This study was carried out in the phase I Clinical Research Center of the First Hospital of Jilin University for 1°year from 2018–11–21 to 2019–12–16 (Chinese Clinical Trial Registry, Registration No. CTR20181,658; http://www.chinadrugtrials.org.cn/index.html). The protocol for this study was approved by the ethics committee of our hospital. Our study was conducted in agreement with the guidelines of the Declaration of Helsinki and the International Conference on Harmonization Good Clinical Practice Guidelines and fulfilled local regulatory requirements. The number of the protocol is QX001SA-01. Written informed consent was obtained from all participants in this study.

This study was designed as a randomized, double-blind, single-dose, two-arm, parallel clinical trial with healthy Chinese men. The purpose of the study was to evaluate the bioequivalence of the proposed biosimilar (QX001S, 45 mg) with its reference drug (ustekinumab, Starino®, a trademark in Europe). The bioequivalence study/investigation included the evaluation of the tolerance, immunogenicity, and PK of the biosimilar and reference drugs. Overall, 178 eligible participants were randomly assigned and stratified into two treatment groups based on body weight (50 to <67.5 kg and ≥67.5 to ≤85 kg) at a 1:1 ratio. The participants in each group received a single subcutaneous injection of 45 mg QX001S or ustekinumab. Individuals in each of the prespecified weight groups were equally assigned to the two treatment groups through randomization (Figure 1).

The inclusion criteria for participants were as follows: 1) healthy men aged between 18 and 50°years; 2) body mass index of 18.0–28.0 kg/m²; 3) total body weight between 50 and 85 kg; and 4) normal blood and urine routine as well as hepatic and renal function test results during enrollment or slight abnormality without clinical significance. Only participants (including their partners) with no fertility plans for the next 6°months who voluntarily used effective contraceptive measures were included.

The main exclusion criteria were as follows: 1) allergic constitution (a variety of drugs and food allergies) or allergic to the study drugs; 2) blood donation or substantial blood loss (>450 ml) within 3°months prior to the study, or plans for blood donation or surgery during the study period; 3) abnormal electrocardiograph (ECG) with clinical significance; 4) abnormal clinical laboratory examination results with clinical significance; 5) a history of invasive systemic fungal infection or other opportunistic infections within 2°months prior to screening; 6) systemic or local infections, such as risk of sepsis and/or known active inflammation, within 2°months prior to or during the screening showing C-reactive protein (CRP) levels above 1.5× the normal upper limit; 7) positive results for either hepatitis (including hepatitis B and C), AIDS, or syphilis; 8) use of any biological products within 3°months prior to the screening; 9) use of any immunoglobulins in the previous year; 10) receiving live (attenuated) vaccine within 12°weeks prior to the screening or during the period of this study; and 11) planning to receive any biological drugs within 3°months or monoclonal antibody drug within 9°months after administration of the study drug.

Participants were admitted to the research center the day before the injection of the test drugs. For safety assessment, each participant was asked to stay in the research center for at least 5°days after receiving the investigational product (IP). All participants were also followed up for the assessment of PK, immunogenicity, and safety for 113°days after administration. For safety reasons, the sentinel staggered dosing method was adopted in this study. This dosing method is often adopted during early phases of single ascending dose (SAD) trials and involves sequential IP administration in staggered cohorts with an appropriate period of observation between dosing individual participants. In our study, the participants were initially administered the IP in staggered cohorts, with the first and second cohorts comprising two and four participants, respectively. These participants were asked to stay at the research center for at least 9°days after the initial dose of IP for safety evaluation. Based on the sentinel safety results, the principal investigator determined whether subsequent participants would be monitored in either sentinel mode or routine follow-up mode.

All participants were administered (by subcutaneous injection) a single dose (45 mg) of QX001S or ustekinumab produced by Qyuns Therapeutic Co. Ltd (batch number: F20180101) and Starino®, Janssen-Cilag International N.V. (batch number: IASIDMS), respectively. Screening was performed 14°days before the drug dosing date. The participants were asked to fast for at least 8 h before the administration of the drugs.

For PK evaluation, blood samples were collected from the participants at 1 h before the administration (predose) and at 4, 12, 24 h (day 2), 48 h (day 3), 96 h (day 5), 120 h (day 6), 144 h (day 7), 168 h (day 8), 192 h (day 9), 216 h (day 10), 240 h (day 11), 288 h (day 13), 336 h (day 15), 504 h (day 22), 672 h (day 29), 1,008 h (day 43), 1,344 h (day 57), 1,680 h (day 71), 2016 h (day 85), 2,352 h (day 99), and 2,688 h (day 113) after administration.

The serum levels of ustekinumab and QX001S were determined using enzyme-linked immunosorbent assay (ELISA) provided by the Junke Zhengyuan (Beijing) Pharmaceutical Research Co., Ltd. (Beijing, China) (Supplementary material).

PK parameters were calculated using a non-compartmental analysis model, for which the concentration-time data were the time to peak (T _max ), maximum observable serum concentration (C _max ), clearance (CL), volume of distribution (Vz), half-life (t _1/2), and area under the curve (AUC) from zero to the final quantifiable concentration (AUC_0–t ) and to infinity (AUC_0–∞). The actual collection time for each sample was used for the analysis of PK parameters. The PK parameters were analyzed using Phoenix WinNonLin®, an internally validated software system (v8.0, Pharsight Corporation, Certara, L.P. Princeton, New Jersey, United States).

LambdaZ was the first-order rate constant associated with the terminal (log-linear) portion of the curve, estimated from the linear regression of time vs. log concentration using the best fit method. The clearance (CL) = dose/AUC0-∞; volume of distribution (Vz) = dose/(AUC0-∞* lambdaZ); and half-life = ln (2)/lambdaZ. AUC0–∞ was calculated as follows: AUC from the time of dosing extrapolated to infinity, based on the last observed concentration (Clast). AUC0-∞ = AUC0-t + Clast/lambdaZ.

The appropriate sample size for each group was estimated using the following parameters: 1) geometric mean ratio (GMR of C_max and AUC for the test drug vs. the reference drug), 2) power (1−β), 3) significance level (two-sided α = 5%), and 4) inter-subject variability (inter-CV) of the C_max and AUC. The appropriate sample size (initial: 142; inter-CV of C _max for ustekinumab: 35.8%, data from (Centocor Ortho Biotech Inc, 2020) was determined with a computer program (NQuery 8.3.0.0, Boston, United States) according to recent FDA guidelines. The GMR of the C _max and AUC for the test drug against the reference drug was set at 95% to obtain 90% power (1−β) at a significance level (two-sided α) of 5% (Julious, 2004). The inter-CV was represented by the coefficient of variation (CV). The final sample size for this study was 178 (89 participants per group) based on the above calculations and the calculated drop-out rate was 20%.

The variance of PK parameters, such as Cmax, AUC_0–t, and AUC_0–∞ was analyzed using the least squares method, but the data required logarithmic transformation before variance analysis. If the 90% confidence intervals (CI) for the ratio (test/reference) of the mean AUC and Cmax (log-transformed) fall within the range of 80–125%, a biosimilar is considered to have pharmacokinetic biosimilarity to its reference product. The statistical analysis methods used in this study were the chi-squared test for categorical variables, t-test for normally distributed data, and Wilcoxon rank-sum test (Mann–Whitney test) for data with an unknown distribution. SAS 9.4 software (SAS Institute Inc. Cary, NC, United States) was used for the statistical analysis of data in this study. A p-value of <0.05 was considered statistically significant.

A total of 612 healthy men were screened and 178 participants were finally recruited for the study. However, only 177 received the assigned study drug because one participant in the ustekinumab group withdrew from the study due to family issues (Fig. 1). Finally, 177 participants were included in the analysis of safety, PK, BE, and immunogenicity (ADA) (Figure 1). The demographic and baseline characteristics of the participants in the two treatment groups were similar, with p > 0.05 (Table 1).

