Professional Documents
Culture Documents
Analytical Biochemistry
journal homepage: www.elsevier.com/locate/yabio
A B S T R A C T
Antibody-based therapeutics targeting membrane proteins have evolved as a major modality for the treatment of cancer, inflammation and autoimmune diseases.
There are numerous challenges, ranging from desired epitope expression to reliable binding/functional assays which are associated with developing antibodies for
this target class. Specifically, having a robust methodology for characterizing antibody interaction with a membrane protein target is essential for providing guidance
on dosing, potency and thus expected efficacy. Fluorescence-activated cell sorting (FACS) has been commonly used to characterize antibodies binding to membrane
protein targets. FACS provides information about the antibody-receptor complex (antibody bound to cells) and the apparent equilibrium dissociation constant (KD ) is
′
elucidated by fitting the antibody-receptor binding isotherm as a function of total antibody concentration to a nonlinear regression model. Conversely, Kinetic
Exclusion Assay (KinExA) has been used to measure solution-based equilibrium dissociation constant (KD) of antibodies. Here, KD is determined by measuring the free
antibody concentration at equilibrium in a series of solutions in which the antibody is at constant concentration and the receptor (either in the membrane or the cell)
is titrated. We measured the binding affinity of the anti-CD20 antibody, Rituximab, using both FACS and KinExA. There was ~25-fold difference in the binding
affinity measured by these two techniques. We have explored this discrepancy through additional experiments around the mathematical framework involved in the
analysis of these two different binding assays. Finally, our study concluded that KinExA enables accurate measurement of the KD for strong protein-protein in
teractions (sub-nanomolar values) compared to FACS.
internalization and recycling, effector function across different levels of binding isotherm as a function of total antibody concentration to a
cell expression, and inhibition of signaling. Here, we present a gener nonlinear regression model.
alized methodology to overcome some of the traditional limitations of This model makes two assumptions, first that the cells have reached
Abbreviations: CBP, Constant Binding Partner; CBPC, Constant Binding Partner Concentration; FACS, Fluorescence-Activated Cell Scan; KinExA, Kinetic Exclusive
Assay; KD, Equilibrium Dissociation Constant; K’D, Apparent Equilibrium Dissociation Constant.
* Corresponding author.
E-mail addresses: qchen@amgen.com, qchen@amgen.com (Q. Chen).
https://doi.org/10.1016/j.ab.2020.113974
Received 22 January 2020; Received in revised form 15 September 2020; Accepted 18 September 2020
Available online 30 September 2020
0003-2697/© 2020 Elsevier Inc. All rights reserved.
A. Vaish et al. Analytical Biochemistry 609 (2020) 113974
equilibrium with the titrated primary antibody, and second that the 2. Results and discussions
concentration of cell bound antibody is an insignificant fraction of the
total antibody concentration [9]. These assumptions can only be true 2.1. Quantification of target protein expression on the cell surface using
when the total concentration of antigen expressed on the cell surface is FACS
less than the equilibrium dissociation constant, KD. However, due to the
sensitivity limitations, accurate FACS measurements are difficult to Membrane-associated protein CD20 was used as the target for
obtain for strong interactions involving sub-nanomolar KD values. binding analysis with the anti-CD20 antibody, rituximab [20]. Ritux
The Kinetic Exclusion Assay (KinExA), developed by Sapidyne, has imab is a chimeric mouse/human therapeutic monoclonal antibody
evolved as a leading biophysical technique to measure solution-based approved for the treatment of B-cell derived lymphoma [20]. It binds to
equilibrium dissociation constant (KD) of high-affinity binding part CD20, a ~35 kD transmembrane protein found on the surface of a subset
ners [10–12]. In KinExA, KD is determined by measuring the free anti of B-cells. FACS analysis was performed to measure the number of CD20
body binding sites at equilibrium in a series of solutions in which the molecules per WIL-2 cell as a function of passage number using
antibody at a constant concentration is titrated with the binding partner. fluorescently-labeled rituximab. Fig. 2A displays FACS spectra of CD20
For soluble proteins, KinExA is a relatively straightforward tool to study expression on WIL-2 cells after the first harvest (passage number 0). 10
