The Importance of Accurate Determination of ADCC Activity During Biosimilar Development
19th August 2015
By: Stefan Termen, Technical Services Scientist
Cells infected with enveloped viruses will express viral proteins on their cell membranes, and many cancer cells will express specific surface antigens. These proteins are recognized by antibodies of the humoral immune system as targets for Antibody-Dependent Cell-mediated Cytotoxicity (ADCC). In this process, effector cells of the immune system recognize the bound antibodies through their Fc receptors, and subsequently mediate lysis of the target cells. Different leukocytes act as effector cells in ADCC: natural killer (NK) cells, macrophages and neutrophilic and eosinophilic granulocytes. NK cells also target and kill cells that fail to express ‘self’ peptides correctly through the MHC Class I pathway.
Effector cells used in vitro for ADCC bioassays can be stable cell lines or freshly isolated cells from a blood donor. The key cell type are NK cells. These may be specifically isolated for use in the assay or used as part of a mixed population of cells termed Peripheral Blood Mononuclear Cells (PBMCs). Freshly isolated cells always exhibit donor variation in terms of cell yield and response. Genetics also plays a role here; the donor Fc gamma receptor IIIa (CD16a) protein exists as two types with a valine residue (V) or a phenylalanine residue (F) at position 158. A valine residue leads to a higher binding affinity for IgG. Donors with the genotype V/F (or F/V) are the most prevalent in the population and usually preferred for development studies, but at some point the ADCC activity of an antibody needs to be assessed using both V/V and F/F donors.
Therapeutic monoclonal antibodies can be used to enhance the ability of NK cells to kill tumour cells. For this purpose they are selected or designed to target a surface protein, and their administration results in the tumour cells becoming clearer targets for ADCC. Rituxan is a monoclonal antibody that targets CD20, which is expressed on B cells, and is used for treatment of B cell malignancies and rheumatoid arthritis. Herceptin is another example that targets HER2 (a.k.a. erbB‑2), a receptor often over-expressed by tumour cells, and is primarily used in breast cancer therapy. ADCC is believed to be important for the mechanism of action of both Rituxan and Herceptin, and has been suggested to play a part also for other therapeutic antibodies.
Many articles have discussed the variability of glycosylation profile between different batches of the same therapeutic antibody and its impact on observed ADCC. Whilst a product has a defined amino acid sequence, it will consist of many different glyco-forms which will contribute to the variability within the drug lot population. More afucosylated glycans present on an antibody confer a higher ADCC activity because of a higher affinity for Fc gamma receptor IIIa. Such functional differences have been seen between different batches of Rituxan (Schiestl et al., Nat. Biotech. (2001) 29, 310-312).
Remicade and Enbrel both target tumour necrosis factor alpha. Remicade has afucosylated glycans, whereas Enbrel has a higher degree of fucosylation. Remicade is also more potent in ADCC (Diagram A). Remsima, a biosimilar to Remicade, has elevated levels of fucosylation compared to the innovator molecule, which confers a lower ADCC activity. However, the manufacturer found that ADCC activity was only affected when specific assay conditions were applied and not observed in a standard PBMC assay. Analytical similarity was confirmed in clinical trials (Weise et al., Blood (2014)124, 3191-3196) in rheumatoid arthritis patients with extrapolation to other indications.
It is important to identify ADCC activity at an early stage of product development so that undesirable clones can be weeded out. Using an ADCC method that is highly sensitive to differences in glycosylation will improve the ease of demonstrating analytical similarity. As the manufacturing process evolves, assessment of ADCC forms an essential part of determining analytical similarity, potentially evaluated across a comprehensive range of different in vitro ADCC assays.
There are numerous variables that should be considered in the design of ADCC methods and within the sphere of analytical similarity. This may represent a departure from the traditional focus of defining method suitability (such as those in ICH(Q2R(1)) to the primary objective of creating methods that are exquisitely sensitive to differences that may be present between the biosimilar and innovator molecules, primarily, but not limited to, differing glycan species.