Sensitivity of generic HCP ELISAs can vary
Due to the time and costs associated with the development of highly sensitive and process-specific host cell protein (HCP) assays, generic host cell protein assays are routinely used during product development. In this article, we show that the sensitivity of commercially available generic HCP assays varies with the source of HCP preparation. For some biomanufacturers, multiproduct HCP assays may provide a way out of the dilemma posed by cost and sensitivity.
Manufacturing biopharmaceuticals using genetically modified microorganisms and cell lines is usually associated with process-related contamination. Among the most important contaminants are the host cell proteins that are inherent to all production processes relying on bacterial, yeast or mammalian cell culture. HCPs can reduce drug efficacy and even delay or kill promising drug candidates due to adverse patient reactions. Therefore, biopharmaceutical manufacturing demands suitable process-accompanying analytical methods that ensure accurate determination of residual HCPs and a maximum degree of purity. Host cell proteins in the drug preparation should be “below detectable levels using a highly sensitive analytical method”. As a rule of thumb, that amount should not exceed about 100 ppm.
Host cell protein enzyme-linked immunosorbent assays
Enzyme-linked immunosorbent assays (ELISAs) with polyclonal antibodies directed against host-cell proteins are typically used to detect HCP populations. Such HCP assays are widely applied for example during process development, to demonstrate consistency of manufacturing, for routine quality control and final release testing.
The first step in the development of an HCP assay is the preparation of an appropriate host cell protein-containing solution that can be used as the antigen for the generation of polyclonal antibodies. The development of a very sensitive HCP assay capable of detecting HCP at levels as low as 10-100 ppm is challenging, and requires extensive experience in antibody development.
Another challenge is the generation of antibodies against host cell proteins that are only weakly or nonimmunogenic. The number of immunogenic HCPs can be increased by chemically modifying the antigen or by employing special immunization protocols. Various cascade immunization protocols that report improvements in immunogenicity have been described in the literature[3-4]. Several rounds of animal immunizations are often necessary to secure broad antigen recognition.
Weakly or nonimmunogenic proteins can be detected by an alternative method. One- or two-dimensional SDS-PAGE combined with sensitive silver staining is followed by protein identification using mass spectrometry and/or western blotting. Through MS, the identity of any contaminating HCP can be determined, but the method is not as sensitive as ELISA. This is because an HCP ELISA measures the collective sum of immunoreactive proteins. In other words, a signal is obtained from a number of different proteins present at low concentrations. Therefore, even when a recombinant product appears to be sufficiently pure by polyacrylamide gel electrophoresis, residual host cell proteins could still be detected by HCP ELISA. Overall, HCP ELISA is more sensitive than methods monitoring individual HCP.
Generic and process-specific HCP ELISA antibodies
In early phases of clinical development, when the risk of product failure is still high, commercially available generic HCP assays are often used to reduce costs. When the drug has passed the proof-of-concept hurdle, process-specific HCP assays that display full specificity are developed.
Generic HCP assays are supposed to detect all host cell proteins that are present in a given cell line – such as Chinese hamster ovary cells (CHO) – independent of the downstream processes employed. Such assays are based on polyclonal antibodies that were developed against a representative HCP preparation, and therefore broadly react with as many HCPs as possible. Process-specific HCP assays use polyclonal antibodies produced against HCPs prepared from a mock run of a specific fermentation and purification protocol. They are intended to recognize all process-specific HCPs, such as cell line characteristics, and also individual process parameters such as fermentation strategy and product isolation. Therefore, they are generally expected to be not as broadly reactive as generic HCP antibodies, and hence not suitable for the sensitive detection of HCPs from different production processes.
Comparison of generic and process-specific HCP assays
Despite their intended general applicability, several biomanufacturers have discovered that generic HCP assays lack sufficient sensitivity and immunoreactivity with certain HCPs. To investigate this issue, we compared the performance of five process-specific CHO HCP ELISAs developed by BioGenes GmbH [designated #1, 2, 3, 4 and 5] to that of a commercially available generic CHO HCP ELISA kit (Cygnus CHO HCP 3G ELISA Kit , catalogue number F 550; designated #6).
