This protocol describes how to assess synergism between an anticancer antibody and antineoplastic drugs in preclinical models by using the combination index equation of Chou and Talalay.
Potentiation of hostile monoclonal antibodies (mAb) by chemotherapeutic agents constitutes a valuable strategy for designing effective and safer therapy against cancer. Here we provide a protocol to identify a rational combination at the preclinical step. First, we describe a cell-based assay to assess the synergism between anticancer mAb and cytotoxic drugs, that uses the combination index equation of Chou and Talalay1. This includes the measurement of tumor cell drug- and antibody-sensitivity using an MTT assay, followed by an automated computer analysis to calculate the combination index (CI) values. CI values of <1 indicate synergism between tested mAbs and cytotoxic agents1. To corroborate the in vitro findings in vivo, we further describe a method to assess the combination regimen efficacy in a xenograft tumor model. In this model, the combined regimen significantly delays tumor growth, which results in a significant extended survival in comparison to single-agent controls. Importantly, the in vivo experimentation reveals that the combination regimen is well tolerated. This protocol allows the effective evaluation of anticancer drug combinations in preclinical models and the identification of rational combination to evaluate in clinical trials.
The conventional approach in the treatment of a large number of different types of cancer was based on monotherapy. Even if it is still used in many cases, this method met several obstacles leading to opting for combined therapies2. Particularly, cancer cells are more susceptible to develop resistance when treated with a single drug by inducing alternative survival mechanisms3, resulting in therapeutic failure in patients4. Moreover, in monotherapy, drugs are usually administrated at a high dose. This situation often results in the occurrence of strong dose-dependent side effects that can be intolerable and force physicians to stop the treatment2. For these reasons, the association of anticancer molecules is now preferred to monotherapy.
Ideal drug combinations would be those that act in synergy against tumor cells, without increased toxicity against normal cells. Synergism refers to the interaction of two or more drugs that produces a therapeutic effect greater than the sum of each individual drug acting separately. Such interactions may result in enhanced clinical therapeutic efficacy2. It limits treatment resistance, increases efficacy, and can also reduce toxicity2. In fact, the dosage of each drug can be reduced to lower their side effects by targeting different pathways. In addition, one of the molecules can also serve as a sensitizing agent against cancer cells. The effect of the second drug may be enhanced on sensitized cells and fewer dosages can be used5.
Combined therapy can include two or more chemotherapeutic drugs and/or biologics, such as monoclonal antibodies6. These mAbs specifically target cells expressing a cell surface antigen of interest and are able to kill tumor cells through immunological pathways including antibody-dependent cell-mediated cytotoxicity (ADCC), with the involvement of immune effector cells7, and complement-dependent cytotoxicity (CDC)6. They can also act via a non-immunological mechanism mediated by apoptosis8,9,10,11. In this case, the induction of the process of programmed cell death may sensitize cancer cells, weaken their function, and make the associated chemotherapeutic drug more effective at a lower dosage. As such, proapoptotic mAb are good candidates for designing combination regimens with antineoplastic drugs.
Different mathematical models have been described to assess drug synergism; one of them is based on the combination index method1. This method is based on the median-effect principle developed by Chou1. The median-effect equation correlates the drug dose and drug effect as follows.
Here, D is the drug dose; Dm is the median-effect dose; Fa is the fraction affected by the dose; m is an exponent that signifies the shape of the dose-effect plot1. The median-effect dose is used to calculate the dose Dx of a drug that inhibits or kills "x" percent of cells. The CI value is then calculated to assess the additive effect of the drug combination, as follows1.
A CI value of 1 indicates an additive effect and a CI value of <1 indicates a synergistic effect, while a CI value of >1 indicates antagonism1. The application of this method is further facilitated by the availability of a computer program, CompuSyn, that determines synergism and antagonism at all doses or effect levels simulated automatically12.
