MONOCLONAL ANTIBODIES IN IMMUNODIAGNOSIS AND IMMUNOTHERAPY Volume 32, Number 4, 2013 ª Mary Ann Liebert, Inc. DOI: 10.1089/mab.2013.0003

A Dual-Chamber System for Screening Cytotoxic Effects of Hybridoma-Produced Antibodies Ton That Ai Long,1 Gavreel Kalantarov,1 Alexandra Chudner,2,* and Ilya Trakht1

There are many methods for evaluating the cytotoxic effect of monoclonal antibodies (MAbs) against cancer cells. Most of these methods require either purified MAbs or biological solutions (e.g., cell culture supernatants, ascitic fluids) containing high concentrations of MAbs. This makes the primary screening of antibody-producing hybridomas for specific cytotoxic antibodies a challenging task. Addressing this issue, this work introduces a high throughput screening method, which enables the identification of cytotoxic antibodies using primary hybridoma populations without prior antibody concentration and/or purification. The method is comprised of a dual-chamber system, where antibody-producing hybridomas and target cancer cells are co-cultured but separated by a porous membrane in which the pore size is sufficient for the diffusion of antibody molecules. The MAbs produced in the system continuously diffuse through the membrane between the two chambers and interact with the target cells placed on the other side of a membrane, resulting in death or proliferation arrest of these cells, if MAbs are cytotoxic or cytostatic. The cytotoxic/cytostatic effect can be registered by measuring the viability of target cells. The advantage of this method is that purification or concentration of antibodies secreted by hybridomas is not required. In addition, this method does not require MAb-secreting hybridomas, which are subcloned or have a high level of MAb production. The method may serve as an effective primary high throughput screening for cytotoxic antibodies.

Introduction

A

ntibodies can cause a direct cytotoxic effect through apoptotic mechanisms: complement-dependent cytolysis (CDC) or antibody-dependent cellular cytotoxicity (ADCC).(1–5) Cytotoxic antibodies are the most preferable candidates for therapeutic applications.(6) The use of "naked" antibodies is more advantageous than the use of ones that require chemical modification as in immunotoxins,(7–9) immunoradioconjugates,(10,11) or prodrugs.(12–14) Measurement of the cytotoxic activity of antibodies must meet certain requirements. For instance, antibodies should be present at high concentration in biologic solutions (e.g., culture supernatants, ascites) or they have to be purified from such solutions. In addition, the incubation time should be long enough (up to 2–3 days) to yield a significant cytotoxic effect and the culture media has to sustain enough nutrients to support the survival and growth of the target cells during such long incubation times. For massive high throughput screening of antibodies generated by hybridoma technology, those requirements may face some

practical challenges. Antibody concentration may be low in primary hybridoma populations, in which antibody-producing cells are not yet cloned and their population may contain only a small fraction of cells that are able to produce cytotoxic antibodies. To elicit a potential cytotoxic effect, antibodies in primary hybridoma supernatants need to be purified or enriched by repeated cloning. However, this becomes practically unreasonable when screening a large panel of hybridomas for the presence of cytotoxic antibodies, especially at the initial stages of hybridoma generation. These matters behooved us to look for alternative screening methods that require neither cloning of primary hybridoma populations nor purification of specific antibodies from primary supernatants. In this report, we demonstrate that the cytotoxic effect of monoclonal antibodies (MAbs) can be detected by using a dual-chamber system in which MAbproducing hybridomas and target cancer cells are simultaneously co-cultured. This system can be developed into a high throughput format for primary screening of hybridoma populations that are producing cytotoxic antibodies.

1 Division of Experimental Therapeutics, Department of Medicine, College of Physicians and Surgeons, Columbia University in New York City, New York. 2 Brooklyn Technical High School, Brooklyn, New York. *Current Address: SUNY Downstate Medical Center, Brooklyn, New York

