Abstract
Background: Multiple factors affect the in vivo antitumor activity of antibody-based therapeutics; however, the influence of cell surface retention on antibody-dependent cellular cytotoxicity (ADCC) is not fully understood. Here we evaluated the importance of cell surface antibody retention in antitumor activity mediated by ADCC in vivo. Materials and Methods: Two mAbs against tumor-associated calcium signal transducer 2 (TACSTD2/TROP2), AR47A6.4.2 and Pr1E11, were used. Antitumor activities against BxPC3 and Colo205 cells were investigated through in vitro and in vivo assays. Results: Pr1E11 showed better cell surface retention than AR47A6.4.2 in vitro although Pr1E11 and AR47A6.4.2 showed equivalent ADCC activity. Complement-dependent cytotoxicity and antiproliferative activity were not observed for either antibody. Pr1E11 exhibited higher antitumor activity than AR47A6.4.2 in vivo. Conclusion: Our results suggest that high cell surface retention can result in potent ADCC activity in vivo. This observation could provide novel insight into how effectively screen for antibodies with strong in vivo antitumor activity.
Monoclonal antibodies (mAb) represent an attractive format for cancer therapeutics due to their specificity and biological activity. For tumor antigen-targeting mAbs, naked human IgG1 (specifically, an IgG1 antibody that is not conjugated to an effector) has been widely used because this isotype can elicit effector functions such as antibody-dependent cellular cytotoxicity (ADCC) and complement-dependent cytotoxicity (CDC) (1). Evidence from several studies suggests the importance of these effector functions in mAb therapeutics. In particular, ADCC activity is believed to be important for antitumor action in the clinical setting because functional polymorphisms in FcγRIIIA (CD16), a major receptor that triggers ADCC, are associated with the therapeutic efficacy of mAbs (2). It has been reported that multiple factors affect ADCC activity, including antibody density on the cell surface, binding affinity, binding valency (monovalent or bivalent), and the binding epitope (3, 4).
In order to screen for mAbs with strong ADCC, in vitro assays are generally performed to select for candidate clones because, in most cases, in vitro ADCC activity generally correlates with in vivo ADCC activity. However, for other antibodies, especially those specific for highly internalized antigens, in vitro ADCC assays do not reflect in vivo ADCC activity.
Tumor-associated calcium signal transducer 2 (TACSTD2/TROP2), a type I glycoprotein that has high homology with epithelial cell adhesion molecule (EPCAM/TROP1), is a suitable target antigen for studying this issue (5). Increasing evidence suggests that TROP2 is an attractive target for mAb-based therapeutics because it is abundantly expressed in malignant tumors and contributes to tumor aggressiveness (6-8). RS7 is a well-known mAb to TROP2 that was established by Stein et al. (9, 10). This clone was found to have significant ADCC activity against TROP2-positive cancer cells in vitro but did not elicit in vivo antitumor activity in its naked format (11). Since RS7 is rapidly internalized after target antigen binding, these observations led us to speculate that antibody retention on the cell surface (after target antigen binding) is important for in vivo ADCC activity.
In this study, we compared the in vitro and in vivo antitumor activity of two mAbs to TROP2, AR47A6.4.2 and Pr1E11, in order to determine the importance of cell surface antibody retention on antitumor activity. To date, AR47A6.4.2 is the only TROP2 mAb clone that has been found to display significant antitumor activity in vivo (12). Pr1E11 was established by our group from a modified adenovirus-based screening system. Pr1E11 was shown to bind to TROP2 with higher affinity than AR47A6.4.2, and recognized a unique epitope. In addition, cell surface retention of Pr1E11 was higher than that of AR47A6.4.2 in vitro (Table I) (13).
Materials and Methods
Cell lines. The human pancreatic cancer cell line BxPC3 was purchased from the American Type Culture Collection (Manassas, VA, USA). The human colon cancer cell line Colo205 was purchased from DS Pharma Biomedical (Osaka, Japan). These cells were maintained in RPMI-1640 medium (Nakalai tesque, Kyoto, Japan) supplemented with 10% heat-inactivated fetal bovine serum (FBS) (Gibco, Grand Island, NY, USA) and 50 μg/ml gentamicin (Nakalai tesque). The Chinese hamster ovary cell line CHO/DG44 was a kind gift from Dr. Lawrence Chasin (Columbia University, New York, NY, USA) and maintained in iscove's modified dulbecco's medium (Nakalai tesque) supplemented with 10% dialyzed FBS (Gibco), HT supplement (Gibco) and 50 μg/ml gentamicin. CHO-K1 was purchased from RIKEN (Tsukuba, Japan) and maintained in EX-CELL325PF CHO serum-free medium (Sigma, St. Louis, MO, USA) supplemented with 6 mM L-glutamine (Nakalai tesque) and 50 μg/ml gentamicin. An α-1,6-fucosyltransferase (FUT8)-knockout CHO cell line, FUT8−/− CHO, for defucosylated antibody production was developed at Kyowa Hakko Kirin Co., Ltd., as previously described (14).