Our group has developed the mAb 8B6 specific for O-acetyl-GD2 ganglioside (OAcGD2) neuroblastoma antigen13 and further demonstrated that this mAb is able to induce cell death with attributes of apoptosis11. To test whether mAb 8B6 can sensitize neuroblastoma cells to the antineoplastic agent topotecan, we adapted the above-mentioned method developed by Chou1. First, we determine the effective dose 50 (ED50) values of mAb 8B6 and topotecan. Next, the neuroblastoma cells with equipotent ratios of the two compounds based on ED50 values are exposed to determine the CI values using the above-mentioned simulation software. This method allows us to demonstrate synergism between mAb 8B6 and topotecan in vitro. Next, we describe a protocol to further assess the potency and the safety of this combination regimen in vivo. This protocol can be easily applied to select potent and safe anticancer mAb and chemotherapeutic agent combinations in preclinical studies. A schematic representation of this study is provided in Figure 1.
Animal housing and experimental procedure were approved by the French Government (agreements #C44-278 and #APAFIS 03479.01). Animal care and procedures were conducted under directive EU 2010/63/EU and French Law #2013-118 on the protection of animals used for scientific purposes.
1. Evaluation of the Drug Interaction Between mAb 8B6 and Topotecan In Vitro
2. Generation of Human Neuroblastoma Xenografts in Nonobese Diabetic NOD Scid Gamma Mice (NSG Mice)
NOTE: Exclude any contamination of the cell culture. Since the basement membrane matrix forms a gel above 5 °C, all cultureware or media coming in contact with the basement membrane matrix reagent should be prechilled/ice-cold. Keep the basement membrane matrix on ice during the entire process.
3. Drug and Antibody Administration in Mice
The representative results and figures are adapted with permission from earlier published work14.
Anti-OAcGD2 mAb 8B6 Synergistically Enhances the Inhibitory Effects of Topotecan on Neuroblastoma Cell Line Growth:
To establish the drug and the antibody concentrations to be used for assessing synergism between topotecan and mAb 8B6, the drug and the antibody sensitivities of human IMR5 neuroblastoma cells were measured first, using an MTT assay. Exposure to either mAb 8B6 or topotecan alone for 72 h resulted in a concentration-dependent inhibition of IMR5 cell viability (Figure 4A). Dose-response curves allowed the calculation of the ED50 values for each compound. To this end, Fa values were computed using the analysis simulation software. The calculated ED50 values were found to be 10 nM ± 1 for topotecan (Figure 4B) and 18 µg/mL ± 3 for mAb 8B6 (data not shown).
Based on these ED50 values, next, the potency of six combinational equipotent ratios of mAb 8B6 and topotecan were tested (Figure 2). The shift of the combination dose-response curve toward the sensitive side of the graph indicates that the combination regimen is more potent than each monotherapy (Figure 4A). To obtain the corresponding ED50 and CI values, Fa values were computed. The ED50 values of topotecan were significantly lower in the presence of mAb 8B6 (p < 0.05, Figure 4B), indicating that mAb 8B6 sensitizes the tumor cell to topotecan. Importantly, the CI values were significantly less than 1.0 (p < 0.05), demonstrating a synergistic interaction (Figure 5). Thus, mAb 8B6 has the potential as an adjuvant therapeutic agent for topotecan chemotherapy.
Anti-OAcGD2 mAb 8B6 Enhances Antitumor Activity of Topotecan In Vivo
Because in vitro models neither take into account the drug half-life nor the drug metabolism, it is necessary to corroborate the in vitro findings in vivo. Moreover, in vivo models are very useful to assess the combination regimen safety. To confirm that the synergism between topotecan and mAb 8B6 was specific to cancer cells, the antineuroblastoma effects of the combination therapy in a tumor xenograft model were assessed. For this, the severe immunodeficient NSG mouse strain was selected. These mice lack both an innate and an adaptive immune system15, and therefore, make it possible for researchers to exclude any immunomodulatory effects induced by topotecan therapy that can affect mAb potency. Treatment was started once the tumors displayed a mean volume of 50 ± 2.5 mm3 (Figure 6A). The treatments consisted of either mAb 8B6 (150 µg i.v., day 7 and day 11), topotecan (0.36 mg/kg i.p., days 7 – 11), or a combination of mAb 8B6 and topotecan. The tumor volumes were monitored during the course of the experimentation. All therapies led to tumor growth retardation compared with control groups (Figure 6A). The combination regimen, however, induced the strongest effect (Figure 6A).
