Pharmacokinetic Similarity of ABP 654, an Ustekinumab Biosimilar Candidate: Results from a Randomized, Double‐blind Study in Healthy Subjects

ABP 654 is a proposed biosimilar to ustekinumab reference product (RP) which works through antagonism of interleukin‐12 and interleukin‐23. Ustekinumab RP is used for the treatment of chronic inflammatory conditions, including some forms of plaque psoriasis, psoriatic arthritis, Crohn's disease, and ulcerative colitis. A randomized, double‐blinded, single‐dose, 3‐arm, parallel‐group study was conducted to assess the pharmacokinetic (PK) similarity of ABP 654 with ustekinumab RP sourced from the United States (US) and the European Union (EU); the PK similarity of ustekinumab US versus ustekinumab EU; and the comparative safety, tolerability, and immunogenicity of all 3 products. A total of 238 healthy subjects were randomized 1:1:1 and stratified by gender and ethnicity (Japanese versus non‐Japanese) to receive a single 90 mg subcutaneous injection of ABP 654 or ustekinumab US or ustekinumab EU. PK similarity was established based on 90% confidence intervals (CIs) for the primary endpoints of area under the concentration‐time curve from time 0 extrapolated to infinity (AUCinf) and maximum observed serum concentration (Cmax) being contained within the prespecified margin of 0.8–1.25. No clinically meaningful differences in immunogenicity were found among the 3 products. Adverse events were similar between treatment groups and consistent with the safety profile of ustekinumab RP. Results indicate that ABP 654, ustekinumab US and ustekinumab EU share similar PK and safety profiles.

ABP 654 is in development as a biosimilar candidate to ustekinumab, a human IgG1κ monoclonal antibody (mAb) that binds with high specificity and affinity to the p40 protein subunit used by cytokines interleukin (IL)-12 and IL-23. 1 IL-12 and IL-23 are involved in immune system responses, such as natural killer (NK) cell activation and CD4 + T-cell differentiation and activation, which play a key role in the pathophysiology of chronic inflammatory diseases. 2 Ustekinumab reference product (RP) blocks IL-12 and IL-23 from binding to their receptor, IL-12Rβ1, on the surface of NK and T cells. 3 Ustekinumab RP is used for the treatment of chronic inflammatory diseases, including some forms of plaque psoriasis, psoriatic arthritis, Crohn's disease, and ulcerative colitis.
While biologics have transformed the treatment of chronic inflammatory disorders, their high costs have increased healthcare expenditure and been an impediment to their widespread use. [4][5][6] To improve access, the US Food and Drug Administration (FDA), Euro-pean Medicines Agency (EMA), and other regulatory agencies worldwide have created abbreviated development and approval pathways for biosimilars, a category of biologic therapeutics that are highly similar to an approved RP. [7][8][9] With the approaching expiration of the exclusivity period for ustekinumab RP, there is increasing interest in ustekinumab biosimilars as safe and efficacious potential alternatives. Comprehensive analytical characterization has found that ABP 654 has the same amino acid sequence as ustekinumab and is similar to it in terms of structure, function, purity, and potency. 10 The clinical study described herein was conducted in healthy subjects to evaluate the similarity of single-dose pharmacokinetics (PKs), safety, and immunogenicity of ABP 654, US FDA-licensed ustekinumab RP, and European Union (EU)-authorized ustekinumab RP, hereafter referred to as ustekinumab US and ustekinumab EU, respectively. For approval in the US or EU, a biosimilar candidate must demonstrate biosimilarity to the RP approved in the respective region. Establishment of PK similarity between ustekinumab US and ustekinumab EU, along with previously established analytical similarity, would complete the scientific bridge that justifies the use of a single-sourced ustekinumab RP in future comparative clinical studies and would support approval in countries where the local RP was not used in the clinical studies.

Methods Subjects
Eligible subjects were healthy men and women between 18 and 45 years of age at the time of screening, with normal or clinically acceptable physical examination, clinical laboratory test values, vital signs, and electrocardiogram (ECG) as determined by the investigator at screening. Subjects were to have a body weight of 50-90 kg at screening which occurred within 28 days of dosing. Subjects were excluded from participation if they were pregnant, breastfeeding, or planning to become pregnant. Subjects were to have no prior exposure to ABP 654, ustekinumab RP (or any other ustekinumab biosimilar), IL-12 antagonists, IL-23 antagonists, or related compounds.
This study was conducted in accordance with the International Council for Harmonization Good Clinical Practice guidelines. The study protocol and all amendments, the informed consent form, and any accompanying materials provided to the subjects were reviewed and approved by a central Institutional Review Board (Advarra IRB; Columbia, Maryland).