binding interactions, since the concentration of a soluble protein target or 20 μg/mL of Alexa 488-labeled rituximab was used to determine
can be accurately controlled. Here, the equilibrated solutions containing maximum fluorescence shift from that of untreated cells. The similar
doubly unliganded antibody and singly liganded antibody are intro intensity shift of WIL-2 cells incubated with 10 or 20 μg/mL Alexa
duced to a ligand-immobilized solid phase (beads) for binding charac 488-labeled rituximab indicates that the CD20 on cell surface was
terization [13,14]. Considering the bivalency of monoclonal antibodies, saturated by the labeled antibody. As the negative control, incubation of
the total antibody concentration used in KinExA binding model is the these cells with an Alexa 488-labeled non-target antibody resulted in a
sum of doubly unliganded antibody with 2 free binding sites, singly minimal intensity shift, demonstrating the specificity of Alexa
liganded antibody having 1 free binding site, and doubly liganded 488-labeled rituximab binding to the target antigen. As shown in Fig. 2B,
antibody [15]. In case of membrane targets, not only is the concentra the mean intensity of anti-CD20 bound cells (geo-mean) was used in the
tion of membrane proteins often unknown, it is also required to remove calibration curve derived from standards to calculate antibody binding
the membrane containing fractions from the equilibrated solutions capacity (ABC), which is essentially the number of receptors per cell.
before measuring concentration of the membrane free antibody. This ABC decreases with passage number as illustrated in Fig. 2C. While cell
separation process leads to removal of singly liganded antibody from viability at harvest remained at 91%, cell-surface CD20 density dropped
doubly unliganded antibody in testing samples. To account for the dif to ~50% by the 6th passage. Reported previously [18], this kind of
ference between the sampling in analysis of antibodies for soluble pro variability in expression level can directly impact the reproducibility
tein and membrane protein targets, Sapidyne provided “whole cell and reliability of the KD determination. To minimize the variability of
equilibrium binding” model in the KinExA-Pro software. For years, antigen concentration presented in the cell binding analysis, we pre
KinExA has provided a reliable methodology [16] to determine KD of pared isolated cellular membranes from a large batch of
antibodies for binding with membrane targets in their natural environ FACS-characterized cells for long term storage and repeated binding
ments using whole cells [3,4,17,18]. However, it is imperative to have a analyses that are required for regulatory registration of therapeutic
consistent receptor expression level in the cells used to obtain a repro molecules. Using the same batch of the CD20 containing cellular
ducible binding affinity determination. We, along with others, have membranes, we established a streamlined workflow of binding affinity
observed that the number of times cells have been transferred determination of membrane protein targets using KinExA.
batch-to-batch during cell culture, which is commonly referred as
“passage number”, can decrease the membrane protein expression level
significantly [19]. To circumvent this issue and ensure reproducibility, 2.2. Measurement of actual equilibrium dissociation constant (KD) of
Rituximab using a Kinetic Exclusion Assay
we isolated and froze a large quantity of target-bearing membranes at
− 80 ◦ C for use in antibody titration experiments using KinExA.
Following a modified version of the typical KinExA whole cell
Here we describe an integrated methodology to determine KD using
isolated cell membranes containing target proteins for analysis on the analysis method [17,18], we titrated a fixed concentration of antibody
into tubes containing a serial dilution of CD20 membrane. After over
KinExA platform (Fig. 1). Traditional FACS analysis was used to measure
target protein density on the cell-surface prior to membrane prepara night incubation (>16 h) at room temperature for the reactions to
approach equilibrium, the incubated samples are subjected to centrifu
tion, and to measure KD . By comparing KD and KD for the same pair of
′ ′
Fig. 1. Schematic illustration of KinExA-based binding affinity determination using cellular membrane. Cells were quantified via FACS for receptor expression,
followed by membrane isolation, and KinExA-based binding analysis.