246

SCREENING CYTOTOXIC ANTIBODY PRODUCED BY HYBRIDOMA Materials and Methods Inserts and plates for cell culture Transwell Permeable Support inserts (hereafter called ‘‘inserts’’), which were used in our experiments, were purchased from Corning (Tewksbury, MA). Each insert has a thin, translucent polycarbonate membrane, which is 6.5 mm in diameter, permeated with 3 mm pores at a density of 2 · 106 pores/cm2, and designed for use with the conventional 24well tissue culture plates. According to the manufacturer, the insert (called the ‘‘upper chamber’’) is suspended inside a well of the tissue culture plate (called the ‘‘lower chamber’’) in which the cells are seeded and cultured in both chambers. The membrane of the insert separates the cells between the two chambers, but allows biological molecules or chemical reagents present in the culture media to diffuse through it, accumulate in the other chamber, and interact with the target cells in this chamber. Cancer cell lines and hybridomas Human breast cancer cell lines SK-BR-3 and MDA-MB-231 were obtained from American Type Culture Collection (ATCC, Manassas, VA) and maintained in Dulbecco modified Eagle medium (DMEM) with standard supplements and 10% heat-inactivated fetal calf serum (FCS). Human ovarian cancer cell line SK-OV-3 and human lung cancer cell line A549 were received from ATCC and cultured in McCoy-10 media with a standard set of supplements and 10% FCS. All the cell cultures were split 1:5 every 3 days. These cell lines were used as targets for monoclonal antibodies in order to screen the antibodies’ direct cytotoxic effect. The clones of human antibody-producing hybridomas used in this study were 13.74, 102.G2, 102.G12, 104.3G3, 104.2H2, 104.3G12, 69.1C12, 27F7, and 27B1. All these hybridoma clones (called ‘‘test hybridomas’’) were discovered and produced in our laboratory. Hybridoma clones and respective human antibodies 13.74, 27F7, and 27B1 were described previously.(16) The method of generation of these hybridomas is described below. Hybridoma cells were seeded in 100 mm Petri dishes and maintained in RPMI-1640 media supplemented with vitamins, amino acids, and 10% FCS. The negative control for the test hybridomas in our experiments was the MFP-2 cell line, which is a human hybridoma fusion partner originally generated and maintained in our laboratory as previously described.(15) MFP-2 cells do not secret any immunoglobulins or individual light chains. These cells were also maintained in RPMI-1640 media supplemented with vitamins, amino acids, and 10% FCS. The cells were split 1:10 every 7 days. Antibodies and reagents Human monoclonal antibody (huMAb) clone 13.74 (IgM, l) was produced in our laboratory and described previously.(16) All other antibodies were commercial products purchased from Thermo Fisher Scientific (Rockford, IL). Goat anti-human IgM antibody-coated agarose beads were purchased from Sigma (St. Louis, MO). Culture media such as RPMI-1640 and DMEM were purchased from Corning Cellgro (Manassas, VA), unless otherwise indicated. For the measurement of cell viability, the stock solution of 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyltetrazolium bromide (MTT)

247

dye and its stop/solubilization buffer were purchased from Promega (Madison, WI). Preparation of MAb-producing hybridomas The method of generating fully human hybridomas, capable of producing huMAbs against antigens expressed in cancer cells has been described previously.(15,16) Briefly, lymphocytes were isolated from lymph nodes or from the peripheral blood of cancer patients and fused with MFP-2 cells. The resulting hybridomas were cultured in RPMI-1640 media and supplemented with 10% FBS, hypoxanthine-aminopterin-thymidine (HAT) medium, and basic supplements. All hybridomas were first screened for production of nonspecific immunoglobulins (Ig) by standard sandwich enzymelinked immunosorbent assay (ELISA). Then, supernatants collected from cultures of human Ig-secreting hybridomas were screened for presence of cancer-associated huMAbs by using cellular ELISA (cELISA) and flow cytometry on cancer cell lines. Measurement of cell viability Cell viability was assessed by a colorimetric MTT assay. This assay is based on the ability of living cells to converse a tetrazolium salt into a formazan product, which is a colored chemical compound and therefore can be quantified using spectrophotometry. Briefly, cells were incubated with 3(4,5-dimethyl-2-thiazolyl)-2,5-diphenyltetrazolium bromide (MTT) dye solution (Promega) for 4 h in 5% CO2 at 37C, followed by 1 h incubation with solubilization/stop solution (Promega) to solubilize the formazan product. Each sample was analyzed in either duplicates or triplicates. Results were recorded by Multiskan ELISA Reader (Thermal Corp., Vantaa, Finland) using a 570 nm filter. Cytotoxicity assays and analysis Selected huMAbs were tested for their cytotoxic effect on cancer cells by using the respective hybridomas and a dualchamber system consisting of an insert and a well of tissue culture plate as described above. The format and procedure of the experiments are illustrated in Figure 1. Briefly, both test hybridomas and target cancer cell lines (called ‘‘target cells’’) were resuspended in DMEM medium supplemented with 2 mM L-glutamine, 10% FCS, and basic supplements (called the ‘‘culture media’’). Between 3.2 · 103 and 2 · 106 test hybridomas were resuspended in 500 mL of the culture media and seeded into the wells of a 24-well plate (i.e., the lower chambers), depending on the experiment. Between 1 · 103 and 4 · 103 target cells were resuspended in 200 mL of the culture media and seeded into each insert (i.e., the upper chamber), depending on the experiment. The volumes of cell suspensions used in the chambers were selected according to the manufacturer’s instructions (Corning). Both test hybridomas and target cells in the dual-chamber system were co-cultured in 5% CO2 at 37C for 24 to 72 h without adding or refreshing the culture media. The MFP-2 cell line was used as negative control for test hybridomas, as described above. In all experiments, test samples were set in triplicates, unless indicated otherwise. After co-culturing, the upper chambers (i.e., the inserts) containing target cells were removed, drained off the culture

248

LONG ET AL.

