The immunogenicity of biosimilar infliximab: can we extrapolate the data across indications?

Abstract

Biopharmaceuticals or ‘biologics’ have revolutionized the treatment of many diseases. However, some patients generate an immune response to such drugs, potentially limiting clinical efficacy and safety. Infliximab (Remicade^®) is a monoclonal antibody used to treat several immune-mediated inflammatory disorders. A biosimilar of infliximab, CT-P13 (Remsima^®, Inflectra^®), has recently been approved in Europe for all indications in which infliximab is approved. Approval of CT-P13 was based in part on extrapolation of clinical trial data from two indications (rheumatoid arthritis and ankylosing spondylitis) to all other indications, including inflammatory bowel disease. This review discusses the validity of extrapolating immunogenicity data across indications – a process adopted by the EMA as part of their biosimilar approval process – with a focus on CT-P13.

Keywords:

• biologics
• biosimilars
• CT-P13
• extrapolation
• immunogenicity
• inflammatory bowel disease
• infliximab

Biopharmaceuticals or ‘biologics’ have revolutionized the treatment of many diseases. Over the past decade or so, several biosimilars have been introduced into medicine with the goal of reducing treatment costs and increasing accessibility to therapy for patients. More recently, the first monoclonal antibody (mAb) biosimilar, CT-P13 (Remsima^®, Inflectra^®), an infliximab biosimilar, was approved for use in Europe and elsewhere. All biologic drugs, including biosimilars, have the potential for immunogenicity, that is, they can elicit an immune response to the drug in the recipient. Immunogenicity primarily manifests in the production of anti-drug antibodies (ADAs), which may lead to reduced clinical efficacy and increased risk of adverse events (AEs). However, the optimal methodology for validating equivalent immunogenicity of biosimilars with their reference medicinal products (RMPs) has been debated. The present review will explore the data on immunogenicity assessment of biosimilars, with particular focus on CT-P13, and will consider the validity of extrapolating immunogenicity data across different clinical indications.

Regulatory guidelines for extrapolation of biosimilars

A biosimilar can be defined as a ‘biotherapeutic product that is similar in terms of quality, safety, and efficacy to an already licensed reference biotherapeutic product’ [1]. According to various regulatory guidelines, it may be possible to extrapolate the demonstrated efficacy and safety data of a biosimilar from one indication to other indications – provided that certain criteria are met [2–5]. Although caution should be exercised, the rationale underpinning extrapolation is that the biosimilars will behave in a similar manner to the RMP in all clinical scenarios [6]. EMA requirements for possible extrapolation of a biosimilar from one indication to another are as follows [2]:

• A convincing demonstration of biosimilarity based on totality of the evidence obtained through a comprehensive comparability exercise.

• Extensive characterization and comparison of physicochemical properties and biological activity between the biosimilar and the RMP.

• Demonstration of comparable clinical efficacy and safety in a sensitive patient population.

• Similarity of mechanism of action across indications.

CT-P13 has the same amino acid sequence as the infliximab RMP (Remicade^®), and both drugs are derived from the same type of cell line (Sp2/0-AG14), although the exact cell line used is specific to the two companies producing either drug [7,8]. CT-P13 was perceived by the EMA to meet the aforementioned biosimilarity criteria, a decision partly based on data from clinical trials performed in patients with rheumatoid arthritis (RA) [9] and ankylosing spondylitis (AS) [10]. Thus, CT-P13 is now approved in Europe across all indications for which the infliximab RMP is licensed, including inflammatory bowel disease (IBD), that is, Crohn’s disease (CD) and ulcerative colitis (UC).

Immunogenicity: definition & predisposing factors

Although biologics are generally considered to be well-tolerated and non-toxic, repeated administration can induce immunogenicity [11], a phenomenon defined as the potential for an antigen to induce an immune response [12]. With biologics, immunogenicity primarily manifests as the formation of ADAs, a response that has been documented in numerous studies with different drugs across a range of indications [11–15]. The mechanisms by which ADAs form in response to biologics are described in detail elsewhere [11,13].