Peripheral blood mononuclear cells (PBMCs). Peripheral blood was collected from healthy volunteers registered at Kyowa Hakko Kirin Co., Ltd. All donors gave written informed consent before blood collection, in accordance with the process approved by the Institutional Ethical Committee of Kyowa Hakko Kirin (approval number #2009_024_00). PBMCs were prepared from heparinized blood using Lymphoprep (Axis Shield, Dundee, UK).
Mice. C.B-17/lcr-[severe-combined immunodeficient (SCID)] Jcl mice were purchased from CLEA Japan (Shizuoka, Japan) and maintained under specific pathogen-free conditions. All animal studies were performed in accordance with the Standards for Proper Conduct of Animal Experiments at Kyowa Hakko Kirin Co., Ltd., under the approval of the company's Institutional Animal Care and Use Committee (approval number #10-015).
Anti-TROP2 antibodies. Recombinant fucosylated [Fuc(+)] and defucosylated [Fuc(−)] chimeric AR47A6.4.2 (cAR47A6.4.2) and Pr1E11 (cPr1E11) antibodies, with a human IgG1 constant region, were produced by CHO/DG44 or FUT8−/− CHO cells as previously described (8). Recombinant mouse IgG1 AR47A6.4.2 (mAR47A6.4.2) and Pr1E11 (mPr1E11) antibodies were generated as follows. Polymerase chain reaction amplicons of each antibody variable region and mouse IgG1 constant region were inserted into pKANTEX93 plasmid vectors, and transfected into CHO/DG44 and CHO-K1 cells. Antibodies were purified from culture supernatants using Hitrap protein G HP (GE Healthcare, Tokyo, Japan).
ADCC assay. An ADCC assay was performed as previously described (10). Briefly, target cells (1×104 cells/well) and PBMCs (2.5×105 cells/well, as effector cells), were seeded in a 96-well U-bottom plate and incubated with different concentrations (0.1-1,000 ng/ml) of antibodies for 4 h at 37°C. Supernatants were collected for a lactose dehydrogenase (LDH) release assay using an LDH-Cytotoxic test (Wako, Osaka, Japan) in order to evaluate cytotoxicity. Assay measurement was performed according to the manufacturer's protocol. The percentage of cytotoxicity was calculated according to the following formula: where E was the experimental release (supernatant activity from target cells incubated with antibody and effector cells), SE was the spontaneous release in the presence of effector cells (supernatant activity from target cells incubated with effector cells), ST was the spontaneous release of target cells (supernatant activity from target cells incubated with medium alone), and M was the maximum release of target cells (activity released from target cells lysed with 9% Triton X-100).
Complement-dependent cytotoxicity (CDC) assay. A CDC assay was performed as previously described (10). Briefly, target cells were seeded at 1×104 cells/well in a 96-well flat-bottom plate with different concentrations (1-10,000 ng/ml) of antibodies and incubated with a human complement serum (Sigma) at a 1/6 dilution in medium. After a 2-h incubation at 37°C, WST-1 reagent (Roche Applied Science, Mannheim, Germany) was added and the plate was further incubated for 4 h to detect formazan dye production by living cells. The absorbance of each well was measured in accordance with the manufacturer's protocol. The percentage of cytotoxicity was calculated according to the following formula: where C was the absorbance of target cells cultured in medium with complement alone, E was that in the presence of TROP2 mAb, and B was the absorbance of a blank well that contained medium and complement without target cells.
