By Stefan Persson, PhD, Resistentia Pharmaceuticals Monday, November 23, 2009

An SPR-based method, Calibration Free Concentration Analysis can be used to accurately determine the concentration of active protein in a sample, relating to the specific binding activity of the protein, and without the need for a standard.

There are many stages of drug development where determining the concentration of active protein in a given sample is critical. By measuring the concentration of an active protein and analyzing its specific binding activity, it is possible to optimize production process of most biologicals and increase the amount of active drug. In addition, reducing the proportion of inactive substance improves the safety profile of a biological drug. In the past, it has been difficult to obtain accurate concentration measurements and comparison of active to total protein concentration has been a complex process.

click to enlarge   Figure 1: Initial binding rate of the protein

Traditionally, protein concentration measurement has required comparison of the sample to a known standard. For example, in colorimetric assays, a spectrophotometer is used to measure light absorbance of a sample, which can in turn, give a measure of concentration. However, as measuring absorbance of light tells very little on its own, the sample must be compared with a standard.

HPLC may be used to determine the concentration of a sample, as the detector responds to the concentration of a compound band as it passes through a flow cell. This produces a strong signal and a large peak. Concentration is calculated by measuring the area under a peak, and comparing to a reference peak.

Both methods can give accurate results, but the need for a standard/calibrant means that the measured value will only be as accurate as the standard/calibrant used. Further problems occur when factors causing variation lead to an unreliable standard, and, in many cases, a standard for a protein may not be available.

A UV absorbance assay (often termed OD280) is a faster and more convenient method for protein concentration measurements, as additional reagents and incubations are not required, nor is a standard. The method is based on the fact that a protein in solution will absorb ultraviolet light at 280 and 200 nm, and the relationship of absorbance to protein concentration is linear. However, the method can only be used to give a rough estimation, as any conclusion is prone to error due to the fact that different proteins have varying absorption characteristics. In addition, any non-protein component of the solution that absorbs ultraviolet light will interfere with the assay and affect the reading.

click to enlarge   Figure 2: Initial binding rate

The main drawback of the methods described above is that they only measure the total amount of protein in a sample and are unable to distinguish between active and inactive molecules. One tool that has recently been developed to address this issue is Calibration Free Concentration Analysis (CFCA) with Biacore T100 software version 2.0, designed for use with Biacore T100. The instrument is part of the range of Biacore label-free interaction analysis instruments, using SPR (Surface Plasmon Resonance) technology to characterize molecules in terms of their specificity of interaction, binding kinetics, and affinity.

CFCA does not require a standard curve for measuring concentration. Instead the concentration determination relies on changes in the binding rate of a protein in a sample with varying flow rates, when the transport of molecules to the sensor surface is limited by diffusion from the sample.

click to enlarge   Figure 3: Measurement of specific binding activity of the protein during exposure to stress in the absence of a reliable standard.

Although it has some limitations—for example it can only be used for analyte-ligand interactions with fast association rates (k_a > 5*10⁴ M^-1s^-1) and/or high affinity (KD ≤~1*10^-7 M)—CFCA can bring great value to the discovery and development of therapeutic proteins. The method enables concentration analysis where no satisfactory calibration standard is available, and can also be used to check the validity of specified concentrations in standards. Most importantly, CFCA can be used to measure the concentration of active protein in a sample accurately, relating to the specific binding activity of the protein.

Materials and Methods
During development of a therapeutic protein, we observed some variations in activity of the protein, and differences between batches. Initially a spectrophotomic method was used to measure concentration of the protein in each batch, however as no well-characterized standard was available, the CFCA method was trialed to accurately assess the concentration of active protein, binding activity of the protein against a range of ligands, and its stability under stress.

A test preparation of the protein, a specifically designed, recombinant Immunoglobulin E (IgE)-derived immunotherapeutic protein, was taken during development and exposed to a form of stress. Three batches of the therapeutic protein were tested against three different interaction partners (ligands) bound to the sensor surface of a Biacore sensor chip CM5. The chip is formed of carboxymethylated dextran covalently attached to a gold surface and is the most versatile of Biacore sensor chips. As the immunotherapeutic protein induces self anti-IgE antibodies and prevents IgE from binding to its receptor, FcεRI, the alpha-chain of the IgE receptor was used as one of the ligands. Also used as ligands were mouse anti-human IgE antibody and Xolair, an IgE blocker containing the active ingredient omalizumab.