2
A. Vaish et al. Analytical Biochemistry 609 (2020) 113974
Fig. 2. Analysis of cell-surface receptor density using the Quantum Simply Cellular (QSC) kit. (A) Alexa Fluor 488-labeled anti-CD20 monoclonal antibody (Rit
uximab) was used to generate histograms for both CD20 cells and quantified anti-huFc coated reference beads under identical incubation conditions. (B) A calibration
curve was obtained using the reference beads binding to 488-labeled Rituximab for antibody binding capacity (ABC) quantification. (C) Cell-surface CD20 density as
a function of passage number was obtained by calculating ABC from the calibration curve for different passage numbers.
determination using the “whole cell equilibrium binding” model pro method described above (Fig. 2B). FACS analysis was performed using
vided in the KinExA-Pro software. The model takes into account the 106 cells/ml. The CD20 receptor concentration was calculated as 4.8 nM
bivalency of monoclonal antibodies, which can exhibit mono and using the following equation.
bivalent binding with their targets [18].
(Cell number/ml) X (Receptors/Cell)
As shown in Fig. 3, serially diluted CD20 membrane was titrated Receptor Concentration = (1)
Avogadro’ s constant
against four different concentrations of rituximab (10, 30, 80, and 200
pM) to generate four independent sigmoidal curves covering a wide Fig. 4 shows rituximab binding after four hours incubation at room
potential affinity range. All these curves were generated using the same temperature with the cells containing 4.8 nM CD20. Data fitting using a
set of serially diluted CD20 membrane concentrations, which enabled one-site binding model (Y=Bmax*X/(KD +X)) or a dose-response model
′
global fitting using n-curve analysis with the “whole cell equilibrium (Y=Bottom+(Top-Bottom)/(1 + 10^((LogEC50-X)))) available in
binding” model for reliable KD determination. In contrast to the model GraphPad Prism 7, we obtained 1.9 nM as the value for the apparent
for soluble protein targets, where global fitting is performed using ti equilibrium dissociation constant (KD ) or the effective antibody con
′
trants concentration, “whole cell equilibrium binding” model globally centration at 50% of antibody/cell complex formation (Fig. 4 B&C).
fits data with the four total antibody concentrations and the fixed serial Here, X is the initial total antibody concentration used in the titration,
membrane dilution ratio of the four titration curves to derive KD and the and EC50 is the effective antibody concentration eliciting the 50%
receptor expression level. KD of rituximab was determined as 79 pM with binding signal of the maximal response of an antibody-cell complex, and
a 95% confidence interval of 50–115 pM. The CD20 expression level was the calculated KD is the total antibody concentration where half the
′
estimated as 1.5e5 receptors/cell, with a 95% confidence interval of receptor binding sites are occupied by antibodies via forming the
1.0e5 – 2.2e5 receptors/cell. antibody-cell complex.
Comparing receptor expression determined by FACS with the one
derived from the KinExA analysis, we noted that the expression level
2.3. Determination of apparent equilibrium dissociation constant (KD ) of
′
3
A. Vaish et al. Analytical Biochemistry 609 (2020) 113974
Fig. 3. (A) Schematic of KinExA experimental setup: Sample preparation involving incubation of membrane with antibodies for >16 hours at room temperature,
followed by separation of free antibodies from membrane-bound antibodies using centrifugation and filtration. Samples were introduced to a flow cell containing
anti-huFc coated beads. The captured free antibodies were detected by a fluorescently tagged anti-human antibody [IgG (H + L)]. (B) The binding signal is pro
portional to the concentration of free antibody in solution for different membrane concentration. Rituximab at four different concentration (10 pM, 30 pM, 80 pM,
and 200 pM) was titrated against CD20 membranes. The top concentration of the membrane contained 2.5 mg/mL total membrane protein. The equilibrium
dissociation constant (KD) was obtained from n-curve “Equilibrium, whole cell” modeling provided in the KinExA™ Pro software.
