Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Bioanalytical Challenges due to Prior Checkpoint Inhibitor Exposure: Interference and Mitigation in Drug Concentration and Immunogenicity Assays

Bioanalytical Challenges due to Prior Checkpoint Inhibitor Exposure: Interference and Mitigation... AAPS J (2021) 23:109 Vol.:(0123456789) https://doi.org/10.1208/s12248-021-00643-4 Research Article Bioanalytical Challenges due to Prior Checkpoint Inhibitor Exposure: Interference and Mitigation in Drug Concentration and Immunogenicity Assays 1 1 1 1 1 1,2 Andrew F. Dengler · Rachel Weiss  · Tiffany Truong · Susan C. Irvin · Nidhi Gadhia · Mohamed Hassanein · 1 1 3 1 1 1 Camille Georgaros · Jessica‑Ann Taylor · Anne Paccaly · Giane Sumner · Matthew D. Andisik · Albert Torri · 1,4 Michael A. Partridge Received: 1 June 2021 / Accepted: 24 August 2021 Abstract Monoclonal antibodies (mAbs) are a leading class of biotherapeutics. In oncology, patients often fail on early lines of biologic therapy to a specific target. Some patients may then enroll in a new clinical trial with a mAb specific for the same target. Therefore, immunoassays designed to quantify the current mAb therapy or assess immunogenicity to the drug may be susceptible to cross-reactivity or interference with residual prior biologics. The impact of two approved anti-PD-1 mAbs, pembrolizumab and nivolumab, was tested in several immunoassays for cemiplimab, another approved anti-PD-1 mAb. The methods included a target-capture drug concentration assay, a bridging anti-drug antibody (ADA) assay and a competitive ligand-binding neutralizing antibody (NAb) assay. We also tested bioanalytical strategies to mitigate cross-reactivity or interference in these assays from other anti-PD-1 biologics. Both pembrolizumab and nivolumab cross-reacted in the cemiplimab drug concentration assay. This was mitigated by addition of antibodies specific to pembrolizumab or nivolumab. ADA specific for pembrolizumab and nivolumab did not interfere in the cemiplimab ADA assay. However, pembrolizumab and nivolumab generated a false-positive response in a target-capture NAb assay. Our results demonstrate that similar exogenous pre-existing anti-PD-1 mAbs (biotherapeutics) such as pembrolizumab and nivolumab are detected and accurately quantified in the cemiplimab drug concentration assay. However, once steady state is achieved for the new therapy, prior biologics would likely not be detected. Cross-reactivity and interference in immunoassays from previous treatment with class-specific biotherapeutic(s) pose significant bioanalytical challenges, especially in immuno-oncology. KEY WORDS clinical impact · drug concentration assay · immunogenicity · monoclonal antibody therapeutic · prior biologic exposure Andrew F. Dengler, Rachel Weiss, Tiffany Truong, and Susan C. INTRODUCTION Irvin are equal first-author contributions. Biotherapeutics and especially monoclonal antibod- Regeneron Pharmaceuticals, Bioanalytical Sciences, 777 Old Saw ies (mAbs) are one of the fastest growing segments of Mill River Rd, Tarrytown, New York 10591, USA. the pharmaceutical industry (1). This trend has acceler - Present Address: Pfizer, 401 N Middletown Rd, Pearl River, New York 10965, USA. ated with the approval of approximately 30 novel mAbs Regeneron Pharmaceuticals, Pharmacometrics (DSP), 777 Old Saw Mill between 2017 and 2020 alone ( https:// w ww. fda. gov/ River Rd, Tarrytown, New York 10591, USA.drugs/ new- drugs- fda- cders- new- molec ular- entit ies- and- To whom correspondence should be addressed. (e–mail: michael.partridge@ new- thera peutic- biolo gical- produ cts/ novel- drug- appro regeneron.com) vals- 2017,2). Abbreviations: ADA, Anti-drug antibody; CPI, Check-point inhibitor; IO, Immuno-oncology; I&I, Immunology and inflammation; mAb, Monoclonal Two therapeutic areas in particular have witnessed antibody; NAb, Neutralizing antibody; PC, Positive control; PD-1, Programmed rapid growth in recent years: cancer immunotherapies, cell death protein 1; PK, Pharmacokinetics; TE, Treatment emergent Vol.:(0123456789) Vol.:(0123456789) AAPS J (2021) 23:109 also referred to as immuno-oncology (IO), and immunol- be susceptible to cross-reactivity from different therapies ogy and inflammatory diseases (I&I, e.g., anti-TNF-α) (2 , directed to the same target. In cases where patients change to 3). This has resulted in an increasingly competitive and a new therapy of the same class before the prior therapy has crowded treatment landscape. In some cases, hundreds of been cleared, this may result in the detection of preceding biotherapeutics engaging the same targets are under inves- therapeutics (6). tigation in clinical studies (see Table I) (4, 5). Bioanalysis of samples collected from patients treated This redundancy is likely beneficial for both patients and with cemiplimab in two oncology trials revealed unexpect- health care providers as it provides multiple therapeutic options. edly high concentrations of drug detected in baseline (pre- However, it also increases the potential for patients to receive dose) samples in the target-capture cemiplimab drug con- different biotherapeutics (approved or investigational) against centration assay. The measurable drug concentrations of up the same target. Patients experiencing disease recurrence, non- to 95 µg/mL, similar to steady state cemiplimab concentra- responders, and those developing resistance may enroll in tions (12–14), could not be explained by high background, clinical trials involving related therapies; for example, patients matrix interference, analytical errors, or sample collection may be anti-PD-1 experienced when enrolling in another errors. The baseline samples with detectable drug were from anti-PD-1 study (6, 7). In the last few years, the Food and Drug patients enrolled in studies that allowed prior treatment with Administration (FDA) and European Medicines Agency (EMA) an anti-PD-1 biotherapeutic, including pembrolizumab and have approved seven immune check point inhibitors (CPIs): nivolumab. one monoclonal antibody targeting the CTLA-4 pathway Like cemiplimab, pembrolizumab, and nivolumab are (ipilimumab), three targeting PD-L1 (atezolizumab, avelumab both human IgG4 mAbs specific for PD-1 and are approved and durvalumab), and four targeting PD-1 (cemiplimab, for a variety of oncology indications (15). Since the cemipli- dostarlimab, nivolumab, and pembrolizumab), for the treatment mab drug concentration assay uses PD-1 as the capture rea- of patients with multiple cancer types (8–10). gent, and a non-specific anti-IgG4 as the detection reagent, In life-threatening diseases like cancer, patients failing to we investigated the potential for these two similar anti-PD-1 respond to one CPI may not be able to wait for clearance therapies to interfere with or cross-react in the cemiplimab of the therapeutic before enrolling in a clinical trial with a drug concentration or immunogenicity assays (14, 16, 17). treatment regimen consisting of a different CPI in combination Analysis of in vitro serum samples spiked with these therapies established that both pembrolizumab with a novel experimental therapy (11). With the limited and nivolumab could be detected in the target- number of immunoassay formats commonly used in clinical capture cemiplimab drug concentration assay. We also trials to measure drug concentrations, anti-drug antibodies demonstrated that the addition of antibodies specific to (ADA), and neutralizing antibodies (NAb), residual systemic either pembrolizumab or nivolumab could block detection drug that binds the same target may cross-react or interfere of the drugs in the assay, thus providing a potential in these assays. For example, target-capture immunoassays strategy to mitigate this cross-reactivity. In addition, we that measure the concentration for one mAb therapeutic might demonstrated that pembrolizumab, nivolumab, or anti-drug antibodies to either of these mAbs do not interfere in the cemiplimab bridging ADA assay. However, pembrolizumab Table I List of mAbs Approved or in Clinical Development for the and nivolumab generate a false-positive response in the Top 10 Targets in I&I and IO target-capture competitive ligand-binding NAb assay. With the increasing number of clinical trials with different Target Approved In clinical develop- Total biotherapeutics that engage the same targets, we anticipate ment that cross-reactivity or interference in immunoassays from EGFR 4 207 211 these biotherapeutics will be an ongoing bioanalytical HER-2 8 186 194 challenge, in assays that use target-capture and generic anti- PD-1 5 152 157 IgG detection reagents. CD20 9 130 139 PD-L1 3 132 135 TNF-α 15 118 133 CD19 6 114 120 MATERIALS AND METHODS IL-6 3 48 51 CD38 2 26 28 Materials and Reagents IL-5 3 10 13 EGFR epidermal growth factor receptor, HER-2 human epidermal Cemiplimab (Libtayo®), the recombinant human PD-1 growth factor receptor 2, PD-1 programmed cell death protein 1, PD- (extracellular domain), the monoclonal anti-cemipli- L1 programmed death ligand-1, TNF-α tumor necrosis factor α, IL-6 mab antibody (ADA