Research A high throughput electrochem i-C Wo c., 1 D Received 1 December 2005; received in revised form 22 May 2006; accepted 22 May 2006 Journal of Immunological Methods 1. Introduction CD20 is a 33–36 kDa B cell-specific cell surface phosphoprotein with four transmembrane domains (Tedder et al., 1988). CD20 is expressed on B cells ranging from pre- to mature B cells, but not on plasma cells. More than 90% of B cell lymphoma expresses CD20 (Anderson et al., 1984), which makes it a good target for immunotherapy of B cell malignancies. Rituximab, a chimeric anti-CD20 monoclonal antibody, binds to CD20 antigen on the surface of both normal and malignant B cells. This binding event recruits the body's natural defenses to kill rituximab-bound B cells through mechanisms including antibody dependent cell cytotoxicity, complement dependent cytotoxicity, and Abbreviations: CHO, Chinese hamster ovary; CV, coefficient of variation; FBS, fetal bovine serum; PBS, phosphate-buffered saline; ELISA, enzyme-linked immunosorbent assay; HRP, horseradish A cell-based ELISA using suspension WIL2 cells in 96-well format was previously developed for measuring relative binding affinities of humanized anti-CD20 variants. We further developed a new cell-binding assay that uses high binding capacity carbon electrode plates for rapid attachment of suspension WIL2 cells and electrochemiluminescence for detection. Compared to the cell- based ELISA, which requires centrifugation for the manual wash steps, significant improvement in assay throughput was achieved by using a microplate washer. The assay can be performed on both 96- and 384-well plates with a standard curve range of 2.74– 2000 ng/ml, which is wider than the range of 15.6–1000 ng/ml for the cell-based ELISA. Using CD20 expressing CHO cell clones, surface expression of ≥33,000 CD20 molecules was sufficient to obtain a dose–response curve in 384-well format. Relative affinities of 15 humanized variants correlated well (r2=0.94) between electrochemiluminescent cell-binding assay and cell-based ELISA. A competitive assay format, using mouse anti-CD20 antibody as the tracer, with a dose–response range of 27.4– 20,000 ng/ml was also developed. The new cell-binding assay method can be used to efficiently support humanization process for selection of anti-CD20 antibody drug candidates and to characterize antibody binding to other cell surface proteins. © 2006 Elsevier B.V. All rights reserved. Keywords: Anti-CD20; Humanized antibody; WIL2 cells; Carbon surface plate; Electrochemiluminescence; Cell-binding assay Abstract Available online 19 June 2006 assay for therapeutic ant Yanmei Lu, Wai Lee Assay and Automation Technology, Genentech In peroxidase; VEGF, vascular endothelial growth factor; HER2, epidermal growth factor receptor 2; ECL, electrochemiluminescence. ⁎ Corresponding author. Tel.: +1 650 225 8289; fax: +1 650 225 1770. E-mail address:
[email protected] (Y.G. Meng). 0022-1759/$ - see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.jim.2006.05.011 paper iluminescent cell-binding D20 antibody selection ng, Y. Gloria Meng ⁎ NA Way, South San Francisco, CA 94080, USA 314 (2006) 74–79 www.elsevier.com/locate/jim apoptosis (Smith, 2003). Since its approval for treating non- Hodgkin's lymphoma, rituximab has also been used for autoimmune diseases (Kneitz et al., 2002; Boye et al., 2003), which require more prolonged dosing regimens. To ologic minimize immunogenicity for long-term treatment of non- malignant B cell disorders, a fully humanized anti-CD20 antibody is desirable. CD20 has four transmembrane regions and a short extracellular domain of approximately 42 amino acids (Tedder et al., 1988). Since solubleCD20was not available, a live suspension cell-based ELISA in 96-well format was previously developed to measure relative binding affinities of humanized anti-CD20 antibodies to CD20molecules on the surface of WIL2 cells (Hong et al., 2004). The use of live cells avoids potential alterations in the CD20 epitope caused by drying or fixation associated with the use of formaldehyde or glutaraldehyde (Baron et al., 1977; Schlosser et al., 1991; Sedgwick and Czerkinsky, 1992). The suspension cell-based ELISA method was later adapted for characterizing antibody binding to other cell surface proteins such as Fc receptors in different