4
A. Vaish et al. Analytical Biochemistry 609 (2020) 113974
(1) equilibrium and (2) a very low ratio of cell bound antibody to total KD
=1+
[Ag]0
(10)
antibody [9]. We tested the impact of less equilibrated samples on KD 2KD
determining KD values using a shorter incubation time at room tem Fig. 5 is a simulated plot of KD /KD vs. [Ag]0/KD using eq (10). As
′ ′
perature or slower equilibrium process at 4 ◦ C. Decreasing the illustrated in Fig. 5, KD and KD are similar for the cases where [Ag]0 is
′
cell-antibody incubation duration from four hours to one hour (a typical less than KD. However, as alluded earlier, FACS sensitivity limitations
incubation time for FACS analysis) at room temperature resulted in requires a critical cell number (~105, which is equivalent to ~5 nM of
~two-fold increase in the KD from 2 nM as shown in Fig. 4 to about 5 nM
′
CD 20) [9] to generate a measurable signal. As a result, KD supersedes KD
′
(data not shown). Similarly, binding analysis of the Rituximab incubated as [Ag]0 is much higher than KD. Considering KD of rituximab as 79 pM,
at 4 ◦ C for four hours resulted in ~two-fold increase in the KD of 4 nM
′
the ratio of KD /KD is ~25, which is approximately half of the 50-fold
′
(data not shown). We cannot absolutely confirm that our assay per difference with the previously published KD [22]. While no informa
′
formed at room temperature has reached equilibrium at four hours, but tion was provided on the equilibrium time or cell expression level per
due to the cell viability constraint this was the longest incubation time
taining to the reported KD , it might be possible that the measurement
′
( ) √̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
( )2
defined in the literature [22] for a traditional cell-based assay as the [Ab]0 − [Ag]0 − KD + [Ab]0 − [Ag]0 − KD + 4[Ab]0 KD
concentration of antibody which results in binding of 50% of the re =
2[Ab]0
ceptors during titration of antibody against a fixed concentration of
(12)
receptor-containing cells. Adapted from the previous work [11], when
50% of the receptor has been used to form a complex with the antibody; Eq (12) can be further rearranged as:
it results in [Ag]/[Ab*Ag] = 1, therefore eq. (2) can be written as:
KD = [Ab] (5)
At equilibria, substituting eq (5) in eq (3):
KD = [Ab]0 − [Ab * Ag] (6)
[Ag]0
(9)
′
KD = KD +
2
Fig. 5. Simulated curve illustrating the correlation between apparent equilib
Eq. (9) transforms into a normalized form after dividing it by KD
rium dissociation constant (KD ), actual equilibrium dissociation constant (KD)
′
5
A. Vaish et al. Analytical Biochemistry 609 (2020) 113974
̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
√( ̅
( ) )2 yeast or phage), where it could measure KD with sub-nanomolar values
′
[Ag]0
1− [Ab]0
− KD
[Ab]0
+ 1 − [Ag][Ab]
0
− KD
[Ab]
+ 4 KD
[Ab] [23,24] using lower number of cells with high surface display levels.
0 0 0
Free Ab (%) = (13) However, binding assay for antibody against a membrane-bound target
2
requires higher number of cells (~105) to generate a measurable signal
Eq (13) can be expanded via Binomial series expansion as discussed in due to the lower expression of these proteins, thus making FACS chal
Appendix for an approximate solution, which can be used to illustrate lenging to estimate sub-nanomolar KD of antibody/membrane protein
′
[Ab]0 − [Ag]0 KD
Free Ab (%) ≈ + (15) KD determination if the constant binding partner concentration is below
[Ab]0 [Ab]0 − [Ag]0
the KD.
And when [Ab]0 < [Ag]0 ,Free Ab (%) is expressed as Recent structural studies have shown that rituximab binds to CD20
dimer on cell surface and forms clustered antibody/antigen complex
KD
Free Ab (%) ≈ (16) [25,26]. Considering the bivalent nature of monoclonal antibody (mAb),
[Ab]0 − [Ag]0
avidity effect could play an important role during binding analysis.
As eqs (15) and (16) are mainly dominated by [Ab]0 , the CBPC, which Using monovalent rituximab Fab one may characterize the avidity of
makes it a CBPC-controlled titration curve (Fig. 6). mAb with comparing to the binding affinity of Fab [17].
In KinExA, binding assay can be modulated in both regimes; KD- In summary, the described methodology from membrane prepara
controlled and CBPC-controlled titration curves. Global fitting of multiple tion to affinity measurement can be applied to any therapeutic antibody
titration curves that include full or partial KD-controlled ([Ab]0 < 5 KD) against a membrane target. In particular, KinExA enables accurate high-
and CBPC-controlled titration curves derives accurate KD and CBPC affinity antibody KD determination. The accurate affinity determination
values. has broader implications both in the assessment of in vivo target
coverage and for lead selection or engineering efforts in therapeutic
3. Conclusions discovery and development.
Here we have demonstrated use of cellular membranes in combina 4. Materials and methods
tion with KinExA for determining the binding affinity of a therapeutic
antibody against a membrane-bound target. Additionally, we measured 4.1. FACS analysis to measure target protein density on the cell surface
prior to membrane preparation
the anti-CD20 antibody, rituximab, KD using FACS, which was compa
′
6
A. Vaish et al. Analytical Biochemistry 609 (2020) 113974
monoclonal antibody containing human Fc, which binds to the anti- membrane were incubated for >16 hours (overnight) at room temper
human Fc coated beads. During FACS analysis, the same fluorescently ature. The membrane-free antibody was separated from the antibody-
labeled anti-target monoclonal antibody must be applied to both the membrane complexes and antibody-free membrane by centrifugation
anti-huFc coated beads and target-bearing cells in the same condition at at 2800xg for 15 minutes. The supernatant was then filtered through 0.2
the same time. Recognized by the same fluorescently labeled anti-target μm 24-well filter plate. The free Ab concentrations in the filtered su
monoclonal antibody, antibody binding capacity (ABC), which repre pernatants were quantified on the KinExA by passing each solution over
sents number of the target receptor/cell, can be quantified with standard the goat-anti-huFc-coated beads, followed by measuring the bead-bound
curve derived from the anti-huFc coated beads. Histogram channel (geo- antibody with fluorescent (Alexa Fluor 647) labeled goat anti-huIgG (H
mean of fluorescence) values corresponding to the specified level of anti- + L) antibody (Jackson Immuno Research). The equilibrium dissociation
human Fc for the bead populations are recorded in a QuickCal® tem constant (KD) was obtained by global fitting of the four titration curves
plate for calculation of ABC values (http://www.bangslabs.com/ using the n-curve “Equilibrium, whole cell” model provided in KinExA™
products/flow-cytometry/fluorescence-quantitation). Noteworthy, this Pro software (Sapidyne Instruments Inc., Boise ID). The KinExA Pro
QuickCal® template is designated to the particular QSC kit with the software provides a choice of two equilibrium analysis models, standard
specific lot # as indicated in each template. “Equilibrium” and “Equilibrium, Whole Cell”. The “Equilibrium” model
is applicable to a soluble system in which it has been shown that the
4.2. CD20 membrane preparation using cell disruptor (constant systems) KinExA signal is proportional to the sum of the free binding sites present
in mixture of unbound, singly bound, and doubly bound bivalent anti
To generate consistent supply of target membrane protein for affinity body [15]. In the case of cells, either intact whole cells or cell membrane
determination, we collected several batches of total 59 billion cells that fragments, all singly and doubly bound Ab is removed by centrifugation
express CD20 between 5e5 to 1e6/cell, averaged at 7.8e5 CD20/cell. On and/or filtration, and only the doubly free Ab is detected. Choosing a
the day of the membrane preparation, the cell pellets (WIL2-S (CD20) model appropriate for the experimental data is important to prevent
cells, total 5.9e10, ~80 g) were weighed and resuspended (approxi significant errors in the estimated parameters. In addition, the whole cell
mately 20% or less) in Hypotonic Buffer (10 mM HEPES, pH 7.4, 1 mM analysis provides values (and confidence intervals) for the KD and
MgCl2, 1.8 mM CaCl2) with Roche protease inhibitors (4 tablets/50 mL) expression level (EL – titrant molecules per cell).