FIG. 1. Design and procedure for cytotoxicity assay using the dual-chamber system. (1) Lower chamber (individual well of 24-well tissue culture plate); (2) upper chamber (insert); (3) target cells; (4) MAb-producing hybridomas; (5) culture media; (6) 96-well ELISA plate. Test hybridomas in the lower chamber and target cells in the insert are co-cultured for a certain period of time (Step A). The inserts are then removed from the system and target cells are assessed for viability using MTT dye (Step B). After MTT staining, culture supernatant is transferred to a 96-well ELISA plate for data quantification using an ELISA reader (Step C).

media, transferred to a new 48-well tissue culture plate (Corning), and immediately filled with 200 mL of the fresh culture media. Each insert was subsequently supplemented with 15 mL MTT stock solution, followed by 4 h incubation; then supplemented with 100 mL of stop/solubilization solution, followed by 1 h incubation, as recommended by the manufacturer. Finally, 100 mL of supernatant from each removed upper chamber (i.e., insert) was transferred to a 96well ELISA plate (Corning) for reading on the Multiskan ELISA Reader using a 570 nm filter. In this study, cytotoxicity index was calculated as the percentage of dead target cells, which were co-cultured with test hybridomas, in comparison to total viable target cells, which were co-cultured with MFP-2 cells (i.e., the negative control for test hybridomas). Since the results of the MTT assay indicated the quantity of viable target cells, cytotoxicity index was calculated by the following formula: Cytotoxicity index ¼ ½(viable control  viable test)=viable control · 100% where viable control was the optical density (OD) value for target cells co-cultured with MFP-2 cells and viable test was the OD value for target cells co-cultured with test hybridomas. Depleting human IgM from system of cytotoxicity assay In some experiments, huMAbs produced during the cytotoxicity assay were depleted from the system by using agarose beads coated with goat anti-human IgM antibodies (called ‘‘anti-human IgM immunobeads’’). Briefly, target cells (SKBR-3 cells, 1 · 103 cells per upper chamber) were co-cultured with test hybridomas (clone 13.74 and MFP-2 cells, 2 · 106 cells per lower chamber), in which the test hybridomas were mixed with different amounts of anti-human IgM immunobeads. Cytotoxicity effect was evaluated after 48 h of incubation as described above. Measurement of antibody concentration Supernatants from the co-culture in the dual-chamber system were collected in order to measure the concentration of antibodies secreted by test hybridomas using ELISA as

described previously.(16) Briefly, wells of a 96-well ELISA plate (Corning) were coated with goat anti-human IgM. Serially diluted supernatants of test hybridomas were then dispensed into coated wells of an ELISA plate and incubated for 2 h, followed by 1 h incubation with HRP-conjugated goat anti-human IgM. Commercial human IgM was used as positive control and for creating the standard calibration curve. Colorimetric reaction was developed by the addition of 3,3’,5,5’ tetramethylbenzidine substrate (Sigma) and stopped with 10% HCl. The assay was read on a Multiskan ELISAReader at 450 nm. Ascent Software (Thermal Corp.) was used to process the data and calculate the antibody concentration. Flow cytometry Unfixed fresh cancer cell lines, including SK-BR-3, A549, and SK-OV-3, were used in flow cytometry for evaluation of antibody binding to the surface targets. A typical procedure of staining cells for flow cytometry included 30 min blocking with cold 10% FCS, followed by 30 min incubation on ice with tested hybridoma supernatant containing human antibody. For all antibodies, the concentration was adjusted to 5 mg/mL, unless otherwise indicated. The antigen-antibody binding was detected by 30 min incubation on ice with FITC-labeled goat F(ab)2 anti-human IgM ( Jackson ImmunoResearch, West Grove, PA). Commercial human IgM ( Jackson ImmunoResearch) was used as a negative isotype control. Stained cells were read on FACScalibur flow cytometer (Becton Dickinson, Franklin Lakes, NJ). Prior to reading, 1 mL propidium iodide (5 mg/mL) was added to the cell samples for excluding dead cells from analysis. Acquired data were analyzed on either CellQuest Pro software (Becton Dickinson, v4.0.2) or FlowJo software (v7.6.3, TreeStar, Ashland, OR). Results of flow cytometry were displayed as the percentage of cells stained positive. Data analysis and statistics One-way ANOVA and independent-sample t-tests were performed to compare the means of the cytotoxicity indexes among groups and between the two groups, respectively. Paired-samples t-test was performed to compare the means of antibody concentrations and cytotoxicity indexes in some experiments. All data were processed and analyzed using

SCREENING CYTOTOXIC ANTIBODY PRODUCED BY HYBRIDOMA

249

Results

FIG. 2. Effect of number of target cells on cytotoxicity assay. Target cells (SK-BR-3) in the insert were co-cultured with test hybridomas (clone 13.74) and negative control (MFP-2), which were placed in the lower chamber. Cytotoxic effect on target cells and concentration of 13.74 antibody in the insert were assessed after 48 h. Data are presented as mean – SEM. Bars indicate cytotoxicity index. Diamonds indicate the antibody concentration. One-way ANOVA was applied to compare means of cytotoxic index (F(2,3) = 13.92; p = 0.03) and antibody concentration (F(2,3) = 1.400; p = 0.372).

