Development of an In-House, Process-Specific ELISA for Detecting HCP in a Therapeutic Antibody, Part 2

During biopharmaceutical manufacturing, final drug products can get contaminated with host-cell proteins (HCPs) derived from a production cell line. HCPs can elicit adverse immune responses, so regulatory authorities require accurate monitoring of their presence and concentration in final drug products. Because they are robust and offer good throughput, enzyme-linked immunosorbent assays (ELISAs) are the first choice for HCP detection to monitor product quality. Generic ELISA kits are commercially available for HCP detection with a number of commonly used biopharmaceutical production cell lines. Use of such kits could obviate the necessity of generating new HCP-reactive antibodies for each new manufacturing process. Although use of a universal kit could save time and money, universal HCP reactive antibodies may not adequately react with all HCPs potentially present in every biopharmaceutical product and unique process.

PRODUCT FOCUS: PROTEINBIOLOGICS

PROCESS FOCUS: MANUFACTURING

WHO SHOULD READ: PRODUCT AND PROCESSDEVELOPMENT, ANALYTICAL, FORMULATIONS, AND QA/QC PERSONNEL

KEYWORDS: HOST-CELL PROTEIN, SP 2/0 CELLS, IMMUNOASSAYS, DATA ANALYSIS SDS-PAGE, STAINING, WESTERN BLOTTING

LEVEL: ADVANCED

Centocor’s recently approved Simponi IgG product for treatment of rheumatoid arthritis (RA) is generated using the common production cell line known as Sp 2/0. Here we conclude our report on development of an in-house, process-specific HCP detection ELISA for measuring HCP content in purification process intermediates and final Simponi drug product. This is the first report on an in-house–developed process-specific HCP detection ELISA supporting late-stage development of a therapeutic antibody for market approval.

We compared the specificity and sensitivity of our in-house ELISA with a commercially available generic Sp 2/0 HCP detection ELISA to determine suitability of the latter for HCP detection in Simponi products. Our in-house ELISA detected as little as 3.9 ng/mL HCPs or 0.78 ng HCPs/mg IgG product (0.78 ppm) and exhibited no cross-reactivity with our IgG drug product. The commercial kit, however, showed very poor HCP detection sensitivity for our in-house Sp 2/0 lysate and cross-reacted with human IgG, giving a false readout of high HCP concentration in our IgG drug product. We concluded that our in-house, process-specific ELISA was more accurate and reliable than the generic kit for detecting HCP in our IgG drug product. The principles and methods described here could be applied to other manufacturing processes (e.g., fed-batch) with different production cell lines.

Part 1 of this report (BioProcess International, March 2011) describes materials and methods for generation of antibodies and lysates; electrophoresis, staining, and blotting; ELISA development; and data analysis. Part 1 also presented results of lysate and antibody characterization and ELISA development. Here, we describe results of comparing our in-house ELISA’s performance with that of our chosen commercial ELISA kit: the Sp 2/0 HCP Western blot detection kit from Cygnus Technologies (www.cygnustechnologies.com).

Discussion

Protein therapeutics such as monoclonal antibodies are typically generated using engineered cell culture systems, in which a protein of interest is expressed at high levels. HCPs are process-related impurities generated during biopharmaceutical manufacturing, so they must be monitored with sensitive analytical methods. In the interest of patient safety, regulatory authorities require that host-cell contaminants be quantitatively measured in final drug products (2, 3). Furthermore, consistent removal of HCPs must be demonstrated and validated for all manufacturing processes (1). Different types of processes (e.g., perfusion and fed-batch cultures) can involve different levels and different species of HCP contaminants. Among all analytical methods used to detect HCP levels, immunoassays are the preferred format because of their high sensitivity, specificity, robustness, and throughput.

Here, we have described our development of an in-house, process specific ELISA assay for quantification of HCPs in a biological product made using the Sp 2/0 mammalian production cell line. The assay demonstrated high accuracy and sensitivity, with a QL of 3.9 ng/mL, or 0.79 ng HCP/mg (ppm) IgG product. We validated utility of this assay by monitoring sequential removal of HCP from Simponi purification process intermediates. When applied to determine the HCP concentration in our Simponi primary reference standard, the assay could detect no HCP above its QL.

During characterization of the anti-HCP antibody preparation and Sp 2/0 cell lysates by 2D gel electrophoresis and Western blotting, the anti-HCP antibody demonstrated strong immunoreactivity and broad recognition to >87% of the HCP protein species on Western blot compared with silver stain. This result provided assurance that critical reagents used in the in-house ELISA are highly specific and detect sufficient amounts of protein species. We used the subsequent validated ELISA assay to successfully monitor removal of HCPs during purification by sampling from different product intermediates from five different phase 3 batches and two process validation lots. Consistency of HCP removal by our manufacturing process demonstrated no detectable level of HCPs present in all four subsequent consistency lots.

We next sought to compare our HCP detection ELISA with a commercially available assay kit. Similar performance between the two would have afforded us the luxury of using the commercial assay for future sample analyses. Use of a commercial kit would be very convenient and relieve us from continually generating and qualifying new batches of anti-HCP antibody preparations. To this end, we evaluated an ELISA for Sp 2/0 cell line HCP detection from Cygnus Technologies by directly comparing it with our in-house assay. But the commercial kit exhibited poor sensitivity in detecting HCPs from our in-house generated Sp 2/0 lysates, and our in-house ELISA detected no HCPs above the QL in HCP control lysates supplied with the kit.