Study Design
This was a randomized, double-blinded, single-dose, 3-arm, parallel-group study conducted in the United States at Pharmaceutical Research Associates, Inc. in Lenexa, KS and Salt Lake City, UT, and at WCCT Global, Inc. in Cypress, CA. Following screening, eligible subjects were admitted to the Clinical Pharmacology Unit (CPU) and were randomized in a 1:1:1 ratio to receive a single 90 mg subcutaneous (SC) injection of ABP 654 (Amgen Inc., Thousand Oaks, CA, USA), ustekinumab US (Stelara; Janssen Biotech, Inc., Horsham, PA, USA), or ustekinumab EU (Stelara; Janssen-Cilag International NV, Belgium). A sample size of 231 subjects (n = 77 per treatment group) was chosen to provide >90% power to demonstrate the similarity of the primary PK endpoints based on the assumptions of between-subject variability (as measured by coefficient of variation) of 40% for ABP 654, ustekinumab US, and ustekinumab EU; a true geometric mean ratio (GMR) of 1 between ABP 654, ustekinumab US, and ustekinumab EU; a bioequivalence margin of 0.8-1.25, two 1-sided tests at α = 0.05, and a 5% dropout rate. The dose was selected based on the product labeling for ustekinumab RP, a history of safe use in healthy subjects, and the expected PK profile. Study drug administration occurred on day 1 after predose baseline procedures were completed. Subjects remained in the CPU until day 3 for safety evaluations and PK assessments, then discharged after the inpatient study procedures were completed. Subjects returned to the CPU on days 7,9,11,13,21,28,35,49,56,70,98, and 112 for PK, safety, and immunogenicity assessments. After their discharge from the CPU, site staff contacted subjects daily up to day 15 to inquire about the subject's well-being and to monitor adverse events.
The prespecified equivalence criteria for the primary PK endpoints area under the concentration-time curve from time 0 extrapolated to infinity (AUC inf ) and maximum observed serum concentration (C max ) was the 90% confidence interval (CI) for the GMR within 0.80-1.25. Secondary endpoints included other PK parameters, safety, tolerability, and immunogenicity of ABP 654 in healthy adult subjects compared with ustekinumab US and ustekinumab EU.

Pharmacokinetic Assessments
The PK concentration analysis set (n = 237) consisted of all subjects who were randomized and received any study drug, and who had at least 1 reported serum concentration of ABP 654 or ustekinumab (US or EU). Serum samples were collected over 112 days post-treatment. Concentrations of ABP 654 and ustekinumab (US and EU) were measured using the Meso Scale Discovery platform (Meso Scale Diagnostics LLC, Rockville, MD, USA). This sensitive electrochemiluminescent (ECL) assay uses a ruthenium-labeled anti-idiotype mAb to detect a signal from antibodies bound to the 3 test products after an electrical current is applied to the detection plate. The assay was validated and qualified based on established quantitative PK methods described in the published literature. [11][12][13] The assay range was 20-32,000 ng/mL, which provided adequate sensitivity to reliably detect ABP 654 ustekinumab US and ustekinumab EU in serum. PK parameters for ABP 654 and ustekinumab (US and EU) were estimated with noncompartmental methods with Phoenix WinNonlin version 8.1 or later using best-fit regression. The PK parameters were estimated from concentration-time profiles, and AUCs were calculated using the linear trapezoidal method. All PK parameters were listed by subject and summarized descriptively using the PK parameter analysis set, which included all subjects with sufficient concentration-time data to identify a valid C max and/or calculate a valid AUC. Baseline body weight was included as a covariate in the primary statistical analysis as well as for analyses using the binding antidrug antibodies (ADAs) negative group.
The point estimates and 90% CIs for the GMR (testto-reference) of PK parameters were calculated from an analysis of covariance model adjusted for baseline body weight using the PK parameter analysis set. To establish PK similarity, the 90% CIs for the GMR between ABP 654 and ustekinumab US or ustekinumab EU for AUC inf and C max were to be within the standard equivalence criteria of 0.8-1.25. To establish PK bridging of the 2 RPs, the same margin was used for the evaluation of ustekinumab US versus ustekinumab EU. The consistency of the PK similarity in the Japanese subgroup versus the overall population was determined based on whether the point estimate of the GMR in the Japanese subgroup and the non-Japanese subgroup fell within the prespecified margin of 0.8-1.25. The 90% CIs for the Japanese subgroup are presented as information only and were not used for the assessment of similarity.

