Tackling Analytical Method Development for ADCs

Martin McCarthy/Getty Images; Dan Ward

The targeted therapy possible with antibody-drug conjugates (ADCs) makes them an attractive class of drugs. Their composition--a biologic linked to a small-molecule cytotoxic, which enables targeted delivery--also creates challenges for analytical method development, both for characterization and lot release purposes. Expertise in both chemical and biological analysis is required, and techniques for both types of molecules must be used. In addition to performing typical characterization studies, the linker chemistry and its impact on heterogeneity and applicable analytical techniques must be understood. The influence of conjugation on antibody binding must also be evaluated, and a wide range of stability studies must be conducted. Fortunately, both improvements in analytical technology and increased experience with ADCs are leading to improved strategies for analytical method development and validation.

Both small and large
As for any drug, analytical methods for the characterization of ADCs must be developed, and separate quality control (QC) tests must be established for lot release of the final drug substance. Because ADCs are both small drug and biologic compounds, characterization and validation need to be appropriate for both types of products. “Testing requirements will still be identity, purity, impurities, activity, concentration, and stability as outlined in the International Conference on Harmonization’s ICH Q5C (1) and ICH Q6B (2) biologics guidelines; however, the testing must cover both functional and physiochemical properties, including process control methods and release and stability-indicating assays for both the large and small molecule,” says Lisa McDermott, principal scientist at SAFC.

Because ADCs are made of three different components, McDermott believes it is crucial to have an advanced control strategy for each of the intermediates, with testing profiles determined as if each of the components is being developed as a stand-alone drug substance. “With this approach, many of the quality parameters can be controlled in the release of these intermediates and allow the final release strategy to focus on the quality of the ADC.”

Main methods
“Conjugation usually results in a mixture of ADCs with different drug-to-antibody ratios (DARs), free drug, and naked antibody concentrations,” says Harpreet Kaur, synthetic chemistry team leader with Dalton Pharmaceutical Services. This increased heterogeneity associated with all ADCs, even site-specifically linked ones, requires the development of robust methods with sufficient resolution to characterize and measure the diversity of product-related species and potential impurities, according to Fred Jacobson, staff scientist and Kadcyla technical development leader with Genentech. 

The critical properties for ADC characterization include target site-specificity and binding properties, stability of the linker and drug species, drug potency and free drug, site(s) of conjugation, DAR, heterogeneity, and solubility. In general, Jacobson notes that most modern chromatographic, electrophoretic, and spectroscopic (ultraviolet-visible [UV-vis] and mass spectrometry [MS]) methods have proven adequate to the task.
UV-Vis spectrophotometry has traditionally been used to measure ADC and free drug concentrations and average DAR. The challenge with this method, according to Kaur, is that the extinction coefficient (λ_max) of the drug may change when conjugated to the antibody or in a different buffer. In addiiton, the drug and antibody should have different λ_max values. 

Chromatographic methods based on hydrophobicity (hydrophobic interaction chromatography [HIC]) and size (size-exclusion chromatography [SEC], SEC-multi-angle laser light scattering (SEC-MALLS) can provide information about the number and location of conjugation sites, average DAR, and free drug content, although these methods are not suitable for purification and characterization of ADCs produced using linkage through lysine residues due to their high degree of heterogeneity, according to Kaur. 

Mass spectrometry (MS) methods are often introduced early in the development process to better understand the product and conjugation or to act as an orthogonal tool to standard chromatographic techniques that will ultimately be required for product release, according to Allan Davidson, analytical development manager for Piramal Healthcare. “While processes can be developed to provide a consistent and accurate drug load, most products remain highly heterogeneous and as such there is the need to better understand the complex picture of antibody structure and drug distribution,” he says. 

Electrospray ionization (ESI)-MS, liquid chromatography (LC)-MS/MS, and matrix-assisted laser desorption/ionization-time of flight (MALDI-TOF) have been used extensively  to analyze the DAR, free drug and metabolite concentrations, and linker stability. The validation of MS methods can be a challenge, however, due to differences in ionization of ADCs with different DAR values and ADC linker hydrolysis under acidic LC-MS conditions or by the acidic matrices used for MALDI, according to Kaur. 

Finally, bioanalytical immunoassays, such as enzyme-linked immunosorbent assays (ELISA) are used for quantitation of the ADC, naked antibody, and free drug content, determination of the extent of ADC binding to the target antigen, and to establish the stability of linker and drug and the immunogenicity, while cell-based mechanism of action assays are used to assess the target-killing ability of the drug. “The biggest challenge with these methods is that the binding of the antibody to the target antigen can be altered by the site and degree of conjugation,” says Kaur.

Analysis of the small-molecule components (linker and drug) is relatively straightforward with well-defined expectations in line with what would be required for an API, according to Jacobson. “One important difference, however, is the requirement for understanding the impact of impurities in the small molecule on the process and quality of the ADC,” he observes. 