positive control), the neutralizing interleukin 6, IL-5 interleukin 5 AAPS J (2021) 23:109 Vol.:(0123456789) monoclonal anti-cemiplimab antibody (NAb positive con- further diluted in the labeled reagent solution. After incuba- trol), the blocking anti-cemiplimab idiotypic antibody, and tion for 1 h, samples were transferred to blocked (5% BSA) the anti-human IgG4 monoclonal antibody, were produced Streptavidin-coated plates and incubated for 1 h, before by Regeneron Pharmaceuticals, Inc. (Tarrytown, NY). addition of Read Buffer and analysis on a QuickPlex SQ Labeling of antibodies and targets with biotin using EZ-link 120 reader (MSD, Gaithersburg, Maryland). Sulfo-NHS-LC-Biotin (Thermo Fisher Scientific), and with ruthenium NHS ester (MSD), was performed according to Target-Capture NAb Assay the manufacturer’s instructions. Anti-nivolumab and anti-pembrolizumab monoclonal anti- Microplates were coated with recombinant human PD-1 bodies were purchased from BioRad (Hercules, CA). Pem- (0.5 µg/mL) and blocked with 5% (w/v) BSA. Unless oth- brolizumab (Keytruda®) was manufactured by Merck Sharp erwise specified, serum samples were diluted tenfold in & Dohme Corp. (Kenilworth, NJ). Nivolumab (Opdivo®) 300 mM acetic acid and incubated for a minimum of 10 min was manufactured by Bristol Myer Squibb (New York, NY). and then neutralized using a capture reagent solution con- All solutions, unless otherwise specified, were prepared in taining 250 mM Tris, 20 ng/mL biotinylated-cemiplimab, assay buffer (1% BSA in 1X PBS). Read Buffer T (4X) and and 5% BSA for 1 h. This was followed by incubation of the Streptavidin-coated microplates were purchased from 100 ng/mL Neutravidin-HRP for 1 h, and finally incubated Meso Scale Discovery (MSD, Gaithersburg, Maryland). with SuperSignal ELISA Pico Chemiluminescent Sub- Glacial acetic acid (17.4 M), NeutrAvidin-HRP, SuperSig- strate, prepared according to manufacturer’s instructions, nal ELISA Pico Chemiluminescent Substrate, and 96-well for 10 min. Microplates were read on a luminescence reader microplates were purchased from Thermo Fisher Scientific (Biotek, Winooski, VT). (Waltham, MA). Trizma base (1.5 M) was purchased from Sigma (St Louis, MO). Wash solution and 10% BSA were Clinical Sample Collection purchased from SeraCare, (Milford, MA). Human serum, including that which was used for the negative quality con- Clinical serum samples were obtained from Phase 1 oncol- trols (NQCs), was purchased from BioIVT (Westbury, NY). ogy studies. Studies included patients with documented anti-PD-1 medication use prior to enrollment in the trials Immunoassay Procedures with cemiplimab. The serum samples were collected from patients prior to dosing of study drug at baseline. All incubations noted below were performed at room tem- perature (RT) unless otherwise specified. Drug Concentration Assay RESULTS Microplates were coated overnight at 4 °C with recombinant Cross‑reactivity of Anti‑PD‑1 Antibodies human PD-1 (0.5 µg/mL) and blocked with 5% (w/v) BSA in the Cemiplimab Target‑Capture Drug Concentration for a minimum of 1 h. After blocking, human serum (2%) Assay samples containing the indicated proteins were added to the microplates and incubated for 1 h. Subsequently, micro- The cemiplimab enzyme-linked immunosorbent assay plates were incubated with 100 ng/mL biotinylated mouse (ELISA) uses recombinant PD-1 as the capture reagent and anti-human IgG4 mAb for 1 h, followed by incubation with a biotinylated anti-IgG4 mAb as the detection component 100 ng/mL NeutrAvidin-HRP for 1 h, and finally incubated (Fig. 1a). Other biologics that are also PD-1 specific and with SuperSignal ELISA Pico Chemiluminescent Substrate, constructed with an IgG4 framework could potentially be prepared according to manufacturer’s instructions, for 10 detected in this target-capture method (Fig. 1b). Therefore, to 30 min. Microplates were read on a luminescence reader we set out to establish whether drug detected at baseline in (BioTek, Winooski, VT). clinical study samples was due to cross-reactivity from other IgG4 anti-PD-1 therapeutics. Bridging Immunogenicity Assay To determine whether other anti-PD-1 mAbs cross- react in the assay, twofold serial dilutions of each of the Unless otherwise specified, serum samples were diluted three mAbs (cemiplimab, pembrolizumab, and nivolumab) tenfold in 300 mM acetic acid and incubated for 30 min. were prepared in human serum at concentrations of 5 to Biotin and ruthenium labeled cemiplimab (2 µg/mL) were 0.078 µg/mL (100 to 1.56 ng/mL after minimum required prepared in assay buffer containing 150 mM Tris (unless dilution) and analyzed in the assay. The signal gener - otherwise specified), and acid-treated serum samples were ated from the serial dilutions of all three mAbs was very Vol.:(0123456789) AAPS J (2021) 23:109 of cemiplimab. This indicated that within the quantita- tive range of the assay, all anti-PD-1 mAbs present in the sample would be detected and accurately quantified with similar sensitivity. Anti‑idiotypic Blocking Antibodies Mitigate Cross‑reactivity in the Drug Concentration Immunoassay in Spiked Samples and in Baseline Clinical Trial Samples A potential strategy to minimize cross-reactivity of pembrolizumab and nivolumab in the cemiplimab ELISA is to use anti-idiotypic antibodies to block binding of the other therapeutic mAbs to PD-1 on the plate (Fig. 3a). To test this strategy, mock serum samples were created by spiking serum with cemiplimab, pembrolizumab, and nivolumab at the middle quality control level (MQC; 1 µg/mL). The mock samples were then tested in the presence or absence of anti-cemiplimab, anti-pembrolizumab, and anti-nivolumab antibodies at 100 × (100 µg/mL) the MQC concentration. The results demonstrate the anti-idiotypic blocking antibodies specifically inhibited binding of the corre- sponding drug to PD-1 on the plate diminishing detection in the cemiplimab ELISA (Fig. 3b through d). The anti- idiotypic antibodies did not cross-react or interfere with quantification of the other mAbs in the assay. This strategy could also be used to confirm the identity of any anti-PD-1 mAb in baseline clinical samples from patients previously treated with an anti-PD-1 mAb. To evaluate this approach, baseline samples collected from patients with prior anti-PD-1 exposure to either pem- Fig. 1 Illustrations of the functional cemiplimab drug concentration assay and cross-reactivity in serum. a Cemiplimab (cemi), captured brolizumab or nivolumab were analyzed in the presence on a PD-1 coated microplate is detected by a biotinylated anti-human and absence of each of the three anti-idiotypic antibod- IgG4 mAb followed by NeutrAvidin-HRP b Other anti-PD-1 human ies (Fig. 3e). In every sample, assay signal was markedly IgG4 mAbs such as nivolumab (nivo) or pembrolizumab (pembro) inhibited (greater than 80%) by only the anti-idiotypic could also be detected antibody that corresponded to each patient’s anti-PD-1 medication history (Fig. 3e). similar, indicating that they can all be detected in the Potential Impact of Other Anti‑PD‑1 Biologics assay. Furthermore, analyte recovery of pembrolizumab on the Specificity and Selectivity of the Cemiplimab  and nivolumab concentrations when interpolated from the ADA Assay cemiplimab standard curve generated values within 20% of the nominal values, thus demonstrating that the mAbs Prior exposure to the same class of anti-PD-1 mAb raises can be accurately quantified in this assay (Fig. 2a). the possibility that some patients previously treated with To determine if there was an additive effect of pem- pembrolizumab or nivolumab may generate ADA that brolizumab or nivolumab in the cemiplimab ELISA, cemi- cross-react in the anti-cemiplimab ADA assay. The bridging plimab at the high quality control level (HQC; 75 ng/mL) cemiplimab ADA assay uses a mouse anti-cemiplimab or the lower level of quantification (LLOQ; 1.56 ng/mL) antibody as the positive control and biotinylated-cemiplimab was added to the serial dilutions of pembrolizumab and and ruthenium-labeled cemiplimab as bridge components nivolumab (Fig. 2b and c). When interpolated off the cemi - (Fig. 4a). plimab standard curve, the concentration of detected drug Serum positive control samples were prepared con- was equal to the sum of pembrolizumab or nivolumab plus taining specific anti-cemiplimab, anti-pembrolizumab, or the HQC (Fig. 2b and c) or LLOQ (data not shown) level AAPS J (2021) 23:109 Vol.:(0123456789) Fig. 2 Other human IgG4 anti-PD-1 mAbs can be detected and accurately quantified in the functional cemiplimab drug concentration assay in serum. a Quantitation of serial dilutions (100 to 1.56 ng/mL) of cemiplimab (blue), pembroli- zumab (green), and nivolumab (orange) in the assay. b Drug concentrations of HQC samples spiked with serial dilutions of pembrolizumab, interpolated from the cemiplimab standard curve, in the cemiplimab drug concentration ELISA c Drug concentrations of HQC samples spiked with serial dilutions of nivolumab, interpolated from the cemiplimab standard curve, in the cemiplimab drug concen- tration assay anti-nivolumab