The purification of recombinant therapeutic proteins from cell culture involves two basic ideas: capture (usually chromatography) to enrich product, and various filtration, chromatography, and membrane-based unit operations to “polish” or remove impurities from the process fluid. Significant impurities of concern arise from host cells: metabolites and genetic materials, which are typically cleared during the capture and polishing steps, plus host cell proteins (HCPs), which often co-purify with the drug substance. Mammalian cells express many hundreds of HCPs along with the product protein. Most HCPs are benign but some interfere with the drug’s activity or stability, while others are immunogenic. The identities and concentrations of residual HCPs therefore become critical quality attributes, with related assays part of the program’s overall quality and risk-mitigation strategy.
During the early stages of a development program, when the identities and concentrations of HCPs are not yet known, liquid chromatography-mass spectrometry is the method of choice for characterizing HCPs. As the protein is promoted to preclinical and clinical stages, and as more is known about a cell line’s host protein profile, developers use more rapid methods such as enzyme-linked immunosorbent assay to check the efficiency of HCP removal.
To determine the suitability of an ELISA for downstream process monitoring, investigators collect and compare data on dilutional linearity, accuracy, precision, and lower limit of quantitation from upstream and downstream samples. Importantly, coverage assessment of the HCP Ab is performed to make sure that ELISA is based on an antibody that has broad reactivity to the HCPs present, to inform on those that are immunoreactive, detectable, and quantifiable by ELISA. This step is critical since validation is costly and time-consuming, and once approved by regulators difficult to change. Developers must therefore source validated ELISAs from trusted vendors who can continue to support HCP assays for the product’s entire lifecycle.
Since this is not always possible, and because ELISA kits and antibodies may show variability even from batch to batch, sponsors should be prepared to conduct bridging studies that compare the antibody coverage of new or resupply reagents with previously validated reagents.
Because of the great variability in HCP expression and the uncertainties of quantifying such low-abundance species, biopharmaceutical developers require sensitive, specific, orthogonal methods to identify HCPs that persist the purification process. These methods include Antibody Affinity Extraction (AAE™), a method offered by Cygnus Technologies, combined with 2D polyacrylamide gel electrophoresis (PAGE) or mass spectrometry. AAE methods are suitable for both in-process and drug substance HCP assays.
AAE was developed in response to challenges and limitations of common orthogonal methods such as 2D Western blot and 2D-differential in-blot electrophoresis in assessing coverage of total HCPs by polyclonal antibody reagent kits.
In AAE, the antibody is covalently immobilized on a chromatography support, then conditioned to minimize antibody leaching, which would result in nonspecific binding. The native, undenatured HCP sample is passed through the resin, where individual HCPs bind and are subsequently eluted with acid. The sample is recycled through the column and eluted until no additional HCP binds. The collected HCP elution fractions are pooled, buffer-exchanged, and concentrated back to the original sample volume. The HCPs are then separated by 2D PAGE, and compared with levels in the original sample through comparison via silver stain or by differential gel electrophoresis using Cy3 or Cy5 fluorescent labeling of HCPs in both the extracted and starting, unextracted samples.
AAE predicts how anti-HCP antibodies will perform in ELISAs, and is sufficiently sensitive for evaluation of individual HCPs that co-purify with the product. By contrast, 2D Western and differential blot electrophoresis are limited in terms of loading capacity, denaturing of native HCP epitopes, failure to transfer HCPs from the gel, sterically inhibited HCPs bound to the membrane, difficulty in aligning electrophoresis gels to Western blot images, and poor specificity. Due to these shortcomings, 2D Western blot methods significantly underestimate the true antibody coverage of upstream HCPs. Additionally, 2D Western blot does not predict how an antibody will react to those highly significant HCPs that co-purify with the drug substance.
As an illustration of the power of AAE, Cygnus Technologies recently applied the method to demonstrate the functional identity between a discontinued HCP antibody kit, F550, with its replacement panel, the F550-1 CHO 3G ELISA kit. The goal of this comparative coverage analysis was to enable developers to transition seamlessly from the original F550 kit to the F550-1 using AAE-MS for HCP detection.
Table 1. Percent Antibody Coverage and CHO HCP Numbers of F550 CHO Master Antigen (CMA) and F550 and F550-1 CHO 3G antibodies.
Cygnus scientists performed separate AAE-MS analysis on the F550 CHO master antigen panel using both the F550 and F550-1 AAE columns. By way of comparison, LC-MS detected 1,673 CHO HCPs in this collection, the F550 found 1,649 matching HCPs, and the F550-1 uncovered 1,639 HCPs (Table 1). Percent antibody coverage was identical, 97%, for both kits and the similarity between the results was 96%. Further analysis determined that the enrichment or detection of the same CHO HCPs in F550 and F550-1 kits by AAE-MS is equivalent (Figure 1).
The insights gained from AAE-MS studies comparing in-process and downstream samples may be further used for assay qualification and downstream process validation, and for the evaluation of problematic HCPs. For example, the enrichment of a single HCP during purification may only be apparent as a loss of dilutional linearity. With MS, that protein may be identified by comparing samples with good linearity with those that have lost linearity. Independent of AAE, MS also provides isoelectric point and molecular weights of analyzed species, information that process development engineers can exploit through process modifications to target the problematic protein(s) for removal or to prevent their expression.
All these benefits, plus others, can assist in formulating HCP analytics from the IND application through post-marketing, when evaluating the effects of process or reagent changes as part of general risk assessment.
Figure 1. Fold enrichment of CHO HCPs Phospholipase B Like 2 (PLBL-2), Lysosomal Phospholipase A2 (LPA2), Lipoprotein Lipase (LPL) and Serine protease (HTRA1) in CMA by F550 and F550-1 CHO 3G ELISA antibodies.