Employing genetically modified biological systems to synthesize therapeutic proteins has been widely successful. However, in addition to producing the recombinant therapeutic protein, host organisms generate endogenous proteins associated with regular cell functions. These host cell proteins (HCPs) represent most bioprocess-related impurities and have the potential to affect the safety and efficacy of the generated product. As such, strict regulations are in place regarding HCP detection, quantitation, and removal.
HCP composition and abundance are distinctive to each host system and production process used for therapeutics manufacture. Many HCPs are present in low amounts, making them challenging to detect and measure. Due to the diversity and complexity of HCPs, each with varying physiochemical and immunological properties, immunoassays (typically a sandwich enzyme-linked immunosorbent assay (ELISA)) are currently the industry standard for HCP analysis, providing high sensitivity and high throughput.
The most critical and first experiment needed to assess the accuracy and specificity of an HCP immunoassay is dilution linearity (DL). While ELISAs are sensitive enough to detect nanogram levels of HCPs present in a substrate, they are limited by the polyclonal antibodies used for the assay. If antibodies are not specific enough or well-correlated with the number of HCPs in a sample, lack of antibody excess, or antigen excess, and lack of dilutional linearity can occur. This can be caused by a subset of HCPs that are not proportionately removed during purification due to some affinity for the drug substance or the purification columns used. Additionally, the product protein itself or certain components in the sample matrix may interfere with the ability of the assay to detect HCPs or other contaminants, causing poor DL.
Image: Integration of orthogonal methods for HCP analysis
DL is performed by assaying a series of doubling dilutions of a sample over the analytical range of the assay. Ideally, the dilution corrected value should be constant within the precision limits of the assay. When properly performed, DL can identify and even discriminate sample matrix interference from antigen excess. This initial assessment can direct the approach taken to solve any assay inaccuracies.
In the case of matrix interference, components in the sample (pH, salts, detergents, solvents, excipients, or the drug substance) can interfere in antibody:antigen binding. Provided the sample has antigen concentrations well above the lower limit of quantitation (LLOQ) of the assay, dilution of the sample in an assay compatible diluent can be the easiest way to overcome the matrix interference. In most cases matrix interference will initially manifest as an increase in dilution corrected values. A dilution will then be reached after which subsequent dilutions yield the same dilution corrected value, known as the minimum required dilution or MRD. When dilution is not an option because the antigen levels fall below the LLOQ, other methods such as pH adjustment, assay protocol modifications, or buffer exchange may overcome the interference.
For sandwich immunoassays to be accurate, there must be an excess of antibody for each antigen. When the antigen is in excess of antibody, the assay can significantly underreport HCP concentration. Cygnus Technologies provides additional in-depth information on antigen excess (“Hook Effect”) and the stoichiometry of sandwich immunoassays when applied to HCP detection in a number of online resources. As with matrix interference, DL will show an increase in dilution corrected values with increasing dilution. Typically, an MRD can be demonstrated where the antigens are no longer in excess and the dilution corrected values become constant. The table below shows representative HCP data for an in-process sample that was underestimated due to lack of antibody excess until it was diluted to its MRD at 1:8.
While DL can be used to demonstrate antigen excess, even at the apparent MRD the true level of total HCP may still be underestimated. This can be intensified by HCP assortment, with a mix of high abundance/high immunogenicity proteins that take over ELISA signals and antibody recognition over low abundance/low immunogenicity proteins. For this reason, good purification process characterization and assay qualification are strongly supported when downstream samples, whether they achieve an MRD or not, undergo orthogonal analysis by methods like 2D PAGE and mass spectrometry to determine if there are any abundant so-called “hitch-hiker” HCPs. The integration of DL analysis together with orthogonal methods can guide purification process improvements to reduce these hitch-hiker impurities. Even if DL cannot be achieved within the analytical range of the assay this does not mean the assay is invalid. Rather poor DL is an indication the purification process needs to be further optimized to remove the most abundant impurities.
With the successful determination of an MRD for a given sample type, an additional step to ensure assay accuracy is Spike-Recovery (SR) analysis performed by spiking various levels of antigen into the sample diluted to its MRD. The analysis of both DL and SR experiments can help discriminate matrix interference from antigen excess and ultimately help determine the most effective approach to improve assay accuracy and specificity. Cygnus Technologies is pleased to offer expert interpretation of DL and SR data as well as orthogonal analytical services to identify individual HCPs.
A comprehensive list of Cygnus ELISA Kits can be found here.