Immunogenicity and safety of a proposed pegfilgrastim biosimilar MSB11455 versus the reference pegfilgrastim Neulasta® in healthy subjects: A randomized, double‐blind trial

Abstract MSB11455 is a proposed biosimilar to the currently licensed reference pegfilgrastim (Neulasta®). This study was designed primarily to compare the immunogenicity of MSB11455 and Neulasta®. As secondary objectives, the safety and tolerability of MSB11455 and Neulasta® were also compared. Healthy adult subjects were randomized to either MSB11455 or Neulasta®, stratified by antipolyethylene glycol (PEG) antibody status at screening and study site. Subjects received a single subcutaneous dose of MSB11455 or Neulasta® (both 6 mg/0.6 mL) on day 1 of each of two study periods (same product in both periods), separated by a washout of 28‐35 days. Immunogenicity samples were taken predose and up to day 84 post–first dose. Noninferiority was confirmed if the upper limit of the exact one‐sided adjusted 95% confidence interval (CI) for the difference in antidrug antibody (ADA)‐positive rates was < 10%. Safety was assessed throughout the study. Overall, 336 subjects were randomized and treated (N = 168 in each group). Noninferiority of MSB11455 over Neulasta® was demonstrated for immunogenicity; the difference in confirmed treatment‐induced ADA‐positive rate between MSB11455 and Neulasta® was −0.6% (upper limit of the exact one‐sided adjusted 95% CI: 6.25%). ADAs were mostly directed against the PEG moiety of pegfilgrastim. No filgrastim‐specific neutralizing antibodies were detected in either treatment group. Safety and tolerability were as expected for pegfilgrastim, and comparable between treatments. This study supports and strengthens the available evidence for the biosimilarity of MSB11455 to Neulasta®.


| INTRODUC TI ON
Myelosuppressive chemotherapy is associated with the development of neutropenia, the consequences of which can be serious, with even minor infections potentially becoming life threatening.
Clinically, chemotherapy-induced neutropenia is defined as an absolute neutrophil count (ANC) of < 1.0x10 9 /L; at this ANC, the risk of infection begins to rise. 1 Febrile neutropenia is defined as a temperature of > 38.2ºC on two determinations with severe neutropenia (ANC < 0.5 × 10 9 /L), and this event indicates a high likelihood of localized or systemic infection. 1 Thus, persistent or severe neutropenia can be chemotherapy dose limiting, affecting the efficacy of these regimens. Treatment with a recombinant human granulocyte colony-stimulating factor (G-CSF), such as filgrastim, stimulates the proliferation, differentiation, and activation of neutrophils and reduces neutrophil maturation time. 2 Pegfilgrastim is a pegylated form of filgrastim that requires administration only once per chemotherapy cycle. 3 High-level evidence indicates that prophylactic use of either filgrastim or pegfilgrastim improves the likelihood of completing dose-dense and dose-intense chemotherapy and allows a broader range of patients to be treated. 4 Prophylactic use of G-CSF in patients receiving myelosuppressive chemotherapy also reduces the risk of early mortality (ie, during the chemotherapy period), including that related to infection, in addition to reducing the risk of febrile neutropenia. 5 The risk of febrile neutropenia, and its associated complications, is generally lower in patients receiving prophylaxis with pegfilgrastim than in those receiving filgrastim. 6 Prophylactic use of these agents is, therefore, recommended in patients receiving a chemotherapy regimen with a high risk of febrile neutropenia and in situations where dose-dense or dose-intense chemotherapy strategies have survival benefits or when reductions in chemotherapy dose intensity or density are known to be associated with a poor prognosis. 4,7 Neulasta ® (pegfilgrastim; Amgen, Inc), the US reference product, is indicated to decrease the incidence of infection, as manifested by febrile neutropenia, in patients with nonmyeloid malignancies receiving myelosuppressive anticancer drugs associated with a clinically significant incidence of febrile neutropenia. 3 MSB11455 is a proposed biosimilar to the currently licensed pegfilgrastim, Neulasta ® .
As with all therapeutic proteins, there is a risk of immunogenic reactions associated with administration of filgrastim or pegfilgrastim.

