LC-MS Detection of Host Cell Proteins: Antibody Affinity Extraction Sample Preparation Outperforms Native Digestion

Host cell proteins (HCPs) can copurify with biological drug substance (DS) and pose potential risks for both patients and drug manufacturers by causing immunogenicity, diminishing DS efficacy, and/or compromising DS stability. Thus, the quantity and nature of residual HCPs in DS generally are considered to be critical quality attributes (CQAs). Enzyme-linked immunosorbent assays (ELISAs) are the gold-standard analytical method for measuring total HCP levels to support both in-process testing and product release. Liquid chromatography with mass spectrometry (LC-MS) has emerged as an orthogonal tool for better understanding of HCP properties and thus more effective removal.

Figure 1: Quantitative Venn diagram displays unique identifications of host cell proteins (HCPs) in drug substance 1.

MS is a powerful complementary method to demonstrate ELISA suitability. When combined with antibody affinity extraction (AAE) technology using an HCP ELISA antibody, MS can be used to demonstrate that an HCP ELISA is suitable for monitoring purification process consistency and product lot release (1).

A major challenge for LC-MS–based HCP identification methods is that there can be a >5× difference in concentration between HCPs and DS (e.g., a monoclonal antibody, mAb) in solution. That precludes the effective identification of low-abundance HCPs. To overcome this challenge, efforts have been made to optimize sample preparation and improve the dynamic range (e.g., with on-line or off-line fractionation) for mAb removal by affinity depletion or molecular weight cut-off (MWCO) and HCP enrichment.

The native digestion (ND) protocol introduced by Huang et al. has become a popular method (2). ND is simpler, faster, and more robust than a traditional “bottom-up” proteomics approach (denaturing digestion), with a limit of detection (LoD) <10 ppm when combined with analytical flow-through LC. However, the ND technique carries an associated risk of recovering a small subset of low-level HCPs.

Developed by Cygnus Technologies in 2013, AAE sample preparation is highly effective for enriching HCPs and depleting DS. When comparing the relative abundance of DS and HCPs with extracted ion chromatograms (data not shown), most peptides in a pre-AAE sample belong to DS. Following AAE enrichment, the relative abundance of HCPs dramatically increases while DS heavy and light chains decrease.

Below we compare ND and AAE enrichment of HCPs to fully characterize HCP profiles for two separate DS samples. Results indicate that AAE-based sample preparation is superior to ND enrichment.

Figure 2: Quantitative Venn diagram displays unique identifications of host cell proteins (HCPs) in drug substance 2.

Materials and Methods

AAE Method: A Chinese hamster ovary (CHO) 3G polyclonal antibody from an F550-1 ELISA kit (Cygnus Technologies) was immobilized covalently on a separate chromatography support. The column was conditioned to prevent significant antibody leaching and to minimize nonspecific binding. HCP-containing DS samples were passed over the column on an ÄKTA 25-L fast protein liquid chromatography (FPLC) system (Cytiva) for binding HCPs and collecting elution fractions. The latter were neutralized, pooled, buffer-exchanged, and concentrated.

Table 1: Total, shared, and unique host cell proteins (HCPs) identified using each sample-preparation strategy with drug substance #1 (Figure 1); AAE = antibody affinity extraction, ND = native digestion.

ND Method: This procedure was adapted from Huang et al. (2).

LC-MS: Pre-AAE (pre-enrichment DS sample) and post-AAE proteins were dissolved, reduced, alkylated, tryptic-digested, desalted, and concentrated. For LC-MS analysis, we used a Vanquish Horizon ultrahigh-performance liquid chromatography (UHPLC) system coupled to an Orbitrap Eclipse Tribrid mass spectrometer (Thermo Fisher Scientific). A survey spectrum (m/z range 350–1700) at 240K resolution preceded MS/MS (375–2000 m/z) of the most intense, multiply charged ions using collision-induced dissociation. Pre-AAE and post-AAE samples were analyzed independently in technical triplicate (randomized sequence). To minimize sample carryover, blank washes ran between sample injections.