SPSS for Windows software (v10.0.1, Chicago, IL). A change was considered to be significant when p < 0.05. In the ANOVA procedure, when a comparison between two specific means was required, the p value for that comparison was determined by either Scheffe or Tamhane post-hoc test. Results were displayed as mean – standard error of mean (SEM).

To evaluate potential application of the proposed dualchamber system for cytotoxic assay, we used a well-developed huMAb 13.74, which is a human IgM. It was previously shown that this huMAb specifically bound to human breast cancer cell lines SK-BR-3 and MCF-7 and caused apoptosis.(16) In order to determine the effect of the number of target cells on the results of the cytotoxicity assay, SK-BR-3 cells were used as target cells and seeded into the upper chambers (or inserts) at different numbers. The hybridoma producing 13.74 huMAb (clone 13.74) was used as the test hybridoma and seeded into the lower chambers at a constant number (2 · 106 cells per chamber). The MFP-2 cell line was placed in the lower chamber as a negative control. Both target cells and test hybridomas were co-incubated for 48 h without refreshing the culture media. As shown in Figure 2, when the number of target cells increased from 1 · 103 to 4 · 103 cells per insert, the cytotoxicity index dropped slightly but was statistically significant (55.07 – 2.32%, 45.56 – 3.53%, and 33.21 – 2.85% for 1 · 103, 2 · 103, and 4 · 103 SK-BR-3 cells per insert, respectively, with F(2,3) = 13.92, p = 0.03). The concentration of 13.74 antibodies in the lower chambers was not much affected by the change in the number of target cells (16.50 – 0.71 mg/mL, 16.00 – 0.00 mg/mL, and 15.00 – 1.41 mg/mL for groups of 1 · 103, 2 · 103, and 4 · 103 target cells per insert, respectively, with F(2,3) = 1.40, p = 0.372). Correlation between the number of test hybridomas and the level of the cytotoxicity index was also evaluated. SK-BR-3 cells were used as target cells and seeded in the inserts at a constant number (1 · 103 cells per chamber); while test hybridomas (clone 13.74) were seeded at different numbers in the lower chambers (3.2 · 103, 16 · 103, 80 · 103, 400 · 103, and 2 · 106 cells per chamber). MFP-2 cells were used as negative

FIG. 3. Effect of number of hybridoma cells on cytotoxicity assay. Target cells (SK-BR-3) in the insert were co-cultured with test hybridoma (clone 13.74) and negative control (MFP-2), which were placed in the lower chamber. Cytotoxic effect on target cells and concentration of 13.74 antibody in the lower chamber were assessed after 48 or 72 h. Data are presented as mean – SEM. Light gray and dark gray bars indicate the cytotoxicity index at 48 and 72 h, respectively. Solid and dashed lines indicate the antibody concentration at 48 and 72 h, respectively. One-way ANOVA was applied to compare means of cytotoxicity index (F(4,11) = 9.29, p = 0.011 for 48 h incubation; F(4,5) = 14.25, p = 0.006 for 72 h incubation) and antibody concentration (F(4,5) = 512.87, p < 0.001 for 48 h incubation; F(4,5) = 2955.21, p < 0.001 for 72 h incubation).

250 control. Both target cells and test hybridomas were co-cultured for 48 and 72 h without changing the culture media. The results are shown in Figure 3. One-way ANOVA showed increase in the number of test hybridomas was significantly proportional with increase in both the cytotoxicity index (F(4,11) = 9.29, p = 0.011 for 48 h incubation; F(4,5) = 14.25, p = 0.006 for 72 h incubation) and the antibody concentration (F(4,5) = 512.87, p < 0.001 for 48 h incubation; F(4,5) = 2955.21, p < 0.001 for 72 h incubation). In comparison to 48 h incubation, 72h incubation did not induce significant increase in the cytotoxicity index. For example, the cytotoxicity index in the group containing 2 · 106 test hybridomas was 45.23 – 2.92% after 72 h incubation vs. 36.73 – 5.6% after 48 h incubation (t(1) = - 2.7; p = 0.226). Paired-samples t-test showed that antibody concentration in the lower chambers was significantly higher after 72 h incubation at the p < 0.05 levels. This corresponded to the accumulation of secreted antibodies in these chambers. In our experiments, the number of hybridoma cells cocultured with target cells was high and apparently caused significant exhaustion of nutrients in the media after prolonged incubation. As a result, this might affect the viability of the target cells. To evaluate this effect, SK-BR-3 target cells (1 · 103 cells per insert) were co-cultured for 48 h with test hybridoma cell (clone 13.74, 2 · 106 cells per lower chamber) and different numbers of MFP-2 negative control cells (4 · 105 and 2 · 106 cells per lower chamber). The culture media was not refreshed or changed during this 48 h co-culture. The viability index of the target cells was calculated as the percentage of viable target cells that were co-cultured with test hybridomas in comparison with the total viable target cells that were cultured alone without any other type of cells. As shown in Figure 4, viability of target cells started to decrease when they were co-cultured with numbers of MFP-2 negative