antibodies and were analyzed in the ADA Collectively, these results demonstrate the specificity and assay. Anti-cemiplimab positive control samples generated suitability of our cemiplimab ADA assay for the detection a strong signal in the assay, while the anti-pembrolizumab of anti-cemiplimab antibodies in the presence of other anti- or anti-nivolumab samples generated signal approximately PD-1 ADA or residual anti-PD-1 therapeutics. equivalent to the negative control samples (Fig. 4b). This suggests that anti-pembrolizumab or anti-nivolumab anti- bodies generated in patients treated with these drugs likely Interference from Other PD‑1 Therapies will not interfere with the detection of anti-cemiplimab on a Target‑Capture NAb Assay antibodies. To test whether residual concentrations of pembroli- Confirmed ADA positive samples were further assessed zumab or nivolumab in circulation can impact the detection in a NAb assay to evaluate the ability to neutralize the of cemiplimab ADA, samples containing an anti-cemiplimab biological activity of the drug. A competitive ligand- monoclonal antibody (500 ng/mL) were tested in the pres- binding NAb assay was developed that uses recombinant ence of increasing concentrations of either pembrolizumab PD-1 as the capture reagent and biotinylated-cemiplimab or nivolumab. The highest concentration of each drug tested and streptavidin-HRP as the detection components (2 mg/mL), was greater than the Cmax levels observed in (Fig. 5a). When present in a serum sample, NAbs will the clinical study samples (12, 13). These results demon- bind to biotinylated cemiplimab, preventing binding to strated that even at high concentrations of pembrolizumab or the PD-1 coated microplate and inhibiting the assay signal nivolumab, detection of the anti-cemiplimab antibody was (Fig. 5a). However, in this assay format, the presence not impacted by the presence of either antibody in serum of other anti-PD-1 biologics could also compete with (Fig. 4c). biotinylated cemiplimab for PD-1 binding, potentially As a control, cemiplimab was also spiked at high generating a false-positive NAb result (Fig. 5b). concentrations in the assay. As expected, this reduced the To t est this, cemi pl im ab, pe mbroli zumab, and anti-cemiplimab antibody assay signal, although control nivolumab were serially diluted in serum from 4000 samples (500 ng/mL) remained positive in the assay when to 31.3 ng/mL and analyzed in the target-capture NAb spiked with cemiplimab at concentrations greater than assay. As demonstrated in Fig. 5c, a false-positive 500 µg/mL, confirming the cemiplimab drug tolerance level NAb signal was detected when approximately 155 ng/ of the assay (Fig. 4c). These experiments demonstrate that mL of any of these anti-PD-1 drugs were added to the cemiplimab ADA assay is specific only for anti-drug the competitive ligand-binding NAb assay, which is antibodies directed to the variable domain of cemiplimab approximately 1000-fold lower than steady state drug and is not impacted by the presence of other anti-PD-1 mAbs. concentrations (12–14). In contrast, excess cemiplimab Vol.:(0123456789) AAPS J (2021) 23:109 Fig. 3 Anti-idiotypic antibod- ies block binding of anti-PD-1 mAbs in the cemiplimab target- capture drug concentration ELISA. a Schematic depicting the anti-idiotypic antibodies (checkerboard pattern) blocking each drug (solid colors) from binding to the PD-1 capture in the cemiplimab ELISA. b Mock sample serum control prepared with cemiplimab at the MQC concentration and tested in the presence or absence of 100 × concentration of all three anti-idiotypic antibod- ies in the cemiplimab ELISA. c Mock sample serum control prepared with pembrolizumab and tested in the presence of 100X concentration of the anti-pembrolizumab and the anti-cemiplimab antibodies in the cemiplimab ELISA. d Mock sample serum control prepared with nivolumab and tested in the presence of 100 × concentra- tion of the anti-nivolumab and the anti-cemiplimab antibod- ies in the cemiplimab ELISA. e Baseline clinical samples with detectable responses in the cemiplimab ELISA from patients with prior exposure to pembrolizumab or nivolumab were evaluated in the presence of each of the three anti-idio- typic antibodies to demonstrate specific signal inhibition (or other anti-PD-1 mAbs) do not generate false-positive DISCUSSION responses in a drug-capture competitive ligand-binding NAb assay, as excess therapeutic from the previous There are numerous therapeutic targets for which multiple capture step is washed away before addition of labeled biologic drugs are approved, including products target- target as detection reagent (not shown). ing TNF-α, PD-1, or PD-L1 (Table I) (15). Because these AAPS J (2021) 23:109 Vol.:(0123456789) Fig. 4 The cemiplimab ADA assay is specific only for anti-cemi - cemiplimab, anti-nivolumab or anti-pembrolizumab antibodies at plimab antibodies, and other anti-PD-1 mAbs do not interfere in three concentrations (X, Y, and Z ng/mL) in serum. c Signal in the the method. a Schematic of the cemiplimab bridging ADA assay in ADA assay of anti-cemiplimab antibody control samples (500 ng/ which ADA in the samples bridge between biotin- and ruthenium- mL) tested in the presence of serial dilutions of cemiplimab, pem- labeled cemiplimab generating signal in the method. b Signal-to- brolizumab, and nivolumab at the indicated concentrations noise ratio in the ADA assay for control samples containing anti- targets are clinically validated, many biotherapeutics are the respective drugs. Using this approach, we identified being investigated in novel combinations (or as biosimi- patients that had previously received either pembrolizumab lars) for the treatment of new indications or to improve effi- or nivolumab, which aligned with patient’s prior medica- cacy in existing populations. During development of these tion history. new drugs, in particular for oncology therapies in the PD-1 In addition to investigating the functional cemiplimab and PD-L1 classes, trial participants may transition to the drug concentration assay, we wanted to understand whether investigational biotherapy (or combination) while still hav- pre-existing pembrolizumab or nivolumab, or ADA directed ing detectable systemic concentrations of their prior therapy against either pembrolizumab or nivolumab, could interfere (10). These prior drugs have the potential to cross-react or in the cemiplimab immunogenicity assays. In the bridging interfere in bioanalytical assays for the new therapy, espe- ADA assay, the presence of pembrolizumab or nivolumab cially with target-capture-based methods. did not interfere with the detection of the anti-cemiplimab In cemiplimab clinical studies, enrollment of patients positive control. Furthermore, only antibodies specific for who had received prior anti-PD-1 therapy was permitted cemiplimab were detected in this assay, with no cross- in some cohorts. Baseline samples (taken prior to cemi- reactivity from specific anti-pembrolizumab or anti- plimab administration) for some of these patients had nivolumab antibodies. Although these results collectively high drug levels in the cemiplimab drug concentration confirm the specificity of our anti-cemiplimab ADA assay, assay. We demonstrated that the assay was able to detect it does not negate the possibility that ADA against other pembrolizumab or nivolumab in spiked samples, with anti-PD-1 therapeutics may already exist but cannot be approximately equivalent quantitation to cemiplimab. In detected. addition, we were able to specifically inhibit assay signal In the competitive ligand-binding NAb assay, neither by addition of anti-idiotype mAbs directed against each of pembrolizumab nor nivolumab interfered in the assay Vol.:(0123456789) AAPS J (2021) 23:109 Fig. 5 Anti-PD-1 biologics generate a false-positive response in a b Schematic of a false-positive NAb response in the presence of target-capture NAb assay in serum. a Schematic depicting the target- anti-PD-1 mAbs that bind to the PD-1 coated plate preventing bioti- capture NAb assay. In the absence of NAb, biotinylated cemiplimab nylated-cemplimab from generating signal in the assay. c Assay sig- binds to a PD-1 coated plate, followed by streptavidin conjugated nal inhibition (%Inhibition) in a cemiplimab target-capture NAb after to HRP, generating signal in the assay. The presence of NAb inhib- addition of serially-diluted cemiplimab, nivolumab, or pembroli- its biotin-cemiplimab binding to PD-1, resulting in signal reduction. zumab, at concentrations ranging from 4000 to 31.25 ng/mL when configured in a drug-capture format (not shown), CONCLUSION since these molecules were washed away before addition of the labeled target. However, in a target-capture format, the We demonstrated that prior exposure to anti-PD-1 mAbs, presence of either pembrolizumab or nivolumab generated pembrolizumab and nivolumab, can be detected and quan- a false-positive NAb response due to target binding and titated in the cemiplimab drug concentration immunoassay. the resulting competition with the biotinylated-cemiplimab