The case for calibration-free concentration analysis

ABSTRACT

Accurate measurement of protein concentration is crucial in biological research, where protein-based reagents play a key role in assay performance and reliability. Traditional methods of protein quantification often measure total protein concentration, failing to account for the active portion capable of binding to its intended target. Calibration-free concentration analysis (CFCA), which uses surface plasmon resonance (SPR) technology, offers a solution by specifically measuring the active protein concentration of the sample. CFCA leverages binding under partially mass-transport limited conditions to directly quantify the functional protein in a sample, overcoming variability associated with recombinant protein production. This article provides a background on CFCA and the case for its more widespread use in protein reagent characterization, as it offers a way to reduce lot-to-lot and vendor-to-vendor variability while improving reproducibility and standardization of assays. This perspective was informed by searching and selecting pertinent articles from PubMed (Nov 2024–March 2025) and by examining the reference lists of key papers.

Plain Language Summary

Measuring protein concentration accurately is important in biological research due to the key role that protein-based reagents play. Traditional methods for protein concentration measurement look at the total amount of protein present but are unable to distinguish how much of the protein is active and able to bind to its target. Calibration-free concentration analysis (CFCA) is a method that uses a technology called surface plasmon resonance (SPR) to measure only the active protein, not just the total amount. It works by focusing on how the protein binds under specific conditions, allowing scientists to directly measure the functional protein. This method helps to avoid some of the inconsistencies that can happen during the manufacturing of these proteins. This article explains CFCA and argues that it should be used more widely to check protein reagents. Using CFCA can make experiments more reliable and reproducible by reducing differences from one batch of protein to another and from different suppliers.

1. Introduction

Protein-based reagents are indispensable in various fields of biological research, diagnostics, and drug development. Accurate determination of protein concentration is essential for ensuring the reproducibility and reliability of experimental results. However, traditional methods of measuring protein concentration are flawed because they measure the total protein in the sample, not the active portion that can bind to its intended binding partner. This distinction is crucial, especially when dealing with recombinant proteins, as they often exhibit variability in production, leading to differences in activity that can significantly impact assay performance.

Calibration-free concentration analysis (CFCA) offers a solution to this problem by specifically measuring the active portion of the protein in a sample. CFCA is a method based on surface plasmon resonance (SPR) technology, which allows the quantification of the active protein concentration in a partially mass-transport limited system. CFCA can be advantageous in many areas of biological research, especially in regulated bioanalysis, as an additional tool for critical reagent characterization. The case for the more widespread use of CFCA in biological research will be explored in this article. To support this, a literature search was conducted in PubMed using the keywords “calibration-free concentration analysis, SPR, CFCA protein concentration, bioanalytical critical reagents” during the time frame spanning November 2024 to March 2025. References were included if they focused on CFCA for protein concentration analysis relevant to the bioanalytical community. The search particularly emphasized bioanalytical papers that covered critical reagents. Additionally, the reference lists of key articles were used to identify other relevant papers.

2. Background on CFCA

CFCA is a specific application of SPR. In SPR, plane-polarized light is directed through a prism onto a thin layer of metal, usually gold. When the light strikes the metal at a specific angle, its energy is absorbed by the metal’s electrons, generating surface plasmons-waves that travel along the surface [1]. The interaction between binding partners is detected by injecting the analyte over an immobilized binding partner under laminar flow conditions. This binding interaction changes the refractive index at the interface, which causes a shift in the resonance angle [1]. The magnitude of this shift is directly proportional to the amount of material bound to the surface.