Safety Analyses
Safety endpoints were summarized descriptively using the safety analysis set (n = 237), which consisted of subjects who were randomized and who received any amount of study drug. Assessments conducted included physical examination, vital sign measurement, recording of concomitant medications and adverse events, clinical laboratory testing, and 12-lead ECG. All reported adverse events were coded to the appropriate system organ class and preferred term according to Medical Dictionary for Regulatory Activities (Med-DRA) version 23.1; the severity of each adverse event was graded by the investigator per Common Terminology Criteria for Adverse Events (CTCAE) version 4.03 criteria.
The number and percentage of subjects reporting any adverse events, any grade ≥3 adverse events, any fatal adverse events, any serious adverse events, any adverse events leading to discontinuation from study, any events of interest (EOIs), and each individual EOI were summarized. Observed and change from baseline clinical laboratory values and vital sign data were summarized descriptively by treatment and timepoint.

Immunogenicity Analyses
Blood samples for ADA assessments were collected at predose and days 11, 35, and 112. All samples were tested in a validated binding, acid-dissociation ECL assay. The screening and confirmatory thresholds were statistically derived during validation. 14 The screening cut point of 1.11 was determined using a 95% prediction interval resulting in a 5% false reactive rate. Samples were analyzed in the screening assay and those having responses equal to or above the screening cut point were identified as potential positive (reactive). Reactive samples were analyzed in a confirmatory assay by spiking with drug and determining the percentage depletion of signal. The confirmatory cut point, defined as the ECL inhibition of 19.2%, was also statistically derived during validation using the 99% prediction interval resulting in a 1% false-positive rate. If equal to or above the confirmatory cut point, samples were confirmed positive and then analyzed in a titer assay in which positive samples were diluted until a negative response was obtained. The assay was sensitive to detect 4.24 ng/mL of ADA and detected 100 ng/mL of ADA in the presence of 50 μg/mL of drug.
Samples testing positive for binding antibodies were evaluated for neutralizing ADA in a validated ECLbased affinity capture elution assay. The screening cut point was statistically determined during validation to be 0.912 based on a 99% prediction interval resulting in a 1% false-positive rate. 14 Samples were subjected to acid treatment and then incubated with excess levels of biotinylated ABP 654. Any neutralizing ADA presented in the sample bound to the biotin-ABP 654 complex, which was then captured onto a streptavidincoated plate. With the addition of acid to the plate, the neutralizing ADA were eluted and transferred to a neutral solution of ruthenylated ABP 654. After incubation, samples were applied to a blocked Meso Scale Discovery streptavidin plate coated with biotinylated p40 monomer protein. In the presence of a neutralizing ADA, the ruthenylated ABP 654 was unable to bind to the p40-coated plate, resulting in a lower signal. The assay was sensitive to detect 79.2 ng/mL of ADA and detected 211 ng/mL of ADA in the presence of 16 μg/mL of drug.
The number and percentage of subjects who developed binding and neutralizing ADAs were tabulated by nominal visit and treatment using the safety analysis set. Transient antibody incidence was summarized through end of study (EOS) for subjects who had negative or no antibody results at baseline and was defined as the number of subjects with a positive postbaseline result with a negative result at the subject's last visit, tested through EOS and divided by the number of subjects with at least 1 post-baseline result through the EOS. Demographic and baseline characteristics were generally comparable across the groups (Table S1). Overall, 50.6% were female, 56.5% were White, and 80.6% were not Hispanic or Latino; 40 (16.9%) were first-or second-generation Japanese (13 in the ABP 654 treatment group, 13 in the ustekinumab US treatment group, and 14 in the ustekinumab EU treatment group). The mean (SD) age was 32.0 (7.31) years, with a range of 18-45 years.