The United States Food and Drug Administration (FDA) and European
Medicines Agency (EMA) have both issued guidance on the development of biosimilars. 8,9 In this guidance, the clinical development program of a biosimilar must include a comparative clinical immunogenicity assessment. MSB11455 and Neulasta ® have analytical similarity of structural and functional attributes (data on file, Fresenius Kabi SwissBioSim GmbH, Switzerland), and have shown pharmacokinetic and pharmacodynamic equivalence in healthy volunteers. 10 Therefore, this biosimilar and the reference product were expected to demonstrate a similar immunogenicity and safety profile.
This study, therefore, compared the immunogenicity of MSB11455 and Neulasta ® as the primary objective. As secondary objectives, the study also compared the safety and tolerability of the two pegfilgrastim products.

| Study design
This was a double-blind, two-dose, randomized, parallel-group, group-sequential, immunogenicity study (NCT03251339) in healthy subjects. The study was conducted at two sites in New Zealand during the period August 2017 to May 2018. Subjects were resident at a study site from day −1 to day 3 of each treatment period.
Subjects were randomized in a double-blind manner to one of the two pegfilgrastim products, MSB11455 or Neulasta ® ( Figure S1), stratified by antipolyethylene glycol (PEG) antibody status at screening (as preexisting PEG antibodies may be a confounding factor for antidrug antibodies [ADAs]). The study consisted of two treatment periods separated by a 28-to 35-day washout period; during the first treatment period, subjects received either MSB11455 or Neulasta ® , and during the second treatment period they received the same product. Subjects received a single subcutaneous injection of pegfilgrastim 6 mg/0.6 mL in the morning of day 1 of each of the two treatment periods, for a total of two injections. The endof-treatment assessment visit occurred 28-35 days after study drug administration in Period 2, whereas the end-of-study assessment visit was 84 ± 3 days after study drug administration in Period 1 or an early termination assessment visit.
The study was conducted in accordance with ethical principles of the International Council for Harmonisation (ICH) guideline for Good Clinical Practice (GCP) and the Declaration of Helsinki, as well as with applicable local regulations. All subjects provided written informed consent before study entry.

| Study population
To be included in the study, participants (healthy men or healthy nonpregnant women) had to be aged ≥ 18 to ≤ 55 years with body mass index (BMI) ≥18.0 to ≤ 29.9 kg/m 2 . Eligible participants were in good health based on comprehensive medical history, physical examination, vital signs, and clinical laboratory tests, and had no known hypersensitivity to any component of either pegfilgrastim product. All subjects were required to comply with the contraception requirements specified in the clinical study protocol.

K E Y W O R D S
antidrug antibodies, immunogenicity, MSB11455, pegfilgrastim, safety, tolerability Potential participants were excluded based on usual exclusion criteria and the following: signs or symptoms of chronic obstructive pulmonary disease, smoking > 10 cigarettes/day, splenomegaly (spleen size > 13 cm in the craniocaudal dimension by ultrasound), prior exposure to any colony-stimulating or growth factor in the 3 months before randomization, and prior exposure to therapeutic monoclonal antibodies (MAbs) targeting the bone marrow or blood cells; exposure to MAbs not affecting bone marrow or blood cells was allowed if the MAbs were discontinued > 3 months or five half-lives (whichever was longer) before screening. Blood (≥500 mL) or plasma donation within 3 months, and stem cell or bone marrow donation within 12 months before screening were also not allowed.