LONG ET AL. control cells that do not secret any antibodies (F(3,8) = 30.41, p < 0.001 for 48 h incubation; and F(3,7) = 37.57, p < 0.001 for 72 h incubation). When the number per lower chamber of MFP-2 cells alone increased from 4 · 105 to 2 · 106 cells, viability of target cells seemed to incline downwards (81.47 – 7.85% vs. 68.41 – 2.99% for 48 h incubation; 82.78 – 6.72% vs. 62.40 – 3.40% for 72 h incubation). However, posthoc comparisons by Tamhane test indicated such tendency was statistically insignificant ( p = 0.262 and p = 0.055 for 48 and 72 h incubation, respectively). However, viability of target cells co-cultured with 2 · 106 MFP-2 cells per lower chamber was significantly higher than those co-cultured with the same amount of clone 13.74 hybridomas (i.e., 2 · 106 cells per lower chamber; 68.41 – 2.99% vs. 44.87 – 1.94% for 48 h incubation with p = 0.026; 62.40 – 3.40% vs. 35.48 – 2.15% for 72 h incubation with p = 0.027, in which the p values were determined by post-hoc Tamhane test). The lowest viability level of target cells co-cultured with test hybridomas (clone 13.74) is certainly explainable by the presence of cytotoxic 13.74 antibodies. These data suggested that the depletion of media nutrients or toxic products of cellular degradation during cell culture might have caused cell death, which might be misinterpreted as real cytotoxic effect caused by antibodies. Therefore, throughout this study, the death rate of target cells co-cultured with MFP-2 cells was used as a baseline or background value. All antibody-caused cytotoxic effects were evaluated in reference to this background value. The idea behind the development of this assay was to screen for cytotoxic MAbs, which were secreted by antibodyproducing hybridomas at the very early stage of hybridoma development, where hybridomas were usually present as mixed populations of several clones. Since only one or some clones of such mixed populations might be able to secrete specific antibodies, the concentration of these antibodies in the

FIG. 4. Viability of target cells treated by exhausted culture media. Target cells (SK-BR-3) placed in the insert were cocultured with or without test hybridomas (clone 13.74) and negative control (MFP-2), which were placed in the lower chamber. Cell viability of target cells was assessed after 48 or 72 h. Data are presented as mean – SEM. Dark gray and light gray bars indicate the viability index at 48 and 72 h, respectively. One-way ANOVA was applied to compare means of viability of target cells during each incubation time with F(3,8) = 30.41, p < 0.001 for 48 h incubation and F(3,7) = 37.57, p < 0.001 for 72 h incubation.

SCREENING CYTOTOXIC ANTIBODY PRODUCED BY HYBRIDOMA

FIG. 5. Effect of mixed population of hybridoma clones on the cytotoxicity assay. Target cells (SK-BR-3) seeded with 1000 cells per upper chamber were co-cultured with 2 · 106 test hybridoma cells per lower chamber, which were the mixtures of negative control cells (MFP-2) and antibodyproducing cells (clone 13.74) at ratios of 4:0, 3:1, 2:2, 1:3, and 0:4. Cytotoxic effect on target cells and concentration of 13.74 antibody in the insert were assessed after 48 h. Data are presented as mean – SEM. Bars indicate cytotoxicity index. The solid line shows the antibody concentration. One-way ANOVA was applied to compare means of cytotoxic index (F(4,10) = 16.34, p < 0.001) and antibody concentration (F(4,10) = 1911.85; p < 0.001).

hybridoma culture may be very low. Thus, cytotoxic effect caused by the antibodies may not register, if only the supernatant of the hybridoma culture was used. However, direct use of the hybridoma cells as shown in the proposed dualchamber system would enable continuous supply of antibodies and eventually their accumulation to levels at which a

251

significant cytotoxic effect can be registered. To demonstrate this idea, cytotoxicity assay using the dual-chamber system was tested on a model simulating a non-cloned mixed hybridoma population, in which cloned hybridoma cells (clone 13.74) were mixed with MFP-2 cells at different ratios. Figure 5 shows the results of experiments in which each upper chamber was seeded with 1000 SK-BR-3 cells and each lower chamber was seeded with 2 · 106 test hybridomas as the mixtures of MFP-2 cells and hybridoma cells (clone 13.74) at ratios of 4:0 (i.e., 2 · 106 MFP-2 cells per 0 hybridoma cells); 3:1 (i.e., 1.5 · 106 per 0.5 · 106); 2:2 (i.e., 1 · 106 per 1 · 106); 1:3 (i.e., 0.5 · 106 per 1.5 · 106); and 0:4 (i.e., 0 per 2 · 106). The cytotoxic effect was assessed after 48 h of co-culture. Comparison of cytotoxicity indexes indicated that the difference between ratios of MFP-2 per 13.74 hybridoma cells was significant with F(4,10) = 16.34, p < 0.001. In the condition of these experiments, the cytotoxic effect of the 13.74 huMAb was detectable, when the antibody-producing cells constituted as little as 25% of total cells (i.e., 0.5 · 106 cells) present in the lower chamber (t(4) = - 4.09; p < 0.001, in comparison to cytotoxicity index between MFP-2 cell-hybridoma cell ratios of 4:0 and 3:1). The inserts’ thin and permeable membrane enables diffusion of antibodies from the lower chamber to the insert. However, this process is not instant and may depend on the number of hybridoma cells seeded in the lower chamber, as well as the incubation time. As shown in Figure 6, when 8 · 104 test hybridoma cells per lower chamber were used, the cytotoxicity index was not significantly different between 24 and 48 h incubation periods (6.18 – 0.92% and 8.61 – 1.45%, respectively, with t(2) = - 1.28, p = 0.328). Meanwhile, antibody concentration in the inserts and the lower chambers were significantly different after 24 h of incubation (0.53 – 0.01 mg/mL vs. 1.73 – 0.01 mg/mL, with t(2) = 19.00, p < 0.001), but not significantly different after 48 h of incubation (1.98 – 1.52 mg/mL vs. 4.43 – 0.06 mg/mL, with t(2) = 2.89, p = 0.115). However, when 2 · 106 test hybridoma cells per lower chamber were used, differences in the cytotoxicity