The binding process of SPR takes place in two steps. The first step is referred to as mass transport, where the analyte in solution must travel through the solution to reach its immobilized binding partner on the sensor surface. The second step is the actual binding event between the analyte and its immobilized binding partner [2]. To understand how CFCA works in SPR, let us first examine a simple 1:1 binding model (Figure 1(a)). As the name implies, this model assumes a one-on-one binding interaction between two partners and is commonly used for kinetic analysis. It aims to avoid a mass-transport limited system, and this is typically achieved by having a low concentration of immobilized ligand on the sensor surface. In this setup, the rate at which the analyte moves through the solution is faster than the ligand binding to the analyte. As a result, the concentration of the analyte at the surface equals its concentration in the bulk solution, minimizing the influence of mass transport. In contrast, CFCA uses at least a partially mass-transport limited system (Figure 1(b)) [2,3]. This occurs when the analyte moves from the bulk flow to the immobilized ligand on the surface at a slower rate than its binding to the ligand [2]. To create this system, a high concentration of ligand on the sensor surface is used along with low concentrations of analyte. This setup leads to the formation of a “depletion zone,” where the analyte in solution is fully bound to the capture ligand [2]. Additionally, there is a requirement that the analyte in solution be run over the immobilized binding partner using two flow rates. The binding data can then be fitted if the diffusion coefficient, molecular weight of the analyte, and flow cell dimensions are also known, and then the active concentration value can be solved (Figure 2) [2,4–7].

Figure 1.

Comparison of 1:1 binding model and mass transport-limited model. (a) Equation detailing a 1:1 binding model which is normally used for affinity analysis. The top part of the equation shows the reversible interaction of two binding partners A and B through their association (k_a) and dissociation (k_d). The bottom equation details the differential equation used to calculate the rate of change of the complex formation over time. In this model the effect of mass transport is negligible and A_bulk=A _surface [3]. (b) Equation showing how mass transport impacts the binding model. A high concentration of antibody on the surface creates a concentration gradient and “depletion zone” as antigen at the surface is all bound to the capture antibody. A_bulk moving to the surface is slower than the association of the antigen to the antibody. A_bulk can be solved for [2,4–7].

Figure 2.

Calibration-free concentration analysis differential equations showing that the only unknown is A _bulk which is the active concentration in the sample (1–4).

CFCA has an interesting history, and it is not a new concept. The first paper showing the experimental possibility of using SPR for active concentration analysis was published by Karlsson et al. in 1993 [8]. In this paper, the authors sought to prove that active concentration could be determined by using the part of the binding curve that is diffusion-limited. This idea originated from work done to understand the theoretical models of mass transfer in thin-layer flow cells initially published in 1967 by Matsuda and later in 1991 by Sjolander and Urbaiczky [9,10]. To prove this, Karlsson et al. used different amounts of antigen on the biosensor surface and concluded that when a high density of ligand on the surface created a system that was fully mass transport limited, the antibody concentration could be determined. This was shown by spiking known concentration of antibody into cell culture media and rabbit serum and then successfully calculating back the recovered concentration with the use of a calibration curve [8]. In 1997, L.L.H. Christensen built upon this and introduced the theory of active concentration without the use of a calibration curve, with an accompanying experimental dataset applying the theory in a paper coauthored with Richalet-Secordel et al. [2,6]. Christensen’s theory pointed out limitations of the Karlsson method, which were primarily the need for a fully mass transport system, which is difficult to achieve, and the use of a calibration curve. Christensen overcame these limitations by showing that if the molecular weight and the diffusion coefficient of the analytes are known, the active concentration could be determined without the use of a calibration curve when using a system that is partially mass-transport-limited [2]. In the Richalet-Secordel et al. paper, this method was applied to successfully determine the active concentrations of monoclonal antibodies and Fab fragments spiked into E. coli crude extracts, as well as rabbit serum [6].

3. Issues with critical protein reagents and how CFCA can help

The critical role of protein-based reagents (antibodies and non-antibody proteins) in biological research and drug development is demonstrated by their widespread use in various assays. It is well recognized that the poor quality of commercially available protein reagents has resulted in extreme waste in scientific research, costing an estimated $350 million USD (2015) annually in the US alone [11]. In fact, a study examining the rate of low reproducibility in preclinical data highlighted the poor quality of biological reagents and reference materials as one of the primary causes [12]. In regulated bioanalysis, antibodies are used for both capture and detection in pharmacokinetic and biomarker ligand binding assays (LBAs) and can also be used in LCMS-based assays for immunocapture. Additionally, in biomarker LBAs, proteins are used as calibrators to determine the concentration of unknown samples. In cell-based assays, recombinant proteins are often used to stimulate or block some downstream activity. For drug release assays, the protein may be used to measure biotherapeutic activity. In all these examples, the quality and concentration of the critical protein reagents, whether antibodies or non-antibody proteins, are essential to the performance of the assay. The decisions made based on the data generated in these assays are crucial, as they can have clinical consequences. If the protein used is of subpar quality, it can lead to misleading or inaccurate results, underscoring the need for accurate protein quantification methods like CFCA.