| Assessments
At screening, samples to determine anti-PEG status were collected and evaluated using a kit-based enzyme-linked immunosorbent assay (ELISA). Samples for evaluating immunogenicity were taken predose on day 1 of Period 1, and postdose on day 13 of Period 1, day 28 of Period 1 (which also provided the predose sample for Period 2), days 13 and 28 of Period 2, and at the end-of-study assessment visit (84 ± 3 days after first drug administration in Period 1) or early termination assessment visit. Subjects who were positive for treatment-induced ADAs and who did not have two subsequent ADA samples at baseline levels by the end-of-study assessment visit continued into immunogenicity follow-up. During this follow-up, samples were taken every 5 weeks ± 7 days until two consecutive ADA samples had returned to baseline.
All immunogenicity samples were evaluated in a bioanalytical laboratory using a validated ELISA for ADA status. In this ELISA, labeled MSB11455 was employed as capture and detection reagent.
The ability of the assay to detect antibodies against MSB11455 and Neulasta ® was demonstrated during validation. All collected samples underwent a multitiered analytical approach to confirm the presence of ADA, collect semiquantitative information, and determine the binding specificity. First, a screening assay was used to detect potential ADA, followed by a confirmation assay, using MSB11455 as the competitive inhibitor, to reliably confirm the presence of ADA.
Next, a titration assay was performed to collect semiquantitative information about the confirmed ADA, and finally a specificity assay was used to distinguish between ADAs binding to PEG and those binding to filgrastim. Relative sensitivity of the assay was determined using affinity-purified polyclonal rabbit antipegfilgrastim and monoclonal mouse anti-PEG antibodies. The determined relative sensitivities were 10 and 19 ng/mL, respectively. The For ADA-positive samples, NAb status was determined using a validated cell-based assay that employed the cytokine-dependent cell line M-NFS-60, a murine lymphoblastoid cell line derived from a myelogenous leukemia cell line and adapted to become filgrastim dependent (therefore, neutralization of filgrastim or pegfilgrastim will inhibit cell proliferation). However, macrophage colony-stimulating factor (MCSF) neutralization also inhibits proliferation of this cell line. Thus, a tiered approach, as per regulatory guidance, was used to detect antibodies neutralizing pegfilgrastim and confirm their specificity toward filgrastim. In the first tier, ADA-positive samples were assessed for neutralizing activity toward pegfilgrastim. In the second tier, neutralizing activity toward MCSF was tested to confirm the specificity of pegfilgrastim inhibition. Finally, samples neutralizing pegfilgrastim but not inhibiting MCSF-induced cell proliferation were tested for their capacity to inhibit filgrastim. The determined DTL of this assay was 0.5 ng/mL MSB11455 and, based on the reasoning already presented for the ADA assay, the drug tolerance of the assay was considered adequate.
To avoid the confounding effect of pre-existing antibodies, a subject was considered to have a confirmed treatment-induced ADA-positive status if one of two situations occurred. Firstly, if the predose sample was ADA negative and at least one postdose sample was positive in the ADA confirmatory assay. Secondly, if the predose sample was positive and at least one postdose sample was positive in the ADA confirmatory assay with a titer-fold increase compared to predose above the minimum significant ratio (the minimum ratio between titers needed to declare a statistically significant difference), which was determined to be 3.6 during assay validation.
Safety and tolerability, including treatment-emergent adverse events (TEAEs), injection site reactions, physical examination findings, vital signs, results of routine laboratory testing and 12-lead electrocardiograms (ECGs), and concomitant medication data, were assessed from the time of signing informed consent and throughout the study. TEAEs of special interest were acute hypersensitivity (occurring within 48 hours after study drug administration); clinically significant increase in spleen size; ANC ≥ 75 × 10 9 /L (or white blood cell [WBC] count ≥ 90 × 10 9 /L); or signs and symptoms of hyperviscosity syndrome. Abdominal ultrasound was performed to assess spleen size before each dose and at the end-of-treatment assessment visit (or early termination assessment visit, as appropriate); further assessments of spleen size were allowed during the study if clinical signs or symptoms suggestive of splenic enlargement were noted.

| Endpoints
The primary study endpoints were confirmed treatment-induced ADA-positive status to pegfilgrastim from predose on day 1 of Period 1 until the end-of-study assessment visit, and confirmed neutralizing antibody (NAb) status to pegfilgrastim from predose on day 1 of Period 1 until the end-of-study assessment visit.
Secondary endpoints included safety and further immunogenicity endpoints.
Safety endpoints were occurrence of TEAEs, serious AEs (SAEs), AEs of special interest, and abnormal laboratory values and vital signs in subjects from the first administered dose until the end-of-study assessment visit. Additional immunogenicity endpoints were ADA status by time point, ADA titer over time, and NAb status by time point.

| Sample size and analysis population
A group-sequential design with an unblinded interim analysis was implemented. A maximum of 404 subjects were to be randomized 1-to-1 to each treatment arm, unless the study was stopped at the interim analysis, which was planned after 336 subjects had been randomized. This sample size was calculated to have 90% power to declare MSB11455 no worse than Neulasta ® , either at the interim or final analysis if the true confirmed ADA-positive rate by the end-of-study assessment visit in the Neulasta ® arm was 12%, based on findings of a previous study (NCT02205320), and assuming a true difference in rates of 0%. Other key assumptions for these sample size calculations included a noninferiority margin of 10% and a type I error rate of ≤5% for a one-sided test for noninferiority; this noninferiority margin corresponded to a maximum true confirmed ADA-positive rate with MSB11455 of 22% under the assumption that the true confirmed ADA-positive rate with Neulasta ® was 12% by the end-of-study assessment visit.
The unblinded interim analysis for futility (nonbinding) and noninferiority took place when exactly 336 subjects were randomized and had completed the end-of-study assessment visit/early termination visit. After review of the interim analysis data, an Independent Data Monitoring Committee recommended that the study be stopped for noninferiority, and the interim analysis became the final analysis.
The primary analysis of treatment-induced ADA-positive status up to end-of-study assessment visit/early termination visit was performed on the intent-to-treat (ITT) population, defined as all randomized patients. Other immunogenicity outcomes and safety were analyzed in all subjects who received at least one administration of study drug (MSB11455 or Neulasta ® ).

| Main analyses
Baseline subject demographics and characteristics were summarized descriptively. For the primary analysis of treatment-induced ADApositive status up to end-of-study assessment visit/early termination visit, the difference in confirmed treatment-induced ADA-positive rates between MSB11455 and Neulasta ® was estimated along with the corresponding exact one-sided adjusted 95% confidence interval [CI]. 11,12 Given that the primary endpoint was binary and the associated rates could be less than 10%, the power and type I error rate calculations of the study design were assessed using exact calculations.
Secondary immunogenicity endpoints were reported as the proportion of subjects with confirmed treatment-induced ADA-positive status at each visit, with corresponding two-sided 95% CI (using Clopper-Pearson method); the denominator was the number of subjects with samples available at that visit. Summary statistics (median values) were used for ADA titers by visit. The proportion of subjects NAb positive to filgrastim for each visit was also calculated, the denominator being the number of subjects with samples analyzed (ie, subjects who were ADA positive). TEAEs were summarized by treatment and overall, as the number and proportion of subjects affected.

| Sensitivity analyses
Sensitivity analyses for the primary immunogenicity endpoint were conducted in the per protocol population (all subjects who received both doses of pegfilgrastim, had at least one ADA status available after day 1 of Period 2, and had no clinically important protocol deviations) and using a stratified analysis allowing for anti-PEG antibody status at screening.

| Patients
A total of 336 subjects were randomized and treated (Figure 1).
Baseline subject demographics and characteristics were well balanced between treatment arms (Table 1); overall, 4.8% of included subjects were anti-PEG antibody positive at screening.

| Immunogenicity
Noninferiority of MSB11455 over Neulasta ® was demonstrated for immunogenicity (upper limit of the exact one-sided adjusted 95% CI < 10%). The confirmed treatment-induced ADA-positive rate was similar with both pegfilgrastim products (Figure 2), with a difference in rate between MSB11455 and Neulasta ® of −0.6% (upper limit of the exact one-sided adjusted 95% CI: 6.25). The overall postdose ADApositive rate, not constrained to a treatment-induced ADA-positive status, was also comparable between treatments. Over the entire study period, 15 subjects (8.9%) who received MSB11455 and 18 (10.7%) who received Neulasta ® had an ADA-positive status at any time.
Treatment-induced ADA-positive rates over time were also comparable between treatments (Figure 3). The highest positivity rate was observed on day 13 of Period 1, and rates decreased thereafter.
Although there were numerically more ADA-positive subjects in the MSB11455 group than in the Neulasta ® group in period 2, the overall positivity rate in period 2 was low (<5.0% in each group). Overall, there were no relevant differences in median ADA titers over time between the two treatment groups among subjects with an ADApositive sample (Figure 3). At most visits, the ratio of median titers in each treatment group was below the minimum significant ratio (3.6), and hence not significantly different. On day 13 of Period 2, the difference for MSB1145 vs Neulasta ® was above the minimum significant ratio, but that a single subject was ADA positive in the Neulasta ® group prevents any conclusions regarding the clinical relevance of this finding. ADAs were mostly directed against the PEG moiety of pegfilgrastim, with no relevant differences in specificity observed between the two treatment groups ( Table 2)

TA B L E 1 Subject demographics and characteristics at baseline by treatment group and overall
baseline by the end-of-study assessment visit, and therefore underwent immunogenicity follow-up, consisting of a scheduled visit every 5 weeks (four subjects in the MSB11455 group and three in the Neulasta ® group). All six evaluable subjects (one subject treated with Neulasta ® was lost to follow-up) had two consecutive ADA samples that returned to baseline (ie, subject completed follow-up) within a maximum of five visits. In the MSB11455 group, one subject required a maximum of five follow-up visits, two required three visits, and one required two visits. In the Neulasta ® group, the two evaluable subjects each required one follow-up visit.