Pharmacokinetics
The mean serum concentration-time profiles after a single SC injection of ABP 654, ustekinumab US, or ustekinumab EU are presented in Figure 1. The results show a high degree of similarity across the 3 treatment groups, both overall and for the Japanese subgroups. Exposure was similar across the 3 treatment groups. Peak concentrations were observed on average about 6-8 days following treatment. After achieving peak concentrations, concentrations declined in a monophasic manner, with a t 1/2 of 21.9-23.9 days. The estimated t 1/2 was consistent with the reported half-life for ustekinumab. 15 For the comparisons of ABP 654 to ustekinumab US, ABP 654 to ustekinumab EU, and ustekinumab US to ustekinumab EU, the point estimates and 90% CIs of the GMRs were fully contained within the prespecified margin of 0.8-1.25 for both the primary PK endpoints (AUC inf and C max ) and the secondary PK endpoint of AUC from zero to the last sampling period (AUC last ) ( Table 1). The GMR of AUC inf was 1.01 between ABP 654 and ustekinumab US, 0.94 between ABP 654 and ustekinumab EU, and 0.93 between ustekinumab US and ustekinumab EU. GMR of C max was 1.00 between ABP 654 and ustekinumab US, 0.95 between ABP 654 and ustekinumab EU, and 0.94 between ustekinumab US and ustekinumab EU. Additionally, for each pairwise comparison in both the Japanese and the binding ADA negative subgroups, the point estimates of the GMRs were contained within the prespecified margin of 0.8-1.25 for AUC inf , C max , and AUC last , thus demonstrating consistency of PK similarity in these subgroups versus the overall population (Table 1).
In general, the geometric means (GMs) of AUC inf and C max were similar following a single SC injection of ABP 654, ustekinumab US, or ustekinumab EU ( Table 2). The mean percentage of AUC inf due to extrapolation from the last quantifiable concentration observed to infinity (AUC%Extrap) was <6% for each treatment group, confirming the adequacy of the duration of PK sampling for the evaluation of AUC inf . In addition, other PK parameters, including C last and AUC last , were also comparable for ABP 654, ustekinumab US, and ustekinumab EU. Moreover, the PK parameters of ABP 654, ustekinumab US, and ustekinumab EU were similar for both the Japanese and the binding ADA negative subgroups ( Table 2). Taken together, these results support the conclusion that there are no clinically meaningful differences in the PK profiles of ABP 654 and ustekinumab RP.

Immunogenicity
As with all biologic agents, the risk for developing binding or neutralizing ADAs must be carefully assessed. A summary of immunogenicity results is presented in Table 3. All 237 subjects in the safety analysis set received a single 90 mg SC injection of ABP 654, ustekinumab US, or ustekinumab EU. All subjects had at least 1 on-study ADA result post-dose. One subject (1.3%) in each in the ABP 654 and ustekinumab EU treatment groups tested positive for pre-existing binding ADAs prior to dosing; no subjects in any treatment group tested positive for pre-existing neutralizing ADAs at or before baseline.
Of the subjects with a post-baseline result through the end of the study, 12 (15.4%) in the ABP 654 treatment group, 30 (38.0%) in the ustekinumab US treatment group, and 29 (36.3%) in the ustekinumab EU treatment group tested positive for binding ADAs. Of these, binding ADA positivity was transient (ie, negative results at the subject's last time point tested within the study period) for 8 (10.3%), 14 (17.7%), and 10 (12.5%) subjects, respectively. Of the subjects with a post-baseline result through the end of the study, 2 (2.6%) subjects in the ABP 654 group, 10 (12.7%) subjects in the ustekinumab US group, and 6 (7.5%) subjects in the ustekinumab EU group tested positive for neutralizing ADAs. The results were transient for 1 (1.3%) subject in the ustekinumab US group.
Of the 2 subjects who tested positive for pre-existing binding ADAs at or before baseline, the subject in the ABP 654 group tested positive for treatment boosted antibodies (ie, binding antibody positive at baseline with a ≥4× increase in magnitude post-baseline) by the  Overall, while the development of binding and neutralizing ADAs was numerically lower for the APB 654 treatment group as compared with ustekinumab (US and EU), these results support the conclusion that there is no meaningful impact on drug exposure.