In addition, bioanalytical methods are required to determine ADC potency, both with respect to and target binding (ELISA-based). Characterization should also cover the impact of conjugation on antibody binding properties towards specific Fc-receptors, because they may affect pharmacokinetic (PK) or secondary mechanisms of action, according to Jacobson.

Priority assays, according to McDermott, include determination of the potency, drug load and distribution, and size variants. Early development of cytotoxicity assays and chromatographic assays such as HIC, reduced reverse phase high-performance liquid chromatography (RPHPLC), and SEC is therefore important. She adds that additional analytics are often used to underpin the accuracy of these methods in early development stages. LC-MS, for example, is often used to assign structural information to the individual components of a mixture for confirmation of drug load and distribution.The presence of hydrophobic linkers and drugs occasionally leads to problems, however. 

As an example, Jacobson points to the difficulty associated with charge-based assays such as ion exchange HPLC (IEX). “Although several new stationary phases have appeared, none have proven great for ADCs due either to nonspecific interactions or inadequate resolution. Capillary isoelectric focusing (cIEF), such as imaged cIEF, has been a reasonable substitute, but experience with monoclonal antibodies (mAbs) suggests that there may be differences in the information obtainable by each method,” he comments. Jacobson also notes that it is difficult to do preparative collection of charge variant species from capillary electrophoresis (CE) for characterization.

Bioanalysis in biological fluids
From an immunoassay perspective, the development and validation of methods for the bioanalysis of ADCs require the consistent availability of high affinity, high specificity, anti-toxin antibody reagents to ensure appropriate selectivity in the analysis, according to Michael Brown, director of ICON’s Bioanalytical Laboratories. “Generation of these types of antibodies to small molecular weight toxins is considered an art, and often times their availability can be limited,” he observes. 

Given the potency of the toxins, it can be a challenge getting an MS method developed and validated with the required lower limit of quantitation (LLOQ). Identifying the metabolites of the toxin can be even more difficult because they are generally present in even lower concentrations than the parent. In addition to the determination of very low concentrations of released drug in physiological fluids in the presence of relatively high ADC levels, interaction of reactive intermediates with albumin or other biomolecules, changing DAR values, and the inability to use assays developed for the parent antibody for the corresponding ADC with a different architecture are also issues that must be considered when evaluating the systemic exposure of ADCs as part of drug pharmacodynamic and pharmacokinetic analyses, according to Kaur from Dalton.

One approach is to use hybrid techniques such as affinity capture LC-MS/MS, but that adds complexity to the analysis. A combination of immunoassay and LC-MS/MS techniques can also be used in addition to the hybrid techniques. “There is also increasing interest in the use of high resolution accurate mass spectrometry to support various facets of ADC bioanalysis given the complex and dynamic nature of these molecules,” says Mario Rocci, senior vice-president of ICON Bioanalytical Laboratories.

Generally, five different analyses are required to characterize the in-vivo performance of an ADC, according to Paula Jardieu, senior vice-president and general manager of ICON’s Bioanalytical Laboratories, including quantitation of the time course of the intact ADC, the total antibody, and the toxin with and without linker, plus the immunogenicity of the ADC. Changes in the DAR over time in vivo, as well as the stability of the ADC in the matrix, also must be evaluated. It is also important to establish the stability of the ADC through the analytical procedure, because instability could introduce artifacts, according to Jardieu. 

Free drug issues are typically monitored using reverse-phase chromatography with UV detection, or if the levels of detection must be very low, using LC-MS/MS against a drug linear curve, according to Dan Peckman, biochemistry manager with Eurofins Lancaster Laboratories. He also notes that confirmation that the DAR does not change with time as a function of molecular stability is typically achieved using LC or UV approaches. Mass spec analysis of the intact ADC and the naked antibody can give relevant information about the stability of the ADC.

The platform approach
For many small-molecule drugs, a platform of generic screens is often used for drugs based on their structures and mechanism of action. For ADCs, however, it is difficult to generate ‘platform’ methods that are suitable to a variety of ADCs because there are many non-specific interactions with stationary phases post-conjugation, according to Davidson. McDermott adds that while many of the analytical methods for ADCs are based on similar techniques, each construct is unique and requires an understanding of the basic chemical or physical properties that must be assessed. “Asking the right question is as important as getting an answer,” she says.

Even given these challenges, SAFC has been able to develop screens for drug load using HIC and reverse-phase chromatography, and a screen for monomer purity using SEC. “These platform methods allow us to move quickly through the development phase and focus on more challenging assays early in the project,” she explains. Other techniques such as iCE, CGE, ELISA, cytotoxicity and methods for residuals are developed by subject matter experts and optimized for each product. Testing for safety and quality attributes (bioburden, endotoxin, pH, osmolality, excipients, and appearance) are either verified using compendial-based platform methods or developed per product. 