Cell proliferation assay. Cells were seeded at 500 cells/well in a 96-well flat-bottom plate with different concentrations (31.6-10,000 ng/ml) of antibodies. After 5-day incubation at 37°C, WST-1 regent was added and further incubation was performed for 4 h. The independent effect of secondary antibody (ThermoFisher Scientific, Waltham, MA, USA), used for crosslinking TROP-2, was also evaluated. The effect on cell proliferation was calculated according to the following formula: where C was the absorbance of target cells cultured in medium alone, E was that in the presence of TROP2 mAb, and B was the absorbance of a blank well containing medium without target cells. In vivo antitumor activity of TROP2 antibody in mouse xenograft models. BxPC3 or Colo205 cells (5×106 cells) were inoculated s.c. into 5- to 6-week-old male SCID mice. After 5 to 6 days, a tumor volume of approximately 100 mm3 was achieved; the mice were randomly assigned to groups of five to eight animals each, based on tumor volume, and treated with antibody solution or phosphate-buffered saline (PBS)(−) by i.p. administration twice per week. The tumor volume was calculated according to the following formula:
Statistical analysis. All in vitro experiments were performed in triplicate and results are reported as the mean with standard error (SEM). Statistical analysis of in vivo experiments was performed with SAS software (release 9.2; SAS Institute Inc., Cary, NC, USA). Statistical significance between the PBS-treated group and the mAb-treated groups was determined by the Kruskal–Wallis and Steel tests.
Results
AR47A6.4.2 and Pr1E11 showed equivalent ADCC activity in vitro. The ADCC activity of chimeric TROP2 mAb was evaluated using BxPC3 (Figure 1A) and Colo205 (Figure 1B) cells, as TROP2 is highly expressed in these cell lines (8). PBMCs were prepared from four independent donors and used as effector cells. When fucosylated chimeric mAbs were used, weak ADCC activity was observed for cAR47A6.4.2/Fuc(+) and cPr1E11/Fuc(+) against the two cell lines. A high antibody concentration (greater than 100 ng/ml) was needed to induce apparent cytotoxicity. In contrast, when defucosylated chimeric mAbs were used, cAR47A6.4.2/Fuc(−) and cPr1E11/Fuc(−) exhibited ADCC activity from 1 ng/ml. The activity reached a maximum at 100 ng/ml and was higher than those of fucosylated mAbs. Maximum ADCC activity for the two mAbs was comparable.
AR47A6.4.2 and Pr1E11 did not elicit CDC and antiproliferative activity in vitro. CDC assays were performed using defucosylated TROP2 mAbs and human complement. When BxPC3 and Colo205 cell lines were used as target cells, neither cAR47A6.4.2/Fuc(−) nor cPr1E11/Fuc(−) had any cytotoxic activity (Figure 2).
In order to assess in vitro antiproliferative activity, BxPC3 and Colo205 cells were cultured with chimeric TROP2 mAbs in the presence or absence of anti-human IgG polyclonal antibody, as cross-linker. As shown in Figure 3, neither cAR47A6.4.2/Fuc(−) nor cPr1E11/Fuc(−) affected proliferation in the two cell lines.
In vivo antitumor effect of TROP2 antibody using mouse xenograft models. The in vivo antitumor effects of defucosylated chimeric mAbs were examined in mouse xenograft models. In the BxPC3 xenograft model (Figure 4), cPr1E11/Fuc(−) showed stronger anti-tumor activity than cAR47A6.4.2/Fuc(−). The minimum tumor growth inhibition ratio (T/Cmin) at day 36 for cPr1E11/Fuc(−) and cAR47A6.4.2/Fuc(−) was 0.36 and 0.47, respectively. Comparable to this result, cPr1E11/Fuc(−) had superior antitumor activity in the Colo205 xenograft model compared with cAR47A6.4.2/Fuc(−) and showed significant antitumor activity. The T/Cmin at day 44 for cPr1E11/Fuc(−) and cAR47A6.4.2/Fuc(−) was 0.29 and 0.48, respectively.
In order to elucidate the in vivo mechanism of action for Pr1E11 and AR47A6.4.2, we compared the antitumor activity of defucosylated human IgG1 mAbs and mouse IgG1 mAbs. Human IgG1-type mAbs were expected to exert in vivo activity through both Fc-dependent and independent functions. In contrast, the effector function of mouse IgG1 is known to be weak (15, 16); therefore, in vivo activity mostly depends on its Fc-independent function. Both mAR47A6.4.2 and mPr1E11 retained their binding activity to TROP2-expressing cell lines (data not shown). However, as shown in Figure 5, mouse IgG1 mAbs were completely devoid of antitumor activity.
