Guest Commentary: A tale of safety and consistency
A tale of safety and consistency
The challenges of bringing ADCs to market
By Drs. Ulrike Herbrand, Christopher Sucato and Alvaro Jorge Amor of Charles River Laboratories
For decades, cancer treatment was driven by chemotherapeutic agents that triggered cell death in tumors, but unavoidably imbued the entire body with a cytotoxic dose of the active compound. Small-molecule strategies were later joined by monoclonal antibody (mAb) therapeutics, which in theory had more focused targets, but displayed limited efficacy in oncological cases, as measured by only limited tumor cell death.
Antibody-drug conjugates (ADCs), which expropriate tumors and unlock a payload of toxic material, are a way around this. ADCs combine the best of protein- and small molecule-based therapies; antibodies are covalently joined to toxins via a linker in order to act as drone-like weapons against tumors.
ADCs are a booming area of development. The non-profit Antibody Society reported last year that there were 32 clinical trials of ADCs ongoing, many in Phase 2 or Phase 3 trials. The ADC market was valued at $1.3 billion in 2016 and is expected to hit $3.1 billion by 2022, according to the pharmaceutical analyst firm Allied Market Research.1 While the pharmaceutical market is still strong in terms of numbers and total revenues with respect to small molecules, 2014-2016 saw record levels of approvals for biologics. This number is expected to rise even more rapidly as the patents for the originator mAbs expire and more biosimilars come onto the market.
ADCs are also highly complex drugs which present significant bioprocessing challenges compared to standalone mAbs. Both the U.S. Food and Drug Administration (FDA) and the European Medicines Agency (EMA) treat ADCs as biologics. However, because ADCs are both drug and biological molecule—a mAb paired with a conjugate that is typically a small molecule or other chemical entity such as an oligonucleotide or peptide—the analytical tests need to be appropriate for the protein and the conjugate species.
One of the biggest challenges facing drug developers is settling on the right drug-antibody ratio. The drug-to-antibody ration (DAR) is the average number of cytotoxic drug groups that are linked to each antibody for specific delivery to cancer cells. But reaching the optimal DAR is not an easy or clear-cut process. For a drug to be viable, the DAR needs to be maintained at specific levels in reproducible fashion, which poses regulatory issues for releasing ADC lots because the biological material is highly sensitive to different stresses and chemistries in a varied manner.
Because distribution of the drug linkage on the polypeptide chains of the mAb can impact the therapeutic efficacy of an ADC, as well as drug clearance, pharmacokinetics and biodistribution, it becomes imperative for analytical scientists to determine DAR with high precision for all marketed biotherapeutics.2
The reasons why a suitable DAR can be hard to achieve are varied. It could be choice of labeling, variability of the chemistry, purification of free drug, prevention of aggregation or the ability of analytics to confirm that the DAR is correct.
Thus, robust analytics must be applied to these drugs to ensure they are produced consistently, carry a well-defined amount of small-molecule payload, maintain specific biological function through in vivo and in vitro analytical techniques, and that no unlinked payload ends up in the final drug. Here, we discuss the broad range of challenges and solutions present in the field of ADC release, including aggregation, consistent drug production and the role bioassays can play—early on—in identifying critical differences between test item and reference product.