Due to the vital role of protein-based reagents in bioanalytical assays, many papers have been published detailing the critical quality attributes (CQAs) of proteins and how they should be assessed [13–16]. Concentration is one of these and is essential for proper reagent protein characterization. According to the FDA Bioanalytical Method Validation Guidance for Industry, concentration should be designated on the datasheet where applicable [17]. Common methods for protein concentration include absorbance at 280 nm, Bradford assay, and bicinchoninic acid (BCA) assay, to name a few [18]. While each of these methods has caveats specific to the method itself, they all have one common flaw: they determine the total protein concentration in the sample, not the active protein. By ‘active protein,’ it means the portion of the protein that can actively bind to a binding partner. For example, in a ligand binding assay, this would be the epitope on the protein of interest, which binds to the paratope of the capture or detection antibodies. In the case of cell-based assay, it may be the portion of the protein that can actively bind to a receptor of interest.

Since protein reagents are often produced recombinantly through biological processes, they are susceptible to variability in production, which can significantly impact the assay’s performance [14,19]. The abundance of misfolded or denatured proteins and differences in post-translational modifications can vary, and these changes are not detectable when measuring the total protein of a sample or not even necessarily through purity analysis. In fact, in our lab, we observed that purity analysis provides little insight into the level of active protein present in the sample (data not published). If the same lot of protein is used throughout a study or the lifecycle of a program, it may not have significant consequences. However, in most cases, this is impractical, and additional lots need to be used. Lot-to-lot variability often plagues recombinant proteins and can create quite a headache when having to bridge a new lot. As stated by Cowan et al., in most cases, there is no “true” or “gold standard” for recombinant proteins that serve as calibrators in biomarker assays [19]. To overcome this, the bioanalytical community’s best practices suggest using an in-assay comparison of the new material with an older, well-characterized lot, either through a bridging method or a commutability method [19]. While both methods are acceptable, issues can arise if the initial lot has been exhausted or if the assay is being run by multiple labs, each with its own initial lot for comparison. In contrast, CFCA provides an understanding of the active protein concentration intrinsic to each lot or batch and provides a more concrete assessment of the material. Therefore, the idea of a “true” or “gold standard” is not even necessary, as the true activity of each batch of material is clearly defined by the CFCA method.

The value of CFCA in overcoming lot-to-lot variability of a recombinant protein calibrator was demonstrated by Harvey et al., and the authors refer to this method as epitope-specific CFCA [5]. In this paper, multiple lots of recombinant human soluble lymphocyte-activation gene 3 (sLAG3) protein were used for calibration curves and analyzed in a ligand binding assay. When using the total protein concentration for three lots obtained via Bradford assay, they exhibited an overall %CV of 42.4% [5]. Upon employing CFCA using a single epitope CFCA with the capture or detection antibody on the chip, the %CV decreased to 19.0% and 18.5%, respectively. Since the biomarker LBA relies on both the capture antibody and detection antibody epitopes to detect the protein of interest, the authors then employed a novel experiment looking at what is referred to as the intersection between the active epitope of the capture and detection antibodies. In this experiment, the capture antibody was immobilized on the chip, followed by sLAG3 and then the detection antibody. When tested this way, the overall %CV was reduced to 6.7% [5]. Another important note in this paper is the introduction of a calibration of the Biacore sensor flow cells. While CFCA is considered to be “calibration-free,” the authors observed differences between flow cells and sensor types. This variability results from the Biacore software using a default form factor for the flow cells to perform the CFCA calculation. To overcome this, calibration using the highly characterized NISTmAb was applied as a method of normalization. This was achieved by immobilizing NISTmAb to different Series S sensors and using the NISTmAb concentration as the gold standard to adjust the form factor for each flow cell. This issue was also called out in Karlsson’s 2016 review paper, and it was suggested that a calibration kit, if made available by the manufacturer, might help address this issue [4]. A calibration kit does not exist at this time, but if CFCA becomes more widely used, perhaps the vendor may take this into consideration.