Discussion
ADCC activity is mediated by multiple factors such as binding affinity, epitope, and antigen expression level; therefore, we often observe that different clones exhibit different biochemical characteristics but have equivalent ADCC activities in vitro. In addition, in vitro ADCC activity sometimes does not reflect in vivo ADCC activity. Thus, these problems make the process of how to select strong ADCC-inducing clones unclear. In this study, we evaluated the in vitro and in vivo antitumor activities of two anti-TROP-2 mAbs, AR47A6.4.2 and Pr1E11, in order to identify the key factor important for in vivo ADCC activity. To date, AR47A6.4.2 is the only antibody clone that has shown in vivo antitumor activity in its naked format. Pr1E11 was established by our group, and has different biochemical properties (high affinity, different epitope, and high cell surface retention) compared to AR47A6.4.2.
We reveal that cPr1E11 and cAR47A6.4.2 have equivalent ADCC activity in vitro; however, the in vivo antitumor activity of cPr1E11 was higher than that of cAR47A6.4.2. Truong et al. reported that AR47A6.4.2 exerts CDC-mediated cell killing of BxPC3 cells (12). In order to confirm their observations, we evaluated the CDC activity of cPr1E11 and cAR47A6.4.2 in two different cell lines, including BxPC3. However, under our experimental conditions, we did not observe CDC activity for either antibody. In Truong et al.'s study, rabbit serum was used for the CDC assay to maximize CDC activity. In contrast, we employed human complement. This difference in complement source might explain the discrepancy in CDC activity. It was also reported that AR47A6.4.2 partially suppressed mitogen-activated protein kinase phosphorylation in response to serum stimulation, possibly contributing to its antiproliferative activity. However, in our study, neither cPr1E11 nor cAR47A6.4.2 showed any antiproliferative effect. Antibody hyper-crosslinking, using secondary polyclonal antibodies, was shown to enhance antiproliferative activity for some antigens (17, 18). This reaction condition mimics FcγR-mediated antibody crosslinking in vivo. We evaluated the combined effect of TROP2 mAbs and a secondary antibody but antiproliferative activity was not observed. In order to confirm the in vivo mechanism of action, we prepared mouse IgG1-type mAbs and evaluated their in vivo activities. This was because mouse IgG1 is incapable of mediating Fc-dependent cytotoxicity. Using the Colo205 xenograft model, we found that both mPr1E11 and mAR47A6.4.2 IgG1-type mAbs were completely devoid of antitumor activity. These results indicate that the in vivo efficacies of cPr1E11 and cAR47A6.4.2 are mostly dependent on ADCC.
From these results, it is apparent that sustained exposure to antibody might affect the in vivo antitumor activity for antibodies that primarily function through ADCC. It has been reported that target antigen-mediated internalization reduces the in vivo half-life and local concentration of antibodies (19, 20). In addition, antibody-mediated antigen internalization reduces the antigen density on target cell surfaces, which restricts antibody-mediated effector functions. Therefore, a high level of antigen–antibody internalization potentially relucts ADCC activity in vivo, but not in vitro. In this study, we did not evaluate the antibody half-life and changes in antigen levels in antibody-treated tumors in vivo. Further studies are needed to evaluate this hypothesis.
In conclusion, we demonstrated that Pr1E11 exerts higher ADCC in vivo than AR47A6.4.2, despite their in vitro ADCC activities being comparable. Our results led us to the hypothesis that cell surface retention of the antibody is an important factor for in vivo ADCC activity, and would affect antigen and antibody down-regulation in vivo. If this were true, the assessment of cell surface antibody retention activity in vitro would represent an easy approach to selecting clones capable of eliciting strong ADCC in vivo. Cell surface retention is the net result of dissociation and internalization of antibodies bound to cell surface antigens, and is easy to measure using appropriate methods (specifically, flow cytometry and cell-based enzyme-linked immunosorbent assay). We also demonstrated that TROP2 is a suitable target not only for antibody-drug conjugate but also for ADCC-dependent naked antibody-based therapeutics. To our knowledge, Pr1E11 is the best TROP2-targeting clone for this application to date. Further evaluation is needed to confirm its efficacy and safety for clinical use.
Acknowledgements
The Authors would like to thank Dr. Mitsuo Sato, Dr. Munetoshi Ando, and Masao Asada for their valuable comments and suggestions.
- Received July 28, 2016.
- Revision received August 17, 2016.
- Accepted August 23, 2016.
- Copyright© 2016 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved