Today, biologics account for more than 50% of all drugs in development and, according to industry reports, are expected to grow to $400 billion over the next few years.¹ The introduction of these targeted therapies, coupled with the rising adoption of patient-centric personalized medicine approaches, have contributed to this rise. Biologics offer high efficacy and provide a more targeted and specific response, with less toxicity and fewer side effects. 

Although the success of biologics has been well demonstrated, there are inherent operational and technological challenges associated with this complex class of drugs. One of these challenges-immunogenicity-has become a key area of regulatory concern. Immunogenicity (IG) is defined by FDA as the propensity of the therapeutic protein to generate immune responses to itself and to related proteins, or to induce immunologically-related adverse clinical events.² In a recent FDA review of 121 approved biological products, 89% of the products had reported IG, and in 49% of the cases, those responses impacted the drug’s efficacy (see figure 1).³

Until recently, IG has mostly been tackled at the preclinical stage by using bioinformatics approaches or in vitro assays followed by protein engineering. However, these methods have several limitations in delivering a qualitative “go/no-go” assessment, rather than quantitative prediction of the impact of immune response on the amount of drug at the site of action, in a target population of clinical subjects. 

This article focuses on how a quantitative systems pharmacology (QSP)-based approach can be used to predict and better manage IG, taking advantage of bioinformatics data as an input to the model. QSP combines computational modeling and experimental data to examine the relationships between a drug, the biological system, and the disease process. QSP simulation can predict the IG of biologics and its impact on the drug’s pharmacokinetics (PK), efficacy, and safety in diverse patient populations, thereby managing its impact in development and manufacturing. Now encouraged by global regulators, QSP is a valued translational tool to guide clinical and regulatory decision-making in biologics drug development.

Understanding immunogenicity

Understanding IG requires us to fill the gaps in our understanding of how therapeutic proteins interact with the body’s immune system. Despite being “biological,” most therapeutic proteins are engineered-even fully humanized biologicals exhibit properties that can potentially be recognized as “non-self” and therefore have an increased risk of promoting an antigenic response. The IG response typically takes place in the form of the production of anti-drug antibodies (ADAs). ADAs may be an inevitable consequence of using biological drugs, but a given ADA level with respect to its binding may be manageable provided certain parameters are correctly optimized (e.g., dose, frequency, route of administration, target patient population, tolerability strategy, co-medications). Finding the optimum for each drug will require a quantitative approach, hence, the interest in QSP modeling.

There are several factors that contribute to IG and, once understood and quantified, can by managed to reduce safety risk:

• The complexity and heterogeneity of unwanted immune responses to drugs can be explained in large part by the fact that these responses against biological agents proceed through the network of cellular and molecular interactions in the immune system. These are highly complex, dynamic interacting systems and can give rise to a range of outcomes, depending upon initial state and patient specific circumstances. Thus, the state of the patient’s immune system and genetics at the individual level (e.g. HLA genotype) may be an important factor in IG.⁴

• ADAs are a heterogeneous population according to affinity, isotype, and neutralizing ability. ADAs may develop early while clinical efficacy is still present, be present at low levels, exist only within immune complexes, or be transient. Non-neutralizing ADAs are possible, and the impact of these on drug pharmacokinetic/pharmacodynamics (PK/PD) can be hard to predict. Improved quantitative understanding of these aspects of IG will be critical to improving the development of safer biological drugs.

• Other factors that contribute to the complexity of IG include the route and frequency of drug administration, the duration of treatment, formation of aggregates, and the co-administration of immunosuppressive agents. The potential for the formation of protein aggregates is an important issue for quantitative prediction of IG.⁵ Immune complexes can be either small, which are cleared rapidly from the system, or large, which persist longer. Since different sized complexes will vary with individuals, the potential for varied PK outcomes and immune complexes may also have implications for safety. 

Quantitative predictions of the impact of the ADA production and its effect on the PK/PD of biologicals will help to manage the IG. When immunosuppressive agents, e.g., methotrexate, are co-administered, the immune reaction against the therapeutic protein is reduced-evaluation of combination PK/PD of biologics with immune-suppressive agents will be necessary to optimize combination therapies.

Quantitative systems pharmacology: Bridging pharmacokinetics and systems biology

Systems biology has been used to create mechanistic models of how molecular networks underpin the behavior of living cells. These models apply a nonlinear, integrative, quantitative, and holistic approach that uses biology, computational modeling, engineering, bioinformatics, and other sciences (including high-quality “omics” information) to understand complex biological systems and how perturbing these systems can cause disease.⁶ QSP integrates these models with the models of drug absorption, distribution, metabolism, and excretion (ADME), which have been informing drug development for decades. Thus, QSP creates an in silico framework for mechanistic, mathematical modeling of diseases and drug action. 

A quantitative framework that integrates diverse computational methodologies can be used to assimilate data across scales to understand the interacting network elements and bridge molecular-to-systems level scales.⁷

QSP models are used to design first-in-human clinical trials, inform the mechanisms of drug efficacy and safety, and confirm drug target binding and modulation. Once we know how much drug is at the site of action, QSP can help answer questions such as, “How will the drug modulate cellular signaling to exert a pharmacological effect?” or “What pharmacological action will it have at that particular organ?” Answering these questions will provide insight into the mechanisms of drug efficacy. This approach can be used to predict how drugs modify cellular networks and how drugs impact and are impacted by human pathophysiology. Further, QSP can facilitate the evaluation of complex, heterogeneous diseases such as cancer, immunological, metabolic, and CNS diseases that commonly require combination therapies to control disease progression. 

Using QSP models to predict and manage immunogenicity 

The high prevalence of IG not only impacts the clinical utility of existing treatments for patients but also the development of novel biologicals. The latter issue is exacerbated by the fact that IG often manifests itself relatively late in the drug development cycle, where the economic impact of attrition is at its greatest. Despite advances in the development of assays and techniques to assess IG at various stages of biological drug discovery and development, it is notoriously difficult to predict, and it seems unlikely that the occurrence of IG in clinical development will be dramatically reduced by modification of drug properties in the foreseeable future. F

Given the complexity of processes involved in IG, the development of quantitative, mechanistic models of humoral and cellular responses involved in IG will be invaluable in supporting development decisions and the regulatory approval process. Like the management of drug-drug interactions (DDI) with physiologically-based PK (PBPK), the focus of the QSP approach is on how IG can be managed, rather than avoided. Just as PBPK models depict ADME processes, QSP models add in biological pathways that are relevant for disease modification. And just as DDI cannot always be “engineered-out,” it can be predicted and managed effectively through clinical trial simulation using PBPK.

The mechanistic elements of the immune biology are understood and can be informed with human patient input data from clinical and potentially in vitro and ex vivo sources. Theoretical mechanistic, multiscale mathematical models of IG represented by the subcellular, cellular, and whole-body levels have been developed, and serve as good starting points.

QSP software focused on managing immunogenicity 

A QSP software tool capturing this mechanistic understanding of the immune biology that can simulate virtual populations with inter-individual variability based on input data will enable an improved clinical development path for biologicals. Optimal dosing routes and regimens can be explored in virtual populations giving input to likely success rates for given biologicals and patient populations. Early emerging clinical data can be included, such as PK profiles matched with ADA titers and drug affinity, enabling optimal decision-making regarding dose and enriching the quality of the models. With rich clinical data, the confidence in the tool predictions can be verified, and scope and mechanistic detail optimized, potentially enabling increased understanding of key IG processes and biomarkers. 

