Scrubbing Host Cell Proteins Out of Antibody Drugs A Two-Acid Elution and a Depth Filter

This application covers a purification recipe for antibody drugs grown in cells. The claimed method pulls the antibody off an affinity column with a mixture of a weak acid and a strong acid, raises the pH of the eluate above about 6.0, and then passes the result through a depth filter, with the aim of leaving far fewer of the production cell’s own proteins in the product. It was filed by Eli Lilly and Company and names Brian David Bowes and Lara Ellen Krebs as inventors. The application is a divisional of Australian application 2021355518.

The Problem

Drugs such as monoclonal antibodies are made by growing engineered cells, commonly Chinese hamster ovary cells, and then fishing the product out of the broth. The broth also contains the host cell’s own proteins. The specification explains that these host cell proteins threaten product quality and patient safety by causing aggregation, fragmenting the product through enzyme activity, or provoking an immune reaction, and so they are treated as a critical quality attribute of any formulation.

The usual answer is more purification steps, and the specification says these cost yield and money. It also notes that existing approaches can remove host cell proteins incompletely or inconsistently, can drag the wanted protein out along with the contaminants, and are sometimes limited to one molecule or format. The stated need is a way to cut host cell proteins without hurting stability, yield or cost.

What This Invention Does

Claim 1 is a method of reducing host cell protein content in a preparation containing a recombinantly produced protein of interest. There are four steps. The preparation goes onto an affinity chromatography column. The protein is eluted with a combination of acids made up of a weak acid and a strong acid. The pH of the eluate is raised to above about pH 6.0. The eluate is then put through a depth filter to give a filtered preparation.

The dependent claims narrow this. The column can be a Protein A, Protein G or Protein L column, and Protein A is the one used in the worked examples. The weak acid is acetic acid at about 20 mM, paired with phosphoric acid at about 5 mM to 10 mM or lactic acid at about 5 mM. An optional viral inactivation step holds the eluate below about pH 4.0 for up to 180 minutes, with claims putting the target at about pH 3.3 to 3.7 and then about pH 3.5. Claim 34 is a fully specified version for an anti-SARS-CoV-2 antibody: Protein A capture, acetic acid with phosphoric or lactic acid, about 20 mM HCl to hold the eluate at pH 3.3 to 3.7, about 250 mM Tris buffer to bring it to pH 6.5 to 7.5, then a depth filter. Further dependent claims name bamlanivimab, etesevimab and bebtelovimab by sequence.

The numbers in the examples are the point. In Example 1, using the antibody etesevimab, total host cell protein measured by mass spectrometry was 23,299 ppm after Protein A elution. In a later run quoted in the specification the figure fell from 179,964 ppm after elution to 77 ppm after depth filtration through an X0SP filter. In Example 2, using bamlanivimab, three elution buffers were compared at 20 mM acetic acid plus 5 mM of a second acid. After the depth filter, the citric acid version left 469 ppm, while phosphoric acid left 7 ppm and L-lactic acid left 16 ppm. The specification notes that the phosphoric and lactic acid combinations also took a troublesome contaminant, phospholipase B-like 2, below the limit of quantification.

A further example looks at ionic strength. With a Protein L elution and a Tris neutralisation to pH 7.0, mixtures applied to the depth filter at below about 45 mM ionic strength gave under 50 ppm of host cell protein, for example 38 ppm at 38 mM and 26 ppm at 23 mM, and the specification reports a correlation between lower ionic strength and lower contamination. It adds that diluting to lower the ionic strength works but the extra volume is a drawback in manufacturing.

Key Features

• Two acids, not one. Elution uses a weak acid plus a strong acid, typically 20 mM acetic acid with 5 to 10 mM phosphoric or 5 mM lactic acid, rather than a single acid.
• Neutralise, then filter. The eluate is taken above about pH 6.0, and in the narrower claims to pH 6.5 to 7.5 with about 250 mM Tris, before it reaches the depth filter.
• Depth filtration as the clean-up step. The specification names several depth filter media, including X0SP, COSP, X0HC and Zeta Plus grades, and shows total host cell protein dropping below 100 ppm, and in places below 20 ppm.
• Low ionic strength. Embodiments specify an eluate ionic strength of about 10 mM to 45 mM at the filter, with some below 30, 20 or 15 mM.
• Built-in viral inactivation. The low pH hold at about pH 3.5 for up to 180 minutes doubles as a viral inactivation step, so the process does not need to add a separate one.
• Not tied to one molecule. The specification says the approach applies to a broad range of proteins, and illustrates it with several antibodies and mammalian host cells.

Who Is Behind It

The applicant is Eli Lilly and Company, the Indianapolis pharmaceutical group founded in 1876. Its Wikipedia entry records that in 2020 it began work with AbCellera on a monoclonal antibody for COVID-19, which became bamlanivimab. That history explains why the claims and examples centre on anti-SARS-CoV-2 antibodies.

On priority, the specification states that this divisional comes from Australian application 2021355518, which is the national phase of international application PCT/US2021/053318, which in turn claims the benefit of US Provisional Application 63/086,915 filed 2 October 2020. The priority country is therefore the United States. The title page names Spruson and Ferguson of Sydney as the attorney.

Why It Matters

Removing host cell proteins is a routine, expensive part of making any biologic. Each extra column or filter step adds time, buffer and product loss, so a recipe that reaches single or double digit ppm using steps the process already contains, an elution, a pH adjustment and a filter, is attractive to a manufacturer. The specification says as much, stating that some embodiments let additional purification steps be dropped, reducing batch processing time and cost.

It is worth reading the claims with care. They are written for protein preparations in general, and the examples are laboratory scale tests on a handful of antibodies. The document does not say how the process performs at commercial scale, and the divisional form means the broader or narrower scope will depend on how examination goes.

Related Concepts

• Monoclonal antibody – the class of drug whose manufacture this purification method is aimed at.
• Protein A – the bacterial protein used in the capture column that binds antibodies before the acid elution.
• Affinity chromatography – the separation technique behind the first step of the claimed method.
• Depth filter – the final filtration step that captures contaminating proteins throughout a porous medium.
• Chinese hamster ovary cell – the main host cell line for antibody manufacture and the source of the contaminating proteins.

---

AU 2026202259 was published in the Australian Official Journal of Patents on 16 April 2026 and is open for public inspection. Patent applications represent inventions that are sought to be protected and do not necessarily reflect commercially available products.