One major issue with ADC production is aggregation, which occurs during the purification, processing or storage of the product. Charles Johnson, CEO of ADC Bio, compared aggregation to scrambling an egg, by which both processes cause loss of access to the original functional form.3
Uncontrolled aggregation can diminish the clinical efficacy in vivo or, in extreme cases, trigger a serious immunogenic response in patients after taking the ADC. Regulatory agencies have recognized the importance of screening for protein aggregation and put pressure on the biotherapeutics industry to find ways that adequately address it. In order to demonstrate clinical safety and efficacy, immunogenicity testing is now a key component of biotherapeutic drug development.4
Dynamic light scattering (DLS) is one of the most effective techniques for submicron size analysis of proteins, their aggregates and other nanoparticles. DLS provides rapid measurements of hydrodynamic radius, degree of polydispersity, temperature of aggregation onset and colloidal stability. Formulation screening with DLS helps drug developers rationalize variables such as temperature, pH or concentration in terms of their impact on stability, solubility or propensity to aggregate. Identifying the formulation conditions that are most likely to deliver ideal behavior during early-phase screening accelerates the development process and greatly reduces the risk of downstream failure.5
Size exclusion chromatography (SEC) is another tool for monitoring aggregation of ADCs. While numerous techniques have been developed to monitor protein aggregation, SEC—also referred to as gel filtration or gel permeation chromatography depending upon whether aqueous or organic solvents are used—has been predominantly favored for routine and validated analyses due to its speed and reproducibility. It is the standard approach for characterization and purification under native physiological conditions.
However, SEC may not be sufficient to fully evaluate protein aggregation, and experts warn it should be used cautiously. ADCs are complex molecules that display different hydrophobic characteristics compared to conjugated antibodies. These characteristics can lead to anomalous results when analyzed with SEC. This method is potentially prone to inaccuracies due to aggregate loss by adsorption or outright filtration of any aggregated species exceeding the SEC resin pore size.6
An alternate technique is analytical ultracentrifugation (AUC), which can probe the behavior of biological molecules in complex mixtures and at high solute concentrations. A study conducted by our laboratory explored AUC as a tool to monitor ADC aggregation characteristics as a function of formulation and ADC.7 In the study, varying amounts of aggregate from an ADC sample were measured by SEC as a function of protein load. While the SEC study produced results for aggregate peak abundance that was not linear with respect to load, the same molecule assayed by AUC under the formulation conditions displayed aggregate levels that were in fact correlated with the load amount, as expected. The poor trending showed by SEC may be attributable to the non-ideal behavior of the hydrophobic regions introduced by conjugation, and possible interactions of these groups with the stationary phase of the column—a factor not encountered by AUC, which can be done in purely aqueous phase, directly in the formulation conditions.
Balancing DAR and free drug
With dozens of ADCs now in preclinical or clinical development, one of the biggest challenges for developers is determining the concentration of unconjugated drug to ensure the safety of the product. Finding the correct DAR is also important for therapeutic efficacy. In cancer treatment, for instance, low DAR could reduce antitumor efficacy, while high DAR may affect antibody structure, stability and antigen binding, leading to a loss of activity.8 Depending on the properties of the drug, the ideal DAR for most ADCs is between 2 and 4.9
These twin goals of consistent drug production and patient safety are often what regulators look for when reviewing a biologic. Clinically approved ADCs have used conjugation chemistries with broad group specificity, targeting naturally occurring amine (lysine) or thiol (cysteine) amino acid side chains.10 Multiple and variable drug incorporation into the ADC is possible, but have to be controlled to maximize efficacy and to meet the regulatory requirements regarding drug entity definition.11
But as the ADCs evolve, meeting the expectations of regulators is becoming more challenging. For one, the amount of label can be inconsistent in the conjugation reaction of the mAb with the small-molecule drug. Inconsistencies also occur when the antigen binding site is blocked during the labeling process, which renders the antibody ineffective.
The cancer drug Kadcyla (trastuzumab emtansine), approved in 2013 for the treatment of HER2-positive metastatic breast cancer, is a good example of older generation technology, where the use of fairly nonspecific labeling is allowed to free lysine residues. This strategy presented a problem, however. Because numerous lysine residues exist on the surface of a given antibody, this reaction has poor definition of where drug conjugates are linked to. Each ADC species had the potential to exhibit different toxicities and properties relating to the absorption, distribution, metabolism and excretion of the molecules; in other words, each batch of a drug was different.
