Purpose: Complex sample matrixes present a significant challenge for most biomolecular binding assays, and, thus, require several purification steps for obtaining reliable assay results. This leads to lengthy assay development times, a need for time-consuming purification steps, and ultimately increased operational costs. Flow induced dispersion analysis (FIDA) is a novel capillary-based technology for characterizing ligand-protein interactions and quantifying biomolecules in complex solutions (e.g., plasma). FIDA is based on Taylor dispersion in a pressure driven flow of a ligand (termed indicator, here Protein-A) interacting selectively with the analyte of interest (here human IgG). FIDA and Taylor dispersion analysis is performed in narrow capillaries and is used for measuring diffusion coefficients and hydrodynamic radii of molecules. In FIDA, the apparent change in hydrodynamic radius of the indicator molecule upon binding to the analyte is exploited for determination of analyte concentration. Here, we present a FIDA method for quantification of human IgG directly in complex F12 fermentation media used in antibody production with CHO cells, without any need for purification.


Methods:

Alexa488-labelled protein-A (Product no. 11047; Molecular Probes, Eugene, OR, US) served as the indicator (100 nM) for quantification of IgG from human serum (analyte; Product no. I 4506; Sigma) spiked (0-1000 nM) into 80% (v/v) F12/DMEM fermentation medium. The advanced F12/DMEM fermentation medium (Product no. 12634, Thermo Fischer Scientific) is commonly used in the production of therapeutic proteins using Chinese Hamster Ovary (CHO) cells as the host cells.

FIDA analyses were performed using a FIDAlyzer instrument with LED fluorescence excitation at 480 nm and emission collection at λ > 510 nm (FIDA-Tech ApS, Copenhagen, Denmark) using a coated fused silica capillary (i.d.: 75 µm, L_T: 100 cm, L_eff: 88 cm) at 25 ^oC.

Sample analysis was performed by filling the capillary with the analyte solution (80% (v/v) F12 fermentation medium with a spiked amount of human IgG), followed by injection of 39 nL of the protein-A-Alexa488 indicator solution (by pressure), which was mobilized towards the detector with analyte solution at 400 mbar. A 67 mM phosphate buffer, pH 7.4, was used as rinse and working buffer. Diffusion coefficients were determined by fitting the Taylorgrams (peaks) to a Gaussian function. The determined peak variances were used to calculate the diffusion coefficients allowing hydrodynamic radii to be calculated from the Stokes-Einstein equation.


Results:

Figure 1 shows how the peak width (variance) of the protein-A-Alexa488 indicator increases with increasing IgG concentration. Figure 1 shows that human IgG can be detected in the F12  fermentation medium without sample preparation even though the medium is complex containing amino acids, vitamins, salts, proteins and sugars. The auto-fluorescence of the medium is apparent as base line shift but does not affect the peak shape. The FIDA approach is characterized by a short analysis time (3 min) and selective fluorescence detection allowing the protein-A-Alexa488 indicator to be determined at a relatively low concentration (100 nM).  The indicator (MW 425000 Da) is small relative to the IgG antibody, but upon binding, the diffusivity will decrease corresponding to a larger apparent hydrodynamic radius. The apparent hydrodynamic radius of protein-A-Alexa488 is plotted as a function of IgG concentration  in Figure 2. The change in apparent size forms the basis for a measure of the interaction and the IgG (analyte) concentration. The binding of protein-A to IgG was expected (protein A binds to the Fc portion of many subclasses of immunoglobulins) and is apparent from the size increase. The unbound Protein-A had a hydrodynamic radius of 5.5 nm, but upon binding to one or two IgG molecules it increased gradually to approximately 11 nm. The data points were fitted to a 1:1 binding isotherm (red line in Figure 2), and a dissociation constant (K_d) for the protein-A-IgG complex in 80% F12 fermentation media of 60 nM was determined. The results document the ability of FIDA to characterize interactions in native medium. Furthermore, the binding curve serves essentially as a calibration curve allowing the quantification (working range 200 – 1000 nM) of the IgG in F12 fermentation media upon determination of the protein-A-Alexa488 apparent hydrodynamic radius.

Conclusion: The FIDA methodology is applicable for biomolecular quantification and for measuring binding affinity directly in complex samples, thereby reducing sample purification requirements. Further, studies are planned to investigate the feasibility of quantifying IgG in CHO cell containg medium upon expression of the IgG.

Co-Author(s)

• MP

  Morten Pedersen

  – FIDA-Tech ApS
• SG

  Sarah Gad

  – FIDA-Tech ApS
• BS

  Brian Sørensen

  – FIDA-Tech ApS

Presenting Author(s)

• JO

  Jesper Ostergaard

  – Associate Professor, University of Copenhagen

Main Author(s)

• JO

  Jesper Ostergaard

  – Associate Professor, University of Copenhagen

Primary/Principal Investigator(s)

• JO

  Jesper Ostergaard

  – Associate Professor, University of Copenhagen

Co-Author(s)

• HJ

  Henrik Jensen

  – University of Copenhagen
• JO

  Jesper Ostergaard

  – Associate Professor, University of Copenhagen