The bioequivalence study was a randomized, double blind, single-center phase I study with three treatment arms: ONS-3010 (Oncobiologics Inc., Cranbury, NJ, USA) and reference products Humira^® EU (AbbVie, Berkshire, UK) and Humira^® US (AbbVie, North Chicago, IL, USA, EudraCT registration # 2013-003551-38). The study was performed in a total number of 198 volunteers, who were deemed healthy and screened negative for (latent) tuberculosis, acute infectious disease, malignancy, and autoimmune disorders. All women of child bearing potential and all males had to practice effective contraception during the study and had to be willing and able to continue contraception for at least 5 months after dose administration of study treatment to be able to participate in the study. In addition, the subjects had to have a negative screening result for Hepatitis B/C and HIV, and smoking was prohibited during the complete study period. Sixty-six subjects per treatment arm received a single subcutaneous dose of 40 mg TNFα antibody according to a randomization schedule, which was stratified for gender. PK, immunogenicity, and safety analyses were conducted for all 198 study participants. The PD effects of these three TNFα blockers were assessed by LPS/aluminum hydroxide whole blood challenges, for which 36 subjects were randomly selected from the 3 treatment arms (6 male and 6 female subjects per arm). Before the LPS/aluminum hydroxide whole blood challenges were implemented as a PD bioassay in the clinical study, the effect of TNFα blockade on the LPS/aluminum hydroxide-induced inflammatory response was explored in an in vitro experiment (see below).

The phase I clinical trial was conducted in accordance with the Declaration of Helsinki and Guideline for Good Clinical Practice and was approved by the Independent Ethics Committee of the Foundation “Evaluation of Ethics in Biomedical Research,” Assen, The Netherlands.

Plasma concentrations of adalimumab (ONS-3010, Humira^® EU, or Humira^® US) following a single 40 mg subcutaneous dose administration were assessed in blood samples collected in 4 mL SST™ Gel and Clot Activator tubes (Becton Dickinson), from day 1 (pre-dose) up to day 71 (post-dose). An electrochemiluminescence assay (Meso Scale Discovery, MSD) was developed and validated to measure the concentration of adalimumab in human serum. Data values below the limit of quantification (BLOQ, 250 ng/mL) were set to 125 ng/mL (50% of the limit of quantification, LOQ) for summarizing and graphical purposes. If more than one-third of all values were BLOQ for a specific time point/treatment combination, the data point was not included in the summary/graph. For PK analysis, BLOQ data points were set to zero when not embedded between two measurable data points (i.e., >BLOQ) and occurring prior to T_max. BLOQ data points were excluded from PK analysis when not embedded between two measurable data points and occurring after T_max, or when embedded between two measurable data points.

Preceding the phase I clinical trial, the effect of in vitro TNFα blockade on LPS/aluminum hydroxide-induced cytokine release in whole blood samples was explored. Blood specimens were obtained from four healthy male volunteers in sodium heparin tubes (Becton Dickinson). Fresh whole blood was (pre-)incubated in 96-well plates with Humira^® EU at 37°C at 5% CO₂. The Humira^® concentration ranged from 0.3125 to 10 μg/mL, which was based on the expected maximum concentrations of 4.7 ± 1.6 μg/mL, observed after a single administration of 40 mg (11). After 90 min, 2 ng/mL LPS (Sigma-Aldrich, St Louis, MO, USA) and 100 μg/mL Alhydrogel 85 (aluminum hydroxide, Brenntag, Frederikssund, Denmark) were added to the samples followed by an additional incubation period of 20 h. In the clinical study, LPS/aluminum hydroxide whole blood challenges were performed in a subset of study participants, pre-dose (day 1) and on day 5, day 15, day 29, and day 64. Conditions for the whole blood challenges were similar as for the in vitro experiment above, except for the fact that the incubations were performed in a larger volume (2 mL), and no pre-incubation with Humira^® EU was performed. Cytokine release was measured in the culture supernatants by an electrochemiluminescence assay (MSD, V-plex; TNFα, IL-1β, IL-6, IL-8, and IFNγ, intra- and inter-assay variation: <10%) or ELISA (R&D Systems; IL-18).