and kept on ice/water. One thousand six hundred (1600) mL of the cell
suspension was passed through Constant Cell Disruption Systems with Disclosure of potential conflicts of interest
the disruption pressure 15 psi. Samples were transferred into Beckman
Coulter Ti45 centrifuge tubes and centrifuged at 35,000 rpm Amit Vaish, Peter Grandsard, and Qing Chen are employees of and
(~96000xg) for 20 min at 4 ◦ C. After carefully removing the superna own stock of Amgen, Inc. Joanne Ho and Helen McBride are former
tants, the membrane pellets were resuspended in ~50 mL of Membrane employees of Amgen, Inc. No other potential conflicts of interest were
Buffer (10 mM HEPES, pH 7.4, 1 mM MgCl2, 1.8 mM CaCl2, 10% su disclosed.
crose) with Roche protease inhibitors (1 tablet/20 mL) and homoge
nized with 25 strokes using a motor driven homogenizer on ice/water. CRediT authorship contribution statement
Total membrane protein concentration was determined with Bio-Rad DC
Protein Assay Kit (Bio-Rad). Amit Vaish: wrote the manuscript.: analyzed the data. Joanne S.
Lin: performed the experiments, analyzed the data. Helen J. McBride:
4.3. Determination of the equilibrium dissociation constant (KD) of Writing - original draft, critically reviewed and helped writing the
Rituximab binding with CD20 membrane using KinExA® technology manuscript. Peter J. Grandsard: Writing - original draft, critically
reviewed and helped writing the manuscript. Qing Chen: wrote the
UltraLink Biosupport (Thermo Fisher Scientific cat# 53,110) was manuscript, performed the experiments, analyzed the data.
coated with goat-anti-huFc (Jackson Immuno Research) in 50 mM
NaCO3, pH 9.6, followed by blocking with 10 mg/mL BSA in 1 M Tris- Acknowledgments
HCl, pH 8.0. To set up the binding analysis, 10 pM, 30 pM, 80 pM, and
200 pM of rituximab was incubated with titrated CD20 membrane We thank Larry Daugherty for assistance with flow cytometry and
fragments in assay buffer containing 0.1 mg/mL BSA, Roche Protease Tom Glass from Sapidyne for providing insight on KinExA technology
Inhibitor Cocktail (EDTA free), and PBS (without Ca2+, Mg2+). Samples and specific binding analysis of cell membrane targets.
containing the mixture of Rituximab and CD20-expressing cell
Appendix
√̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
([Ab]0 − [Ag]0 − KD ) + ([Ab]0 − [Ag]0 − KD )2 + 4[Ab]0 KD
Free Ab (%) = (A1)
2 [Ab]0
√̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
X + X2 + C
Free Ab (%) = (A3)
2
7
A. Vaish et al. Analytical Biochemistry 609 (2020) 113974
Considering the absolute value of X, |X|, in the square root, and rearranging eq (A3):
√̅̅̅̅̅̅̅̅̅̅̅̅̅̅
X + |X| 1 + |X|C2
Free Ab (%) = (A4)
2
Performing Binomial series expansion of (1 + |X|C2 )1/2 in eq (A4):
[
C C2
X + |X| 1 + 2|X| 2 −
8|X|4
+…
Free Ab (%) = (A5)
2
This expansion is valid for. C
|X|2
<1
Rearranging eq (A5) and ignoring the higher order terms:
X |X| C
Free Ab (%) = + + (A6)
2 2 4|X|
In eq (A6), when [Ab]0 < KD, X would a negative number, whereas |X| is a positive number, which would transform eq (A6) as:
C
Free Ab (%) = (A7)
4|X|
After substituting values of C and |X| in eq (A7):
KD
Free Ab (%) = (A8)
|{[Ab]0 − [Ag]0 − KD }|
whereas, when [Ab]0 > KD, X is a positive number for [Ab]0 > [Ag]0 and Free Ab (%) is expressed as:
C
Free Ab (%) = X + (A10)
4|X|
Substituting value of C and |X| in the expression:
{[Ab]0 − [Ag]0 − KD } KD
Free Ab (%) = + (A11)
[Ab]0 {[Ab]0 − [Ag]0 − KD }
For a very narrow concentration regime when [Ab]0 < [Ag]0 , Free Ab (%) expression will not have X component (X is a negative number and |X| is a
positive number, for [Ab]0 < [Ag]0 which will cancel out each other), as a result eq (A12) Free Ab (%) will be for [Ab]0 < [Ag]0 as follows:
KD
Free Ab (%) ≈ (A13)
[Ab]0 − [Ag]0
In summary, for [Ab]0 > KD, Free Ab (%) is primarily expressed by eq (A12) except for a very narrow range [Ab]0 < [Ag]0 where it can be expressed by
eq (A13).