species. However, the assay involves multiple manual washes after sample and detection antibody incubation steps. Eachwash procedure requires centrifugation of the plates followed by immediate aspiration of the wash buffer using a hand-held vacuum manifold. The substantial hands-on time limited the assay throughput to only two 96-well plates per run. Although the wash procedure can be simplified by using adherent transfected cells on tissue culture plates, develop- ing a stable cell line expressing a physiologically relevant number of surface molecules may take several months. Recently, an electrochemiluminescent (ECL) assay methodology using microwell plates with carbon electro- des built into the bottom for coating capture molecule and a ruthenium labeled molecule for generating signals was developed (Meso Scale Discovery, Gaithersburg, MD; Best et al., 2005). The ECL signals were generated by oxidation of the ruthenium label on the electrode. A high- energy electron transfer from the tripropylamine radical, which was generated by oxidation of coreactant tripro- pylamine, placed ruthenium in an excited state. Relaxa- tion of the excited state ruthenium to the ground state resulted in chemiluminescent emission at 620 nm (Liang et al., 1996). The non-smooth high binding surface of the carbon electrode was found to bind cells tightly, allowing the plates to be washed on a microplate washer. Here we report the development and characterization of a high throughput ECL cell-binding assaymethod for measuring antibody binding to live suspension and adherent cells. 2. Materials and methods 2.1. Reagents and cell lines Full-length chimeric and humanized anti-CD20 anti- Y. Lu et al. / Journal of Immun body variants with different affinities for CD20 and mu- tations on Fc were generated and purified in Genentech (South San Francisco, CA) as described previously (Hong et al., 2004; Shields et al., 2001). Goat F(ab′)2 anti-human IgG Fc antibody (Jackson ImmunoResearch,West Grove, PA) was labeled with ruthenium (II) tris-bipyridine-(4- methylsulfone) NHS ester (Meso Scale Discovery, Gaithersburg, MD) according to the manufacturer's protocol. Ruthenium-labeled streptavidin was purchased from the same manufacturer. The human B-lymphoblastoid cell line WIL2 was purchased from American Type Culture Collection (ATCC, Rockville, MD). Cells were cultured in RPMI 1640 supplemented with 2 mM L-glutamine and 10% fetal bovine serum (FBS). CHO cell clones expressing different amounts of CD20 surface molecules were grown in 50/50 F12/DMEMmedium supplemented with 2 mM L-glutamine, 10 μg/ml glycine, 15 μg/ml hypo- xanthine, 5 μg/ml thymidine and 5% dialyzed FBS as described previously (Hong et al., 2004). 2.2. ECL cell-binding assay For direct ECL cell-binding assay, WIL2 or CD20 ex- pressingCHOcells werewashedwith phosphate-buffered saline (PBS) and seeded at 25,000 cells/well in 25 μl PBS on 96-well or 3125 cells/well in 25 μl PBS on 384-well MULTI-ARRAY High Bind plates (Meso Scale Discov- ery). The High Bind plates are microplates with carbon electrodes integrated into the bottom of plates.WIL2 cells were incubated for one hour at room temperature to allow cell attachment to the carbon surface. To block non- specific binding sites on the plates, 25 μl of 30% FBS in PBS was added to each well to yield 50 μl of 15% FBS in PBS. The plates were then incubated at room temperature for 30 min with mild agitation. Blocking solution was decanted and excess liquidwas removed by tapping plates on paper towels. Chimeric anti-CD20 human IgG1 with mouse variable regions was used as a standard, and humanized anti-CD20 variants were assayed as samples. Three fold serial dilutions of antibodies (2.74–2000 ng/ ml) in PBS containing 10% FBS (assay buffer) were added to the plates (25 μl/well for both 96- and 384-well plates). After one hour incubation at room temperature with mild agitation, the plates were washed three times with PBS (300μl for 96-well or 90μl for 384-well plates). Cells on plates were initially washed by manual addition of PBS and flicking of plates before a microplate washer (ELx405 Select, Bio-Tek Instruments, Inc., Winooski, VT) was found to yield a similar level of signal (data not shown). The wash program parameters were set at a dispense flow rate of 1 and an aspiration rate of 3 mm/s. 75al Methods 314 (2006) 74–79 Bound antibodies were detected by adding 25 μl of 0.25 μg/ml ruthenium-labeled F(ab′)2 goat anti-human IgG Fc in assay buffer. After a final one hour incubation at room temperature with mild agitation, free detection anti- body was removed by performing another three washes. Tris-based Read Buffer T (1×, without surfactant) (Meso Fig. 1. Titration curves of chimeric anti-CD20 antibody (circle), huma- nized anti-CD20 variant 1 (square) and anti-CD20 variant 2 (triangle) in direct (A) and competitive (B) ECL cell-binding assays. For the direct cell-binding assay, WIL2 cells in 96-well high binding capacity carbon electrode plates were incubated with serial dilutions of anti-CD20 anti- 76 Y. Lu et al. / Journal of Immunologic bodies. Bound antibodies were detected with ruthenium-labeled anti- human IgG Fc. The ratio of ECL signal at 2000 ng/ml versus 0 ng/ml of chimeric anti-CD20 antibody was 318. The background ECL signal of cells without antibody was 46±4 (A). For the competitive cell-binding assay, cells were incubated with a mixture of 500 ng/ml of biotinylated mouse version of anti-CD20 variant 1 and serial dilutions of anti-CD20 antibodies. Bound biotinylated mouse anti-CD20 antibodies were detected with ruthenium-labeled streptavidin (B). Scale Discovery) containing tripropylamine, a coreactant for the ruthenium label, was added to the plates (150 μl/ well for 96-well or 40 μl/well for 384-well plates). Upon electrochemical stimulation, ruthenium label bound to the carbon electrode emitted luminescence light at 620 nm. ECL signals were captured by a built-in charge-coupled device (CCD) camera in Sector Imager 6000 reader (Meso Scale Discovery) and recorded as counts. For competitive ECL cell-binding assay, WIL2 cells were seeded at 25,000 cells/well on 96-well High Bind plates and blocked as described above. A final concen- tration of 500 ng/ml of biotinylated mouse IgG2a with variable regions of humanized anti-CD20 variant 1 was used as a tracer. Anti-CD20 antibodies were serially diluted 1/3 (final concentrations of 27.4–20,000 ng/ml) in assay buffer, mixed with tracer, and added to the plates. The plates were incubated for one hour at room tem- perature. After three washes with PBS, 25 μl of 0.25 μg/ ml ruthenium-labeled streptavidin was added to detect bound tracer. The plates were then incubated for one hour at room temperature. After another three washes to re- move free ruthenium-labeled streptavidin, plates were read as described above. 2.3. Data analysis Chimeric anti-CD20 antibody was used as a standard for the relative binding affinity analysis. Midpoint ECL signal of the standard titration curve was calculated by first dividing the difference in ECL counts at 2000 ng/ml and 2.74 ng/ml by two, and then adding the count at the 2.74 ng/ml. The corresponding concentrations of standard and samples at this midpoint ECL count were determined from the titration curves using a four-parameter regression curve-fitting program XLfit (ID Business Solutions Ltd., Guildford, Surrey, UK). The relative binding affinity was calculated by dividing the midpoint count concentration of standard by that of sample. Coefficient of variation (CV) was calculated by ANOVA analysis using StatView software (SAS Institute, Cary, NC). Error bars in the figures represent the standard deviations. 3. Results and discussion 3.1. ECL cell-binding assay Previously, a live suspension cell-based ELISA for measuring relative binding affinities of humanized anti- bodies was reported (Hong et al., 2004). After incubation of sample or detection antibody,WIL2 suspension cells in 96-well round bottom plates were washed manually by al Methods 314 (2006) 74–79 first centrifuging the plates, and then carefully aspirating ologic the supernatant. To improve throughput, we developed a new ECL cell-binding assay method using high binding capacity carbon electrode plates. WIL2 cells at 25,000 cells/well were incubated on the plates for one hour to allow cell attachment to the carbon surface. Serially di- luted anti-CD20 antibodies (2.74–2000 ng/ml) were incubated with WIL2 cells that adhered to the 96-well plates. After removing unbound antibodies, bound anti- bodies were detected with ruthenium-labeled anti-human IgG Fc. Data of ECL signal generated from the ruthenium