Tumor growth can be further analyzed to study the survival rate upon treatment. To this end, the event resulting in mouse euthanasia was defined as the tumor volume ≥ 1 cm3 for ethical reasons. This allowed us to perform a Kaplan-Meyer survival analysis and to calculate the median event-free survival time for each experimental group. As shown in Figure 6B, both monotherapies improved the event-free survival (EFS) substantially when compared to control groups (the median EFS of the vehicle-treated group was 21 days; of the control antibody group, 22 days; of the topotecan group, 26 days; of the mAb 8B6 group, 29 days [Figure 6B]). Yet, the combined therapy had the strongest effect on the mice survival, with a median EFS extended to 39.5 days (p < 0.05, mAb 8B6 vs. combination; p < 0.01, topotecan vs. combination [Figure 6B]).
Because synergistic interactions can result in increased toxicity, the combination regimen safety needs to be assessed. As such, weight loss is commonly used as a sensitive marker for health monitoring in rodents16. Thus, weight loss-measured as a decline in percentage from the initial weight-was retained as an indicator of the systemic tolerability of each tested regimen. No loss of body weight was observed, suggesting that the treatment was well tolerated (Figure 7). These data suggest that the combination of topotecan plus mAb 8B6 represents a more potent antitumor efficacy in vivo than either agent alone, without detectable toxicity.
Figure 1: Schematic representation of the study. Please click here to view a larger version of this figure.
Figure 2: Layout of the combination experiment on a 96-well plate with ratios of topotecan-to-mAb 8B6, prepared as six solutions. The solutions were prepared as described in the protocol. Wells labeled 1 to 4 serve as mAb 8B6 solutions, wells labeled 5 to 8 serve as topotecan solutions, and wells labeled 9 to 12 serve as mAb 8B6 and topotecan (combination) solutions. Wells labeled A and H serve as controls; wells labeled B serve as 0.125 x ED50 concentration drug solutions; wells labeled C serve as 0.25 x ED50 concentration drug solutions; wells labeled D serve as 0.5 x ED50 concentration drug solutions; wells labeled E serve as ED50 concentration drug solutions; wells labeled F serve as 2 x ED50 concentration drug solutions; wells labeled G serve as 4 x ED50 concentration drug solutions, as indicated. Therapeutic agents with two different units (i.e., µg/mL for mAb 8B6, and nM for topotecan) are analyzed in a fixed-ratio combination (0.075, topotecan/8B6). Please click here to view a larger version of this figure.
Figure 3: Injections. (A) Intravenous injection into the tail vein of a restrained mouse, using an insulin syringe of 27 G x 1/2 in, 1 mL. (B) Intraperitoneal injection to the lower right quadrant of the mouse's abdomen, using a 1 mL syringe mounted with a 25 G needle. Please click here to view a larger version of this figure.
Figure 4: Dose-effect relationship of topotecan, mAb 8B6, and their combination on growth viability inhibition of IMR5 neuroblastoma cells after 72 h of exposure. An MTT assay was performed as described in the protocol. (A) The dose-response curves shown are representative of three independent replicates, each run in quadruplicates. The data are presented as the mean ± SD; *** p < 0.001. (B) ED50 of topotecan used as a single agent or in combination with mAb 8B6. The data are presented as the mean ± SEM. This figure has been modified from Faraj et al.14. Please click here to view a larger version of this figure.
Figure 5: Combination index values. Percentage survival values were transformed into fraction-affected (Fa) values and used to calculate the combination index (measure of synergy, additivity, and antagonism) using computer software, as indicated in the text. In the combination index plots, data are presented as mean ± SD for three independent replicates. Results show that mAb 8B6 had a synergistic effect with topotecan (CI <1). This figure has been modified from Faraj et al.14. Please click here to view a larger version of this figure.
Figure 6: Combination treatment of IMR5 xenografts in NSG mice with mAb 8B6 plus topotecan. (A) Mice bearing human neuroblastoma IMR5 xenografts were treated with vehicle (PBS, i.p.), topotecan alone (0.36 mg/kg, i.p.), control IgG alone (150 µg, i.v.), mAb 8B6 alone (i.v.), or topotecan and mAb 8B6 combined, as indicated. The administration of mAb 8B6 or control antibody treatment started on day 7 after the IMR5 cell inoculation and was repeated once on day 11. The topotecan or PBS treatment was started on day 7 and given for five consecutive days. The tumor growth was monitored, and tumor volumes were calculated. The mean tumor volume ± SEM of each treatment group (PBS group, 9 mice; all other groups, 10 mice) are depicted (* p < 0.05 for mAb 8B6 against mAb 8B6 and topotecan together, ** p < 0.01 for topotecan against mAb 8B6 and topotecan together), as indicated. A significant reduction in xenograft volume was observed for the mAb 8B6/topotecan combination, compared to the drug-alone controls, as indicated (p < 0.05). (B) Event-free survival Kaplan-Meyer curves of different groups treated are shown. This figure has been modified from Faraj et al.14. Please click here to view a larger version of this figure.
Figure 7: Mean weight for each treatment group, as indicated. The mean weight of the mice on day 0 was defined as 100% weight. The weight in each group remained stable for the period of treatment. The data are presented as the mean ± SEM. This figure has been modified from Faraj et al.14. Please click here to view a larger version of this figure.
To predict the effect of drug interactions, three methods can be used: the isobologram methodology17, the nonlinear mixture model18, and the combination index1. Combination index analysis is the most commonly used because its application is simplified by the availability of a user-friendly computer program. For this purpose, we first characterized the dose-effect response of each agent used alone or in combination, by performing an MTT assay19. This methodology relies on the ability of viable cells to reduce the tetrazolium salt in a purple-colored formazan product with an absorbance maximum of 570 nm19. At death, the cells lose the capacity to convert MTT into formazan. Thus, the quantity of the colored formazan product is proportional to the number of viable cells in culture19. Indeed, we chose this methodology as a nonradioactive alternative to the tritiated thymidine incorporation into DNA for measuring cell proliferation19. As such, MTT assays have been widely adopted and remain popular in academic labs as evidenced by thousands of published articles. While the quantity of formazan is measured by recording the absorbance at 570 nm using a plate-reading spectrophotometer, a reference wavelength of 620 nm is sometimes used but not necessary for most assay conditions.In consideration of the critical step described here, we found that the culture conditions used to grow cells can affect the results of MTT assays. We found that the viable cell number, the cell metabolic activity, the age of cultures, and the number of passages and details of the growth medium can all be important factors. Thus, they must be taken into consideration when repeating the assay and analyzing data. Furthermore, cell culture conditions differ for each cell lines. It is, thus, complicated to provide any clear indications about the cell seeding density. As a rule of thumb, an average number of cells between 2,000 – 10,000 cells per well is recommended to achieve an optimum cell density within 72 h. Importantly, MTT reduction reflects viable cell metabolism and not, specifically, cell proliferation or cell death. Therefore, MTT assays should neither be described as measuring cell proliferation nor cell cytotoxicity20,11. The detection of cell death relies on specific cell-based assays. For example, in previous studies, we used western blot analysis to detect caspase 3 activation and/or flow cytometry analysis to detect phosphatidylserine in the outer plasma membrane leaflet, to evidence the proapoptotic activity of mAb 8B613.
