Advertisement

Role of Modeling and Simulation in the Development of Novel and Biosimilar Therapeutic Proteins

  • Yow-Ming C. Wang
    Correspondence
    Correspondence to: Yow-Ming C. Wang (Telephone: +1-301-796-0387).
    Affiliations
    Office of Clinical Pharmacology, Office of Translational Sciences, Center for Drug Evaluation and Research, U.S. Food and Drug Administration, Silver Spring, Maryland 20993
    Search for articles by this author
  • Yaning Wang
    Affiliations
    Office of Clinical Pharmacology, Office of Translational Sciences, Center for Drug Evaluation and Research, U.S. Food and Drug Administration, Silver Spring, Maryland 20993
    Search for articles by this author
  • Sarah J. Schrieber
    Affiliations
    Office of Clinical Pharmacology, Office of Translational Sciences, Center for Drug Evaluation and Research, U.S. Food and Drug Administration, Silver Spring, Maryland 20993
    Search for articles by this author
  • Justin Earp
    Affiliations
    Office of Clinical Pharmacology, Office of Translational Sciences, Center for Drug Evaluation and Research, U.S. Food and Drug Administration, Silver Spring, Maryland 20993
    Search for articles by this author
  • Theingi M. Thway
    Affiliations
    Office of Clinical Pharmacology, Office of Translational Sciences, Center for Drug Evaluation and Research, U.S. Food and Drug Administration, Silver Spring, Maryland 20993
    Search for articles by this author
  • Shiew Mei Huang
    Affiliations
    Office of Clinical Pharmacology, Office of Translational Sciences, Center for Drug Evaluation and Research, U.S. Food and Drug Administration, Silver Spring, Maryland 20993
    Search for articles by this author
  • Issam Zineh
    Affiliations
    Office of Clinical Pharmacology, Office of Translational Sciences, Center for Drug Evaluation and Research, U.S. Food and Drug Administration, Silver Spring, Maryland 20993
    Search for articles by this author
  • Leah Christl
    Affiliations
    Therapeutic Biologics and Biosimilar Staff, Office of New Drug, Center for Drug Evaluation and Research, U.S. Food and Drug Administration, Silver Spring, Maryland 20993
    Search for articles by this author
Published:November 02, 2018DOI:https://doi.org/10.1016/j.xphs.2018.10.053

      Abstract

      Modeling and simulation (M&S) is an important enabler of knowledge integration in novel biological product development programs. Given the volume of data generated from clinical trials and the complexity of pharmacokinetic (PK) and pharmacodynamic (PD) properties for reference products, extending the use of M&S to biosimilar development is logical. Assessing PK and PD similarity is normally a critical part of demonstrating biosimilarity to a reference product. Thoughtful considerations are necessary in study design to minimize the PK and PD variability, thereby increasing the sensitivity for detecting potential differences between products. In addition, the sensitivity of PD biomarkers depends partly on their relevance to the mechanism(s) of action and the dynamic range of PD response(s), including the impact of certain structural differences on PD in the relevant population. As such, opportunities exist for leveraging the available M&S knowledgebase to maximize the efficiency in the design and interpretation of PK and PD similarity studies. This article describes M&S applications which have contributed to and can continue to enhance biosimilar development programs.

      Keywords

      Abbreviations used:

      ANC (absolute neutrophil count), BLA (biological license application), CDER (center for drug evaluation and research), FDA (food and drug administration), M&S (modeling and simulation), PD (pharmacodynamic), PHS (public health services), PK (pharmacokinetic)