Database Search: Using Proteome Discoverer software, version 2.5 (Thermo Fisher Scientific), we searched raw spectra against a proprietary database of HCPs from CHO cells. HCPs were identified with a false-positive detection rate (FDR) 0.1% and two unique peptides.

Table 2: Total, shared, and unique host cell proteins (HCPs) identified using each sample-preparation strategy with drug substance #2 (Figure 2).

Results and Discussion

We used the results from AAE and ND enrichment of two different DS samples to compare the techniques’ strengths and weaknesses. Both methods led to identification of more HCPs than was possible with no treatment at all (standard, denaturing condition), so both effectively enriched for HCPs and depleted DS for deeper proteomic coverage (Tables 1 and 2, Figures 1 and 2). That was true for DS samples containing moderately different totals of HCPs: 16 for DS #1 and 27 for DS #2.

In both cases, AAE outperformed ND in the number of proteins identified. AAE enrichment provided for 32 identified HCPs in DS #1, whereas ND enabled identification of 20 HCPs (Table 1); AAE enrichment provided for 44 identified HCPs in DS #2, whereas ND enabled identification of only 33 HCPs (Table 2). The AAE method led to identification of all HCPs found in unenriched samples, whereas ND captured only a subset (75% for DS #1, ~92.6% for DS #2). AAE sample preparation provided for identification of all HCPs that could be identified using ND, further supporting that although both enrich for similar populations of HCPs, the AAE method does so more effectively.

Discussion: Residual HCPs in biopharmaceuticals are undesired process-related impurities that need to be controlled. Understanding their physicochemical and biochemical properties is important for achieving maximal HCP removal. The key challenge of LC-MS–based methods for HCP characterization in DS is the limited dynamic range (3–4 orders of magnitude) of most high-resolution mass spectrometers compared with the broad dynamic ranges (>5×) required to detect low-level HCPs (<10 ppm). The ND method is a simple and efficient way to deplete therapeutic proteins, thus reducing the dynamic range requirement and improving the sensitivity of HCP detection. Despite its broad adaptation for HCP characterization, however, ND does not enrich low-abundance HCPs that can contribute to HCP totals in final DS as quantified by ELISA. Our results show that many HCPs are not detected when using ND but are detected with AAE sample preparation. The HCPs have been enriched beyond the LoD of LC-MS through their reactivity with the CHO HCP ELISA antibody. Note that those HCPs contribute to total HCP levels as quantified by the corresponding F550-1 CHO HCP ELISA kit (Cygnus Technologies).

AAE enrichment can miss low-level HCPs that do not react to the HCP ELISA antibody. Despite this limitation, AAE sample preparation adequately depleted the DS and enriched for many HCPs, giving it deeper proteomic coverage than ND provides. Our results confirm the superiority of AAE enrichment to detect and possibly quantify HCPs in DS samples.

References

1 Zilberman A, et al. Host Cell Protein Analysis: Immunoassays and Orthogonal Characterization By Antibody Affinity Extraction and Mass Spectrometry Methods. BioProcess Int. 20(9)si 2022: https://www.bioprocessintl.com/sponsored-content/host-cell-protein-analysis-immunoassays-and-orthogonal-characterization-by-antibody-affinity-extraction-and-mass-spectrometry-methods.

2 Huang L, et al. A Novel Sample Preparation for Shotgun Proteomics Characterization of HCPs in Antibodies. Anal. Chem. 89(10) 2017: 5436–5444; https://doi.org/10.1021/acs.analchem.7b00304.

Corresponding author Alla Zilberman is vice president, technical marketing and business development at Cygnus Technologies, LLC, part of Maravai LifeSciences, 1523 Olde Waterford Way, Leland, NC 28451; 1-910-454-9442; https://www.cygnustechnologies.com/custom-development-services/home. Darren Bertin is a research and development laboratory technician, Stephen Stahlschmidt is a laboratory technician II, Johanna Barnhill is a senior associate scientist, Timothy Licknack is a scientist, and Jared Isaac is associate director of chromatography and mass spectrometry, all at Cygnus Technologies. AAE is a trademark of Cygnus Technologies.