When combination testing drugs, they can be delivered together or successively in time. Given the variety of the mechanism of action, it is difficult to provide precise guidelines for the relevant experimental setting. Yet, most investigators utilize direct testing of drug combinations. In addition, concentrations that produce a defined single-agent effect of 50% cell growth inhibition are commonly used. Drug concentrations reflecting plasma concentrations achievable in patients are commonly considered as relevant clinically. As such, the combination dose-effect analysis is performed for several doses, keeping the combination ratio constant with a serial dilution of the (ED50)mAb/(ED50)drug ratio, although it is not an absolute requirement1. Of note, the CI values should be also calculated at various ED tested (e.g., ED50, ED75, and ED90) because they may change with the Fa in a nonlinear manner1. After performing the automated analysis of the CI values, the linear correlation coefficient r-value estimated by the algorithms should be >0.9, to reflect the goodness of the fit of the experimental data. CI values of <1 indicate synergy, CI values of =1 indicate additivity, and CI values of >1 denote antagonism1. Due to the variability in both the experimental methods and the biological system, the calculated CI should be further subjected to statistical consideration. As such, a simple method is to repeat the drug combination experiment several times followed by the calculation of the resulting CI values before determining the statistics of the mean ± SD1. It is also advised to test the combination in the largest possible panel of cell lines, to take into account the variability that exists between them.
Importantly, several parameters in the in vitro study are not taken into consideration for the in vivo extrapolation. For instance, the in vitro viability assays neither take into consideration the in vivo half-life of the drug, nor the in vivo drug metabolism that may result in drug inactivation. Thus, the extrapolation from in vitro to in vivo remains a separate question, and the beneficial drug interaction evidenced in vitro should be further confirmed in vivo. As such, a clear demonstration of synergy in mice bearing tumor xenografts, using the combination index method, has been published21. Yet, in vivo studies require a large number of animals to accurately define synergy and are more expensive and more time-consuming. The alternative effective model F-test still requires substantial resources22. As a different approach to limit the use of a large number of animals consists in demonstrating the drug synergy in cell culture models followed by the evidence of an increased antitumor response of the combination in a limited xenograft study14, we used the human neuroblastoma IMR5 cells as the tumor target for the in vivo study. We retained IMR5 cells because they are tumorigenic in immunocompromised mice23. While human tumor xenograft models provide valuable models for testing drug combinations in vivo24, it can be difficult to establish such models from a variety of cell lines25. To increase the incidence of tumor formation in mice, cells can be injected into mice with a membrane basement matrix as a vehicle25. Importantly, since membrane basement matrix polymerizes and solidifies at temperatures above 5 °C, all cultureware or media coming in contact with the membrane basement matrix should be prechilled/ice-cold, and the membrane basement solutions should be kept on ice during the entire process25. With the use of the membrane basement matrix, we observed a 100% take rate with IMR5 cells, while, in the absence of membrane basement matrix, this cell line demonstrated a <80% take rate (data not shown). In addition, to increase the incidence of tumor graft, membrane basement matrix can also increase the tumor growth rate25. We observed that all xenografts had appeared by day 11. However, during the first seven days following the tumor cell challenge, the presence of lumps could be observed, which may be caused by the presence of the initial inoculum (data not shown). Thus, we did not consider measurements of tumor size meaningful before this time. Using the membrane basement matrix made it possible to reduce the number of mice in the experimental setting and to perform the in vivo experiment within three months.
Drug/antibody dosages may vary depending on the cell line and mouse strain. Due to the limitation of the in vitro viability assay for extrapolating the in vivo drug/antibody dosage described above, clinically relevant dosages were retained for both topotecan and mAb 8B6 in the in vivo experimental setting, based on the data published by others26,27. To extrapolate the human dosage setting to mice, previously published guidelines28 were followed. Per se, we injected 150 µg of antibody 8B6 i.v. at day 7, followed by a second injection five days later (total dose/mouse = 300 µg). Topotecan treatment started the same day as the first mAb 8B6 injection by injecting 0.36 mg/kg topotecan or PBS control i.p. for five consecutive days. In consideration of the critical step described here, we recommend starting antibody infusions when the tumor xenografts reach ~50 mm3. Higher tumor xenograft volumes will result in a lower response rate because tumor burdens clearly correlate inversely with antibody concentration, antibody exposure, and antibody efficacy29. We also retained weight loss as an indicator of systemic tolerability of each tested regimen16. However, the interpretation of collected weights may not be helpful with a large tumor mass. In this case, body condition scoring, as described elsewhere16, is advised.