Safety Evaluation
An overall summary of adverse events is presented in Table 4. Adverse events were reported by 22 (28.2%) subjects in the ABP 654 group, 18 (22.8%) subjects in the ustekinumab US group, and 29 (36.3%) subjects in the ustekinumab EU group. Most adverse events were CTCAE grade 1 or 2 in severity. Across all 3 treatment groups, headache was the most commonly reported Grade ≥3 adverse events were reported in 1 (1.3%) subject each in the ustekinumab US (grade 3 events of Crohn's disease and small intestinal obstruction) and ustekinumab EU (grade 3 event of appendicitis perforated) treatment groups. Each of the grade ≥3 adverse events were also identified as serious adverse events.
No fatal adverse events were reported for any subjects. In addition, there were no adverse events leading to study discontinuation. One (1.3%) subject in the ustekinumab US group experienced an EOI of reversible posterior leukoencephalopathy syndrome (preferred term of grade 1 somnolence, nonserious) and 1 (1.3%) subject in the ustekinumab EU group experienced an EOI of serious infections (including mycobacterial and Salmonella infections) (preferred term of grade 3 appendicitis perforated, serious). There were no serious systemic hypersensitivity reactions, facial palsy, pustular psoriasis, erythrodermic psoriasis, malignancy, cardiovascular events, serious depression including suicidality, or venous thromboembolism EOIs. Overall, the frequency, type, and severity of adverse events were comparable across treatment groups and were consistent with the safety profile of ustekinumab RP.