However, at steady state for the new therapy, these prior detection antibody, even at relatively low concentrations. biotherapeutics would likely not be detected. In addition, Therefore, a drug-capture format is recommended for we demonstrated that these prior biotherapeutics or anti- competitive ligand-binding NAb assays to avoid false- drug antibodies to either of these mAbs do not interfere in positive responses from other biotherapeutics to the same the cemiplimab bridging ADA assay. However, in a target- target (17). capture competitive ligand-binding cemiplimab NAb assay, It is unknown whether true treatment-emergent anti- pembrolizumab and nivolumab generated a false-positive pembrolizumab or anti-nivolumab immunogenicity cor- response. relates to subsequent ADA responses to other anti-PD-1 Immunoassays that use the drug target as a reagent in mAbs. However, ADA to fully human or humanized the assay, especially in the capture step, are potentially mAbs are predominantly directed to the variable domains, susceptible to interference or cross-reactivity with other and ADA directed to the unique regions of one molecule biologics directed to the same target. Some of the most would be unlikely to cross-react with the other molecules widely prescribed pharmaceutical products are biologics (8, 16, 18). Furthermore, immunogenicity to one mAb with the same target, most prominently drugs targeting may not be predictive of an ADA response after subse- TNF-α, PD-1, or PD-L1. In oncology particularly, there quent exposure to a different mAb that binds to the same are additional targets (other than PD-1/PD-L1) that have target. more than one approved therapy, including CD20, EGFR, AAPS J (2021) 23:109 Vol.:(0123456789) 4. Database T-TA. https:// tabs. craic. com/ antib odies. Accessed 16 and HER-2 (Table I) (15). Cross-reactivity or interference Dec 2020. in immunoassays from previous exposure to biotherapeu- 5. Upadhaya S, Hubbard-Lucey VM, Yu JX. Immuno-oncology tics of the same class is an ongoing bioanalytical chal- drug development forges on despite COVID-19. Nat Rev lenge with molecules directed to these targets (17, 19). Drug Discov. 2020;19(11):751–2. ht t ps: / / doi . or g/ 10. 103 8/ d41573- 020- 00166-1. This work highlights the importance of understanding both 6. Fujita K, Uchida N, Kanai O, Okamura M, Nakatani K, patient medication history and assay format, to ensure the Mio T. Retreatment with pembrolizumab in advanced non- most appropriate bioanalytical strategies are implemented. small cell lung cancer patients previously treated with nivolumab: emerging reports of 12 cases. Cancer Chemother Acknowledgements We thank the patients who participated in these Pharmacol. 2018;81(6):1105–9. https:// doi. org/ 10. 1007/ trials and the investigators at the individual study sites. We thank the s00280- 018- 3585-9. rest of the Bioanalytical Sciences team members for their essential 7. Haanen J, Ernstoff M, Wang Y, Menzies A, Puzanov I, Grivas P, contributions to all of the cemiplimab oncology studies. et al. Rechallenge patients with immune checkpoint inhibitors following severe immune-related adverse events: review of the literature and suggested prophylactic strategy. J Immunother Can- Author Contribution AFD, RW, TT, SCI, MH, MDA, and MAP cer. 2020;8(1):e000604. https://d oi.o rg/1 0.1 136/j itc-2 020-0 00604. conceptualized the analysis and performed the data analysis. AP ana- 8. Wongchenko MJ, Ribas A, Dreno B, Ascierto PA, McArthur GA, lyzed the pharmacokinetic data and oversaw pharmacokinetic interpre- Gallo JD, et al. Association of programmed death ligand-1 (PD- tation. AFD, RW, and MDA oversaw the sample analysis. CG, JAT, and L1) expression with treatment outcomes in patients with BRAF GS conceptualized, developed, and validated the cemiplimab PK assay. mutation-positive melanoma treated with vemurafenib or cobi- TT and NG performed the sample analysis. AT oversaw the bioanalyti- metinib combined with vemurafenib. Pigment Cell Melanoma cal plans and interpretation. AFD, RW, TT, MH, SCI, MDA, and MAP Res. 2018;31(4):516–22. https://doi. org/ 10. 1111/ pcmr. 12670. wrote the manuscript. 9. Menzies AM, Johnson DB, Ramanujam S, Atkinson VG, Wong ANM, Park JJ, et al. Anti-PD-1 therapy in patients with Funding This research was funded by Regeneron Pharmaceuticals, advanced melanoma and preexisting autoimmune disorders or Inc. major toxicity with ipilimumab. Ann Oncol. 2017;28(2):368– 76. https://doi. org/ 10. 1093/ annonc/ mdw443. Declarations 10. Gentzler R, Hall R, Kunk PR, Gaughan E, Dillon P, Slingluff CL Jr, et al. Beyond melanoma: inhibiting the PD-1/PD-L1 pathway Disclosure The views expressed in this article are those of the in solid tumors. Immunotherapy. 2016;8(5):583–600. https:// authors and do not reflect official policy of Regeneron Pharmaceuticals. doi. org/ 10. 2217/ imt- 2015- 0029. 11. Murciano-Goroff YR, Warner AB, Wolchok JD. The future of Conflict  of Interest All authors are employees of Regeneron Phar- cancer immunotherapy: microenvironment-targeting combina- maceuticals, Inc., and may hold stock and/or stock options in the com- tions. Cell Res. 2020;30(6):507–19. https:// doi. org/ 10. 1038/ pany. s41422- 020- 0337-2. 12. Papadopoulos KP, Johnson ML, Lockhart AC, Moore K, Fal- chook GS, Formenti SC, et al. First-in- human study of cemipli- Open Access This article is licensed under a Creative Commons mab alone or in combination with radiotherapy and/or low-dose Attribution 4.0 International License, which permits use, sharing, cyclophosphamide in patients with advanced malignancies. Clin adaptation, distribution and reproduction in any medium or format, Cancer Res. 2020;26(5):1025–33. https://doi. or g/10. 1 158/1078- as long as you give appropriate credit to the original author(s) and the 0432. CCR- 19- 2609. source, provide a link to the Creative Commons licence, and indicate 13. Kitano S, Shimizu T, Koyama T, Ebata T, Iwasa S, Kondo S, et if changes were made. The images or other third party material in this al. Dose exploration results from Phase 1 study of cemiplimab, article are included in the article’s Creative Commons licence, unless a human monoclonal programmed death (PD)-1 antibody, in indicated otherwise in a credit line to the material. If material is not Japanese patients with advanced malignancies. Cancer Chem- included in the article’s Creative Commons licence and your intended other Pharmacol. 2021;87(1):53–64. https:// doi. org/ 10. 1007/ use is not permitted by statutory regulation or exceeds the permitted s00280- 020- 04161-6. use, you will need to obtain permission directly from the copyright 14. Yang F, Paccaly AJ, Rippley RK, Davis JD, DiCioccio AT. holder. To view a copy of this licence, visit http:// creat iveco mmons. Population pharmacokinetic characteristics of cemiplimab in org/ licen ses/ by/4. 0/. patients with advanced malignancies. J Pharmacokinet Phar - macodyn. 2021. https://doi. org/ 10. 1007/ s10928- 021- 09739-y. 15. Vaddepally RK, Kharel P, Pandey R, Garje R, Chandra AB. Review of indications of FDA-approved immune checkpoint inhibitors per NCCN guidelines with the level of evidence. References Cancers (Basel). 2020;12(3):738. https://doi .or g/10. 3390/ c ance rs120 30738. 1. Newswire P. Global biopharmaceuticals market expected to 16. van Schouwenburg PA, van de Stadt LA, de Jong RN, van Buren reach $526,008 million, by 2025 |CAGR 13.8% - allied market EE, Kruithof S, de Groot E, et al. Adalimumab elicits a restricted research. PR Newswire. 2016. anti-idiotypic antibody response in autoimmune patients resulting 2. US Department of Health and Human Services UF. Licensed in functional neutralisation. Ann Rheum Dis. 2013;72(1):104–9. Biological Products with Supporting Documents. 2018. Avail- https://doi. org/ 10. 1136/ annrh eumdis- 2012- 201445. able from: ht t ps: / / www . a c c e s sda t a . fda . gov/ sc ri p t s/ c de r/ da f/ 17. de Zwart M, Lausecker B, Globig S, Neddermann D, Le Bras B, index. cfm. Guenzi A, et al. Co-medication and interference testing in bioanal- 3. Dey M, Zhao SS, Moots RJ. Anti-TNF biosimilars in rheumatol- ysis: a European Bioanalysis Forum recommendation. Bioanaly- ogy: the end of an era? Expert Opin Biol Ther. 2021;21(1):29–36. sis. 2016;8(19):2065–70. https://doi. or g/10. 4155/ bio- 2016- 0179 . https://doi. org/ 10. 1080/ 14712 598. 2020. 18024 21. Vol.:(0123456789) AAPS J (2021) 23:109 18. van Schie KA, Hart MH, de Groot ER, Kruithof S, Aarden LA, 19. Wickremsinhe ER, James CA. Bioanalysis and the oncology rev- Wolbink GJ, et al. The antibody response against human and olution. Bioanalysis. 2021;13(5):291–4. https://doi. org/ 10. 4155/ chimeric anti-TNF therapeutic antibodies primarily targets bio- 2021- 0014. the TNF binding region. Ann Rheum Dis. 2015;74(1):311–4. https://doi. org/ 10. 1136/ annrh eumdis- 2014- 206237. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png "The AAPS Journal" Springer Journals

Bioanalytical Challenges due to Prior Checkpoint Inhibitor Exposure: Interference and Mitigation in Drug Concentration and Immunogenicity Assays

Loading next page...