Despite poor sensitivity, however, the commercial ELISA suggested the presence of significantly higher HCP levels in our drug product than was indicated by the in-house ELISA, which detected no HCPs at concentrations above the assay QL. The Cygnus ELISA product data sheet clearly outlined potential limitations of the assay, including the possibility that anti-HCP antibodies would not comprehensively detect every protein present in an end-user’s samples and that certain sample components (including human IgGs) could cause either positive or negative interference. We investigated the possibility that the product itself was causing positive interference (a false-positive signal) in the Cygnus ELISA by performing Western blot analysis of Simponi samples using the Cygnus anti-HCP antibodies. Indeed, those antibodies detected bands corresponding in molecular weight to the IgG heavy and light chains (Figures 8 and 9), which strongly suggests IgG positive interference and erroneous HCP quantification in the Cygnus ELISA.

We also sought to explain the differential sensitivity of these ELISAs in detecting HCPs in different Sp 2/0 cell lysates. We analyzed both in-house and kit lysates and HCP detection antibodies using SDS-PAGE with silver staining and by Western blotting to ascertain whether obvious differences in protein content or immunoreactivity existed. The in-house lysate contained many more protein species and appeared to be considerably less concentrated than the Cygnus lysate (Figure 7). In Western blotting, the in-house antibodies detected a vast array of proteins in the in-house HCP lysate, whereas the Cygnus antibodies detected far fewer protein species in the same lysate (Figure 8).

Limited immunoreacti
vity of the Cygnus antibodies to such a small subset of protein species in the in-house HCP lysate could explain the poor sensitivity of the ELISA kit when measuring HCP content of our in-house lysate. Limited immunoreactivity of the Cygnus antibodies could possibly be explained by the limited protein diversity of the cell lysate that Cygnus may have used as the immunogen for antibody generation and to affinity purify those antibodies. In addition, protein concentration of the Cygnus HCP lysate was dramatically underestimated, as evidenced by comparison with our in-house HCP lysate, which we carefully quantified with the Bradford assay. So the Cygnus antibodies may have been incorrectly “calibrated” when their capacity for HCP detection was measured, which could further explain the kit’s lack of sensitivity for HCP detection in our in-house lysate, which was much less concentrated than the Cygnus lysate.

Furthermore, affinity purification had optimized the Cygnus antibodies for a specific subset of HCPs, and the lack of ELISA sensitivity would be exacerbated if that subset was a minor constituent of our in-house lysate —or of unclarified Sp 2/0 whole-cell lysates in general. The in-house antibodies had much greater diversity because of the greater diversity of the in-house lysate, which did not undergo clarification after generation of whole-cell extracts before use as an immunogen for antibody generation. This would explain why the in-house antibodies had greater immunoreactivity than the Cygnus antibodies to HCPs in the in-house lysates. Furthermore, the limited number of distinct proteins in the Cygnus HCP lysate could explain the poor sensitivity of our in-house ELISA in accurately measuring the HCP content of the Cygnus lysate.

An Adaptable Method

This is the first report of an in-house, process-specific ELISA for detecting HCPs in an IgG product manufactured with a mammalian expression system, which has been subsequently approved by regulatory authorities from the United States, Europe, Japan, and other countries for treating rheumatoid arthritis. With antibodies raised against the complete pool of Sp 2/0 HCPs, our in-house HCP detection ELISA had better sensitivity in detecting HCP in a crude whole-cell extract than did a “generic,” commercial ELISA kit optimized to detect a subset of the HCP pool. Also, the in-house– generated antibodies exhibited no cross-reactivity to human IgG, which demonstrates the appropriateness of using these antibodies for detecting HCP in human IgG drug products.

Developing a universal ELISA for HCP detection in finished biopharmaceutical drug products is certainly a daunting task, considering the molecular complexity of the target and the uniqueness of individual drug production and purification processes.

The process-specific HCP detection ELISA described here can certainly be used for other IgG products produced using same host cells and perfusion bioreactor processes. The process of establishing an in-house HCP assay can also be applied to other common mammalian expression systems and different manufacturing processes.

About the Author

Author Details
Edward Savino is senior associate scientist, Bing Hu is senior research scientist, Jason Sellers is an associate scientist, Andrea Sobjak is senior associate scientist, and Nathan Majewski is research scientist in pharmaceutical development; Sandra Fenton is associate director of biologics research; and corresponding author Tong-Yuan Yang is associate director of biologics clinical pharmacology at Centocor Research and Development Inc., 145 King of Prussia Road, Radnor, PA 19087, 1-610-889-4566; [email protected].

REFERENCES

1.) WHO Technical Report Series No.878 1998.WHO Expert Committee on Biological Standardization: Forty-Seventh Report, World Health Organization, Geneva.

2.) Office of Biologics Research and Review 1995.Points to Consider in the Production and Testing of New Drugs and Biologicals Produced by Recombinant DNA Technology, US Food and Drug Administration, Rockville.

3.).

4.) Schwertner, D, and M Kirchner. 2010. Are Generic HCP Assays Outdated?. BioProcess Int. 8:56-62.

5.) Hoffman, K 2010. Letter to the Editor: A Rational Approach to Process-Specific Host-Cell Protein Detection. BioProcess Int. 8:8-10.