Discussion
The goal of this study was to determine whether there are clinically meaningful differences between ABP 654 and ustekinumab RP in terms of PK profile, safety, and immunogenicity under the conditions of use approved for ustekinumab RP and in accordance with the regulations for biosimilars development. This study was conducted in healthy subjects because they provide a homogenous population for sensitive PK comparisons.  Additionally, ustekinumab was well tolerated in healthy subjects at the studied dose level. 15 To detect potential PK differences between ABP 654, ustekinumab US, and ustekinumab EU, a single 90 mg SC injection was administered to study subjects. This dose provided sufficient exposure for an evaluation of PK between the 3 products and is consistent with the recommended dosage of ustekinumab RP for the treatment of adults with psoriasis. In this study, the primary PK endpoints, AUC inf and C max , were shown to be similar for ABP 654 and ustekinumab RP sourced from the US and EU. In addition, the 90% CIs for GMRs of C max , AUC inf , and AUC last were fully contained within the standard prespecified criteria, confirming a similar PK profile. Results from analyses of the key secondary PK endpoints also supported the conclusion of PK similarity. In addition, consistency of PK similarity was demonstrated in the Japanese subgroup versus the overall population. ABP 654 is being developed as a biosimilar to ustekinumab. Development of a biosimilar utilizes a stepwise approach designed to systematically assess the similarity of a candidate biosimilar in comparison to the RP. 7,8,16 Previously, extensive analytical characterization determined that ABP 654 has the same amino acid sequence and is similar to ustekinumab RP in structure, function, purity, and potency. 10 A fundamental next step of the biosimilar development process is an evaluation of the PK profile. 7,8,16 The general standard for PK studies to establish equivalence is the 90% CI for the GMR being within the prespecified acceptance range of 0.8-1.25 for overall exposure (eg, AUC).
An assessment of immunogenicity is an important component of building the totality of evidence to support biosimilarity. The presence of ADAs after treatment may decrease the efficacy of a biologic by neutralizing it or decreasing its half-life. 17 Although immunogenicity is not a concern for most biologics, some trigger ADAs which may impact efficacy and safety. While the development of binding and neutralizing ADAs was numerically lower for the ABP 654 treatment group compared with ustekinumab US and EU, these results are not expected to be associated with clinically meaningful differences in immunogenicity. Moreover, the presence of anti-ustekinumab neutralizing antibodies is not associated with any adverse safety events across the treatment groups.
We hypothesize that the lower numerical difference in ADA incidence detected against ABP 654 is likely due to the lack of both galactose-α-1,3-galactose (alpha gal) and N-glycolylneuraminic acid (NGNA) nonhuman residues, both of which are present in ustekinumab RP. It is well known that the use of different expression systems will result in differences in the glycan profile of the expressed protein. Ustekinumab is manufactured in the murine mouse myeloma cell line, Sp2/0. ABP 654 does not have these nonhuman glycans because it is expressed in a Chinese hamster ovary cell line, therefore it is not unexpected to observe a numerical difference in ADA incidence between ustekinumab Baseline was defined as the last non-missing assessment taken prior to the first dose of study investigational product. Percentages were calculated using the corresponding category count as the denominator. EOS, end of study; EU, European Union; US, United States. a Subjects considered on-study after signing informed consent. b Negative result at the subject's last time point tested within the study period.
RP and ABP 654. Several studies have demonstrated that the presence of nonhuman glycans on proteins can enhance their immunogenicity. 18 The mechanism behind this immunogenicity is related to the presence of preexisting or developing antiglycan antibodies that result in immune complex formation and enhanced uptake by antigen presenting cells through Fc receptor interactions. Similar results have been described for influenza hemagglutinin, bovine serum albumin, and ovalbumin containing alpha-gal. [19][20][21] The presence of nonhuman glycans on ustekinumab RP likely enhances immune reactivity to its protein backbone. Our results suggest that, despite a numerical difference in ADA incidence, there was no clinically meaningful impact on the PK parameters between the treatment groups.
The safety analysis showed that a single 90 mg SC injection of ABP 654, ustekinumab US, or ustekinumab EU was safe and well tolerated. Most adverse events were grade 1 or 2 in severity. No fatalities or adverse events leading to study discontinuation were reported. No new or unexpected safety signals were noted. Overall, the frequency, type, and severity of adverse events were similar between treatment groups with no clini-cally meaningful differences and were consistent with the safety profile of ustekinumab RP.
Finally, ustekinumab RPs sourced from the US and the EU were shown to be similar with respect to PK profiles, safety, and immunogenicity. Establishing similarity between RPs sourced from different regions is an important regulatory step in the biosimilar developmental process. In an effort to facilitate biosimilar development, the FDA and EMA allow the use of foreign-sourced RPs in comparative clinical studies, provided a "scientific bridge" between the local and the foreign-sourced RP is demonstrated in comprehensive analytical studies and a PK study. In this study, single doses of ABP 654, ustekinumab US, and ustekinumab EU administered to healthy subjects showed comparable PK profiles. Thus, the scientific bridge between the US and EU RPs was established and supports the use of a single RP in an ongoing comparative clinical study.
The effectiveness of ustekinumab RP to control disease progression of patients with plaque psoriasis, psoriatic arthritis, Crohn's disease, and ulcerative colitis is well established [22][23][24][25][26][27][28] ; however, these conditions are associated with significant treatment Only treatment-emergent adverse events were summarized. For each preferred term, subjects were included only once, even if they experienced multiple events in that preferred term. Safety endpoints were summarized descriptively using the safety analysis set (n = 237), which consisted of subjects who were randomized and who received any amount of study drug. COVID-19, coronavirus disease 2019; EU, European Union; US, United States. a Grade ≥3 adverse events were reported in 1 subject each in the ustekinumab US (Crohn's disease and small intestinal obstruction) and ustekinumab EU (appendicitis perforated) treatment groups. Both subjects that experienced the grade ≥3 adverse events were in the non-Japanese subgroup. Grade ≥3 adverse events were also identified as serious adverse events.
costs. [29][30][31][32] With healthcare costs continuing to rise, biosimilars have the potential to improve access to biologics by providing less expensive treatment options with safety and efficacy profiles comparable to their RPs. 5,6 This study demonstrated PK equivalency in healthy volunteers following a single 90 mg SC injection of ABP 654, ustekinumab US, or ustekinumab EU. The overall safety and immunogenicity profiles showed no clinically meaningful differences between ABP 654 and ustekinumab RP. In addition to the evidence from analytical and functional similarity assessments, these comparative clinical results add to the scientific justification that ABP 654 is highly similar to ustekinumab RP and support further clinical development of ABP 654. A clinical study (NCT04607980) comparing the efficacy, safety, and immunogenicity of ABP 654 and ustekinumab RP in subjects with moderate-to-severe plaque psoriasis is currently ongoing.