Another paper showing the value of CFCA for recombinant protein characterization was published in 1999 by Zeder-Lutz et al [20]. In this paper, the authors performed active concentration on four different recombinant proteins using multiple antibodies and other ligands to assess their activity. Interestingly, and as suggested by the term “epitope-specific calibration-free concentration analysis,” coined by Harvey et al., the activity did change depending on the immobilized binding partner, highlighting the epitope specificity of CFCA. However, the most significant takeaway was that the active concentration, in most cases, was much lower than the nominal concentration obtained through traditional total protein concentration methods, with all proteins analyzed being less than 30% active [20].

Another point of concern for recombinant protein differences in assay performance is due to proteins purchased from different vendors [13]. Different vendors may use different cell lines, purification methods, and concentration determination methods, which can lead to differences in active protein present and impact the binding to the capture and detection antibodies used in assay performance [19]. CFCA can provide a similar benefit to overcoming lot-to-lot variability, as each batch of material generated by a different vendor can now have an active concentration assigned to it.

If CFCA is used to minimize the effects of lot-to-lot variability and vendor-to-vendor variability of the protein reagents, it may help with lot-to-lot differences observed between LBA kits. It is estimated that 70% of an assay’s performance is associated with the biologically derived raw materials being used, i.e., antibodies and calibrators [21]. In a 2023 paper by Luo et al. focusing on lot-to-lot variability LBA kits, specifically IVD grade, the authors emphasize the importance of accurate protein concentration measurement and suggest using multiple standard methods of measurement, such as BCA and A280 [21]. Additionally, to obtain the most accurate concentration, it is suggested that individual calibrators be developed specifically for each protein being quantitated [21]. While the authors’ vigorous approach is appreciated, utilizing CFCA may be an easier solution, as again, in this scenario, one would be required to mark one lot as the “gold standard.”

Using this standardization for the reagents may lead to better comparability of the data generated between different assays or kits for the same biomarker. An example that gives some support for this was presented by Westdijk et al. for standard characterization of the Sabin-IPV polio vaccine [22]. For the vaccine to have the required level of efficacy, the amount of D-antigen must be present at a certain level. However, the ELISA methods that were in place to detect the level of D-antigen gave variable results between labs, and additionally, different manufacturers used different methods and reagents to test for the antigen. When the authors utilized CFCA for antigen D concentration, they were able to obtain consistent D-antigen concentrations, regardless of the high-affinity antibody used [22]. Interestingly, this paper saw no difference in active concentration values when different antibodies were used as long as they were high-affinity binders. However, it is unclear if the antibodies had differing epitopes. In this case, CFCA provides a unique solution to standardize testing. Though this case study is vaccine-related, it poses an interesting concept for the application of CFCA to the standardization of biomarker LBAs.