That ADAs could pose a serious health risk was demonstrated when neutralizing antibodies to thrombopoietin were shown to induce severe thrombocytopenia in some patients treated with pegylated recombinant human megakaryocyte growth and development factor (PEG-rHuMGDF) [16]. Subsequently, it has become clear that ADAs may be associated with several important clinical consequences, including alterations in drug pharmacokinetics and bioavailability [17–21], reductions in drug efficacy [14,17,18,22–24], cross-reaction with endogenous proteins and inhibition of the latter’s physiological function [11,22] and allergic drug reactions and other AEs [18,22,23]. Importantly, however, not all biologics elicit the same degree of immunogenicity, and the extent of immunogenicity may even be different among individual patients receiving the same biologic. Indeed, the development of immunogenicity depends on a number of factors that are discussed in detail in several review articles [11,13,22]. These include:

• Treatment-dependent factors such as dose, route of administration, duration of treatment and temporal pattern (e.g. episodic versus regularly scheduled treatment).

• Product-dependent factors, including modifications introduced into the biologic. For example, variations in the amino acid sequence of a biologic mimicking an endogenous protein may make the protein sufficiently unlike ‘self’ to induce an immune response. In addition, certain types of glycosylation have been shown to be more immunogenic [25], while other glycosylation patterns as well as other modifications such as PEGylation can reduce immunogenicity by interfering with antibody binding. The degree to which the biologic is humanized may also be important, although even fully human mAb biologics can be immunogenic [26].

• Process-related factors, including manufacturing, packaging, storage, handling procedures, and the presence of any impurities/contaminants. Aggregate content is also important and should be kept to a minimum. Why and how aggregates induce an immunogenic response are complex topics of ongoing debate [27]. Highlighting the importance of packaging, one of several factors increasing the incidence of pure red cell aplasia in patients treated with recombinant human erythropoietin, was thought to be an increase in immunogenicity to the drug caused by an interaction between the stabilizer polysorbate 80 and leachates from uncoated rubber in the stoppers of prefilled syringes [28].

• Patient-related factors including age, concomitant medications (such as immunomodulators), disease state (e.g. whether the patient is immunosuppressed or not), genetic predisposition, including the presence of susceptible human leukocyte antigen (HLA) alleles [29], and general allelic variation of the target protein. Such factors are likely to account for some of the differences in ADA formation observed between patients with the same disease participating in the same clinical study.

Infliximab is a chimeric IgG1 human (75%)/murine (25%) mAb against TNF that is approved in many regions around the world for use in patients with RA, AS, psoriatic arthritis, plaque psoriasis, CD and UC. Studies have found that the murine component in the Fab fragment of infliximab is implicated in the immunogenicity of the drug, inducing the formation of human anti-chimeric antibodies [14,30]. Multiple studies have shown that the formation of ADAs during infliximab treatment is associated with a loss of clinical response [14,24,31–38], and an increased rate of infusion reactions including potentially serious events such as urticaria, chills, fever, nausea, headache, dyspnea, and even the very rare but potentially fatal anaphylaxis [14,31,33,36,39].

Comparison of the immunogenicity of CT-P13 with RMP

A two-tiered approach is desirable when comparing the immunogenicity of a mAb RMP to that of its biosimilar. First, in vitro assays are used to show similar or cross-immunogenicity equivalence of the RMP and the biosimilar. For CT-P13, an antibody conformational assay was performed, revealing its ‘fingerprint’. This technique uses an array of antibodies raised against a panel of peptides derived from infliximab. When these antibodies were tested for their recognition of 14 batches of RMP versus CT-P13, similar profiles of exposed epitopes on the two molecules was observed, indicating CT-P13 and its RMP have very similar higher-order immuno-dominant structures [7]. In an additional study, ADAs from the sera of 86 infliximab-treated patients with IBD were tested for their cross-recognition of CT-P13. The results showed that virtually all anti-RMP antibodies in these IBD patients similarly recognized CT-P13 [40]. In the second tier of the comparison, the most valid methodology involves in vivo testing of the formation of ADAs against the RMP and biosimilar in patients receiving these agents in a head-to-head trial. Well-validated assays to detect ADAs should be used, and should be the same for the biosimilar and RMP. Assays used to detect antibodies against mAb biologics, such as conventional ELISAs and radio-immunoprecipitation assays, involve the use of anti-immunoglobulin reagents to detect antibodies bound to the drug, but they can also bind to the drug itself. Therefore, bridging-format ELISA, electrochemiluminescence-based detection of ADAs or surface plasmon resonance-based drug-tolerant assays are more appropriate [41].