label were collected. The titration curves of three anti- bodies are shown in Fig. 1A. Chimeric anti-CD20 IgG was used as a standard; affinity matured humanized anti- CD20 variant 1 with a wild type Fc and another affinity matured anti-CD20 variant 2 with four amino acid muta- tions on the Fc portion were tested as samples. For anti- CD20 variants 1 and 2, the calculated binding affinities relative to the chimeric anti-CD20 standard were 0.58 and 1.51, respectively. After one hour seeding of WIL2 cells, the attachment of cells to the high binding capacity carbon surface was tight enough so that the wash procedure can be per- formed on an ELISA microplate washer under gentle dispense and aspiration conditions. The magnitude of signal was equivalent to that obtained frommanual wash (data not shown). Whereas after one hour incubation of WIL2 on polystyrene plates (MaxiSorp, Nunc), only few cells remained under the same wash conditions on the microplate washer (data not shown). To rule out the possibility that mutations on the Fc portion of anti-CD20 variant 2 affected the binding to anti-human IgG Fc detection antibody in the direct bin- ding format, we performed a competitive ECL cell-bin- ding assay. Biotinylatedmouse IgG2a with the anti-CD20 variant 1 variable regions was used as a tracer. A concen- tration of 500 ng/ml, which was in the linear range of the binding curve of serially diluted tracer (data not shown), was used in the competitive assay. Serial dilutions of anti- CD20 antibodies (27.4–20,000 ng/ml) were mixed with the tracer to compete for binding to CD20 molecules on the surface ofWIL2 cells. Bound tracer was detected with ruthenium-labeled streptavidin. Using chimeric anti- CD20 antibody as a standard, the relative affinities of anti-CD20 variants 1 and 2 were 0.45 and 1.8, respec- tively (Fig. 1B). These results agreed with that obtained from the direct ECL cell-binding assay format. A competitive assay is useful for comparing affinities of antibodies with constant regions from different species. In addition, when screening for antibody drug candidates that inhibit downstream intracellular signaling, a direct format onlymeasures binding affinity of an antibody to receptor. A Y. Lu et al. / Journal of Immun competitive assay provides more information on how well an antibody blocks receptor–ligand interaction. Here we demonstrated that the ECL cell-binding assay could be performed in both direct and competitive formats. 3.2. ECL cell-binding assay characterization To test the reproducibility of the ECL cell-binding assay, humanized anti-CD20 variant 1 was run in 18 independent assays, in duplicate, using WIL2 cells up to twenty passages during a period of eight weeks. Themean binding affinity relative to chimeric anti-CD20 antibody was 0.59±0.05, similar to results obtained using WIL2 cell-based ELISA (0.63±0.08) (Hong et al., 2004). The inter- and intra-assay coefficients of variation were 7.91 and 7.01%, respectively, which were also similar to those of WIL2 cell-based ELISA (7.8 and 6.3%, respectively) (Hong et al., 2004). Since the binding of anti-CD20 antibody does not cause internalization of surface CD20 molecules, the in- cubation steps for the ECL cell-binding assaywere carried out at room temperature. The direct ECL cell-binding assay has a sensitivity of 2.74 ng/ml for the chimeric antibody. The ratios of ECL signal at 2.74 ng/ml versus 0 ng/ml were approximately six- and four-fold for chi- meric anti-CD20 and variant 1, respectively (Fig. 1A). In the WIL2 cell-based ELISA (Hong et al., 2004), incu- bation steps were performed at 4 °C. Under the same 4 °C incubation conditions, the assay sensitivity of WIL2 ECL cell-binding assay was equivalent to that of WIL2 cell- based ELISA (approximately two-fold separation in signal between 2.74 ng/ml and 0 ng/ml of anti-CD20 variant 1) (data not shown). Therefore, room temperature incubation resulted in improved assay sensitivity. The ECL cell-binding assay has a wider assay range (2.74– 2000 and 27.4–20,000 ng/ml in the direct and competitive formats, respectively) compared with cell-based ELISA (15.6–1000 ng/ml in the direct and competitive formats) (Hong et al., 2004, and data not shown). To demonstrate specificity of the ECL cell-binding assay, the assay was run in the absence ofWIL2 cells. Only