      Introduction

      To obtain marketing authorization for a biological product, drug developers must submit a biological license application (BLA) either as a 351(a) BLA for novel or “stand-alone” products (as per the Public Health Services [PHS] Act section 351(a)) or as a 351(k) BLA for biosimilar and interchangeable products (as per the PHS Act section 351(k)). The regulatory requirements for these 2 types of BLAs differ. The 351(a) BLAs must contain all required data and information necessary to demonstrate the safety, purity, and potency (i.e., safety and effectiveness) of the proposed biological product, whereas 351(k) BLAs must contain data demonstrating biosimilarity or interchangeability of the proposed biosimilar or interchangeable product to a biological product licensed under 351(a) of the PHS Act in the United States (U.S.-licensed reference product). The 351(k) licensure pathway is an abbreviated pathway created by the Biologics Price Competition and Innovation Act of 2009 in which a manufacturer that shows its proposed biological product is biosimilar to or interchangeable with an U.S. Food and Drug Administration (FDA)–approved reference product may rely, in part, on the FDA's previous determination of safety and effectiveness for the reference product for approval.
      Biological products include a wide range of products such as vaccines, blood and blood components, allergenics, somatic cells, gene therapy, tissues, and recombinant therapeutic proteins. They are currently regulated by either the Center for Biologics Evaluation and Research or the Center for Drug Evaluation and Research (CDER) within FDA. The biological products regulated by the CDER are mostly produced by biotechnology methods and are broadly referred to as therapeutic proteins. Therapeutic proteins include, but are not exclusive to, the following: monoclonal antibodies designed as targeted therapies, cytokines, growth factors, enzymes, such as thrombolytics, and immunomodulators. The scope of our discussions will be limited to therapeutic proteins.
      During development of novel therapeutic proteins, modeling and simulation (M&S) tools are essential in characterizing the pharmacokinetic (PK) profiles and the dose-exposure relationship of investigational products. Population PK (PopPK) modeling analysis,
      • Ette E.I.
      • Williams P.J.
      Population pharmacokinetics I: background, concepts, and models.
      an M&S approach, is almost always a component of recent 351(a) BLA submissions. For products that were approved at a time when PopPK analyses were not a common practice, literature reports of such analyses are often available although they may not be associated with the clinical trial data used to support the product approval. As an example, Chatelut et al.
      • Chatelut E.
      • Rostaing L.
      • Gregoire N.
      • et al.
      A pharmacokinetic model for alpha interferon administered subcutaneously.
      published a PopPK modeling report for interferon alfa-2b in 1999 which is 13 years after the initial U.S. approval. Generally, published literature contains a collection of various M&S analyses conducted by scientists in the industry, academia, and regulatory agencies based on data collected during drug development stages and through real-world experience after regulatory approval. Collectively, these sources constitute an information-rich M&S knowledgebase for the reference product. Hence, the investment in M&S in a biosimilar program, which are likely to be initiated years after commercialization of the respective reference products, can logically focus on applying the available knowledge to optimize their development programs.
      • Dodds M.
      • Chow V.
      • Markus R.
      • Perez-Ruixo J.J.
      • Shen D.
      • Gibbs M.
      The use of pharmacometrics to optimize biosimilar development.
      In this article, we will provide our perspectives on leveraging the publicly available M&S knowledgebase to support biosimilar development. First, we introduce the regulatory pathway for novel or stand-alone biological products and for biosimilar products, respectively. Then, we present a summary of clinical pharmacology studies used to support FDA-approved 351(k) BLA applications; and finally, we discuss the role of M&S approaches in the context of both novel and biosimilar product development.

      Regulatory Approval of Biological Products

       351(a) BLAs—Novel or “Stand-Alone” Products

      To demonstrate that an investigational product is safe and effective for the intended use, drug development programs commonly include multiple clinical trials. These programs generally begin with evaluating PK, pharmacodynamic (PD), safety, and tolerability in ascending-dose phase 1 studies. Subsequent phase 2 studies generally evaluate multiple doses and dose regimens for the treatment effects and PD responses with the aim to identify the appropriate dose(s) and regimen(s) that can be studied in larger scale confirmatory safety and efficacy (or phase 3) trials. Table 1 summarizes these traditional drug development phases and the role of M&S. The cumulative data package in any given 351(a) BLA ultimately needs to demonstrate that the drug is safe and effective for its proposed use(s) and that the benefits of the drug outweigh the risks, which includes providing adequate scientific justifications that support the dosing recommendations.
      Table 1Comparison of Therapeutic Protein Development Programs Under 2 Regulatory Pathways and the Role of Modeling and Simulation (M&S)
      Variable351(a) Novel Product or “Stand-Alone” Development Programs351(k) Biosimilar Product Development Programs
      Potential clinical studies
      • Phase 1 single and multiple ascending-dose study to establish tolerability and characterize PK and PD properties
      • Phase 2 dose-finding study to establish dose-response and identify doses and regimens for evaluation in phase 3
      • Phase 3 safety and efficacy study to provide evidence of safety and effectiveness at the recommended dose and dose regimen
      • Comparative PK (and PD) study to establish PK (and PD) similarity between a biosimilar product and the reference product
      • Comparative clinical study to address residual uncertainties and support biosimilarity, that is, no clinically meaningful differences between a biosimilar product and the reference product
      Utility of M&S
      • 1
        Characterize PK and dose-exposure relationship with PK modeling
      • 2
        Characterize PD properties, evaluate PK-PD and dose-response relationship with PK-PD modeling
      • 3
        Identify covariates contributing to between-subject variability in PK and PD with population PK and PK-PD modeling
      • 4
        Identify alternate primary endpoints or biomarkers for populations where it may not be practical to evaluate other endpoints
      • 5
        Support study design, for example, dose selection, with model-based simulations and trial duration and sample size with disease progression models
      • 6
        Integrate data to justify the recommended dose for clinical use
      • 7
        Alleviate the need for additional clinical efficacy studies through exposure-matching approaches in pediatrics, when appropriate
      • a
        Identify dose for PK similarity study based on dose-exposure relationship
      • b
        Identify dose for PK and PD similarity study based on dose-response relationship
      • c
        Optimize study design based on the understanding of between-subject PK and PD variability in study design, for example, sample size, study subject enrollment criteria
      • d
        Support PD biomarker selection
      • e
        Provide supportive evidence and justification that a study design will have the ability to detect clinically meaningful differences between a biosimilar and the reference product