In vivo immunogenicity of CT-P13 and its RMP was assessed in two clinical trials, using an electrochemiluminescent immunoassay method recommended by both the US FDA and the EMA. Using this assay, two randomized, controlled clinical trials in patients with RA (PLANETRA) and AS (PLANETAS) showed no differences in the incidence of ADAs between RMP and CT-P13. Of note, the two trials used different treatment doses, with RMP and CT-P13 given as a monotherapy to patients with AS or in combination with methotrexate (MTX) to patients with RA. In the Phase III PLANETRA study in RA, following dosing with CT-P13 or RMP at 3 mg/kg in combination with MTX, ADAs were detected in 25.4 and 25.8% of patients in the CT-P13 and RMP groups, respectively, at week 14, in 48.4 and 48.2% of patients at week 30 and in 52.3 and 49.5% of patients at week 54 (Figure 1) [9,42]. The rate of formation of ADAs was comparable between CT-P13 and RMP irrespective of whether antibodies against CT-P13 (CT-P13-tag) or RMP (RMP-tag) were measured [9]. In the Phase I PLANETAS study in AS (monotherapy dose of 5 mg/kg), in the CT-P13 and RMP groups, respectively, ADAs were detected in 9.1 and 11.0% of AS patients at week 14; in 27.4 and 22.5% of patients at week 30; and in 22.9 and 26.7% of patients at week 54 (Figure 2) [10,43]. Also, a high concordance was observed between the CT-P13-tag and RMP-tag tests. There was no difference in the rate of AEs between CT-P13 and the RMP in either study. Taken together, these data strongly indicate that there is no difference in the rate of antibody formation between CT-P13 and RMP in these two conditions. Data from the extension phases of PLANETRA and PLANETAS, in which all patients were treated with CT-P13, have also recently been presented and showed that immunogenicity profiles were similar in a group of patients who switched from RMP to CT-P13 at the start of the extension phase (week 54) and in those that received CT-P13 throughout the main study and extension. In PLANETRA, the proportion of patients with ADAs at the end of the extension (week 102) was 49.6% in the group that switched from RMP to CT-P13 compared with 46.4% in those maintained on CT-P13 throughout; in PLANETAS, these figures were 30.7 and 25.0%, respectively [44,45].

Figure 1. Proportion of patients with anti-drug antibodies to CT-P13 and infliximab reference medicinal product (RMP) in the PLANETRA study in patients with rheumatoid arthritis [9,42].

Figure 2. Proportion of patients with anti-drug antibodies to CT-P13 and infliximab reference medicinal product (RMP) in the PLANETAS study in patients with ankylosing spondylitis [10,43].

Rationale for extrapolation of immunogenicity data with CT-P13

While CT-P13 has been approved for the same indications as the infliximab RMP, there have been some concerns as to whether it is possible to extrapolate immunogenicity data from patients with RA and AS to patients with IBD [4,5,46]. A primary concern is that patients with RA receiving infliximab also receive the immunosuppressant MTX, which suppresses the formation of ADAs and is less frequently used in patients with IBD. Another concern is that patients with AS historically exhibit a lower incidence of ADAs to infliximab than other patient populations. It has therefore been questioned whether the immunogenicity of CT-P13 has been investigated in suitably sensitive populations. These concerns are considered in more detail in the following sections.

