADAMTS-13 activity. A trend for decreasing large multimers, a fraction of which also included UL multimers, and increasing levels of the intermediate fraction was observed during the first 24 hours postinfusion in individual profiles at BAX 930 doses of 20 or 40 U/kg. In addition, during the first 9 hours, the mean postdose VWF:RCo levels decreased by approximately 30%. There was also an increase in the platelet count in all dosing groups. Taken together, these findings provide evidence of in vivo ADAMTS-13 activity after BAX 930 administration.

This study demonstrated that the protein antigen and activity PK parameters of BAX 930 were comparable to those estimated from previous studies of ADAMTS-13 administered to patients with cTTP as a constituent of FFP. Furlan et al²⁴  reported limited PK data in patients treated with FFP and demonstrated that the recovery was nearly 100% and the half-life was 2.1 to 3.3 days. Similar results were reported by Fujimura et al.⁴³  The comparability of plasma-derived and recombinant ADAMTS-13 PKs suggests dosing regimens in future studies of the long-term safety and efficacy of BAX930 can be modeled on the clinical experiences with plasma-based regimens. No patients developed an inhibitor to ADAMTS-13 during this single-dose study. The lack of an apparent immune response will need to be confirmed in the upcoming pivotal study in which patients will be repeatedly exposed to BAX 930 for approximately 1 year.

In conclusion, BAX 930 appeared to be safe and well-tolerated over a dose range of 5 to 40 U/kg in patients with cTTP, and there was no evidence of an immune response to BAX 930 after a single infusion. The activity PK parameters were comparable to those estimated from FFP studies and demonstrated approximate dose proportionality with respect to C_max and AUC. Finally, there was evidence for BAX 930 activity in vivo, including effects on VWF multimers, platelet count, and serum LDH. These data provide the basis for proceeding to a phase 3 pivotal study in cTTP, using previously anticipated prophylactic and treatment dosing regimens.

The authors thank Barbara Plaimauer and Herbert Gritsch for work on ADAMTS-13-related assays and Ulrich Budde for work on the analyses of VWF. Finally, the authors are grateful to the rADAMTS clinical study team for operational support: Alex Bauer, Martin Wolfsegger, Roger Berg, Sandra Raff, Elizabeth Staron, Li Zhang, and Dietmar Hollensteiner.

This study was funded by Shire.

Contribution: B.E. designed the study; C.M., J.D., C.H., L.M., and B.E. interpreted data; M.S., P.K., K.K., L.R., J.W., R.S., J.A.K.H., M.K., Y.F., and B.E. performed research; L.M. performed statistical analysis; J.D., C.H., and B.E. wrote the manuscript; and B.E. supervised the research.

Conflict-of-interest disclosure: J.D., C.H., L.M., and B.E. are employees of Shire. C.M. is an employee of Quintiles, which assisted with the conduct of this study. P.K., K.K., L.R., J.W., R.S., J.A.K.H., M.K., and Y.F. received consultancy fees for conducting the research. M.S. received honoraria for speakers’ fees for Baxalta/Shire and advisory boards.

Correspondence: Marie Scully, University College London Hospitals NHS Trust, 60 Whitfield St, London W1T 4EU, United Kingdom; e-mail: m.scully@ucl.ac.uk.