low background signal was detected at 2 μg/ml of chimeric anti-CD20 antibody (data not shown). In addition, when using WIL2 cells, insignificant signals were observed for bevacizumab (Avastin, anti-VEGF; Presta et al., 1997) or trastuzumab (Herceptin, anti-Her2; Carter et al., 1992) even at concentrations of 10 μg/ml (data not shown). While the anti-CD20 variant 1 generated an expected titration curve similar to that shown in Fig. 1A on the same plate (data not shown). These results demonstrated that bevacizumab or trastuzumab did not bind to proteins on the surface ofWIL2 cells, suggesting the binding of anti-CD20 antibodies was 77al Methods 314 (2006) 74–79 specific to CD20 molecules on WIL2 cells. Similar results were obtained using rituximab in 96- well plates or using anti-CD20 variant 1 in either 96- or 384-well plates (data not shown). In addition, using suspension WIL2 cells which expressed 0.16 million CD20 molecules per cell (Hong et al., 2004) on 384-well plates, variant 1 gave a titration curve (data not shown) with comparable signal to noise ratio to that on 96-well plates (Fig. 1A). 3.4. Evaluation of optimal cell number for ECL cell- binding assay using different cell lines For ECL cell-binding assay, since cells are attached to carbon surface as amonolayer, less cells are used compared to the suspension cell-based ELISA (250,000–300,000 cells/well on 96-well plates). A cell number of 25,000 cells/ well in 96-well plates was found to be optimal for mea- suring binding of anti-CD20 antibodies toWIL2 cells (after testing 3125–100,000 cells/well) and another B cell line BJAB that expresses a comparable level of CD20 asWIL2 cells (data not shown). The best signal to noise ratio was obtained at 37,500 cells/well for assaying binding of a ological Methods 314 (2006) 74–79 Fig. 2. ECL cell-binding assay using CHO cell clones expressing different amounts of CD20 surface molecules. Serial dilutions of rituximab (2.74–2000 ng/ml) were incubated with CHO cell clones 2H3 (circle), 6D7 (square), 4H10 (diamond) and 3G8 (triangle) on 384- well high binding capacity plates. Bound rituximab was detected with ruthenium-labeled anti-human IgG Fc. For clones 2H3, 6D7, 4H10 and 3G8, the ratios of ECL signal at 2000 ng/ml versus 0 ng/ml of rituximab were 346, 225, 31 and 2, respectively; and the ratios of signal at 2.74 ng/ 78 Y. Lu et al. / Journal of Immun 3.3. Evaluation of surface CD20 copy number requirement for ECL cell-binding assay Generation of CHO cell clones with different CD20 surface expression levels has been described previously (Hong et al., 2004). Clones 2H3, 6D7, 4H10 and 3G8 expressed 1.2, 0.19, 0.033 and estimated by Scatchard analysis (data not shown). Signals References 79Y. Lu et al. / Journal of Immunological Methods 314 (2006) 74–79 number of≥50,000 cells/well resulted in slight decrease in signal for theWIL2 andBJABcell lines tested. Thismay be caused, in part, by high density of cells blocking the flow of electric current across the carbon surface. These results demonstrate that the ECL cell-binding assay can be applied to multiple suspension cell lines for different antigen and antibody pairs. 3.5. Correlation of relative binding affinities of anti-CD20 variants measured in cell-based ELISA and ECL cell-binding assay To determine whether the relative binding affinities of anti-CD20 variantsmeasured in the ECL cell-binding assay were comparable to that obtained using the cell-based ELISA (Hong et al., 2004), 15 variants with different affinities to CD20 antigen were assayed using WIL2 cells on 96-well plates. The relative binding affinities measured in both methods showed a high degree of correlation by linear regression analysis (r2=0.94) (Fig. 3). 