The mean serum concentration-time curves after a single dose of ustekinumab and its biosimilar in healthy Chinese men (n = 177) showed a slowly increasing phase after administration, followed by a slowly declining phase at higher concentrations and a slightly faster elimination phase at lower concentrations (Figures 2, 3).

PK analysis of ustekinumab and its biosimilar showed slow clearance, long t _1/2, and small Vz values in the non-compartmental analysis model. The median T _max values were similar between the two groups at 168 h after administration. The estimated geometric mean t _1/2 values were comparable between the two treatment groups, ranging from 577.9 to 580.6 h. The clearance (CL) and Vz values in the two groups were also comparable. Other PK parameters such as the mean C _max , AUC_0–t , and AUC_0–∞ estimates in the two groups were also similar (p > 0.05, Table 2, Supplementary table S1, Figures 2, 3).

As shown in Table 2, the PK parameters were comparable between the two treatment groups; in particular, the key characteristics of PK i.e. the 90% CIs of the ratios for C _max , AUC_0–t , and AUC_0–∞ were 100.90–118.68%, 98.71–115.26%, and 98.49–115.81%, respectively, for QX001S vs. Ustekinumab, which were within the predefined bioequivalence limits, i.e. ranging from 80.00 to 125.00%, indicating pharmacokinetic biosimilarity between those two drugs. Since a larger inter-CV indicates a broader 90% CI, the sample size was re-estimated using the bioequivalence analysis data (GMR and inter-CV). Thus, the re-estimated value was closer to the number of recruited participants.

In this study, before administration of the IPs, ADA was detected in several participants in the QX001S group (two participants) and ustekinumab group (three participants). In total (before and after administration), 17 participants in the QX001S group (19.1%, 17/89) and 36 in the ustekinumab group (40.9%. 36/88) were positive for ADA. Evaluations showed that the ADA-positive rates increased over time, especially by days 57 (1,344 h) and 113 (2,688 h). It is notable that in both the QX001S and ustekinumab groups, the positive rates for the ADA test increased as the drug concentration declined. However, the ADA-positive rates were lower in the QX001S group than in the ustekinumab group (Table 3).

The mean serum concentration-time curves for the ADA-positive and total participants were similar in the two treatment groups (Figure 2). The analysis of ADA-related sensitivity showed the results were consistent with the bioequivalence analysis data as mentioned above, supporting the conclusion regarding the pharmacokinetic biosimilarity between two treatment groups. After removing the data for ADA (+) participants before dosing (five participants) and the data for all ADA (+) participants in this study, the re-calculated 90% CIs for C _max , AUC_0–t, and AUC_0–∞ in the two groups were within the pharmacokinetic biosimilarity limits of 80.00–125.00% (Table 2).

In this study, there were no serious AEs (SAEs), discontinuations, or deaths due to AEs. Overall, 75 (42.4%) participants experienced adverse reactions, specifically 38 participants (42.7%) in the QX001S group and 37 (42.0%) in the ustekinumab group (Table 4). The overall incidence rates for adverse reactions in the two groups were similar (p > 0.05). However, in the QX001S group, an increase in the alanine aminotransferase level was found in one participant with a grade III adverse reaction (n = 1 [1.1%]). Other adverse reactions were classified as Grades I and II according to the common terminology criteria for AEs (CTCAE). The most common adverse reactions (>4%) were upper respiratory infection and an increase in the parameters for the clinical lab tests, such as the alanine aminotransferase level, leukocyte count, neutrophil count, aspartate aminotransferase level, and triglyceride level (Table 4).

One participant in the ustekinumab group experienced injection site swelling after administration on day 1, but the swelling disappeared by the next day without treatment. The rest of the participants had no injection site reactions. To treat the AEs observed, 15 participants (seven in the QX001S group and eight in the ustekinumab group) transiently (one or two times) received medications such as acetaminophen tablets, cefixime, roxithromycin, and oseltamivir phosphate capsules, mostly for the control of upper respiratory infection. Taken together, no relationship between ADA development and adverse reactions was found in this study. All the adverse reactions in this study were reported to the Institutional Review Board of The First Hospital of Jilin University.