       351(k) BLAs—Biosimilar and Interchangeable Products

      A 351(k) BLA needs to contain data and information to demonstrate that the proposed product is biosimilar to or interchangeable with the reference product which is approved by FDA and licensed under 351(a) of the PHS Act. It is important to note that the abbreviated regulatory pathway for biological products is different from the Abbreviated New Drug Application regulatory pathway for generic drugs. Through multiple guidance documents,
      U.S. Food and Drug Administration
      FDA's Biosimilarity Guidances.
      FDA has articulated (a) the step-wise approach to demonstrate biosimilarity; (b) the totality-of-the-evidence approach to regulatory review of biosimilar applications; (c) the general scientific principles in conducting comparative analyses covering structural and functional assessments (the foundation), animal testing, human PK and PD (when applicable) studies, clinical immunogenicity assessments, and additional clinical studies; and (d) the general considerations for conducting switching studies to support a demonstration of interchangeability.
      Clinical pharmacology data (i.e., PK and PD data) are a critical part of demonstrating biosimilarity by providing evidence to support that there are no clinically meaningful differences between the biosimilar product and the reference product.
      U.S. Food and Drug Administration
      FDA's Guidance for Industry: Clinical Pharmacology Data to Support a Demonstration of Biosimilarity to a Reference Product.
      The assessments of PK and PD similarity are appropriate to support a demonstration of biosimilarity because PK and PD endpoints can be sensitive for detecting differences between the products and are mechanistically relevant to the activity of the therapeutic biologic under study. These PK and PD similarity studies can be conducted in patient populations or in healthy subjects, whichever is more sensitive to detect clinically meaningful differences between products. Therefore, biosimilar product development programs generally contain at least 1 PK (and PD when available) similarity study and 1 comparative clinical study that assesses immunogenicity (Table 1). Similarly, PK and PD data are important to support a demonstration of interchangeability between a proposed product and the reference product.

       Clinical Pharmacology Data in FDA-Approved 351(k) BLAs

      As of September 14, 2018, the CDER list of licensed biological products,
      U.S. Food and Drug Administration
      CDER list of licensed biological products.
      published on the FDA's purple book website, has 144 therapeutic proteins approved as 351(a) BLAs and 12 products approved as 351(k) BLAs. The approved 351(k) BLAs include 3 infliximab biosimilars, 2 biosimilars each for filgrastim and adalimumab, and 1 biosimilar each for pegfilgrastim, epoetin alfa, etanercept, trastuzumab, and bevacizumab. Table 2 shows a summary of the clinical pharmacology data supporting the approval of these 12 biosimilar products. Eight of the 12 biosimilar programs included only PK similarity assessments because suitable PD biomarkers were not available for assessing PD similarity. The remaining 4 programs, namely for biosimilars to filgrastim, pegilgrastim, and epoetin alfa, included both PK and PD similarity assessments. Both filgrastim and epoetin alfa biosimilar programs evaluated PD similarity based on 2 distinct PD biomarkers, one following a single dose and the other following multiple doses. The PK and PD similarity studies are either with a crossover design or a parallel design and generally evaluated 1 single-dose level and 1 single-dose regimen in the case of multiple dose studies.
      Table 2Clinical Pharmacology Studies in the BLA of 12 Approved Biosimilar Products as of September 14, 2018
      Study DesignProduct Name (Trade Name [Proper Name])Dose RegimenReference
      PK similarity study
       SD—crossoverErelzi (etanercept-szzs)50 mg SCPMID: 27790726
       SD—parallelInflectra (infliximab-dyyb)5 mg/kg IVPMID: 26395834
      Renflexis (infliximab-abda)5 mg/kg IVPMID: 26577771
      Ixifi (infliximab-qbtx)10 mg/kg IVPMID: 29504427
      Amjetiva (adalimumab-atto)40 mg SCPMID: 27466231
      Cyltezo (adalimumab-adbm)40 mg SCPMID: 27813422
      Mvasi (bevacizumab-awwb)3 mg/kg IVPMID: 28864922
      Ogivri (trastuzumab-dkst)8 mg/kg IVSee FDA
      FDA Briefing Document for the Meeting of Oncology Drug Advisory Committee held on July 13, 2017 (https://www.fda.gov/downloads/AdvisoryCommittees/CommitteesMeetingMaterials/Drugs/OncologicDrugsAdvisoryCommittee/UCM566369.pdf).
      PK and PD similarity study
       SD—crossover (PK & PD)

      MD—parallel (PD)
      Zarxio (filgrastim-sndz)10 mcg/kg SC SD

      2.5-10 mcg/kg SC daily ×7
      See FDA
      FDA's review (https://www.accessdata.fda.gov/drugsatfda_docs/nda/2015/125553Orig1s000ClinPharmR.pdf).
      Retacrit (epoetin alfa-epbx)100 U/kg SC SD