While many of the examples given above look at the active concentration of the recombinant calibrator protein (non-antibody), there is no reason why the analysis cannot be applied to characterize the active concentration of an antibody. Affinity analysis is often used to characterize antibodies. In the 2019 White Paper on Recent Issues in Bioanalysis, critical reagent characterization methods suggested included functional testing in an orthogonal method such as SPR to understand the binding characteristics, and this was further elaborated on by additional papers [23,24]. It has also been noted by Haulenbeek et al. that this type of assessment may be helpful for troubleshooting assay issues [24]. For affinity analysis, concentration is a key component of the equation for an accurate value. The affinity value generated for an antibody can be different depending on the binding partner being used. This is most likely due to different active concentrations of the proteins. In fact, Pol published a paper investigating the effect of total vs. active concentration on affinity analysis in 2010 [25]. In this paper, the author compared the values obtained from affinity analysis using the total concentration obtained (A280) vs. the active concentration of bovine cysteine protease inhibitor cystatin B mutants binding with Papain [25]. Interestingly, the data showed that the association rate and affinity (KD) were greatly impacted by the different concentrations used (total vs. active). In fact, if the total protein concentration had been used, the data would have been misleading and resulted in values that showed little binding differences between the mutants [25]. If affinity analysis is being used as a QC check of antibody lots over time or to qualify new lots of material, using the total protein concentration of the binding target could lead to misleading results. Especially if new lots of material are used or if the binding partner itself is experiencing some degradation over time; therefore, using the active concentration may be a better, more accurate way to determine the affinity of an antibody.

CFCA offers a unique solution for protein reagent characterization for both antibodies and non-antibody proteins. If employed by vendors, it could offer more accurate lot-to-lot assessments and the possibility of harmonizing reagents across different vendors. Since CFCA analysis is epitope-specific, further work would need to be done to understand the details of how vendors might implement this. For example, if two vendors were to perform CFCA on recombinant PDL-1, which antibodies would they use? Additionally, it is not yet clear what impact the quality of the binding partner immobilized on the sensor might have on this analysis. How would lot-to-lot changes in the immobilized binding partner being used for analysis affect the results? While the logistics of how this might work will need to be worked out, the widespread adaptation by vendors could have a profound impact on the quality and reliability of scientific research.

3.1. Additional CFCA use cases

Multiple papers have been published using CFCA for different applications, and Karlsson has done an extensive review of these case studies [4]. Beyond the reagent characterization examples above, these include the measurement of antibodies for vaccine studies, the measurement of anti-drug antibodies for immunogenicity assessment of a large molecule drug, and biomarker analysis [4]. Despite its many applications, the widespread use of CFCA has never really taken off. In support of biomarker LBAs, two main use cases have been published. One use case is where the CFCA method itself is used to quantitate the endogenous biomarker, and the other, as previously mentioned in this article, uses CFCA to characterize the protein used as a calibrator in the LBA.

For direct biomarker measurement, publications have shown the detection of proteins such as B2-microglobulin (B2M) and serum amyloid A in healthy and disease-state serum samples [26]. It was demonstrated that the method could detect serum B2M down to 13 ng/mL in healthy and disease state serum. Additional studies have shown quantitation of immunoglobulin G in bovine and caprine milk samples [27]. The advantages of using CFCA for direct biomarker quantitation include the ability to quantitate in complex matrices and the fact that no calibrator protein is needed. A potential downside to using CFCA for this type of analysis is the throughput limitation on some SPR systems, and sample volume may also be a limitation.

Another interesting application of CFCA is to measure antibody responses after treatment with a biotherapeutic. Pol et al. were the first to utilize this in combination with affinity analysis to characterize complex immune responses in cynomolgus monkeys after dosing with two competing immunotherapeutic proteins [28]. In this case, the understanding of the quality and quantity of the immune response generated allowed the authors to have a better understanding of the potential efficacy of the two drug candidates [28]. Similar to this and of interest to the bioanalytical community, Aniol-Nielsen et al. validated a CFCA assay for use in the detection of anti-drug antibodies (ADA) in clinical samples [29]. The assay was set up by immobilizing endogenous human insulin, insulin degludec (Tresiba®), and turoctocog alfa (NovoEight®) to sensors. They then used various monoclonal and polyclonal antibodies against the targets for their assay validation [29]. The authors validated their method by looking at accuracy, precision, and reproducibility but also introduced some validation parameters unique to the SPR system [29]. These include an acceptable range of protein immobilization on the sensor, epitope masking, and activity of the protein immobilized on the sensor [29]. Overall, the authors found that their method showed precision within 10% and recovery of spiked positive controls within 90–112% [29]. They then tested this method on clinical samples, being able to quantitate the ADAs in the high-titer samples. The advantage of using CFCA for this application is that the method will give an absolute value of the ADA present rather than a titer, which is usually reported for ADA methods. It should be noted that in this paper, the titer levels did not correlate to the concentrations obtained via CFCA [29]. However, a possible explanation for this is that a titer value is a combination of affinity and concentration, whereas CFCA is an absolute concentration. Additionally, CFCA was not able to quantitate samples with low titers as they may have been below the lower limit of the assay [29]. While using CFCA for routine clinical or preclinical ADA analysis might not be ideal due to sensitivity or throughput limitations on the instrument, it does provide an alternate method for a more in-depth look at the ADA response if required.