A platform designed in a modular way, which allows modeling of competitive compound data and models in the full context of the mechanistic knowledge on humoral and cellular immune response, is an objective endpoint. The simulator facilitates incorporation of input data including modality and sequence, in vitro assays of DC activation, T-cell epitope identification/binding, population baseline immune status, genetics, and drug clinical PK/PD data. This tool can then be used to more rationally and quantitatively explore the likelihood of IG from the preclinical stage of drug development. The ability to incorporate clinical data enables the extension of learn-confirm cycles into Phase I, II, and III stages of clinical development, thus giving the potential to not only optimize dose, route, and regimen, particularly critical for special populations, but also to build confidence in the models.

The IG simulator (see diagram) integrates drug and literature data: in vitro, population, and clinical information. The mechanistic model is composed of two parts representing the immune response and biologics PBPK. Distribution of physiological parameters in the population of interest are used to instantiate the model with virtual subject parameters and subsequently run simulation with dosing regimen specified in clinical trial protocols. In addition to population parameters, the IG simulator uses HLA allele frequencies and variability in immune system as a function of age or disease.

Next steps in IG management using QSP

Certara formed a consortium in 2017, bringing together a group of leading biopharmaceutical companies in a precompetitive environment to cooperatively develop an IG simulation platform based on state-of-the art QSP science and methods. The simulator is used to predict IG and its impact on compound PK, efficacy, and safety in diverse patient populations in drug discovery and development. QSP models are applicable across all stages of drug discovery and development-from before the selection of clinical candidates to interpret preclinical databases, to informing biomarker translation, and supporting clinical development of drug candidates. 

The future of QSP and the use of the new IG simulator has already demonstrated its ability to impact go/no-go decisions by identifying those biologic drug candidates that could impact PK, and, hence, dosing, versus those without correlation. This could markedly improve success rates in Phase II attrition of compounds negatively affected by IG. The IG simulator can also help in the development of biosimilars since they are not identical to the reference product and therefore might display different IG responses. 

QSP translates PK or exposure into pharmacological effect and builds on gaining insights from PK/PD and PBPK approaches with systems biology models of biological and biochemical processes. QSP can support personalized dosing, guiding treatment optimization for varying target populations and patient types. QSP is now gaining regulatory and industry acceptance, as is has been demonstrated to fill in the gaps between early-stage PK and late-stage understanding of drug efficacy.

Piet van der Graaf, PharmD, PhD, is the Senior Vice President of Quantitative Systems Pharmacology; and Andrzej Kierzek, PhD, is the Head of Systems Modeling at Certara.

References

1. Biopharma Dive.(2018). The Future of Biologics Development: Insights to Help Biopharma Companies Win the Race to Market. Washington, DC: Industry Dive.
2. US Department of Health and Human Services, Food and Drug Administration. (2014, August). Guidance for Industry. Immunogenicity Assessment for Therapeutic Protein Products. 
3. Wang YM, Wang J, Hon YY, Zhou L, Fang L, & Ahn HY. (2016). Evaluating and reporting the immunogenicity impacts for biological products-A clinical pharmacology perspective. The AAPS Journal, 18(2), 395–403.18 WP
4. Van Schouwenburg PA, Rispens T, & Wolbink GJ. (2013). Immunogenicity of anti-TNF biologic therapies for rheumatoid arthritis. Nature Reviews Rheumatology 9(3), 164–172.
5. Benson N and van der Graaf PH. (2014). The rise of systems pharmacology in drug discovery and development. Future Medicinal Chemistry 6(16), 1731–1734.
6. Yin L, Chen X, Tiwari A, Vicini P, & Hickling TP. (2015). The role of aggregates of therapeutic protein products in immunogenicity: An evaluation by mathematical modeling. Journal of Immunology Research 2015, Article ID 401956, 1–14.
7. Bradshaw E, Spilker, M, et al, (2019). Applications of Quantitative Systems Pharmacology in Model-Informed Drug Discovery: Perspective on Impact and Opportunities. CPT Pharmacometrics Systems Pharmacology, (2019)