      3 times/wk SC ×28 d
      PMID: 27473384

      PMID: 27112531
       SD—crossover (PK and PD)

      MD—crossover (PD)
      Nivestym (filgrastim-aafi)5 mcg/kg SC SD

      5 mcg/kg SC daily ×5
      See FDA
      FDA's review (https://www.fda.gov/downloads/Drugs/DevelopmentApprovalProcess/DevelopmentResources/UCM617301.pdf).
       SD—crossoverFulphila (pegfilgrastim-jmdb)2 mg SCPMID: 29671069
      All studies were conducted on healthy subjects.
      IV, intravenous; MD, multiple dose; SC, subcutaneous; SD, single dose.
      a FDA Briefing Document for the Meeting of Oncology Drug Advisory Committee held on July 13, 2017 (https://www.fda.gov/downloads/AdvisoryCommittees/CommitteesMeetingMaterials/Drugs/OncologicDrugsAdvisoryCommittee/UCM566369.pdf).
      The PK similarity studies for filgrastim, pegfilgrastim, epoetin alfa, and etanercept are in crossover design because these products have a short half-life and low immunogenicity potential. Whereas the remaining products used a parallel study design for the PK similarity studies because they exhibit a long half-life. Whether a crossover design is suitable for the PD similarity studies depends on the temporal profile of PD responses. After a single dose of filgrastim or pegfilgrastim, the absolute neutrophil count (ANC) rises above the baseline and then returns to baseline within a relatively short time which is within the timeframe of a reasonable washout period for the PK profile. Therefore, a study with the crossover design is feasible for simultaneous evaluations of similarity in PK and PD with ANC as the biomarker. On the other hand, repeated dosing of filgrastim is needed to attain a robust response on CD34 + cell count, a biomarker for the mobilization of hematopoietic progenitor cells into the peripheral blood. Therefore, the assessment of similarity in temporal profile of CD34 + cell count requires a parallel study design. Similarly, a crossover study design is suitable for the PD response to epoetin alfa measured by the reticulocyte count which has a quick onset and short duration of response, whereas a parallel study design is suitable for the PD response measured by the hemoglobin level which has a slow onset and requires repeated dosing for a robust response.

      Roles of M&S Dependent on the Regulatory Pathway

       Supporting 351(a) BLA by Acquiring Knowledge

       Integrating Data by PK and PK-PD Modeling

      Throughout a novel product development program, M&S has a focus on generating new knowledge, as shown in Table 1. A basic application of M&S is the integrated analysis of PK data obtained from clinical trials conducted in both healthy subjects and patients throughout the phases of product development, for example, using PopPK analyses. Such PK-focused analyses elucidate the PK characteristics of the drug and potential sources of between-subject PK variability (e.g., body-weight in the case of many therapeutic proteins).
      • Dirks N.L.
      • Meibohm B.
      Population pharmacokinetics of therapeutic monoclonal antibodies.
      When biomarkers for PD responses are available, the M&S tools are also valuable for characterizing the temporal PD profiles, the dose-response relationship, and the relationship between PK and PD (i.e., exposure-response relationship).
      Model-based analyses are particularly useful for therapeutic proteins with nonlinear PK, such as those with target-mediated drug disposition.
      • Dua P.
      • Hawkins E.
      • van der Graaf P.H.
      A tutorial on target-mediated drug disposition (TMDD) models.
      In the presence of target-mediated drug disposition, the level of pharmacological targets can influence the systemic clearance contributing to the PK variability. However, the target-mediated processes are generally saturated when the drug concentrations are in excess compared with the target level. When relevant PD biomarkers are available, model-based analyses can be applied to characterize both PK and PD simultaneously. In a few special cases where an interplay of PK and PD exists, also known as PD-mediated drug disposition or PDMDD,
      • Wang Y.M.
      • Krzyzanski W.
      • Doshi S.
      • Xiao J.J.
      • Perez-Ruixo J.J.
      • Chow A.T.
      Pharmacodynamics-mediated drug disposition (PDMDD) and precursor pool lifespan model for single dose of romiplostim in healthy subjects.
      the drug concentrations (PK profiles) trigger PD responses which subsequently modulate the PK properties. The timing of observed PK-PD interplay varied among products. Several more recent examples show that the reduction in tumor burden after treatment with antitumor immunotherapy results in a decrease in systemic clearance of the immunotherapy agents,
      • Wang Y.
      • Booth B.
      • Rahman A.
      • Kim G.
      • Huang S.M.
      • Zineh I.
      Toward greater insights on pharmacokinetics and exposure-response relationships for therapeutic biologics in oncology drug development.
      • Liu C.
      • Yu J.
      • Li H.
      • et al.
      Association of time-varying clearance of nivolumab with disease dynamics and its implications on exposure response analysis.
      • Li H.
      • Yu J.
      • Liu C.
      • et al.
      Time dependent pharmacokinetics of pembrolizumab in patients with solid tumor and its correlation with best overall response.
      which then exhibit time-dependent PK properties. Therefore, the between-subject PK and PD variability are intricately linked, and integrated PK-PD modeling analyses are particularly necessary to establish the PK-PD relationships and to identify sources of between-subject PK and PD variability. These PK and PD variability data are often used to inform the design of clinical comparability studies, when needed, to support manufacturing changes, specifically, considering within-subject variability for a crossover design and between-subject variability for a parallel design in sample size estimation.