Tumors can be collected at different time points for immunochemistry analysis, to provide further information about the mechanism of action triggered in vivo. In a previous work, the tumor apoptotic index was assessed after mAb 8B6 infusion13. To this end, tumor apoptotic cells were detected using a TUNEL assay13. In addition, the percentage of tumor cell nuclei was scored with Ki67 antigen-positive staining to analyze the tumor proliferation index13.
Severe immunodeficient mice lack both innate and adaptive immunity with the loss of T cells, B cells, and natural killer (NK) cells with reduced macrophage and antigen-presenting cell functions and the absence of circulating complement30. Thus, this protocol does not allow researchers to draw any conclusions regarding putative effects of chemotherapy in potentiating mAb therapy by altering the tumor microenvironment in vivo31. However, both putative immunomodulatory effects of the chemotherapeutic agent and the role of the fragment-crystallizable (Fc) region of the mAb in enhancing the drug antitumor potency merit consideration, but these questions require the availability of a syngeneic model with an intact immune system and a physiological tumor microenvironment.
The authors have nothing to disclose.
Grant support: Fondation de Projet de L'Université de Nantes, les Bagouz' à Manon, La Ligue contre le Cancer comité de Loire-Atlantique, comité du Morbihan, and comité de Vendée, une rose pour S.A.R.A.H, L'Etoile de Martin and la Société Française de Lutte contre les Cancers et les leucémies de L'Enfant et de L'adolescent (SFCE). M.B. and J.F. are supported by La Ligue Contre Le Cancer. The authors thank the UTE-facility of the Structure Fédérative de Recherche François Bonamy. The authors also thank Dr. S. Suzin (Inserm, Paris) for providing the IMR5 cells and Ms. H. Estéphan for her technical assistance.
Cell Proliferation kit (MTT) | Roche | 11-465-007-001 | |
CompuSyn software | ComboSyn | Combosyn can be downloaded for free at http://www.combosyn.com | |
Electric shaver | Bioseb | BIO-1556 | |
Fetal calf serum | Eurobio | CVFSVFF00-01 | 10% heat-inactivated fetal calf serum in RPMI 1640 |
Firefox | Mozilla Corporation | Firefox can be downloaded for free at http://www.mozilla.org/en-US/firefox/ | |
Heat lamp | Verre&Quartz | 4003/1R | |
Human neuroblastoma IMR-5 cell line | Accegen Biotechnology | ABC-TC0450 | IMR-5 is a clone of the human neuroblastoma cell line IMR32 5459762. IMR-5 cells were generously provided by Dr. Santos Susin (U.872, Paris, France) |
L-glutamine | Gibco | 25030-024 | 2 mM in RPMI 1640 |
Lysis solution | Roche | 11-465-007-001 | |
mAb 8B6 | University of Nantes | N/A | |
Matrigel | Corning | 354248 | |
Multiskan FC | Thermofischer Scientific | N08625 | |
Needle 21G 1 ½ | BD Microlance | 304432 | |
Needle 25G 1 | Terumo | NN-2525R | |
NSG mice | Charles River Laboratories | 5557 | |
Nunc MicroWell 96-well microplates | Thermofisher | 167008 | |
PBS | VWR | L182-10 | |
PBS, 0,05% EDTA | Sigma-Aldrich | E9884 | |
PC that runs windows 7 | Microsoft | Windows 7 can be purchased at http://www.microsoft.com/en-gb/software-download/windows7 | |
Penicillin-Streptomycin | Gibco | 15140-122 | 100 units/mL penicillin and 100 mg/mL streptomycin in RPMI 1640 |
Reagent reservoir | Thermofischer Scientific | 8094 | |
Rodent restrainer | Bioseb | TV-150-SM | |
RPMI 1640 | Gibco | 31870-025 | |
Syringe 1 mL | Henke Sass Wolf | 5010.200V0 | |
Topotecan | Sigma-Aldrich | T2705 |