3.2. Other potential uses

Even though the use of CFCA is not widespread, this methodology has the potential to provide a great impact beyond the examples shown here. One such application would be for potency testing in drug release assays. CFCA can be employed to ensure each patient is receiving the proper dosage of the active drug, rather than the total protein concentration dosage, which may only contain a certain fraction of the active species and may differ from lot-to-lot.

Another possible use of CFCA is its unique ability to quantitate a protein in a complex matrix. Therefore, it may be useful in measuring the concentration of individual proteins in a cocktail for use in multiplex ligand binding assays. Antibodies specific for each individual calibrator could be immobilized onto the chip, and then each active concentration could be determined. Additionally, CFCA could be useful in monitoring protein stability over time, and since it can detect within a complex matrix, the presence of bulking agents, such as BSA, would not interfere as it would in some other characterization methods.

4. Implementation

The application of CFCA for the characterization of biomarker LBA calibrators was discussed in the 2023 White Paper on Recent Issues in Bioanalysis [30]. The White Paper states that total protein concentration is acceptable, but CFCA may be used as a method of characterization. However, since CFCA is specific to the antibodies being used in the assay, the use of the calibrator needs to be defined for a specific assay or binding partner. Due to this, it may be best not to state the concentration obtained from CFCA on the certificate of analysis but may be more applicable to defining the use of CFCA in the assay’s validation method. It is also suggested that the use of CFCA be defined in applicable SOPs [30].

Currently, this is the only reference for how to implement CFCA for reagent characterization in the regulated bioanalysis space. As the application is still emerging there are some aspects of CFCA that should be further investigated to ensure that it is being implemented properly. First, does the active concentration value differ among Biacore instruments and can other SPR instruments be used. CFCA can be, theoretically, used on other SPR instruments, but most published papers use the Biacore systems. However, due to the widespread underuse of CFCA, Cytiva, the current manufacturer of the instruments, has decided not to include this analysis capability in its most recent Biacore software. Therefore, at this time, analysis must be done on older instruments, such as the T200, which lacks the throughput capabilities of the new models, such as the 8K. However, with a thorough understanding of the math behind the analysis, it is possible for users to perform their own scripting to do this, though this requires specific expertise and puts additional work on the analyzing lab.

5. Future perspective

CFCA is a high value method that can be employed across various areas of biological research as and drug development. If a more widespread usage of CFCA were to take hold, the scientific field could benefit from higher quality reagents, better understanding of data generated between assays, less cost for crossing in new reagent lots or method re-validation due to changes in lots or vendors, and most importantly more accurate dosing of medicine for patients.

Funding Statement

This paper was not funded.

Article highlights

• Calibration-Free Concentration Analysis (CFCA), based on Surface Plasmon Resonance (SPR) technology, specifically measures the active portion of a protein capable of binding to its target, overcoming the limitations of traditional total protein concentration methods.

• CFCA provides critical insights into protein reagent quality, reducing lot-to-lot and vendor-to-vendor variability by directly quantifying functional protein in the sample.

• Widespread adoption of CFCA could enhance reagent quality, enable consistent assay performance, and minimize waste from the use of low-quality reagents.

Author contributions

Shannon D. Chilewski conducted the research and authored the manuscript.

Disclosure statement

The author was employed at Bristol Myers Squibb (BMS) at the time of conceptualization and manuscript writing/submission.

The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

No writing assistance was utilized in the production of this manuscript.