       Translating PK and PD Data to Practical Use

      Beyond enhancing the scientific understanding of quantitative pharmacology, quantitative model-based analyses
      European Medicines Agency
      Activity Report of the Modelling and Simulation Working Group.
      (e.g., with PK and PK-PD models) have been used to support the (a) design of first-in-human studies based on the principles of interspecies scaling, (b) identification of dose range for dose-finding studies, and (c) selection of doses and dose regimens for the confirmatory safety and efficacy trials based on the established dose-response and exposure-response relationships. Incorporating between-subject PK and PD variability estimated from PopPK and PopPK-PD modeling analyses into model-based simulations can further help with optimizing the study design. At the stage of regulatory review before commercialization, the use of M&S often contributes to risk and benefit evaluations and supports dosing recommendations.
      • Bhattaram V.A.
      • Bonapace C.
      • Chilukuri D.M.
      • et al.
      Impact of pharmacometric reviews on new drug approval and labeling decisions--a survey of 31 new drug applications submitted between 2005 and 2006.
      • Lee J.Y.
      • Garnett C.E.
      • Gobburu J.V.
      • et al.
      Impact of pharmacometric analyses on new drug approval and labelling decisions: a review of 198 submissions between 2000 and 2008.
      • Wang Y.
      • Bhattaram A.V.
      • Jadhav P.R.
      • et al.
      Leveraging prior quantitative knowledge to guide drug development decisions and regulatory science recommendations: impact of FDA pharmacometrics during 2004-2006.
      Overall, M&S functions to integrate and bridge various types of data during the product development and facilitate moving a product toward commercialization.
      • Harnisch L.
      • Shepard T.
      • Pons G.
      • Della Pasqua O.
      Modeling and simulation as a tool to bridge efficacy and safety data in special populations.
      Consequently, multiple legislative initiatives including the Critical Path Initiative, the Prescription Drug User Fee Act VI, and the 21st Century Cures Act have brought modeling and simulation to the forefront of review science and policy development. Ultimately, the M&S knowledgebase supports the regulatory review
      • Bhattaram V.A.
      • Bonapace C.
      • Chilukuri D.M.
      • et al.
      Impact of pharmacometric reviews on new drug approval and labeling decisions--a survey of 31 new drug applications submitted between 2005 and 2006.
      • Lee J.Y.
      • Garnett C.E.
      • Gobburu J.V.
      • et al.
      Impact of pharmacometric analyses on new drug approval and labelling decisions: a review of 198 submissions between 2000 and 2008.
      • Wang Y.
      • Bhattaram A.V.
      • Jadhav P.R.
      • et al.
      Leveraging prior quantitative knowledge to guide drug development decisions and regulatory science recommendations: impact of FDA pharmacometrics during 2004-2006.
      and can provide primary or supportive evidence to support regulatory decisions, including the dosing recommendations.

       Supporting 351(k) BLA by Leveraging Established M&S Knowledgebase

      By the time a biosimilar development program commences, a relatively large M&S knowledgebase on the reference product is often established. Therefore, M&S in a biosimilar program should focus on leveraging existing knowledge to enhance the efficiency and effectiveness in demonstrating PK and PD (when applicable) similarity between a proposed biosimilar product and the reference product. A reasonable strategy would be to, first, identify appropriate PK models and PK-PD models in published literature, followed by modifying the models to fit the intended use and conducting clinical trial simulations to identify appropriate study elements.