Early ADC research found that high DAR was associated with increased clearance, the potential for aggregation and increased toxicity, and interest in ADC development waned. But the field of ADCs is surging once again, due primarily to a more defined labeling system with cysteine engineering and chemistries. These modifications are aimed at helping developers meet regulatory expectations of consistent yield with minimal site homogeneity. Antibodies with site-specific conjugation chemistries are now sought to improve the ADC homogeneity that regulators are seeking.
The primary way of testing the new generation of site-specific homogeneous conjugates is with high-quality mass spectrometry (MS) and protein characterization. MS can analyze the structure of ADCs and liquid chromatography coupled to MS can characterize primary sequence conformation and identification of modifications. LC-MS can also be used to identify sites of attachment of the chemotherapy drug, and calculate the DAR.12 Biomolecular mass spectrometry, sometimes called native MS, is a versatile and efficient tool, capable of providing a direct snapshot of ADC.
There are also new conjugation methods designed to reduce heterogeneity and to focus more on site-selectivity. For example, a customized platform called the THIOMAB antibody drug platform that uses site-directed mutations to incorporate cysteines into the antibody is being used to improve the conjugation linker joining the antibody and the payload. Highly reactive thiols, which play a key role in protein folding, are the favored functional groups to attach payloads to the linker. This strategy will ultimately lead to an ADC with full control over site reactivity and a precise DAR, the authors of the study reported.13
Hydrophobic interactive chromatography (HIC) is also used to determine the DAR. This technique works by separating different species of drug conjugated antibody and calculating the contributing proportions of the drug conjugated antibody from the whole. The method used to determine DAR will vary with respect to the chemical properties of the toxin, linker and the chemistry used to attach the payload to the antibody. Lysine conjugates, because of their highly heterogenous nature, are less amenable to HPLC characterization methods, therefore spectroscopy or MS techniques may be used to measure DAR. Cysteine conjugates or site-specific conjugates on the other hand are better suited to be analyzed by HPLC methods.1
Measuring biological activity
The determination of biophysical properties is crucial during all phases of the life cycle of an ADC. The methods do provide information on single aspects of the drug; by using different methods, the ADC picture becomes clear. However, these techniques do little in displaying the biological activity of a biotherapeutic. This is where bioassay work becomes important. Though in most instances bioassays cannot provide the comprehensive explanation for activity losses, what they can do is cover the impact of all potential differences of a certain test item compared to the reference product caused by protein degradation, whether intentional or otherwise. Thus, it is often recommended to portfolio a given drug using a variety of biochemical, biophysical and bioassay approaches.
While it’s true that bioassays can be tedious to set up, it is nonetheless important to start using them early in the developmental process of an ADC or its biosimilar. The method of choice is dependent on the developmental phase as well as on the testing purpose. Assays intended for lot release need to be robust, reproducible and reliable in order to fulfill validation requirements.
Often the bioactivity assay of choice for ADCs and QC purposes is a cytotoxicity assay, which measures the inhibition of proliferation or (in some cases) a pathway-specific cell death. These assay types, when carefully set up, usually fulfill the requirements of a guideline-compliant method validation and show stability-indicating properties, which complement biophysical data. Furthermore, during the early phase of development, methods that reflect the internalization capabilities and effects of the cytostatic payloads, e.g. cell cycle arrest, deliver valuable information on the drug behavior in vitro.
Ulrike Herbrand, Ph.D., is a scientific director, global in-vitro bioassays for Charles River in Erkrath, Germany. She is an expert in mechanism of action-reflecting bioassays for protein therapeutics, specifically monoclonal antibodies.
Christopher Sucato, Ph.D., is a senior scientist, biophysical characterization for Charles River in Woburn, Mass. Christopher is the lead scientist for biophysical characterization, with research experience including characterization of protein-protein and protein-nucleic acid interactions by a range of techniques, the evaluation of kinetic interaction constants and protein expression and purification
Alvaro Jorge Amor, Ph.D., is a scientist, biophysical characterization for Charles River in Woburn, Mass. He is a leading technical scientist, specializing in protein characterization and overseeing biophysical studies which include AUC, SPR and spectrophotometric analysis.