Blood samples for analysis of adalimumab-associated antibodies against ONS-3010 and Humira^® were collected in 4 mL SST™ Gel and Clot Activator tubes (Becton Dickinson). A bridging ELISA assay was used to measure anti-adalimumab antibodies in serum samples by electrochemiluminescence. Positive immunogenicity samples were further investigated to confirm the specificity of binding. If a sample was confirmed positive for specific anti-adalimumab antibodies, the neutralizing capacity of these antibodies was investigated. Binding specificity was determined by competitive inhibition with 5 mg/mL of unlabeled ONS-3010; samples with an inhibition >20.7% were confirmed positive for anti-drug antibody. Assay sensitivity was 2.463 ng/mL and inter-run assay precision 5.143–11.354%.

Pre-dose (day 1) and on day 2 (hematology only), day 4, day 8, day 36, and day 72, blood and urine samples were collected for routine hematology, biochemistry, and urinalysis safety parameter assessments. In addition, general safety measures such as vital signs, electrocardiography, and symptoms were assessed throughout the study.

Demographic and PD data were summarized and presented by graphical and tabular presentations. For the inhibition of TNFα and IFNγ release after incubation with LPS/aluminum hydroxide and increasing adalimumab concentrations in vitro, a model was fitted to the data in R (v2.15.2, R Foundation for Statistical Computing, Vienna, Austria, 2012) and described by a maximal effect (E_max) and half maximal effect concentrations (EC₅₀).

A non-compartmental analysis was performed to describe the PK of the antibodies using SOFTWARE. The area under the curve (AUC) was computed from zero to the last measurement point (AUC_0–last). If the terminal phase was sufficiently well characterized (at least three data points after T_max and a linear regression r² ≥0.5), the terminal half-life (t_1/2) and the AUC zero to infinity (AUC_0–inf) were estimated. PK profiles of the investigational product (ONS-3010) and the reference products (Humira^® EU and Humira^® US) were compared using general linear model procedures in SAS^®. An analysis of variance (ANOVA) was performed on the untransformed elimination rate constant (K_el), apparent terminal elimination half-life (t_{1/2 el}), and ln-transformed AUC_0–last, AUC_0–inf, and C_max. The ANOVA model included treatment as a fixed effect. The ratio of means with the 90% geometric confidence interval (CI) was calculated for AUC_0–last, AUC_0–inf, and C_max. Bioequivalence was evaluated in accordance with EMA and FDA guidelines on the investigation of bioequivalence (30, 31). The hypothesis of bioequivalence for the PK parameters AUC_0–last, AUC_0–inf, and C_max was supported with 90% CI for the relative geometric means ratio fully enclosed in the equivalence bounds of 80–125%.

A total number of 90 male and 108 female subjects aged between 18 and 55 years, with a BMI of 19–29 kg/m² and a body weight >50 kg were enrolled in the clinical trial with both sexes equally divided over the three treatment groups (ONS-3010 and Humira^® EU/US, Figure 1; Table 1).

Five subjects (four Humira^® US; one Humira^® EU) had an incomplete PK profile as a result of early study termination, and therefore PK parameters could not be determined accurately. These participants were excluded from the statistical analysis on total exposure parameters (e.g., AUC_0–last). Two of these subjects had a PK profile up to 7 and 8 days, based on which C_max and T_max could be determined, and these data were included in the statistical analysis (Table 2).

Subcutaneous administration of adalimumab resulted in a steady increase in plasma concentrations starting at approximately 36 h post-dose to reach a maximum concentration after 6–7 days, followed by a log-linear decrease and a non-linear decrease once the plasma concentration reached lower levels, similar to other monoclonal antibodies (Figure 2). Large variability in antibody concentration was observed between subjects during the elimination phase, with some subjects having adalimumab plasma concentration levels BLOQ (<250 ng/mL) 5 weeks after dosing, while others continued to have measurable levels up to the last study visit (day 71). On the primary PK endpoints, AUC_0–inf and C_max, equivalence was demonstrated based on bounds of 80–125% for the ratio of the geometric means (ONS-3010/Humira^®; Table 3). Equivalence was also demonstrated for secondary PK endpoint AUC_0–last.