       Optimizing the Study Design

      The FDA's guidance
      U.S. Food and Drug Administration
      FDA's Guidance for Industry: Clinical Pharmacology Data to Support a Demonstration of Biosimilarity to a Reference Product.
      recommends maximizing the sensitivity of the study by evaluating PK or PK-PD similarity at a dose that is on the steep part of the dose-exposure or dose-response curves. M&S analyses can be conducted to identify suitable doses for PK and PK-PD similarity assessments, when adequate models can be found in the public domain or can be developed based on available data. For products with a linear PK and a long half-life, an enhancement to the sensitivity of PK similarity study can come from controlling the between-subject PK variability based on understanding of the intrinsic factor(s) or covariate(s) with significant contribution to PK variability. Model-based simulations can illustrate the dose-exposure relationship of products with nonlinear PK, as well as the dose-response relationships which are typically nonlinear in nature. These types of simulations may thereby reveal the dose range on the steep part of the dose-exposure and dose-response curves to support dose selection for PK and PD similarity studies.
      In addition to the impact of dose on exposure, the impact of structural difference(s) between the proposed biosimilar and the reference product on exposure in the relevant population could also be quantified, if such data were available. Other study design elements that may be evaluated with M&S tools include the sampling schedule for PK and PD, and sample size. When the study objective was to demonstrate the similarity in both PK and PD, the dose selection needs to consider the balance between PK variability and PD variability. As an example, we refer to the case of pegfilgrastim illustrated by Zhu et al.,
      • Zhu P.
      • Sy S.K.B.
      • Skerjanec A.
      Application of pharmacometric analysis in the design of clinical pharmacology studies for biosimilar development.
      where the M&S results supported the study design based on the dose-response relationship and the PK and PD variability.
      Through mechanism-based models of drug action, M&S provide insight into understanding the most sensitive population for PD response evaluation. This is particularly relevant when patients and healthy subjects exhibit different baseline PD levels and different magnitude of PD responses. One such example is the effect of insulin on glycosylated or glycated hemoglobin (HbA1c). Indirect PD response models for HbA1c response have shown that when initiating treatment, higher baseline glucose values translate to greater HbA1c responses (i.e., lowering of HbA1c).
      • Woo S.
      • Pawaskar D.
      • Jusko W.J.
      Methods of utilizing baseline values for indirect response models.
      Consequently, patients who have higher baseline glucose levels would be a more sensitive population, as compared to healthy subjects, with respect to the PD response(s) to insulin, as measured by the reduction of HbA1c. Conversely, healthy subjects would be a more sensitive population, as compared to patients, for PD biomarkers such as neutrophils, reticulocytes, or red blood cells after treatment with therapeutic proteins that stimulate the production of such cell types. In general, compared with healthy subjects, a patient population is expected to have a lower number of these types of cells at baseline and may also achieve a smaller increase in the absolute cell count in response to treatment.
      • Wang Y.M.
      • Krzyzanski W.
      • Doshi S.
      • Xiao J.J.
      • Perez-Ruixo J.J.
      • Chow A.T.
      Pharmacodynamics-mediated drug disposition (PDMDD) and precursor pool lifespan model for single dose of romiplostim in healthy subjects.
      • Wang B.
      • Ludden T.M.
      • Cheung E.N.
      • Schwab G.G.
      • Roskos L.K.
      Population pharmacokinetic-pharmacodynamic modeling of filgrastim (r-metHuG-CSF) in healthy volunteers.
      • Ramakrishnan R.
      • Cheung W.K.
      • Wacholtz M.C.
      • Minton N.
      • Jusko W.J.
      Pharmacokinetic and pharmacodynamic modeling of recombinant human erythropoietin after single and multiple doses in healthy volunteers.
      Considerations should include both safety and mechanistic aspects when determining an adequately sensitive population to use in PK and PD similarity studies.
      Performing simulations of virtual clinical trials before designing and conducting a clinical PK or PK-PD similarity study can lead to optimization of the study design and improvement in overall study conduct efficiency to demonstrate similarity based on the primary endpoint(s) of the study.

       PD Biomarker, Endpoint Selection

      To select PD biomarkers for use in the PD similarity assessments, the FDA recommends considering the (a) temporal profile (the onset and the return to baseline), (b) dynamic range of PD responses, (c) sensitivity of the PD biomarkers to potential differences between the proposed biosimilar product and the reference product, and (d) relevance of PD markers to the mechanism of action.
      U.S. Food and Drug Administration
      FDA's Guidance for Industry: Clinical Pharmacology Data to Support a Demonstration of Biosimilarity to a Reference Product.
      Regarding the dose selection for use in a PD similarity study, a dose on the steep part of the dose-response curve would be considered optimal. In addition, the shape of the dose-response curve will likely depend on the choice of PD endpoints and PD biomarkers.
      • Zhu P.
      • Sy S.K.B.
      • Skerjanec A.
      Application of pharmacometric analysis in the design of clinical pharmacology studies for biosimilar development.
      Performing independent M&S analyses to illustrate the nature of the dose-response relationship can serve as part of the justification for the selected PD biomarker(s). In cases where multiple potential PD biomarkers are relevant to the mechanism of action of a product, M&S studies may also serve as a means of comparative evaluations to support the PD biomarker selection. If dose-exposure and dose-response information are available in multiple populations, the M&S analyses may also inform the selection of study population for PK and PD similarity evaluation.