All ELISAs were tested against four different concentrations (12.5, 25, 50 and 100ng total protein/ml; based on the specified starting concentration) of eight different process-specific and one generic CHO HCP standardized preparations. The ELISAs were performed according to BioGenes’ in-house and the manufacturer’s protocols respectively. The relative immunoreactivity of each HCP ELISA antibody to the test HCP preparations was calculated as a percentage of the respective reference protein preparation. In this calculation, it was assumed that the generic and process-specific CHO HCP antibodies detect 100% of their respective CHO HCP reference preparation.
First: how much do HCP populations vary under different culture conditions and processing? Fig. 1 shows the variation between the process-specific and generic CHO HCP preparations used in our study. Although some proteins’ bands may occur in all HCP preparations, the relative amounts varied considerably between preparations. It is clear that process-specific parameters strongly influence the HCP profile.
The mean relative immunoreactivity of the generic CHO HCP ELISA (#6) with CHO HCPs obtained from different production processes was only 40% (Fig. 2). The generic HCP kit showed a relative reactivity above 50% in only three of eight CHO HCP preparations. The relative reactivity for CHO HCP preparation #3 was only 20%. In the case of HCP preparations #4 and #5, the relative reactivity was 40% and 25% respectively. Hence, for most CHO HCPs, the generic CHO HCP ELISA kit did not fulfill the criterion required for sensitive HCP detection. This is in contrast to the manufacturer’s specification that nearly all CHO HCPs could be detected independent of culture conditions and downstream processing. For a generic assay capable of recognizing a broad HCP population, one would expect a relative immunoreactivity of at least 70%.
As was to be expected, the results of the process-specific CHO HCP ELISAs for the detection of different CHO HCP preparations were mixed. Some process-specific CHO HCP antibodies (#3 and #4) surprisingly showed higher mean relative immunoreactivity (about 55%) than that of the generic kit (Fig. 3; data for #4 not shown), while others showed means of only 33% (#1) and 3% (#2) (data not shown). The mean relative immunoreactivity of process-specific HCP antibodies in #5 was comparable to that of the generic assay #6 (data not shown). Thus in some cases, antibodies generated against a specific set of HCPs can be even more sensitive than generic antibodies.
An alternative: multiproduct HCP ELISAs
The generic HCP assay we tested did not perform as specified. Antibody sensitivity varied considerably depending on the production protocol used for obtaining the CHO HCP. Of the five process-specific HCP antibodies tested, one showed comparable and two showed better performance than the generic assay antibodies. In view of the marked differences between HCP preparations, the varying reactivity of the generic antibody to different HCP preparations is perhaps not surprising, even if efforts were made to obtain a balanced spectrum of immunoreactivity to major and minor antigens as well as strongly and weakly immunogenic antigens. The user must be aware of this potential lack of sensitivity and design appropriate controls and tests to take this into account.
Since the development of highly sensitive process-specific HCP antibodies is associated with higher costs, the development of such assays is usually begun only after a successful clinical proof-of-concept study. This means that in Phases I and II of clinical trials, patients can be exposed to host cell protein-related risks if generic HCP assays are used. A possible alternative is developing ‘multiproduct’ HCP assays as a compromise between generic and process-specific assays. Multiproduct HCP assays are suitable for biomanufacturers that use proprietary production cell lines for several products. The basis of such assays is antibodies generated against cell line-specific HCP populations that have been prepared according to optimized process-specific cultivation and fermentation protocols. Once available, such a multiproduct HCP assay can be applied like a generic assay, with the difference that it could be made far more sensitive and specific. Multiproduct HCP ELISAs could ensure very sensitive HCP detection during process development, routine quality control or final release testing, and can be optimized if necessary for a specific production process. Since multiproduct HCP assays can be used for all products derived from a particular cell line or expression system, their development is a worthwhile investment.
 Center for Biologics Evaluation and
Research. Points to Consider in the Manufacture and Testing of Monoclonal Antibody Products for Human Use. US Food and Drug Administration, 28 February 1997.
 Eaton LC., J. Chromatogr. A 705, 1995: 105–114J.
 V. Anicetti et al., Appl. Biochem. and
Biotechnol., 22, 151–168 (1989).
 J. Thalmer and J. Freund, J Immunol.
Methods 100, 245 (1984).
 Anicetti VR, et al. Appl. Biochem.
Biotechnol. 22, 1989: 151–168.
Köpenicker Straße 325
12555 Berlin, Germany