FIG. 6. Diffusion of antibody into the insert. Target cells (SK-BR-3) in the insert were co-cultured with test hybridomas (clone 13.74) and negative control (MFP-2), which were placed in the lower chamber. Cytotoxic effect on target cells and concentration of 13.74 antibody in both chambers were assessed after 24 or 48 h. Data are presented as mean – SEM. Black square and white circle show cytotoxicity index at 24 and 48 h, respectively. Checked and white columns show antibody concentration in the insert and lower chamber after 24 h, respectively. Gray and stripped columns show the antibody concentration in the insert and lower chamber after 48 h, respectively. Paired samples t-test was conducted to compare means of cytotoxicity index and antibody concentration.

252

LONG ET AL.

FIG. 7. Effect of antibody depletion on cytotoxicity assay. Test hybridomas (clone 13.74) and negative control (MFP-2) were mixed with anti-human IgM antibody-coated agarose beads (immunobeads) at different volumes and placed in the lower chamber, then co-cultured with target cells (SK-BR-3) that were placed in the insert. Cytotoxic effect on target cells and concentration of 13.74 antibody in the insert were assessed after 48 h. Data are presented as mean – SEM. Bars indicate cytotoxicity index. Solid line indicates antibody concentration. One-way ANOVA was applied to compare means of cytotoxic index (F(2,6) = 16.44, p = 0.001) and antibody concentration (F(2,6) = 33.01; p = 0.004). indexes and the antibody levels in both chambers were dependent on the incubation time. For instance, the cytotoxic index of 48 h incubation was almost two times higher than that of 24 h incubation (42.92 – 5.17% vs. 23.18 – 2.41%, with t(2) = - 6.55, p = 0.026). After 24 h of incubation, antibody concentration in the lower chamber was approximately two times higher than that in the insert (20.33 – 1.41 mg/mL vs. 9.77 – 0.39 mg/mL, t(2) = 5.93, p = 0.027). After 48 h of incubation, the antibody levels were not significantly different between two chambers (21.15 – 0.45 mg/mL vs. 17.25 – 1.25 mg/ mL, t(1) = 1.71, p = 0.338). It was important to prove if the cytotoxic effect observed in this study was indeed caused by 13.74 huMAb that was pro-

duced in the dual-chamber system during the cytotoxic assay. Thus, cytotoxic assay with antibody depletion was performed in which a constant number of test hybridoma cells (clone 13.74) was seeded in the lower chamber along with various amounts of agarose beads coated with immobilized rabbit antihuman IgM (anti-human IgM immunobeads). Human MAb 13.74, which was an IgM, was captured and immobilized on anti-human IgM immunobeads. MFP-2 cells, serving as negative control, were also incubated with immunobeads in the same manner as test hybridoma (clone 13.74). As shown in Figure 7, there was a proportional correlation between the cytotoxic index and the amount of immunobeads. The cytotoxic index linearly decreased and was fully abolished eventually, along with increasing doses of anti-human IgM immunobeads (30.80 – 5.05%, 27.36 – 4.88%, and 3.20 – 1.65% vs. 0 mL, 5 mL, and 20 mL of immunobeads, respectively, with F(2,6) = 16.44, p = 0.001). These results demonstrate that, once being immobilized by anti-human IgM immunobeads, huMAb 13.74 produced by hybridoma cells were not able to diffuse through the membrane between the two chambers and elicit a cytotoxic effect on target cells. This finding confirmed that the cytotoxic effect on target cells was actually caused by the specific antibodies, which were produced by the test hybridoma cocultured in the dual-chamber system. Based on the results of testing a well-characterized hybridoma (clone 13.74) as shown above, the proposed dual-chamber system was applied to screen other huMAb-secreting hybridomas, which were developed in our laboratory, for cytotoxic effect of their huMAbs on cell lines of breast cancer (SK-BR-3), lung cancer (A549), and ovarian cancer (SK-OV-3). Briefly, primary hybridomas were initially screened for production of human nonspecific Igs as described above. Then, only immunoglobulin-secreting hybridomas (1 · 106 cells per lower chamber) were selected for co-culturing with target cells (1 · 103 cells per insert or upper chamber) using the proposed system. Cytotoxicity effect was accessed after 48 h as described above. Representative results are shown in Table 1. The data generated using flow cytometry, immunohistochemistry, and cell-ELISA confirmed that huMAbs secreted by these hybridomas were able to bind to antigens expressed on those target cell lines. Figure 8 shows representative histograms of flow cytometry analysis.

