In vitro immunotherapy potency assays using real-time cell analysis

Abstract

A growing understanding of the molecular interactions between immune effector cells and target tumor cells, coupled with refined gene therapy approaches, are giving rise to novel cancer immunotherapeutics with remarkable efficacy in the clinic against both solid and liquid tumors. While immunotherapy holds tremendous promise for treatment of certain cancers, significant challenges remain in the clinical translation to many other types of cancers and also in minimizing adverse effects. Therefore, there is an urgent need for functional potency assays, in vitro and in vivo, that could model the complex interaction of immune cells with tumor cells and can be used to rapidly test the efficacy of different immunotherapy approaches, whether it is small molecule, biologics, cell therapies or combinations thereof. Herein we report the development of an xCELLigence real-time cytolytic in vitro potency assay that uses cellular impedance to continuously monitor the viability of target tumor cells while they are being subjected to different types of treatments. Specialized microtiter plates containing integrated gold microelectrodes enable the number, size, and surface attachment strength of adherent target tumor cells to be selectively monitored within a heterogeneous mixture that includes effector cells, antibodies, small molecules, etc. Through surface-tethering approach, the killing of liquid cancers can also be monitored. Using NK92 effector cells as example, results from RTCA potency assay are very well correlated with end point data from image-based assays as well as flow cytometry. Several effector cells, i.e., PBMC, NK, CAR-T were tested and validated as well as biological molecules such as Bi-specific T cell Engagers (BiTEs) targeting the EpCAM protein expressed on tumor cells and blocking antibodies against the immune checkpoint inhibitor PD-1. Using the specifically designed xCELLigence immunotherapy software, quantitative parameters such as KT50 (the amount of time it takes to kill 50% of the target tumor cells) and % cytolysis are calculated and used for comparing the relative efficacy of different reagents. In summary, our results demonstrate the xCELLigence platform to be well suited for potency assays, providing quantitative assessment with high reproducibility and a greatly simplified work flow.

Fig 1. Working principle of the xCELLigence RTCA system and result from NK92 and TALL-104.

(A) Working principle of xCELLigence impedance technology applied to immunotherapy monitoring. The xCELLigence RTCA label-free technology monitors cell number by changes in impedance measured through gold electrodes embedded in proprietary E-Plates. When seeded alone, target adherent cancer cells proliferation rate is registered as increase in the impedance-related Cell Index (CI) parameter over time. Effector non-adherent immune cells produce small baseline level signal due absence of tight surface adhesion over the gold electrodes. When immune cells are added to adherent target cells, their cytolytic activity causes the adherent cells to round up and detach, consequently reducing CI value. (B) Impedance monitoring is validated using NK92 as effector cells over nuclear red-labeled PC3 prostate cancer cells (left) or MCF7 breast cancer cells (right) as targets. When seeded alone, target cells adhere to the plate and proliferate, increasing the CI readout (blue lines). NK92 effector cells seeded alone caused only a small increase in the CI value over the initial background measurement (green lines). When added to target cells, NK92 cause cell cytolysis and subsequent progressive decrease in CI (red lines). Y-axis is the normalized cell index generated by the RTCA software and displayed in real time. X-axis is the time of cell culture and treatment time in hour. Mean values of the CI were plotted ± standard deviation. Time interval is 6 hours for all the figures unless indicated. (C) The same PC3 target cells are treated with a cytotoxic T cell line (Tall-104), showing the suitability of the technology to use different effector cell types. (D) Images taken at different time points after NK92 addition show correlation between CI drop (red line in the plot), reduction in target cells number (red nuclei PC3 in the images) and changes in cell morphology/adhesion in apoptotic cells (red nuclei in images enlargements).

Fig 2. Parameters of % cytolysis and KT₅₀ determined through impedance measurement.

(A) Cell Index plot for nuclear-red labeled PC3 target cells treated with different E:T ratios of NK92 cytolytic cells. Samples have been internally normalized for the Cell Index value measured before NK92 addition (Normalized Cell Index plot). (B) The Cell Index plot is converted to a % Cytolysis plot by the xCELLigence Immunotherapy Software (xIMT). (C) % Cytolysis measured at 6 and 24 hours after NK92 addition for the different E:T ratios. One way ANOVA result indicates significant difference between individual treatment and control at 6 hours (light blue) and at 24 hours (red); (*** p< 0.001) and (**** p<0.0001). (D) 50% Killing Time (KT₅₀) for the same E:T ratios in (C). ND: Not Detected.

Fig 3. Correlation between % cytolysis determined through impedance measurement and flow cytometry.

(A) Replica plates for the same experiment in Fig 2 has been collected and analyzed by Flow Cytometry. Nuclear red-gated PC3s show ratio and time dependent increase of early apoptotic (annexin V+, DAPI-; bottom right of each plot), and late apoptotic (annexin V+, DAPI+; upper right of each plot) cells. (B) Charts show the % apoptotic cells for the flow data. (C) Total apoptosis measured by flow cytometry is similar to the results of impedance analysis.

Fig 4. Impedance assessment of BiTE-mediated cytotoxicity.

(A) Normalized Cell Index plot for PC3 target cells incubated with PBMCs at different E:T ratios without BiTE. (B) Same E:T ratios as (A) but with 1 μg/ml anti-EpCAM/CD3 BiTE. (C) At E:T ratio of 10:1, different BiTE concentrations resulted in varied dynamic cytolysis of the target cells. (D) Same result from (C) showed as % cytolysis plot. (E) Example of BiTE concentration depended % cytolysis from E:T ratio 10:1 and 1.25:1. (F) KT₅₀ comparison for result from (E). Significant analysis performed by one way ANOVA. (*** p< 0.001),); (** p< 0.01);); (* p< 0.05);); (NS Not Significant); (ND Not Detected).

Fig 5. Adaptation of the xCELLigence killing assay to suspension B cells.

(A). Illustration of B cell attachment through anti-CD40 antibody on the surface of the gold sensor embedded in the well. (B). Raji, 60,000 cells per well, were seeded on the E plate and NK92 cells were added to the well at different E:T ratios. Cytolysis and target cells only are used as positive and negative controls for cyutolysis. (C). Same data as (B) are plotted by xIMT software and displayed as % cytolysis. (D). Comparison of the % cytolysis of Raji cells at different E:T ratios measured by RTCA and flow cytometer at 4 hours (left, correlation coefficient R = 0.9967) and 24 hours (right, R = 0.9903).

Fig 6. CAR-T mediated killing assay of tethered Raji B cells.

(A) Raji cells were seeded at 40,000 cells per well on E plate. One day after seeding, effector CAR-T cells were added to the well at E:T ratio of 2:1. Mock CAR-T cells and NK92 were added for comparison. (B) Same data from (A) but displayed as % cytolysis.

Fig 7. PC3 killing by PBMCs is enhanced by anti PD-1 blocking antibody.

(A) Freshly isolated PMBCs shows increased PD-1 expression after SEB stimulation. (B) PC3 at 5,000 cells per well were seeded on E plate and treated with freshly isolated PBMCs two days after initial seeding in presence with increasing concentrations of the anti PD-1 blocking antibody. PC3 cells only and PC3 cells plus PBMCs and without antibody were included as controls. The insert show the dose dependent curve and EC₅₀ calculated at t = 150 hours. (C) Same data from (B) but displays as % cytolysis.