Immunogenicity of infliximab RMP across indications

Several studies have investigated immunogenicity with the infliximab RMP across a range of indications, with some suggesting that there are differences in immunogenicity across these different therapeutic areas [14,32,33,35,47–49]. A summary of the immunogenic data for the infliximab RMP across its licensed indications is provided in the European label for this agent [49]. In clinical studies using single and multiple doses ranging from 1 to 20 mg/kg, ADAs to the infliximab RMP were detected in 14% of patients using any immunosuppressant therapy, and in 24% of patients without immunosuppressant therapy. In RA, 8% of patients who received the recommended scheduled treatment dose regimens combined with MTX developed ADAs to infliximab. In psoriatic arthritis, patients received 5 mg/kg infliximab RMP with and without MTX. Antibodies to infliximab occurred in 4% of patients receiving MTX and in 26% of patients not receiving MTX at baseline, with ADAs occurring in 15% of patients overall. Finally, in the original trial of patients with CD who received maintenance infliximab treatment, ADAs to RMP occurred in 3.3% of patients receiving immunosuppressants and in 13.3% of patients not receiving immunosuppressants [49]. However, it can be misleading to compare the incidence of ADAs between studies in different indications for various reasons:

• Immunogenicity data are highly dependent on the sensitivity and specificity of the assay used to measure the ADAs [50,51]. For example, radioimmunoassays are not susceptible to plastic-plate interference (non-specific binding) as opposed to standard solid-phase ELISA-based methods. Even within ELISA assays, variation exists regarding drug tolerance of the assay to detect ADAs; for example, an anti-lambda-based assay was shown to be more sensitive in this respect than standard ELISA [52]. Moreover, fluid-based assays and the anti-lambda based assay are also able to detect monovalent IgG4 ADAs, which may represent a significant portion of ADAs that go undetected in most ELISAs due to their potential monovalent form. It follows that levels of ADAs detected by some ELISA-based assays may be an underestimate [32]. Furthermore, dissimilar technical protocols, such as sample handling and freeze/thaw cycles, can also impact on resulting immunogenicity data. Therefore, the validity of comparing immunogenicity rates across different studies employing different assays is highly questionable.

• Immunogenicity data are highly dependent on the timing of sampling. A recent study concerning temporal pattern of ADA formation in IBD patients showed that the evolution of ADA is gradual at first, but the majority of subjects form ADAs within the first 12 months of infliximab therapy [52]. Additionally, it is known that some ADAs can be transient and disappear upon subsequent sampling and continued drug administration [52–55]. Taken together, these data indicate that comparing immunogenicity rates between cross-sectional studies that sample ADAs at different and non-standardized single time points is difficult, and can, in fact, be very misleading.

• Immunogenicity data are highly dependent on the dosing and schedule of biologic drug administration. A study specifically addressing the relationship between infliximab ADAs and treatment strategy in CD patients revealed a significantly lower incidence of ADAs with a regular maintenance treatment schedule when compared with an episodic strategy (8 vs 30%) [34]. Usage and type of concomitant immunomodulators can also impact heavily on ADA levels. For example, infliximab treatment in CD patients resulted in the formation of ADAs in 73% of patients who did not receive concomitant immunomodulators, compared with just 46% of patients who did [36].

Considering all of these factors, it is clear that the only robust and valid methodology for comparing rates of immunogenicity across different disease states would be to employ the same assay in the same laboratory, and to prospectively examine serial consecutive samples from patients with different disorders receiving the same dosing of drug, with and without the same immunomodulators. Unfortunately, however, such validated ‘cross-indication’ data are as yet unavailable.