       Evaluating the Relationship Between PD Responses and Clinical Outcome Measure

      PK similarity data add to the totality of the evidence to a support a demonstration of biosimilarity because systemic exposure is generally a driver of treatment responses. When the selected PD biomarkers are relevant to the mechanism of action of the product, and the PD responses measured by the biomarkers are relevant to the clinical outcomes, PD similarity data also add to the totality of the evidence to support a demonstration of biosimilarity. While an established quantitative correlation between PD responses and clinical outcomes is not a requirement in the context of regulatory approval of biosimilars, the existence of any correlation would provide further scientific support and would be a positive attribute to biosimilar programs that use a PK-PD similarity approach to support a demonstration of no clinically meaningful differences as a part of demonstrating biosimilarity. In addition, such analyses may further elucidate the sensitivity of the PD biomarkers. An excellent case example of this type of analysis is the use of ANC as a PD biomarker in the filgrastim biosimilar programs. Li et al.
      • Li L.
      • Ma L.
      • Schrieber S.J.
      • et al.
      Quantitative relationship between AUEC of absolute neutrophil count and duration of severe neutropenia for G-CSF in breast cancer patients.
      showed that the PD measure of change in the area under the effect curve measured by ANC correlated with the clinical efficacy measure of change in the duration of severe neutropenia and contained less between-patient variation as compared with the duration of severe neutropenia. The authors concluded that the results support the use of ANC as a biomarker for dose selection and optimization of clinical trial design with a smaller sample size. This case example illustrates the plausibility of using in silico M&S to evaluate PK and PD responses versus clinical response relationship, as described in the FDA's Biosimilar Action Plan
      U.S. Food and Drug Administration
      FDA's Biosimilar Action Plan.
      released in July 2018.
      As science evolves and technologies advance, newly discovered pharmacologically relevant and sensitive PD biomarkers may emerge for approved therapeutic proteins, for which PK-PD models will need to be established. Depending on the extent of information available in the public domain for any proposed PD biomarker, additional clinical data may be necessary to justify that the proposed PD biomarker is a relevant and sensitive biomarker (refer to the PD Biomarker, Endpoint Selection section). For example, when clinical data for a proposed PD biomarker are only available at a single-dose level, a small pilot study evaluating a range of doses may be necessary to evaluate the dose-response profile and the dynamic range of the PD response, which can serve the purpose of justifying that the selected dose is on the steep part of the dose-response curve.
      When a PD biomarker appears promising and the available public data are amenable to model-based analyses, performing M&S exercises before designing or conducting a PD similarity study can be beneficial, as these analyses may inform the PD dose-response relationships, PD variability, and so forth. The same M&S approaches applied in the development programs of novel products should also be applied in these situations to address pertinent PD-related outstanding questions that are relevant to the proposed PD biomarker(s). However, as was also already stated, M&S approaches alone may not negate the need for additional prospective clinical data collection to support the PD biomarker selection.

      Summary and Future Perspectives

      Scientists in the biopharmaceutical industry, academia, and regulatory agencies have built a very successful M&S framework and routinely apply the technologies to support development programs of novel products. The M&S technology has continually advanced and expanded from basic PK modeling to PK-PD modeling, disease progression modeling, quantitative systems pharmacology with relevant mechanistic specificity, and predictive immunogenicity. The ever-evolving sciences of quantitative pharmacology have played a critical role in addressing important issues in drug development with great success. To address current and emerging needs, innovation and improvement of these technologies would undoubtedly be essential for continued success.
      M&S approaches have contributed to and may continue to enhance biosimilar development programs. The emergence of an abbreviated pathway for regulatory approval of therapeutic proteins as biosimilars presents an arena for further extension of the application of the familiar M&S framework. The focus of M&S studies in support of biosimilar development programs is on maximizing the efficiency of the PK similarity studies or PK-PD similarity studies and simultaneously increasing the effectiveness in supporting similarity assessments for PK or PK-PD. Publicly available PK or PK-PD models are an important resource and a starting point for applying M&S approaches in biosimilar programs. Conceptually, we conclude that the utility of M&S in biosimilar programs is actually not so different as compared with novel product programs.