 
/lp/springer-journals/bioanalytical-challenges-due-to-prior-checkpoint-inhibitor-exposure-MVqxpoOYoS
Publisher
Springer Journals
Copyright
Copyright © The Author(s) 2021
eISSN
1550-7416
DOI
10.1208/s12248-021-00643-4
Publisher site
See Article on Publisher Site

Abstract

AAPS J (2021) 23:109 Vol.:(0123456789) https://doi.org/10.1208/s12248-021-00643-4 Research Article Bioanalytical Challenges due to Prior Checkpoint Inhibitor Exposure: Interference and Mitigation in Drug Concentration and Immunogenicity Assays 1 1 1 1 1 1,2 Andrew F. Dengler · Rachel Weiss  · Tiffany Truong · Susan C. Irvin · Nidhi Gadhia · Mohamed Hassanein · 1 1 3 1 1 1 Camille Georgaros · Jessica‑Ann Taylor · Anne Paccaly · Giane Sumner · Matthew D. Andisik · Albert Torri · 1,4 Michael A. Partridge Received: 1 June 2021 / Accepted: 24 August 2021 Abstract Monoclonal antibodies (mAbs) are a leading class of biotherapeutics. In oncology, patients often fail on early lines of biologic therapy to a specific target. Some patients may then enroll in a new clinical trial with a mAb specific for the same target. Therefore, immunoassays designed to quantify the current mAb therapy or assess immunogenicity to the drug may be susceptible to cross-reactivity or interference with residual prior biologics. The impact of two approved anti-PD-1 mAbs, pembrolizumab and nivolumab, was tested in several immunoassays for cemiplimab, another approved anti-PD-1 mAb. The methods included a target-capture drug concentration assay, a bridging anti-drug antibody (ADA) assay and a competitive ligand-binding neutralizing antibody (NAb) assay. We also tested bioanalytical strategies to mitigate cross-reactivity or interference in these assays from other anti-PD-1 biologics. Both pembrolizumab and nivolumab cross-reacted in the cemiplimab drug concentration assay. This was mitigated by addition of antibodies specific to pembrolizumab or nivolumab. ADA specific for pembrolizumab and nivolumab did not interfere in the cemiplimab ADA assay. However, pembrolizumab and nivolumab generated a false-positive response in a target-capture NAb assay. Our results demonstrate that similar exogenous pre-existing anti-PD-1 mAbs (biotherapeutics) such as pembrolizumab and nivolumab are detected and accurately quantified in the cemiplimab drug concentration assay. However, once steady state is achieved for the new therapy, prior biologics would likely not be detected. Cross-reactivity and interference in immunoassays from previous treatment with class-specific biotherapeutic(s) pose significant bioanalytical challenges, especially in immuno-oncology. KEY WORDS clinical impact · drug concentration assay · immunogenicity · monoclonal antibody therapeutic · prior biologic exposure Andrew F. Dengler, Rachel Weiss, Tiffany Truong, and Susan C. INTRODUCTION Irvin are equal first-author contributions. Biotherapeutics and especially monoclonal antibod- Regeneron Pharmaceuticals, Bioanalytical Sciences, 777 Old Saw ies (mAbs) are one of the fastest growing segments of Mill River Rd, Tarrytown, New York 10591, USA. the pharmaceutical industry (1). This trend has acceler - Present Address: Pfizer, 401 N Middletown Rd, Pearl River, New York 10965, USA. ated with the approval of approximately 30 novel mAbs Regeneron Pharmaceuticals, Pharmacometrics (DSP), 777 Old Saw Mill between 2017 and 2020 alone ( https:// w ww. fda. gov/ River Rd, Tarrytown, New York 10591, USA.drugs/ new- drugs- fda- cders- new- molec ular- entit ies- and- To whom correspondence should be addressed. (e–mail: michael.partridge@ new- thera peutic- biolo gical- produ cts/ novel- drug- appro regeneron.com) vals- 2017,2). Abbreviations: ADA, Anti-drug antibody; CPI, Check-point inhibitor; IO, Immuno-oncology; I&I, Immunology and inflammation; mAb, Monoclonal Two therapeutic areas in particular have witnessed antibody; NAb, Neutralizing antibody; PC, Positive control; PD-1, Programmed rapid growth in recent years: cancer immunotherapies, cell death protein 1; PK, Pharmacokinetics; TE, Treatment emergent Vol.:(0123456789) Vol.:(0123456789) AAPS J (2021) 23:109 also referred to as immuno-oncology (IO), and immunol- be susceptible to cross-reactivity from different therapies ogy and inflammatory diseases (I&I, e.g., anti-TNF-α) (2 , directed to the same target. In cases where patients change to 3). This has resulted in an increasingly competitive and a new therapy of the same class before the prior therapy has crowded treatment landscape. In some cases, hundreds of been cleared, this may result in the detection of preceding biotherapeutics engaging the same targets are under inves- therapeutics (6). tigation in clinical studies (see Table I) (4, 5). Bioanalysis of samples collected from patients treated This redundancy is likely beneficial for both patients and with cemiplimab in two oncology trials revealed unexpect- health care providers as it provides multiple therapeutic options. edly high concentrations of drug detected in baseline (pre- However, it also increases the potential for patients to receive dose) samples in the target-capture cemiplimab drug con- different biotherapeutics (approved or investigational) against centration assay. The measurable drug concentrations of up the same target. Patients experiencing disease recurrence, non- to 95 µg/mL, similar to steady state cemiplimab concentra- responders, and those developing resistance may enroll in tions (12–14), could not be explained by high background, clinical trials involving related therapies; for example, patients matrix interference, analytical errors, or sample collection may be anti-PD-1 experienced when enrolling in another errors. The baseline samples with detectable drug were from anti-PD-1 study (6, 7). In the last few years, the Food and Drug patients enrolled in studies that allowed prior treatment with Administration (FDA) and European Medicines Agency (EMA) an anti-PD-1 biotherapeutic, including pembrolizumab and have approved seven immune check point inhibitors (CPIs): nivolumab. one monoclonal antibody targeting the CTLA-4 pathway Like cemiplimab, pembrolizumab, and nivolumab are (ipilimumab), three targeting PD-L1 (atezolizumab, avelumab both human IgG4 mAbs specific for PD-1 and are approved and durvalumab), and four targeting PD-1 (cemiplimab, for a variety of oncology indications (15). Since the cemipli- dostarlimab, nivolumab, and pembrolizumab), for the treatment mab drug concentration assay uses PD-1 as the capture rea- of patients with multiple cancer types (8–10). gent, and a non-specific anti-IgG4 as the detection reagent, In life-threatening diseases like cancer, patients failing to we investigated the potential for these two similar anti-PD-1 respond to one CPI may not be able to wait for clearance therapies to interfere with or cross-react in the cemiplimab of the therapeutic before enrolling in a clinical trial with a drug concentration or immunogenicity assays (14, 16, 17). treatment regimen consisting of a different CPI in combination Analysis of in vitro serum samples spiked with these therapies established that both pembrolizumab with a novel experimental therapy (11). With the limited and nivolumab could be detected in the target- number of immunoassay formats commonly used in clinical capture cemiplimab drug concentration assay. We also trials to measure drug concentrations, anti-drug antibodies demonstrated that the addition of antibodies specific to (ADA), and neutralizing antibodies (NAb), residual systemic either pembrolizumab or nivolumab could block detection drug that binds the same target may cross-react or interfere of the drugs in the assay, thus providing a potential in these assays. For example, target-capture immunoassays strategy to mitigate this cross-reactivity. In addition, we that measure the concentration for one mAb therapeutic might demonstrated that pembrolizumab, nivolumab, or anti-drug antibodies to either of these mAbs do not interfere in the cemiplimab bridging ADA assay. However, pembrolizumab Table I List of mAbs Approved or in Clinical Development for the and nivolumab generate a false-positive response in the Top 10 Targets in I&I and IO target-capture competitive ligand-binding NAb assay. With the increasing number of clinical trials with different Target Approved In clinical develop- Total biotherapeutics that engage the same targets, we anticipate ment that cross-reactivity or interference in immunoassays from EGFR 4 207 211 these biotherapeutics will be an ongoing bioanalytical HER-2 8 186 194 challenge, in assays that use target-capture and generic anti- PD-1 5 152 157 IgG detection reagents. CD20 9 130 139 PD-L1 3 132 135 TNF-α 15 118 133 CD19 6 114 120 MATERIALS AND METHODS IL-6 3 48 51 CD38 2 26 28 Materials and Reagents IL-5 3 10 13 EGFR epidermal growth factor receptor, HER-2 human epidermal Cemiplimab (Libtayo®), the recombinant human PD-1 growth factor receptor 2, PD-1 programmed cell death protein 1, PD- (extracellular domain), the monoclonal anti-cemipli- L1 programmed death ligand-1, TNF-α tumor necrosis factor α, IL-6 mab antibody (ADA positive control), the