Data from two studies using combination MTX and infliximab RMP therapy allows the comparison of immunogenicity data in RA and CD. In the Combination Of Maintenance Methotrexate-Infliximab Trial (COMMIT) – the only clinical trial of infliximab RMP in CD employing MTX as an immunomodulator – 4% of the 63 treated patients developed antibodies against infliximab in the first 54 weeks of treatment [56]. This rate was slightly lower than the 8.5% ADA rate found in the comparable, but significantly larger, 54-week Anti-TNF Therapy in Rheumatoid Arthritis with Concomitant Therapy (ATTRACT) clinical trial, performed in MTX-treated RA patients [57]. Notably, the doses used in the ATTRACT study, ranging from 3 to 10 mg/kg every 4 weeks, are comparable or higher than the doses used in CD. Also of note is that the assay used in the COMMIT CD trial was the fluid-phase homogenous mobility shift assay, which is more sensitive for ADA detection than the solid-phase ELISA used in the ATTRACT trial. This suggests that RA would be no less suitable or sensitive for examining the immunogenicity of CT-P13 versus the IBD population and that, overall, immunogenicity rates between these two indications are likely to be comparable despite a lack of direct comparison.

Conclusions

Immunogenicity is a well-recognized phenomenon common to most biologics, including infliximab. ADAs to the infliximab RMP have been shown to decrease its clinical effectiveness and may lead to an increase in infusion-related reactions. CT-P13 was approved in Europe across all indications for which the RMP is indicated, based in part on data from clinical trials showing its equivalence in RA and AS. Cross-immunogenicity in vitro tests showed similar immunogenicity of RMP and CT P13. Moreover, in randomized, double-blind trials, there were no differences in immunogenicity between CT-P13 and RMP in RA (3 mg/kg with MTX) or AS (5 mg/kg monotherapy). Notably, despite the fact that patients with AS tend to have reduced immunogenicity, the observation that both agents had similar rates of ADA production supports the contention that they have similar immunogenicity. While acknowledging that only a small amount of high-quality data on the subject are available and that further head-to-head comparisons are necessary for definitive conclusions to be drawn, the data described here suggest there is low likelihood of a meaningful difference in the immunogenicity and incidence of ADAs against CT-P13 and RMP across different diseases, including IBD. Notably, a direct comparison of the immunogenicity of CT-P13 and RMP in an IBD population will be possible upon completion of an ongoing, randomized parallel-group Phase III study recruiting patients with CD (ClinicalTrials.gov identifier: NCT02096861 [58]). These data, along with those from other studies that form part of the pharmacovigilance plan for CT-P13, should allow firm conclusions to be made regarding the relative immunogenicity of CT-P13 and RMP. Due to the inherent variation involved in biosimilar production, it should be remembered, however, that data regarding the immunogenicity of CT-P13 cannot be extrapolated to other, future infliximab biosimilars, or indeed to biosimilars of other RMPs.

Expert commentary

There has been some debate in the literature regarding the validity of ‘indication extrapolation’, that is, the practice of extrapolating data collected for a biosimilar drug in one clinical indication to another in which no clinical trial data are available. This issue is of particular relevance when considering the immunogenicity of a biosimilar, as so many factors are involved in determining the extent to which ADAs are formed in response to a biologic drug. Furthermore, the consequences for the pharmacokinetics, efficacy and safety profile of the drug concerned can be profound. Variations between different studies in the type of ADA measurement assays used, dosing schedules and the timing of ADA sampling can make it difficult to compare clinical trial data on the immunogenicity rate of infliximab in different indications. Despite these caveats, the best validated data regarding the immunogenicity of originator infliximab (or the ‘RMP’) in RA versus IBD indicates a broadly similar rate across these two indications, supporting the validity of extrapolating immunogenicity data from RA to IBD. Crucially, in randomized controlled trials, the in vivo immunogenicity profile of the infliximab biosimilar CT-P13 was shown to be highly comparable to that of the infliximab RMP in both RA and AS. The dosage administered, patient population, indication studied and the usage (or not) of MTX did not lead to differences in immunogenicity between the two biologics. While cautious interpretation is always prudent, these data should instill confidence that the immunogenicity of CT-P13 is highly similar to that of the infliximab RMP, and that data extrapolation was an appropriate method to employ in this instance.