Table 1. Screening Hybridomas for Cytotoxic Antibodies Using Proposed Dual-chamber System and Different Target Cells Cytotoxicity index (%) Test hybridoma 102.G2 102.G12 104.2H2 104.3G3 104.3G12 69.1C12 27.B1 27.F7

Isotype of antibody m, m, m, m, m, m, m, m,

l k l k l k k k

SK-BR-3 (human breast cancer)

A549 (human lung cancer)

SK-OV-3 (human ovarian cancer)

27.61 – 6.71 26.32 – 1.17 16.18 – 8.74 25.71 – 4.76 18.97 – 0.02 2.13 – 8.19 4.79 – 6.78 0.99 – 9.26

18.42 – 0.23 17.06 – 7.85 0.64 – 5.53 20.92 – 3.2 19.22 – 1.85 5.91 – 3.70 7.26 – 4.30 4.39 – 2.80

28.9 – 4.95 33.21 – 8.81 0.00 – 9.41 31.24 – 3.34 26.12 – 18.89 0.08 – 9.03 5.88 – 6.24 8.82 – 0.00

Both target cells (i.e., SK-BR-3, A549, and SK-OV-3 cell lines) and test hybridoma cells (i.e., huMAb-producing hybridomas and MFP-2 cells used as negative control) were co-cultured in the suggested dual-chamber system. Target cells were placed in upper chambers (1 · 103 cells per chamber). Test hybridoma cells were placed in lower chambers (1 · 106 cells per chamber). The cytotoxic effect was assessed after 48 h of co-culture. Data are presented as mean – SEM.

SCREENING CYTOTOXIC ANTIBODY PRODUCED BY HYBRIDOMA

253

FIG. 8. Representative histograms of flow cytometry analysis. Fresh cancer cell lines (SK-BR-3, A549, and SK-OV-3) were stained for 30 min with IgM antibody-containing supernatants harvested from the test hybridoma cultures, followed by staining with FITC-conjugated goat F(ab)2 against human IgM. Stained cells were read by FACScalibur flow cytometer (Becton Dickinson). Dead cells were excluded from analysis by propidium iodide staining. Acquired data were analyzed on FlowJo software (TreeStar, v7.6.5). Results were displayed as the percentage of cells stained positive. Solid curves represent histograms of test antibodies and dashed curves represent histograms of IgM isotype negative control.

Discussion Screening for cytotoxic effects of antibodies is a critical step in searching for MAbs that may be used in immunomodulation of cancer (e.g., antibody-based immunotherapy, vaccine immunotherapy, etc.). Although there are many methods for evaluating antibody cytotoxicity, most of them require purified antibodies and high antibody concentrations, which may not be practically available during the primary screening of antibody-producing hybridomas at a very early stage of their generation. To overcome such limitations, we developed a new method of primary screening for the cytotoxicity of antibodies, which directly used antibody-producing hybridomas, instead of antibody-containing hybridoma supernatants or purified antibodies. The advantage of directly using antibody-producing hybridomas in a cytotoxicity assay is that antibodies are continuously produced and supplied to the assay, thus securing a real-time continuous effect on target cells. This allows for addressing the issues of both the insufficient concentration in primary supernatants and the potential degradation of anti-

bodies in culture media. In addition, the use of hybridomas may minimize a non-specific cytotoxic effect caused by nutrient depletion of the media supernatants, when antibodycontaining culture supernatants are used. We observed that target cancer cells were killed when challenged with supernatants harvested from old and dense cultures of themselves, as well as of cells, which did not produce immunoglobulins or any specific cytotoxic factors (data not shown). A deficiency in nutrients and a high level of cellular metabolic products in old culture supernatants, which may be considered as exhaustion of culture media, may cause such nonspecific cytotoxic effects on target cancer cells. As for the design of the proposed cytotoxicity assay in this work, hybridomas can be technically co-cultured with target cells in the same chamber. However, a disadvantage of this design is that results are often difficult to interpret, as the death of target cells have to be distinguished from the death of test hybridomas. Thus the design in which hybridomas and target cells are simultaneously cultured in two chambers separated by a porous membrane was explored. In this system, the antibodies are produced by hybridomas on one side of a membrane, diffuse through the