neutralizing interleukin 6, IL-5 interleukin 5 AAPS J (2021) 23:109 Vol.:(0123456789) monoclonal anti-cemiplimab antibody (NAb positive con- further diluted in the labeled reagent solution. After incuba- trol), the blocking anti-cemiplimab idiotypic antibody, and tion for 1 h, samples were transferred to blocked (5% BSA) the anti-human IgG4 monoclonal antibody, were produced Streptavidin-coated plates and incubated for 1 h, before by Regeneron Pharmaceuticals, Inc. (Tarrytown, NY). addition of Read Buffer and analysis on a QuickPlex SQ Labeling of antibodies and targets with biotin using EZ-link 120 reader (MSD, Gaithersburg, Maryland). Sulfo-NHS-LC-Biotin (Thermo Fisher Scientific), and with ruthenium NHS ester (MSD), was performed according to Target-Capture NAb Assay the manufacturer’s instructions. Anti-nivolumab and anti-pembrolizumab monoclonal anti- Microplates were coated with recombinant human PD-1 bodies were purchased from BioRad (Hercules, CA). Pem- (0.5 µg/mL) and blocked with 5% (w/v) BSA. Unless oth- brolizumab (Keytruda®) was manufactured by Merck Sharp erwise specified, serum samples were diluted tenfold in & Dohme Corp. (Kenilworth, NJ). Nivolumab (Opdivo®) 300 mM acetic acid and incubated for a minimum of 10 min was manufactured by Bristol Myer Squibb (New York, NY). and then neutralized using a capture reagent solution con- All solutions, unless otherwise specified, were prepared in taining 250 mM Tris, 20 ng/mL biotinylated-cemiplimab, assay buffer (1% BSA in 1X PBS). Read Buffer T (4X) and and 5% BSA for 1 h. This was followed by incubation of the Streptavidin-coated microplates were purchased from 100 ng/mL Neutravidin-HRP for 1 h, and finally incubated Meso Scale Discovery (MSD, Gaithersburg, Maryland). with SuperSignal ELISA Pico Chemiluminescent Sub- Glacial acetic acid (17.4 M), NeutrAvidin-HRP, SuperSig- strate, prepared according to manufacturer’s instructions, nal ELISA Pico Chemiluminescent Substrate, and 96-well for 10 min. Microplates were read on a luminescence reader microplates were purchased from Thermo Fisher Scientific (Biotek, Winooski, VT). (Waltham, MA). Trizma base (1.5 M) was purchased from Sigma (St Louis, MO). Wash solution and 10% BSA were Clinical Sample Collection purchased from SeraCare, (Milford, MA). Human serum, including that which was used for the negative quality con- Clinical serum samples were obtained from Phase 1 oncol- trols (NQCs), was purchased from BioIVT (Westbury, NY). ogy studies. Studies included patients with documented anti-PD-1 medication use prior to enrollment in the trials Immunoassay Procedures with cemiplimab. The serum samples were collected from patients prior to dosing of study drug at baseline. All incubations noted below were performed at room tem- perature (RT) unless otherwise specified. Drug Concentration Assay RESULTS Microplates were coated overnight at 4 °C with recombinant Cross‑reactivity of Anti‑PD‑1 Antibodies human PD-1 (0.5 µg/mL) and blocked with 5% (w/v) BSA in the Cemiplimab Target‑Capture Drug Concentration for a minimum of 1 h. After blocking, human serum (2%) Assay samples containing the indicated proteins were added to the microplates and incubated for 1 h. Subsequently, micro- The cemiplimab enzyme-linked immunosorbent assay plates were incubated with 100 ng/mL biotinylated mouse (ELISA) uses recombinant PD-1 as the capture reagent and anti-human IgG4 mAb for 1 h, followed by incubation with a biotinylated anti-IgG4 mAb as the detection component 100 ng/mL NeutrAvidin-HRP for 1 h, and finally incubated (Fig. 1a). Other biologics that are also PD-1 specific and with SuperSignal ELISA Pico Chemiluminescent Substrate, constructed with an IgG4 framework could potentially be prepared according to manufacturer’s instructions, for 10 detected in this target-capture method (Fig. 1b). Therefore, to 30 min. Microplates were read on a luminescence reader we set out to establish whether drug detected at baseline in (BioTek, Winooski, VT). clinical study samples was due to cross-reactivity from other IgG4 anti-PD-1 therapeutics. Bridging Immunogenicity Assay To determine whether other anti-PD-1 mAbs cross- react in the assay, twofold serial dilutions of each of the Unless otherwise specified, serum samples were diluted three mAbs (cemiplimab, pembrolizumab, and nivolumab) tenfold in 300 mM acetic acid and incubated for 30 min. were prepared in human serum at concentrations of 5 to Biotin and ruthenium labeled cemiplimab (2 µg/mL) were 0.078 µg/mL (100 to 1.56 ng/mL after minimum required prepared in assay buffer containing 150 mM Tris (unless dilution) and analyzed in the assay. The signal gener - otherwise specified), and acid-treated serum samples were ated from the serial dilutions of all three mAbs was very Vol.:(0123456789) AAPS J (2021) 23:109 of cemiplimab. This indicated that within the quantita- tive range of the assay, all anti-PD-1 mAbs present in the sample would be detected and accurately quantified with similar sensitivity. Anti‑idiotypic Blocking Antibodies Mitigate Cross‑reactivity in the Drug Concentration Immunoassay in Spiked Samples and in Baseline Clinical Trial Samples A potential strategy to minimize cross-reactivity of pembrolizumab and nivolumab in the cemiplimab ELISA is to use anti-idiotypic antibodies to block binding of the other therapeutic mAbs to PD-1 on the plate (Fig. 3a). To test this strategy, mock serum samples were created by spiking serum with cemiplimab, pembrolizumab, and nivolumab at the middle quality control level (MQC; 1 µg/mL). The mock samples were then tested in the presence or absence of anti-cemiplimab, anti-pembrolizumab, and anti-nivolumab antibodies at 100 × (100 µg/mL) the MQC concentration. The results demonstrate the anti-idiotypic blocking antibodies specifically inhibited binding of the corre- sponding drug to PD-1 on the plate diminishing detection in the cemiplimab ELISA (Fig. 3b through d). The anti- idiotypic antibodies did not cross-react or interfere with quantification of the other mAbs in the assay. This strategy could also be used to confirm the identity of any anti-PD-1 mAb in baseline clinical samples from patients previously treated with an anti-PD-1 mAb. To evaluate this approach, baseline samples collected from patients with prior anti-PD-1 exposure to either pem- Fig. 1 Illustrations of the functional cemiplimab drug concentration assay and cross-reactivity in serum. a Cemiplimab (cemi), captured brolizumab or nivolumab were analyzed in the presence on a PD-1 coated microplate is detected by a biotinylated anti-human and absence of each of the three anti-idiotypic antibod- IgG4 mAb followed by NeutrAvidin-HRP b Other anti-PD-1 human ies (Fig. 3e). In every sample, assay signal was markedly IgG4 mAbs such as nivolumab (nivo) or pembrolizumab (pembro) inhibited (greater than 80%) by only the anti-idiotypic could also be detected antibody that corresponded to each patient’s anti-PD-1 medication history (Fig. 3e). similar, indicating that they can all be detected in the Potential Impact of Other Anti‑PD‑1 Biologics assay. Furthermore, analyte recovery of pembrolizumab on the Specificity and Selectivity of the Cemiplimab  and nivolumab concentrations when interpolated from the ADA Assay cemiplimab standard curve generated values within 20% of the nominal values, thus demonstrating that the mAbs Prior exposure to the same class of anti-PD-1 mAb raises can be accurately quantified in this assay (Fig. 2a). the possibility that some patients previously treated with To determine if there was an additive effect of pem- pembrolizumab or nivolumab may generate ADA that brolizumab or nivolumab in the cemiplimab ELISA, cemi- cross-react in the anti-cemiplimab ADA assay. The bridging plimab at the high quality control level (HQC; 75 ng/mL) cemiplimab ADA assay uses a mouse anti-cemiplimab or the lower level of quantification (LLOQ; 1.56 ng/mL) antibody as the positive control and biotinylated-cemiplimab was added to the serial dilutions of pembrolizumab and and ruthenium-labeled cemiplimab as bridge components nivolumab (Fig. 2b and c). When interpolated off the cemi - (Fig. 4a). plimab standard curve, the concentration of detected drug Serum positive control samples were prepared con- was equal to the sum of pembrolizumab or nivolumab plus taining specific anti-cemiplimab, anti-pembrolizumab, or the HQC (Fig. 2b and c) or LLOQ (data not shown) level AAPS J (2021) 23:109 Vol.:(0123456789) Fig. 2 Other human IgG4 anti-PD-1 mAbs can be detected and accurately quantified in the functional cemiplimab drug concentration assay in serum. a Quantitation of serial dilutions (100 to 1.56 ng/mL) of cemiplimab (blue), pembroli- zumab (green), and nivolumab (orange) in the assay. b Drug concentrations of HQC samples spiked with serial dilutions of pembrolizumab, interpolated from the cemiplimab standard curve, in the cemiplimab drug concentration ELISA c Drug concentrations of HQC samples spiked with serial dilutions of nivolumab, interpolated from the cemiplimab standard curve, in the cemiplimab drug concen- tration assay anti-nivolumab