In a separate experiment preceding the clinical study, the effect of in vitro TNFα blockade on LPS/aluminum hydroxide-induced cytokine release in whole blood samples was explored. Addition of adalimumab to the LPS/aluminum hydroxide-triggered whole blood samples resulted in dramatically reduced TNFα levels measured in the culture supernatant. An average reduction in TNFα level of 97 ± 0.6% was observed already at the lowest adalimumab concentration evaluated (0.3125 μg/mL), as compared to the LPS/aluminum hydroxide only samples. With increasing concentrations of adalimumab, further reductions in TNFα levels were measured, up to 99 ± 0.1%. In addition, TNFα blockade affected the LPS/aluminum hydroxide-induced release of IFNγ with a maximal reduction of 93 ± 4% observed at an adalimumab concentration of 10 μg/mL. The adalimumab concentration-effect curves for TNFα and IFNγ are presented in Figure 3, with an E_max of 99 and 97% and an EC₅₀ of 0.006 and 0.6 μg/mL for TNFα and IFNγ, respectively. Increasing concentrations of adalimumab did not affect the release of IL-6, IL-1β, and IL-18 following LPS/aluminum hydroxide stimulation of whole blood (data not shown).

In the clinical study, adalimumab treatment resulted in a more than 99% reduction in measurable TNFα levels in LPS/aluminum hydroxide-triggered whole blood samples (Figure 4A). This effect was maximal at the first time point investigated (day 5) and lasted until at least 3 weeks post-dose. Thereafter, TNFα levels started to return to baseline with individual time courses depending on the adalimumab plasma concentration. On average, TNFα levels were still approximately 85% reduced at the last time point investigated (day 64). The reduction in TNFα levels did not differ between the three treatment groups.

Lipopolysaccharide/aluminum hydroxide-induced release of IL-1β and IL-6 was not inhibited by TNFα blockade (Figures 4B,C). For all treatment groups, a slight increase over baseline cytokine levels was observed on day 29, both for IL-1β (50, 25, and 17% for treatment ONS-3010, Humira^® EU, and Humira^® US, respectively) and IL-6 (25, 29, and 19%). Administration of adalimumab resulted in a reduction of LPS/aluminum hydroxide-induced IL-8 release (Figure 4D). Minor differences were observed between treatment groups, with maximal reductions observed on day 5 [ONS-3010, −24% (CV 56%)] and day 15 [Humira^® EU, −29% (CV 28%); Humira^® US, −30% (CV 44%)], and a return to baseline levels between day 29 and day 64. Whereas adalimumab inhibited LPS/aluminum hydroxide-induced IFNγ release in the preclinical in vitro experiment, this effect could not be confirmed in the clinical study. IFNγ release was highly variable between subjects and BLOQ (<114 pg/mL) in the baseline samples of 21 out of the total number of 36 participants. IFNγ release strongly differed between males and females with mean baseline values of 1308 ± 1803 and 144 ± 278 pg/mL, respectively. Also, for the other cytokines (TNFα, IL-1β, IL-6, and IL-8) sex differences were observed, but to a lesser extent (cytokine release on average 1.6-fold higher in the male subjects, data not shown).

The immunogenicity profiles were well comparable between the ONS-3010, Humira^® EU, and Humira^® US treatment groups for the confirmed and neutralizing antibodies (Table 4).

Adalimumab administration resulted in a mild decrease in neutrophil count, amounting maximally 15–20% at day 8, and returning to baseline levels after 5 weeks (Figure 5A). Furthermore, adalimumab administration induced a mild increase in lymphocyte count, peaking 3 days after dose administration (Figure 5B). There were no indications for differences between treatment groups in neutrophil and lymphocyte cell counts over time.