4. Conclusion We developed a high throughput ECL cell-binding assay using 96- or 384-well high binding capacity carbon electrode plates. We demonstrated that the assay was spe- cific, quantitative and reproducible. Using a microplate washer for the wash steps makes the assay compatible with automation. The increased throughput also makes it fea- sible to extend the cell-binding assay to other applications, such as screening or epitope mapping of hybridoma clones for antibodies against multitransmembrane proteins and for characterizing antibody binding to a panel of cell surface Fc receptors from different species. Acknowledgments We thank Pankaj Oberoi at Meso Scale Discovery for his advice in developing the ECL cell-binding assay. The authors would also like to thank the following people at Genentech: Camellia Adams and Ashley Penn for providing anti-CD20 antibodies, Greg Spaniolo for helping with statistics, and Laura DeForge for helpful discussions. Anderson, K.C., Bates, M.P., Slaughenhoupt, B.L., Pinkus, G.S., Schlossman, S.F., Nadler, L.M., 1984. Expression of human B cell- associated antigens on leukemias and lymphomas: a model of human B cell differentiation. Blood 63, 1424. Baron, D., Wernet, P., Schunter, F., Wigzell, H., 1977. Detection and quantitation of human Ia-type antigens with iodinated protein A and specific purification of antibodies against Ia-type alloantigens. Scand. J. Immunol. 6, 385. Best, J.D., Jay, M.T., Otu, F., Ma, J., Nadin, A., Ellis, S., Lewis, H.D., Pattison, C., Reilly, M., Harrison, T., Shearman, M.S., Williamson, T.L., Atack, J.R., 2005. Quantitative measurement of changes in amyloid-β(40) in the rat brain and cerebrospinal fluid following treatment with the γ-secretase inhibitor LY-411575 [N2-[(2S)-2- (3,5-difluorophenyl)-2-hydroxyethanoyl]-N1-[(7S)-5-methyl-6- oxo-6,7-dihydro-5H-dibenzo[b,d]azepin-7-yl]-L-alaninamide]. J. Pharmacol. Exp. Ther. 313, 902. Boye, J., Elter, T., Engert, A., 2003. An overview of the current clinical use of the anti-CD20monoclonal antibody rituximab.Ann.Oncol. 14, 520. Carter, P., Presta, L., Gorman, C.M., Ridgway, J.B., Henner, D., Wong, W.L., Rowland, A.M., Kotts, C., Carver, M.E., Shepard, H.M., 1992. Humanization of an anti-p185HER2 antibody for human cancer therapy. Proc. Natl. Acad. Sci. U. S. A. 89, 4285. Hong, K., Presta, L.G., Lu, Y., Penn, A., Adams, C., Chuntharapai, A., Yang, J., Wong, W.L., Meng, Y.G., 2004. Simple quantitative live cell and anti-idiotypic antibody based ELISA for humanized antibody directed to cell surface protein CD20. J. Immunol. Methods 294, 189. Kneitz, C., Wilhelm, M., Tony, H.P., 2002. Effective B cell depletion with rituximab in the treatment of autoimmune diseases. Immunobiology 206, 519. Liang, P., Sanchez, R.I., Martin, M.T., 1996. Electrochemilumines- cence-based detection of β-lactam antibiotics and β-lactamases. Anal. Chem. 68, 2426. Presta, L.G., Chen, H., O'Connor, S.J., Chisholm, V., Meng, Y.G., Krummen, L., Winkler, M., Ferrara, N., 1997. Humanization of an anti-vascular endothelial growth factor monoclonal antibody for the therapy of solid tumors and other disorders. Cancer Res. 57, 4593. Schlosser, M., Witt, S., Ziegler, B., Ziegler, M., 1991. Influence of target cell preparation on binding of monoclonal islet cell reactive antibodies (mc-ICRA) in cellular enzyme-linked immunosorbent assay (CELISA). J. Immunol. Methods 140, 101. Sedgwick, J.D., Czerkinsky, C., 1992. Detection of cell-surface molecules, secreted products of single cells and cellular prolifer- ation by enzyme immunoassay. J. Immunol. Methods 150, 159. Shields, R.L., Namenuk, A.K., Hong, K., Meng, Y.G., Rae, J., Briggs, J., Xie, D., Lai, J., Stadlen, A., Li, B., Fox, J.A., Presta, L.G., 2001. High-resolution mapping of the binding site on human IgG1 for FcγRI, FcγRII, FcγRIII and FcRn and design of IgG1 variants with improved binding to the FcγR. J. Biol. Chem. 276, 6591. Smith, M.R., 2003. Rituximab (monoclonal anti-CD20 antibody): mechanisms of action and resistance. Oncogene 22, 7359. Tedder, T.F., Streuli, M., Schlossman, S.F., Saito, H., 1988. Isolation and structure of a cDNA encoding the B1 (CD20) cell-surface antigen of human B lymphocytes. Proc. Natl. Acad. Sci. U. S. A. 85, 208. obtained at 37,500 cells/well were up to 1.5-fold higher compared to those obtained using 25,000 cells/well. A cell A high throughput electrochemiluminescent cell-binding �assay for therapeutic anti-CD20 antibod..... Introduction Materials and methods Reagents and cell lines ECL cell-binding assay Data analysis Results and discussion ECL cell-binding assay ECL cell-binding assay characterization Evaluation of surface CD20 copy number requirement for ECL cell-binding assay Evaluation of optimal cell number for ECL cell-binding assay using different cell lines Correlation of relative binding affinities of anti-CD20 variants measured in cell-based ELISA a..... Conclusion Acknowledgments References