antibodies and were analyzed in the ADA Collectively, these results demonstrate the specificity and assay. Anti-cemiplimab positive control samples generated suitability of our cemiplimab ADA assay for the detection a strong signal in the assay, while the anti-pembrolizumab of anti-cemiplimab antibodies in the presence of other anti- or anti-nivolumab samples generated signal approximately PD-1 ADA or residual anti-PD-1 therapeutics. equivalent to the negative control samples (Fig. 4b). This suggests that anti-pembrolizumab or anti-nivolumab anti- bodies generated in patients treated with these drugs likely Interference from Other PD‑1 Therapies will not interfere with the detection of anti-cemiplimab on a Target‑Capture NAb Assay antibodies. To test whether residual concentrations of pembroli- Confirmed ADA positive samples were further assessed zumab or nivolumab in circulation can impact the detection in a NAb assay to evaluate the ability to neutralize the of cemiplimab ADA, samples containing an anti-cemiplimab biological activity of the drug. A competitive ligand- monoclonal antibody (500 ng/mL) were tested in the pres- binding NAb assay was developed that uses recombinant ence of increasing concentrations of either pembrolizumab PD-1 as the capture reagent and biotinylated-cemiplimab or nivolumab. The highest concentration of each drug tested and streptavidin-HRP as the detection components (2 mg/mL), was greater than the Cmax levels observed in (Fig. 5a). When present in a serum sample, NAbs will the clinical study samples (12, 13). These results demon- bind to biotinylated cemiplimab, preventing binding to strated that even at high concentrations of pembrolizumab or the PD-1 coated microplate and inhibiting the assay signal nivolumab, detection of the anti-cemiplimab antibody was (Fig. 5a). However, in this assay format, the presence not impacted by the presence of either antibody in serum of other anti-PD-1 biologics could also compete with (Fig. 4c). biotinylated cemiplimab for PD-1 binding, potentially As a control, cemiplimab was also spiked at high generating a false-positive NAb result (Fig. 5b). concentrations in the assay. As expected, this reduced the To t est this, cemi pl im ab, pe mbroli zumab, and anti-cemiplimab antibody assay signal, although control nivolumab were serially diluted in serum from 4000 samples (500 ng/mL) remained positive in the assay when to 31.3 ng/mL and analyzed in the target-capture NAb spiked with cemiplimab at concentrations greater than assay. As demonstrated in Fig. 5c, a false-positive 500 µg/mL, confirming the cemiplimab drug tolerance level NAb signal was detected when approximately 155 ng/ of the assay (Fig. 4c). These experiments demonstrate that mL of any of these anti-PD-1 drugs were added to the cemiplimab ADA assay is specific only for anti-drug the competitive ligand-binding NAb assay, which is antibodies directed to the variable domain of cemiplimab approximately 1000-fold lower than steady state drug and is not impacted by the presence of other anti-PD-1 mAbs. concentrations (12–14). In contrast, excess cemiplimab Vol.:(0123456789) AAPS J (2021) 23:109 Fig. 3 Anti-idiotypic antibod- ies block binding of anti-PD-1 mAbs in the cemiplimab target- capture drug concentration ELISA. a Schematic depicting the anti-idiotypic antibodies (checkerboard pattern) blocking each drug (solid colors) from binding to the PD-1 capture in the cemiplimab ELISA. b Mock sample serum control prepared with cemiplimab at the MQC concentration and tested in the presence or absence of 100 × concentration of all three anti-idiotypic antibod- ies in the cemiplimab ELISA. c Mock sample serum control prepared with pembrolizumab and tested in the presence of 100X concentration of the anti-pembrolizumab and the anti-cemiplimab antibodies in the cemiplimab ELISA. d Mock sample serum control prepared with nivolumab and tested in the presence of 100 × concentra- tion of the anti-nivolumab and the anti-cemiplimab antibod- ies in the cemiplimab ELISA. e Baseline clinical samples with detectable responses in the cemiplimab ELISA from patients with prior exposure to pembrolizumab or nivolumab were evaluated in the presence of each of the three anti-idio- typic antibodies to demonstrate specific signal inhibition (or other anti-PD-1 mAbs) do not generate false-positive DISCUSSION responses in a drug-capture competitive ligand-binding NAb assay, as excess therapeutic from the previous There are numerous therapeutic targets for which multiple capture step is washed away before addition of labeled biologic drugs are approved, including products target- target as detection reagent (not shown). ing TNF-α, PD-1, or PD-L1 (Table I) (15). Because these AAPS J (2021) 23:109 Vol.:(0123456789) Fig. 4 The cemiplimab ADA assay is specific only for anti-cemi - cemiplimab, anti-nivolumab or anti-pembrolizumab antibodies at plimab antibodies, and other anti-PD-1 mAbs do not interfere in three concentrations (X, Y, and Z ng/mL) in serum. c Signal in the the method. a Schematic of the cemiplimab bridging ADA assay in ADA assay of anti-cemiplimab antibody control samples (500 ng/ which ADA in the samples bridge between biotin- and ruthenium- mL) tested in the presence of serial dilutions of cemiplimab, pem- labeled cemiplimab generating signal in the method. b Signal-to- brolizumab, and nivolumab at the indicated concentrations noise ratio in the ADA assay for control samples containing anti- targets are clinically validated, many biotherapeutics are the respective drugs. Using this approach, we identified being investigated in novel combinations (or as biosimi- patients that had previously received either pembrolizumab lars) for the treatment of new indications or to improve effi- or nivolumab, which aligned with patient’s prior medica- cacy in existing populations. During development of these tion history. new drugs, in particular for oncology therapies in the PD-1 In addition to investigating the functional cemiplimab and PD-L1 classes, trial participants may transition to the drug concentration assay, we wanted to understand whether investigational biotherapy (or combination) while still hav- pre-existing pembrolizumab or nivolumab, or ADA directed ing detectable systemic concentrations of their prior therapy against either pembrolizumab or nivolumab, could interfere (10). These prior drugs have the potential to cross-react or in the cemiplimab immunogenicity assays. In the bridging interfere in bioanalytical assays for the new therapy, espe- ADA assay, the presence of pembrolizumab or nivolumab cially with target-capture-based methods. did not interfere with the detection of the anti-cemiplimab In cemiplimab clinical studies, enrollment of patients positive control. Furthermore, only antibodies specific for who had received prior anti-PD-1 therapy was permitted cemiplimab were detected in this assay, with no cross- in some cohorts. Baseline samples (taken prior to cemi- reactivity from specific anti-pembrolizumab or anti- plimab administration) for some of these patients had nivolumab antibodies. Although these results collectively high drug levels in the cemiplimab drug concentration confirm the specificity of our anti-cemiplimab ADA assay, assay. We demonstrated that the assay was able to detect it does not negate the possibility that ADA against other pembrolizumab or nivolumab in spiked samples, with anti-PD-1 therapeutics may already exist but cannot be approximately equivalent quantitation to cemiplimab. In detected. addition, we were able to specifically inhibit assay signal In the competitive ligand-binding NAb assay, neither by addition of anti-idiotype mAbs directed against each of pembrolizumab nor nivolumab interfered in the assay Vol.:(0123456789) AAPS J (2021) 23:109 Fig. 5 Anti-PD-1 biologics generate a false-positive response in a b Schematic of a false-positive NAb response in the presence of target-capture NAb assay in serum. a Schematic depicting the target- anti-PD-1 mAbs that bind to the PD-1 coated plate preventing bioti- capture NAb assay. In the absence of NAb, biotinylated cemiplimab nylated-cemplimab from generating signal in the assay. c Assay sig- binds to a PD-1 coated plate, followed by streptavidin conjugated nal inhibition (%Inhibition) in a cemiplimab target-capture NAb after to HRP, generating signal in the assay. The presence of NAb inhib- addition of serially-diluted cemiplimab, nivolumab, or pembroli- its biotin-cemiplimab binding to PD-1, resulting in signal reduction. zumab, at concentrations ranging from 4000 to 31.25 ng/mL when configured in a drug-capture format (not shown), CONCLUSION since these molecules were washed away before addition of the labeled target. However, in a target-capture format, the We demonstrated that prior exposure to anti-PD-1 mAbs, presence of either pembrolizumab or nivolumab generated pembrolizumab and nivolumab, can be detected and quan- a false-positive NAb response due to target binding and titated in the cemiplimab drug concentration immunoassay. the resulting competition with the biotinylated-cemiplimab