      References

        • Ette E.I.
        • Williams P.J.
        Population pharmacokinetics I: background, concepts, and models.
        Ann Pharmacother. 2004; 38: 1702-1706
        • Chatelut E.
        • Rostaing L.
        • Gregoire N.
        • et al.
        A pharmacokinetic model for alpha interferon administered subcutaneously.
        Br J Clin Pharmacol. 1999; 47: 365-371
        • Dodds M.
        • Chow V.
        • Markus R.
        • Perez-Ruixo J.J.
        • Shen D.
        • Gibbs M.
        The use of pharmacometrics to optimize biosimilar development.
        J Pharm Sci. 2013; 102: 3908-3914
        • U.S. Food and Drug Administration
        FDA's Biosimilarity Guidances.
        FDA, Silver Spring, MD2018 (Available at: https://www.fda.gov/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/ucm290967.htm. Accessed September 20, 2018)
        • U.S. Food and Drug Administration
        FDA's Guidance for Industry: Clinical Pharmacology Data to Support a Demonstration of Biosimilarity to a Reference Product.
        FDA, Silver Spring, MD2016 (Available at: https://www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/UCM397017.pdf. Accessed September 20, 2018)
        • U.S. Food and Drug Administration
        CDER list of licensed biological products.
        FDA, Silver Spring, MD2018 (Available at: https://www.fda.gov/downloads/Drugs/DevelopmentApprovalProcess/HowDrugsareDevelopedandApproved/ApprovalApplications/TherapeuticBiologicApplications/Biosimilars/UCM560162.pdf. Accessed September 20, 2018)
        • Dirks N.L.
        • Meibohm B.
        Population pharmacokinetics of therapeutic monoclonal antibodies.
        Clin Pharmacokinet. 2010; 49: 633-659
        • Dua P.
        • Hawkins E.
        • van der Graaf P.H.
        A tutorial on target-mediated drug disposition (TMDD) models.
        CPT Pharmacometrics Syst Pharmacol. 2015; 4: 324-337
        • Wang Y.M.
        • Krzyzanski W.
        • Doshi S.
        • Xiao J.J.
        • Perez-Ruixo J.J.
        • Chow A.T.
        Pharmacodynamics-mediated drug disposition (PDMDD) and precursor pool lifespan model for single dose of romiplostim in healthy subjects.
        AAPS J. 2010; 12: 729-740
        • Wang Y.
        • Booth B.
        • Rahman A.
        • Kim G.
        • Huang S.M.
        • Zineh I.
        Toward greater insights on pharmacokinetics and exposure-response relationships for therapeutic biologics in oncology drug development.
        Clin Pharmacol Ther. 2017; 101: 582-584
        • Liu C.
        • Yu J.
        • Li H.
        • et al.
        Association of time-varying clearance of nivolumab with disease dynamics and its implications on exposure response analysis.
        Clin Pharmacol Ther. 2017; 101: 657-666
        • Li H.
        • Yu J.
        • Liu C.
        • et al.
        Time dependent pharmacokinetics of pembrolizumab in patients with solid tumor and its correlation with best overall response.
        J Pharmacokinet Pharmacodyn. 2017; 44: 403-414
        • European Medicines Agency
        Activity Report of the Modelling and Simulation Working Group.
        EMA, London, UK2016 (Available at:) (Accessed September 20, 2018)
        • Bhattaram V.A.
        • Bonapace C.
        • Chilukuri D.M.
        • et al.
        Impact of pharmacometric reviews on new drug approval and labeling decisions--a survey of 31 new drug applications submitted between 2005 and 2006.
        Clin Pharmacol Ther. 2007; 81: 213-221
        • Lee J.Y.
        • Garnett C.E.
        • Gobburu J.V.
        • et al.
        Impact of pharmacometric analyses on new drug approval and labelling decisions: a review of 198 submissions between 2000 and 2008.
        Clin Pharmacokinet. 2011; 50: 627-635
        • Wang Y.
        • Bhattaram A.V.
        • Jadhav P.R.
        • et al.
        Leveraging prior quantitative knowledge to guide drug development decisions and regulatory science recommendations: impact of FDA pharmacometrics during 2004-2006.
        J Clin Pharmacol. 2008; 48: 146-156
        • Harnisch L.
        • Shepard T.
        • Pons G.
        • Della Pasqua O.
        Modeling and simulation as a tool to bridge efficacy and safety data in special populations.
        CPT Pharmacometrics Syst Pharmacol. 2013; 2: e28
        • Zhu P.
        • Sy S.K.B.
        • Skerjanec A.
        Application of pharmacometric analysis in the design of clinical pharmacology studies for biosimilar development.
        AAPS J. 2018; 20: 40
        • Woo S.
        • Pawaskar D.
        • Jusko W.J.
        Methods of utilizing baseline values for indirect response models.
        J Pharmacokinet Pharmacodyn. 2009; 36: 381-405
        • Wang B.
        • Ludden T.M.
        • Cheung E.N.
        • Schwab G.G.
        • Roskos L.K.
        Population pharmacokinetic-pharmacodynamic modeling of filgrastim (r-metHuG-CSF) in healthy volunteers.
        J Pharmacokinet Pharmacodyn. 2001; 28: 321-342
        • Ramakrishnan R.
        • Cheung W.K.
        • Wacholtz M.C.
        • Minton N.
        • Jusko W.J.
        Pharmacokinetic and pharmacodynamic modeling of recombinant human erythropoietin after single and multiple doses in healthy volunteers.
        J Clin Pharmacol. 2004; 44: 991-1002
        • Li L.
        • Ma L.
        • Schrieber S.J.
        • et al.
        Quantitative relationship between AUEC of absolute neutrophil count and duration of severe neutropenia for G-CSF in breast cancer patients.
        Clin Pharmacol Ther. 2018; 104: 742-748
        • U.S. Food and Drug Administration
        FDA's Biosimilar Action Plan.
        FDA, Silver Spring, MD2018 (Available at: https://www.fda.gov/downloads/Drugs/DevelopmentApprovalProcess/HowDrugsareDevelopedandApproved/ApprovalApplications/TherapeuticBiologicApplications/Biosimilars/UCM613761.pdf. Accessed September 20, 2018)