References N. Hawkins, W.J. Boyle, W.S. Simonet, J.K. Sullivan, Denosumab, a fully human
monoclonal antibody to RANKL, inhibits bone resorption and increases BMD in
knock-in mice that express chimeric (murine/human) RANKL, J. Bone Miner. Res. :
[1] J. Li, Z. Zhu, Research and development of next generation of antibody-based
the official journal of the American Society for Bone and Mineral Research 24
therapeutics, Acta Pharmacol. Sin. 31 (2010) 1198.
(2009) 182–195.
[2] F.-F. Lin, R. Elliott, A. Colombero, K. Gaida, L. Kelley, A. Moksa, S.-Y. Ho,
[6] R.B. Dodd, T. Wilkinson, D.J. Schofield, Therapeutic monoclonal antibodies to
E. Bykova, M. Wong, P. Rathanaswami, S. Hu, J.K. Sullivan, H.Q. Nguyen, H.
complex membrane protein targets: antigen generation and antibody discovery
J. McBride, Generation and characterization of fully human monoclonal antibodies
strategies, BioDrugs 32 (2018) 339–355.
against human Orai 1 for autoimmune disease, J. Pharmacol. Exp. Therapeut. 345
[7] C.A. Benedict, A.J. MacKrell, W.F. Anderson, Determination of the binding affinity
(2013) 225.
of an anti-CD34 single-chain antibody using a novel, flow cytometry based assay,
[3] D.H. Bedinger, I.D. Goldfine, J.A. Corbin, M.K. Roell, S.H. Adams, Differential
J. Immunol. Methods 201 (1997) 223–231.
pathway coupling of the activated insulin receptor drives signaling selectivity by
[8] M.L. Jones, M.A. Alfaleh, S. Kumble, S. Zhang, G.W. Osborne, M. Yeh, N. Arora, J.J.
XMetA, an allosteric partial agonist antibody, J. Pharmacol. Exp. Therapeut. 353
C. Hou, C.B. Howard, D.Y. Chin, S.M. Mahler, Targeting membrane proteins for
(2015) 35–43.
antibody discovery using phage display, Sci. Rep. 6 (2016), 26240-26240.
[4] H. Zheng, Y. Bae, S. Kasimir-Bauer, R. Tang, J. Chen, G. Ren, M. Yuan, M. Esposito,
[9] S.A. Hunter, J.R. Cochran, Chapter two - cell-binding assays for determining the
W. Li, Y. Wei, M. Shen, L. Zhang, N. Tupitsyn, K. Pantel, C. King, J. Sun,
affinity of protein–protein interactions: technologies and considerations, in: V.L.
J. Moriguchi, H.T. Jun, A. Coxon, B. Lee, Y. Kang, Therapeutic antibody targeting
Pecoraro (Ed.) Methods in Enzymology, Academic Press2016, pp. 21-44.
tumor- and osteoblastic niche-derived Jagged 1 sensitizes bone metastasis to
[10] R.J. Darling, P.-A. Brault, Kinetic exclusion assay technology: characterization of
chemotherapy, Canc. Cell 32 (2017) 731–747, e736.
molecular interactions, Assay Drug Dev. Technol. 2 (2004) 647–657.
[5] P.J. Kostenuik, H.Q. Nguyen, J. McCabe, K.S. Warmington, C. Kurahara, N. Sun,
[11] N. Ohmura, S.J. Lackie, H. Saiki, An immunoassay for small analytes with
C. Chen, L. Li, R.C. Cattley, G. Van, S. Scully, R. Elliott, M. Grisanti, S. Morony, H.
theoretical detection limits, Anal. Chem. 73 (2001) 3392–3399.