Analysis through the development process
As ADCs move through the drug-development process, different analytical methods are often required. Much of the early-stage development is driven by methods required to support development of the conjugation process, according to Davidson, and only a few methods may be required to support early conjugation process development, for example, SEC for aggregation, HIC for DAR, and RPC for free drug. Once the conjugation process has been established, additional analytical methods are required to ensure that the functionality of the antibody and potency are effective, and other characterization methods are then introduced (e.g., cytotoxicity, binding, charge heterogeneity, residual solvents, excipient testing, etc.).

For the production of early-phase clinical supplies, in fact, McDermott notes that multiple orthogonal methods are required to ensure method accuracy and process consistency. Then, as the phase of development advances further and multiple lots are produced, the number of methods can be reduced to the ones determined to provide information around critical quality attributes. Methods are often optimized as the project progresses, and further understanding of the chemistry is achieved. “A good example for ADCs is the assay for monitoring any residual-free drug equivalents,” she comments. “Typically, this method is first developed for either the drug or drug linker used in production. As stability information or multiple lot information is available, further compounds that are drug-related may be detected, and the method will be modified to track these impurities. Continual monitoring of release and stability data is necessary to ensure adequate methods are available for validation.”

It is important to realize, Jacobson adds, that sometimes knowledge from characterization studies will result in the need for a method for lot-release or may be used to justify its omission. “The availability of analytical methods and product-specific knowledge that may lead to the addition of new assays (or replacement of older technologies) as a clinical candidate approaches licensure,” he says. 

Building a bridge
ADCs present a complex bioanalytical problem that requires leveraging large- as well as small-molecule analytics. Recent developments in LC-MS/MS technology and the use of orthogonal analytical and bioanalytical methods combined with process knowledge support the idea that manufacturing and analysis of ADCs with consistent quality and efficacy is achievable, according to Kaur. He also notes that with the move to adopt a quality-by-design (QbD) approach for defining critical quality attributes for ADC, it is becoming easier to demonstrate the reproducibility of conjugation and ADC analysis. Meanwhile, McDermott asserts that “Advances in method development require subject matter experts in both fields to forge integrated techniques that will bridge the knowledge gap and blend pharma and biotech platforms, and the interface of these analytical techniques allows real understanding of the ADC construct to be achieved.”

Importantly, sharing of information is taking place not only between different groups within pharmaceutical manufacturers, but also between companies. “Cross-fertilization within the biopharmaceutical industry, particularly driven by the proliferation of ADC-focused conferences, workshops, instrument company webinars, etc., has lead to an increased convergence in the methods being used,” observes Jacobson. “While there are clearly some differences, for example those that might result from the specific requirements imposed by a particular company’s unique conjugation technology, in general good ideas are being picked up and incorporated widely as they are introduced,” he continues. Peckman agrees. “I think the most helpful resource is the sharing of information from technical presentations and technical papers. We are at a point where many groups are experiencing the same challenges with characterizing ADCs, and evaluating the research results obtained by others inspires new ideas for method changes in the future."

Evolving ADC chemistry
Adding to the complexity of ADC method development and validation is the growing diversity of bioconjugates being advanced into the clinic. “Traditional cysteine and lysine chemistry is still a significant part of the ADC regime, but we are seeing a greater breadth of linkers, toxins, mAbs, and classes of conjugation chemistry. New analytical methods are needed for each of these new classes, both for characterization of the chemistry and release and stability testing,” says McDermott.

In addition, as more products make it into the clinic and then into the market, Jacobson expects that regulators will become more familiar, and more comfortable, with the ability of current analytical tools to demonstrate the robustness of ADC manufacturing. “One consequence may be a better definition of what really needs to be controlled by product testing. In such a rapidly evolving area, the regulatory requirements are clearly evolving as more knowledge makes it into health authority submissions and into the peer-reviewed literature,” he says. 

On the other hand, as the ADC industry continues to expand at pace and current understanding of the complex interactions between drug payloads and antibodies increases, Davidson notes that new problems are constantly being uncovered that often require analytical solutions. “It can therefore be expected that more issues will be identified and there will be more challenges to come,” he concludes.

References
1. ICH, Q5C Stability Testing of Biotechnological/Biological Products, (1995).
2. ICH, Q6B Specifications: Test Procedures and Acceptance Criteria for Biotechnological/Biological Products, (1999).

About the Author
Cynthia A. Challener, PhD, is a contributing editor to BioPharm International.

Article DetailsBioPharm International
Vol. 28, Issue 2
Pages: 17–21

Citation
When referring to this article, please cite it as:
C. Challener, " Tackling Analytical Method Development for ADCs," BioPharm International 28 (2) 2015.