254 membrane, accumulate, and interact with target cells on the other side of that membrane. Since antibodies are continuously supplied and accumulated during the assay, this method does not require purification, concentration, or any other methods of antibody enrichment. This makes it possible to screen for cytotoxic hybridomas even before they are cloned or when the antibody production is low. As a result, this method can be applied to screen primary hybridoma populations in the earliest stages of development (i.e., right after these hybridomas are generated and not yet cloned) for potential antibody candidates, which are able to elicit a cytotoxic/cytostatic effect on target cells. Another advantage of this method is its simple experimental procedure, which enables the design of experiments using a high throughput format. This work has demonstrated that the proposed method can be successfully applied to detect the direct cytotoxic effect of huMAbs on target cancer cells, which may be realized, for instance, through apoptotic mechanisms. This method can be modified to screen for cytotoxic antibodies, which utilize the mechanism of complement-binding and complementdependent cytotoxicity. In conclusion, the cytotoxicity assay using co-culture of antibody-producing hybridoma and target cancer cells in the dual-chamber system may be a method of choice for the primary screening of cytotoxic antibodies produced by hybridoma technology. We also believe that the application of this method is not limited to screening for cancer-specific antibodies, but can also be expanded to screening antibodies for their inhibitory effect on intracellular pathogens of infectious diseases such as malaria, leishmania, and listeriosis. The latter utilities of the described method are yet to be tested. Acknowledgment This study was funded by Acceptys Inc. (New York, NY). This publication is dedicated to the memory of our late colleague Dr. Irena Kirman. Author Disclosure Statement The authors declare no conflict of interest with this research. References 1. Iwatani Y, Amino N, Mori H, Asari S, Matsuzuka F, Kuma K, and Miyai K: A microcytotoxicity assay for thyroid-specific cytotoxic antibody, antibody-dependent cell-mediated cytotoxicity and direct lymphocyte cytotoxicity using human thyroid cells. J Immunol Methods 1982;48(2):241–250. 2. Eisenbarth GS, Morris MA, and Scearce RM: Cytotoxic antibodies to cloned rat islet cells in serum of patients with diabetes mellitus. J Clin Invest 1981;67(2):403–408. 3. Irie RF, Ollila DW, O’Day S, and Morton DL: Phase I pilot clinical trial of human IgM monoclonal antibody to gangli-

LONG ET AL.

4. 5.

6. 7.

8.

9.

10.

11.

12. 13.

14.

15.

16.

oside GM3 in patients with metastatic melanoma. Cancer Immunol Immunother 2004;53(2):110–117. Adams GP, and Weiner LM: Monoclonal antibody therapy of cancer. Nat Biotechnol 2005;23(9):1147–1157. Shuptrine CW, Surana R, and Weiner LM: Monoclonal antibodies for the treatment of cancer. Semin Cancer Biol 2012;22(1):3–13. Aly HA: Cancer therapy and vaccination. J Immunol Methods 2012;382(1–2):1–23. Kawakami K, Kawakami M, and Puri RK: Overexpressed cell surface interleukin-4 receptor molecules can be successfully targeted for antitumor cytotoxin therapy. Crit Rev Immunol 2001;21(1–3):299–310. Frankel AE, Powell BL, and Lilly MB: Diphtheria toxin conjugate therapy of cancer. Cancer Chemother Biol Response Modif 2002;20:301–313. Iyer U, and Kadambi VJ: Antibody drug conjugates—Trojan horses in the war on cancer. J Pharmacol Toxicol Methods 2011;64(3):207–212. David KA, Milowsky MI, Kostakoglu L, Vallabhajosula S, Goldsmith SJ, Nanus DM, and Bander NH: Clinical utility of radiolabeled monoclonal antibodies in prostate cancer. Clin Genitourin Cancer 2006;4(4):249–256. Pohlman B, Sweetenham J, and Macklis RM: Review of clinical radioimmunotherapy. Expert Rev Anticancer Ther 2006;6(3):445–461. Denny WA: Prodrug strategies in cancer therapy. Eur J Med Chem 2001;36(7–8):577–595. Wu AM, and Senter PD: Arming antibodies: prospects and challenges for immunoconjugates. Nat Biotechnol 2005;23(9): 1137–1146. Singh Y, Palombo M, and Sinko PJ: Recent trends in targeted anticancer prodrug and conjugate design. Curr Med Chem 2008;15(8):1802–1826. Kalantarov GF, Rudchenko SA, Lobel L, and Trakht I: Development of a fusion partner cell line for efficient production of human monoclonal antibodies from peripheral blood lymphocytes. Hum Antibodies 2002;11(13):85–96. Kirman I, Kalantarov GF, Lobel LI, Hibshoosh H, Estabrook A, Canfield R, and Trakht I: Isolation of native human monoclonal autoantibodies to breast cancer. Hybrid Hybridomics 2002;21(6):405–414.

Address correspondence to: Dr. Ilya Trakht Division of Experimental Therapeutics Department of Medicine College of Physicians and Surgeons Columbia University in New York City 630 W 168th Street P & S Building, Room 8-503 New York, NY 10032 E-mail: [email protected] Received: January 15, 2013 Accepted: April 21, 2013