Five-year view

Confidence in the validity of ‘indication extrapolation’ for biosimilars in general and CT-P13 in particular is increasing. Recent data have shown highly similar immunogenicity profiles between patients who switched from RMP to CT-P13 and those who continued on CT-P13. Such data add to the weight of evidence suggesting that the immunogenicity of CT-P13 in vivo is similar to that of its RMP. A clinical trial in patients with RA, spondyloarthritis, UC, CD and psoriasis in Norway (NOR-SWITCH; ClinicalTrials.gov identifier: NCT02148640 [59]) will also allow comparison of immunogenicity data in patients switching from infliximab RMP to CT-P13 versus those who continue treatment with the infliximab RMP.

It is clear that factors such as the type of ADA measurement assay used and the timing of sampling have an important impact on measured immunogenicity data. It is hoped that such factors will be considered in future clinical trials of biosimilars, thereby increasing homogeneity in the field. Such an approach should result in more confidence when comparing clinical trials, especially those intended as a base for extrapolation. When considering new protein therapeutics coming to market, it is possible that immunogenicity can be reduced at the design stage through methods such as deimmunization (e.g. the removal of T-cell epitopes by site-directed mutagenesis), an approach already investigated with some success in the oncology setting [60,61].

Key issues

• The immunogenicity of monoclonal antibody (mAb) protein therapeutics is an important concern as anti-drug antibodies (ADAs) can impact upon clinical efficacy and safety.

• Many factors can influence the formation of ADAs, including whether the drug is administered in combination with immunomodulators. The formation of ADAs may be variable in different indications and different patients, meaning that predicting the extent of their formation is difficult.

• Infliximab is a mAb therapy used in the treatment of several immune-mediated inflammatory diseases. It is known to induce the formation of ADAs in some patients, resulting in a loss of clinical response and increased risk of adverse events.

• The infliximab biosimilar CT-P13 was approved in Europe for use in all indications in which originator infliximab (or the reference medicinal product [RMP]) is licensed, based in large part on data from two key equivalence trials in rheumatoid arthritis (RA) and ankylosing spondylitis (AS).

• There have been concerns regarding the extrapolation of these data to other indications such as Crohn’s disease and ulcerative colitis, particularly with respect to immunogenicity, as the formation of ADAs can be unpredictable and vary between indications.

• Comparing infliximab immunogenicity data between trials in RA and Crohn’s disease indicates that there is probably not a substantial difference in ADA incidence across these two indications. This suggests that RA would be no less suitable or sensitive for examining the immunogenicity of CT-P13 versus other populations. Examining immunogenicity in AS patients, the majority of whom received infliximab monotherapy, provides an additional tier of confidence on which to base this comparison.

• Key data from the PLANETRA and PLANETAS trials comparing CT-P13 and the infliximab RMP in RA and AS, respectively, show there are no noticeable differences in immunogenicity between these two drugs. Importantly, the dosage administered, patient population, indication studied and usage (or not) of methotrexate did not lead to differences in immunogenicity between the two biologics.

• These data can support confidence among clinicians and patients that extrapolation of immunogenicity data for CT-P13 collected in RA and AS to other indications, in particular inflammatory bowel disease, is clinically valid.

Financial & competing interests disclosure

S Ben-Horin has received consultancy fees and/or research support from AbbVie, Janssen, Takeda, Schering-Plough, and Celltrion. G Heap has received unrestricted education grants/travel support from AbbVie, Dr Falk Pharma, and Tillotts Pharma UK. T Ahmad has received consultancy fees and/or research support from AbbVie, Merck, Takeda, NAPP, and Celltrion. Y Chowers has received consultancy and lecture fees and/or research support from AbbVie, Janssen, Takeda, and Schering-Plough. H Kim and T Kwon are full time employees of Celltrion. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed. Editorial support (writing assistance, assembling tables and figures, collating author comments, grammatical editing, and referencing) was provided by Ryan Woodrow (Aspire Scientific Limited, Bollington, UK) and was funded by Celltrion Healthcare Co., Ltd (Incheon, Republic of Korea).

Notes