Direct immobilization (covalent linkage of the ligand to the sensor surface) to sensor chip CM5 is recommended when carrying out CFCA, as this provides the high binding capacity required for the assay. Capture approach—where a capturing molecule is covalently linked to the sensor surface, and the ligand bound to the capture molecule—may be used as a second choice. However, due to the lower binding capacity of capture approach, it is more difficult to reach the high immobilization levels necessary. Direct immobilization using a standard amine coupling procedure was used in this instance. The recommended immobilization level for molecules with M_w ~150,000 Da is above 5000 RU (resonance units), to favor mass transport (or diffusion rate) limited binding. Mass transport limitation arises when the analyte (therapeutic protein) binds to the sensor surface faster than it diffuses from the solution to the surface during injection, or when the analyte doesn’t diffuse fast enough from the surface during dissociation, leading to re-binding. Mass transport limitation is a disadvantage when carrying out kinetic analysis, as it may result in the rate of diffusion being measured rather than the actual interaction rate, but is an advantage in CFCA.

To measure the concentration using CFCA, a relationship between initial binding rate and analyte concentration is used. On a sensor surface with a high immobilization level and under partial or complete mass transport limited conditions, the initial binding rate as shown in Figure 1 can be described as a function of the molecular weight of the analyte (M_W), mass transport coefficient (k_m) and concentration of the analyte (Figure 2). As all parameters apart from the concentration are either known or can be measured, this function may be used to calculate concentration of the sample.

In the laminar flow conditions that apply in Biacore, the mass transport coefficient (k_m) can be calculated automatically by the Biacore T100 software, using the dimensions of the flow cell, flow rate, and diffusion coefficient. The flow cell dimensions are known and included in the software, and the flow rate is entered during assay set-up. The value of the diffusion coefficient (D) must be provided by the user and for many proteins may be found in the literature. The value of D can also be determined by taking the size and shape of the molecule together with the viscosity of the medium and the temperature, and using the Biacore T100 Coefficient Calculator (www.biacore.com).

click to enlarge   Figure 4: Specific binding (%) of three batches of protein, each tested against three different interaction partners.

In our example, the sample protein (recombinant IgE-derived therapeutic protein) was run at two flow rates (5 and 100 μl/min) against each ligand, with a sample injection time of 36 seconds and each injection carried out in duplicate. Biacore T100 determined the initial binding rate (dR/dt) of each run using SPR technology. The software was then able to evaluate the concentration of the sample, setting the analyte concentration as a parameter to fit and k_m as a known constant together with M_W.

To help eliminate systematic variations in the response and improve the robustness of the assay, a blank cycle (injecting buffer instead of sample) was run for each flow rate, and the blank cycles were run with the same flow rate and contact time as the sample cycles from which they were subtracted. When entering values in the CFCA method builder for these blank cycles, M_W and D were left blank.

Review and conclusions
The results of CFCA and kinetics measurements of runs against the alpha-chain of the IgE receptor demonstrated that after two hours exposure to stress, the protein had lost 15% of its binding activity, and, after 24 hours specific binding activity, had dropped to about 60%. However, it was observed that loss of activity following up to eight hours exposure to stress depends exclusively on a decrease in the specific binding activity (Figure 3), with no change to the rate constants. After 24 hours’ exposure, k_a (the rate of complex formation) and K_A (association constant) were also seen to begin to change, although any differences observed were minimal.

Following the application of CFCA with each of the three different ligands (IgE receptor alpha-chain, mouse anti-human IgE antibody, and Xolair), specific binding activity of the protein was determined. It was observed that when compared to the total protein concentration (as determined by spectrophotometry), specific binding activity constitutes between 62-95% of total protein concentration. The results showed consistency between batches when tested against the three different ligand molecules (Figure 4).

Following these initial experiments, a CFCA assay was developed for quantification of the therapeutic protein during development, and for implementation as a Quality Control assay at a CRO/CMO. We found that the assay developed using this method is fast and reliable, with the major advantage of eliminating the need to depend on the performance of standards. Moreover, sample consumption is very low, minimizing the need for scaling up of the drug/protein.

About the Author
Stefan Persson gained his PhD in Pharmacology from Uppsala University, following which he worked as a scientist in Immunology at Lund University. He is currently vice president, Pharmacology, at Resistentia Pharmaceuticals.

This article was published in Bioscience Technology magazine: Vol. 32, No. 11, November, 2008, pp. 14-18.