F I G U R E 2
Proportion of healthy subjects with treatmentinduced ADA-positive status at any time after the first of two single doses of MSB11455 or Neulasta ® administered 28-35 days apart until the end-of-study assessment visit/early termination visit   Abbreviations: ADA, antidrug antibody; mPEG, methoxy-polyethylene glycol. a All ADAs associated with a low titer

| Sensitivity analyses
Sensitivity analyses conducted either in the per protocol population or with stratification for anti-PEG antibody status at screening confirmed the primary analysis results, showing the noninferiority of MSB11455 over Neulasta ® for immunogenicity (data not shown).

| Safety and tolerability
Safety and tolerability were similar between treatments (Table 3).
There were no deaths and the numbers of reported SAEs were small with no notable differences between treatments. TEAEs leading to discontinuation of study drug occurred with similar frequency in both treatment groups (Table 3) and were most commonly an increased WBC count to ≥ 50.0 × 10 9 /L (9.5% of MSB11455 and 11.9% of Neulasta ® recipients), which was initially a protocol prespecified treatment withdrawal criterion. TEAEs were generally mild to moderately severe, self-limiting, and resolved without sequelae. No TEAEs were of National Cancer Institute Common Terminology Criteria for Adverse Events (CTCAE) grade ≥ 4. The most common TEAEs were musculoskeletal and connective tissue disorders and nervous system disorders ( Table 3).
AEs of special interest were reported in similar proportions of subjects in the MSB11455 and Neulasta ® groups (Table 3)

| D ISCUSS I ON
As therapeutic proteins, both MSB11455 and Neulasta ® were expected to induce ADAs. 3 Although, existing data suggest a low immunogenic potential for pegfilgrastim, 3 the ability to stimulate antibody formation can differ between structurally closely related molecules, with even minor changes affecting immunogenicity. 13,14 ADAs have the potential to adversely interfere with the therapeutic action of, or reduce or prolong exposure to, the therapeutic agent and/or predispose to adverse reactions. 13,14 In the case of filgrastim/pegfilgrastim, ADAs with filgrastim neutralizing capacity are particularly critical due to potential cross-reactivity with the endogenous counterpart.
Demonstration of immunogenic similarity of a biosimilar to the reference product in a head-to-head comparison is, therefore, a critical parameter for defining the safety of the biosimilar. 13 The design of this study was appropriate for comparing the immunogenicity and safety profiles of MSB11455 and Neulasta ® , with both the study design and population in line with regulatory guidance.
The parallel-group design of the study ensured that immunogenicity assessments were related to only one treatment. This study was conducted in healthy subjects because they are the most homogenous population for a sensitive assessment of immunogenicity; their immune system is not affected by chemotherapy (thereby improving sensitivity to detect differences in immunogenicity), and animal models have low predictive value with respect to immunogenicity in humans. 13 The Study results also showed that the safety and tolerability of MSB11455 and Neulasta ® were similar, and were as expected for administration of pegfilgrastim. 3 Bone pain and headache were among the most frequently reported TEAEs for both treatments, which is expected with pegfilgrastim treatment. 3 Pegfilgrastim is also associated with musculoskeletal pain because of bone marrow remodeling and increased precursor turnover. 15 Splenic rupture has been reported following administration of filgrastim 16 or pegfilgrastim 3 ; therefore, the spleen was monitored throughout the study. Two subjects experienced a mild-to-moderate increase in spleen size; of these only one was deemed by the investigator as clinically significant on splenic ultrasound examination. In both instances, this spontaneously resolved. A relatively large number of subjects reported an increase in WBC count to ≥ 50.0 x 10 9 /L as an AE of special interest, which resulted in treatment discontinuations in these subjects (as per corresponding withdrawal criterion prespecified in the initial version of the clinical study protocol). However, all these events were self-limiting, returning spontaneously to normal levels within 10 days. The initial WBC count 'safety threshold' defined in the initial version of the clinical study protocol was at a level too low for healthy volunteers who had intact bone marrow. The study protocol was, therefore, amended to reflect the pronounced increment in WBC count expected in healthy subjects following stimulation by pegfilgrastim (to a WBC count ≥ 90.0 x 10 9 /L) and no further events of this nature were reported after this amendment.

| CON CLUS ION
MSB11455 and Neulasta ® showed comparable immunogenicity, safety, and tolerability in healthy subjects, and MSB11455 was noninferior to Neulasta ® for immunogenicity, in terms of treatmentinduced ADA-positive rates. This study supports and strengthens available evidence for the biosimilarity of MSB11455 to Neulasta ® .