However, at steady state for the new therapy, these prior detection antibody, even at relatively low concentrations. biotherapeutics would likely not be detected. In addition, Therefore, a drug-capture format is recommended for we demonstrated that these prior biotherapeutics or anti- competitive ligand-binding NAb assays to avoid false- drug antibodies to either of these mAbs do not interfere in positive responses from other biotherapeutics to the same the cemiplimab bridging ADA assay. However, in a target- target (17). capture competitive ligand-binding cemiplimab NAb assay, It is unknown whether true treatment-emergent anti- pembrolizumab and nivolumab generated a false-positive pembrolizumab or anti-nivolumab immunogenicity cor- response. relates to subsequent ADA responses to other anti-PD-1 Immunoassays that use the drug target as a reagent in mAbs. However, ADA to fully human or humanized the assay, especially in the capture step, are potentially mAbs are predominantly directed to the variable domains, susceptible to interference or cross-reactivity with other and ADA directed to the unique regions of one molecule biologics directed to the same target. Some of the most would be unlikely to cross-react with the other molecules widely prescribed pharmaceutical products are biologics (8, 16, 18). Furthermore, immunogenicity to one mAb with the same target, most prominently drugs targeting may not be predictive of an ADA response after subse- TNF-α, PD-1, or PD-L1. In oncology particularly, there quent exposure to a different mAb that binds to the same are additional targets (other than PD-1/PD-L1) that have target. more than one approved therapy, including CD20, EGFR, AAPS J (2021) 23:109 Vol.:(0123456789) 4. Database T-TA. https:// tabs. craic. com/ antib odies. Accessed 16 and HER-2 (Table I) (15). Cross-reactivity or interference Dec 2020. in immunoassays from previous exposure to biotherapeu- 5. Upadhaya S, Hubbard-Lucey VM, Yu JX. Immuno-oncology tics of the same class is an ongoing bioanalytical chal- drug development forges on despite COVID-19. Nat Rev lenge with molecules directed to these targets (17, 19). Drug Discov. 2020;19(11):751–2. ht t ps: / / doi . or g/ 10. 103 8/ d41573- 020- 00166-1. This work highlights the importance of understanding both 6. Fujita K, Uchida N, Kanai O, Okamura M, Nakatani K, patient medication history and assay format, to ensure the Mio T. Retreatment with pembrolizumab in advanced non- most appropriate bioanalytical strategies are implemented. small cell lung cancer patients previously treated with nivolumab: emerging reports of 12 cases. Cancer Chemother Acknowledgements We thank the patients who participated in these Pharmacol. 2018;81(6):1105–9. https:// doi. org/ 10. 1007/ trials and the investigators at the individual study sites. We thank the s00280- 018- 3585-9. rest of the Bioanalytical Sciences team members for their essential 7. Haanen J, Ernstoff M, Wang Y, Menzies A, Puzanov I, Grivas P, contributions to all of the cemiplimab oncology studies. et al. Rechallenge patients with immune checkpoint inhibitors following severe immune-related adverse events: review of the literature and suggested prophylactic strategy. J Immunother Can- Author Contribution AFD, RW, TT, SCI, MH, MDA, and MAP cer. 2020;8(1):e000604. https://d oi.o rg/1 0.1 136/j itc-2 020-0 00604. conceptualized the analysis and performed the data analysis. AP ana- 8. Wongchenko MJ, Ribas A, Dreno B, Ascierto PA, McArthur GA, lyzed the pharmacokinetic data and oversaw pharmacokinetic interpre- Gallo JD, et al. Association of programmed death ligand-1 (PD- tation. AFD, RW, and MDA oversaw the sample analysis. CG, JAT, and L1) expression with treatment outcomes in patients with BRAF GS conceptualized, developed, and validated the cemiplimab PK assay. mutation-positive melanoma treated with vemurafenib or cobi- TT and NG performed the sample analysis. AT oversaw the bioanalyti- metinib combined with vemurafenib. Pigment Cell Melanoma cal plans and interpretation. AFD, RW, TT, MH, SCI, MDA, and MAP Res. 2018;31(4):516–22. https://doi. org/ 10. 1111/ pcmr. 12670. wrote the manuscript. 9. Menzies AM, Johnson DB, Ramanujam S, Atkinson VG, Wong ANM, Park JJ, et al. Anti-PD-1 therapy in patients with Funding This research was funded by Regeneron Pharmaceuticals, advanced melanoma and preexisting autoimmune disorders or Inc. major toxicity with ipilimumab. Ann Oncol. 2017;28(2):368– 76. https://doi. org/ 10. 1093/ annonc/ mdw443. Declarations 10. Gentzler R, Hall R, Kunk PR, Gaughan E, Dillon P, Slingluff CL Jr, et al. Beyond melanoma: inhibiting the PD-1/PD-L1 pathway Disclosure The views expressed in this article are those of the in solid tumors. Immunotherapy. 2016;8(5):583–600. https:// authors and do not reflect official policy of Regeneron Pharmaceuticals. doi. org/ 10. 2217/ imt- 2015- 0029. 11. Murciano-Goroff YR, Warner AB, Wolchok JD. The future of Conflict  of Interest All authors are employees of Regeneron Phar- cancer immunotherapy: microenvironment-targeting combina- maceuticals, Inc., and may hold stock and/or stock options in the com- tions. Cell Res. 2020;30(6):507–19. https:// doi. org/ 10. 1038/ pany. s41422- 020- 0337-2. 12. Papadopoulos KP, Johnson ML, Lockhart AC, Moore K, Fal- chook GS, Formenti SC, et al. First-in- human study of cemipli- Open Access This article is licensed under a Creative Commons mab alone or in combination with radiotherapy and/or low-dose Attribution 4.0 International License, which permits use, sharing, cyclophosphamide in patients with advanced malignancies. Clin adaptation, distribution and reproduction in any medium or format, Cancer Res. 2020;26(5):1025–33. https://doi. or g/10. 1 158/1078- as long as you give appropriate credit to the original author(s) and the 0432. CCR- 19- 2609. source, provide a link to the Creative Commons licence, and indicate 13. Kitano S, Shimizu T, Koyama T, Ebata T, Iwasa S, Kondo S, et if changes were made. The images or other third party material in this al. Dose exploration results from Phase 1 study of cemiplimab, article are included in the article’s Creative Commons licence, unless a human monoclonal programmed death (PD)-1 antibody, in indicated otherwise in a credit line to the material. If material is not Japanese patients with advanced malignancies. Cancer Chem- included in the article’s Creative Commons licence and your intended other Pharmacol. 2021;87(1):53–64. https:// doi. org/ 10. 1007/ use is not permitted by statutory regulation or exceeds the permitted s00280- 020- 04161-6. use, you will need to obtain permission directly from the copyright 14. Yang F, Paccaly AJ, Rippley RK, Davis JD, DiCioccio AT. holder. To view a copy of this licence, visit http:// creat iveco mmons. Population pharmacokinetic characteristics of cemiplimab in org/ licen ses/ by/4. 0/. patients with advanced malignancies. J Pharmacokinet Phar - macodyn. 2021. https://doi. org/ 10. 1007/ s10928- 021- 09739-y. 15. Vaddepally RK, Kharel P, Pandey R, Garje R, Chandra AB. Review of indications of FDA-approved immune checkpoint inhibitors per NCCN guidelines with the level of evidence. References Cancers (Basel). 2020;12(3):738. https://doi .or g/10. 3390/ c ance rs120 30738. 1. Newswire P. Global biopharmaceuticals market expected to 16. van Schouwenburg PA, van de Stadt LA, de Jong RN, van Buren reach $526,008 million, by 2025 |CAGR 13.8% - allied market EE, Kruithof S, de Groot E, et al. Adalimumab elicits a restricted research. PR Newswire. 2016. anti-idiotypic antibody response in autoimmune patients resulting 2. US Department of Health and Human Services UF. Licensed in functional neutralisation. Ann Rheum Dis. 2013;72(1):104–9. Biological Products with Supporting Documents. 2018. Avail- https://doi. org/ 10. 1136/ annrh eumdis- 2012- 201445. able from: ht t ps: / / www . a c c e s sda t a . fda . gov/ sc ri p t s/ c de r/ da f/ 17. de Zwart M, Lausecker B, Globig S, Neddermann D, Le Bras B, index. cfm. Guenzi A, et al. Co-medication and interference testing in bioanal- 3. Dey M, Zhao SS, Moots RJ. Anti-TNF biosimilars in rheumatol- ysis: a European Bioanalysis Forum recommendation. Bioanaly- ogy: the end of an era? Expert Opin Biol Ther. 2021;21(1):29–36. sis. 2016;8(19):2065–70. https://doi. or g/10. 4155/ bio- 2016- 0179 . https://doi. org/ 10. 1080/ 14712 598. 2020. 18024 21. Vol.:(0123456789) AAPS J (2021) 23:109 18. van Schie KA, Hart MH, de Groot ER, Kruithof S, Aarden LA, 19. Wickremsinhe ER, James CA. Bioanalysis and the oncology rev- Wolbink GJ, et al. The antibody response against human and olution. Bioanalysis. 2021;13(5):291–4. https://doi. org/ 10. 4155/ chimeric anti-TNF therapeutic antibodies primarily targets bio- 2021- 0014. the TNF binding region. Ann Rheum Dis. 2015;74(1):311–4. https://doi. org/ 10. 1136/ annrh eumdis- 2014- 206237. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Journal

"The AAPS Journal"Springer Journals

Published: Oct 4, 2021

Keywords: clinical impact; drug concentration assay; immunogenicity; monoclonal antibody therapeutic; prior biologic exposure

References