L. Tan, F. Asuncion, X. Li, M.S. Ominsky, M. Stolina, D. Dwyer, W.C. Dougall,
8
A. Vaish et al. Analytical Biochemistry 609 (2020) 113974
[12] I.A. Darwish, T.A. Wani, N.Y. Khalil, D.A. Blake, Novel automated flow-based [19] H. Yu, T. Cook, P. Sinko, Evidence for Diminished Functional Expression of
immunosensor for real-time measurement of the breast cancer biomarker CA15-3 Intestinal Transporters in Caco-2 Cell Monolayers at High Passages, 1997.
in serum, Talanta 97 (2012) 499–504. [20] M.D. Pescovitz, Rituximab, an anti-CD20 monoclonal antibody: history and
[13] H. AlRabiah, M.A. Hamidaddin, I.A. Darwish, Automated flow fluorescent mechanism of action, Am. J. Transplant. 6 (2006) 859–866.
noncompetitive immunoassay for measurement of human plasma levels of [21] M. Reff, K. Carner, K. Chambers, P. Chinn, J. Leonard, R. Raab, R. Newman,
monoclonal antibodies used for immunotherapy of cancers with KinExA™ 3200 N. Hanna, D. Anderson, Depletion of B cells in vivo by a chimeric mouse human
biosensor, Talanta 192 (2019) 331–338. monoclonal antibody to CD20, Blood 83 (1994) 435–445.
[14] T.A. Wani, S. Zargar, S.M. Wakil, I.A. Darwish, New analytical application of [22] S. Uchiyama, Y. Suzuki, K. Otake, M. Yokoyama, M. Ohta, S. Aikawa, M. Komatsu,
antibody-based biosensor in estimation of thyroid-stimulating hormone in serum, T. Sawada, Y. Kagami, Y. Morishima, K. Fukui, Development of novel humanized
Bioanalysis 8 (2016) 625–632. anti-CD20 antibodies based on affinity constant and epitope, Canc. Sci. 101 (2010)
[15] T.R. Glass, D.J. Winzor, Confirmation of the validity of the current characterization 201–209.
of immunochemical reactions by kinetic exclusion assay, Anal. Biochem. 456 [23] J.J. VanAntwerp, K.D. Wittrup, Fine affinity discrimination by yeast surface
(2014) 38–42. display and flow cytometry, Biotechnol. Prog. 16 (2000) 31–37.
[16] O. Kusano-Arai, R. Fukuda, W. Kamiya, H. Iwanari, T. Hamakubo, Kinetic [24] A. Frenzel, T. Schirrmann, M. Hust, Phage display-derived human antibodies in
exclusion assay of monoclonal antibody affinity to the membrane protein clinical development and therapy, mAbs 8 (2016) 1177–1194.
Roundabout 1 displayed on baculovirus, Anal. Biochem. 504 (2016) 41–49. [25] L. Rougé, N. Chiang, M. Steffek, C. Kugel, T.I. Croll, C. Tam, A. Estevez, C.
[17] P. Rathanaswami, J. Babcook, M. Gallo, High-affinity binding measurements of P. Arthur, C.M. Koth, C. Ciferri, E. Kraft, J. Payandeh, G. Nakamura, J.T. Koerber,
antibodies to cell-surface-expressed antigens, Anal. Biochem. 373 (2008) 52–60. A. Rohou, Structure of CD20 in complex with the therapeutic monoclonal antibody
[18] L. Xie, R. Mark Jones, T.R. Glass, R. Navoa, Y. Wang, M.J. Grace, Measurement of rituximab, Science 367 (2020) 1224.
the functional affinity constant of a monoclonal antibody for cell surface receptors [26] A. Kumar, C. Planchais, R. Fronzes, H. Mouquet, N. Reyes, Binding mechanisms of
using kinetic exclusion fluorescence immunoassay, J. Immunol. Methods 304 therapeutic antibodies to human CD20, Science 369 (2020) 793.
(2005) 1–14.