Cell-Based Label-Free Technologies for Drug Discovery

Drug discovery is a resource intensive and risky endeavor. High attrition rates, particularly at the preclinical and clinical stages plague the drug development process. One reason is that drug candidate leads are advanced though the early stages of the drug development process – without the full interrogation of the underlying biochemical mechanism(s), toxicological profile, offtarget interactions and pharmacokinetic and pharmacodynamic properties of candidate lead compounds.

What is needed are technologies that can provide incisive and predictive information about the lead candidate drugs early on in the process. This is to reduce the attrition rate and avoid the late development costs of candidate compounds. Cell-based label free technologies have received considerable attention in the preclinical drug development process.

Cell Study

Even though cell-based label-free platform technologies have existed for at least a couple of decades, the essence of label-free technology as applied to cell biology and drug discovery is increasingly being realized. As the name implies, the preclusion of the label allows for the assessment of cells in their native and physiologically relevant environment, circumventing the potential negative impact of the label on cellular processes.

The inclusion of certain labels and reporters – particularly labels for live cells has been shown to impact various aspects of cellular behavior. Labelfree technologies have the added advantage of being non-invasive and therefore live cells that are present in tissue culture wells can be continuously interrogated.

This feature leads to one of the main advantages of label-free technologies, which is real-time kinetic measurement. The realtime monitoring of cellular processes offers advantages over traditional end-point assays. First, the comprehensive representation of the entire length of the assay is possible, allowing the user to make informed decisions regarding the timing of manipulations or treatments.

Second, the actual kinetic response of cells provides important information regarding the biological status such as cell growth, arrest, morphological changes, and apoptosis. There are a number of optical and non-optical technologies which allow for the label-free assessment of cellbased experiments.

The xCELLigence Real-Time Cell Analysis (RTCA) system is a series of three instruments that differ primarily in throughput capacity (Figure1). The RT-CA SP system is composed of four main components: the electronic microtiter plates (E-Plate); the station which accommodates one E-Plate and is placed inside a tissue culture incubator; the analyzer, for sending and receiving the electronic signals and the control unit which operates the software and continuously acquires and displays the data. The other two systems are the RT-CA MP and RT-CA DP.

Principle of Detection

The system utilizes impedance readout to non-invasively quantify cellular status in real-time. The cells are seeded in E-Plates that are integrated with gold microelectrode arrays (Figure 2). The application of a low voltage (less than 20 mV) AC signal leads to the generation of an electric field between the electrodes.

The field interacts with the ionic environment inside the wells of the E-Plates and is differentially modulated by the number of cells in the well, the morphology of the cells and the strength of cell attachment. The impedance readout harnesses and quantifies these unique changes in cell morphology and adhesion, allowing for an unbiased detection of specific cellular processes in realtime.

Many agents that are used for cancer therapeutics modulate the balance between cell viability/proliferation and cell death. Cell death occurs through a spectrum of distinct morphological and biochemical pathways, culminating in apoptosis, necrosis and autophagy.

Reduced viability often results from diminished cell proliferation or cell cycle arrest. It is important to understand the modes of cell death or reduced proliferation that is mediated by cytotoxic agents. This would help to decipher the mechanism of drug action and its possible side effects. However, the direct effect on cell death or cell cycle progression is often transient and difficult to capture by standard end-point assays.

For example, apoptosis occurs only during a short period of time, often within hours, and is frequently followed by secondary necrotic events. Cell cycle arrest is often followed either by resumed progression or cell demise via apoptosis. Therefore, it is crucial to conduct relevant assays, either cell viability or cell death, at optimal time points.

Considering that the kinetics of cyto-toxic compounds is likely to be different between various compounds, it is important to monitor viability/toxicity continuously. This is to pinpoint the optimal time points to conduct endpoint assays and gain additional mechanistic information.

The cell analysis system can be used to assess compoundmediated cytotoxicity. Figure 3A shows the impedance-based monitoring of the effects of several cytotoxic compounds – including kinase inhibitor staurosporine, proteasome inhibitor MG-132, anti-mitotic agent vincristine, Deoxyribonucleic Acid (DNA) damaging agent 5-FU and Histone Deacetylase (HDAC) inhibitor scriptaid, on HeLa cells.

Even though all compounds demonstrated reduced viability effects at the end of study, the kinetic response shows the timedependent cytotoxic effect of different compounds on the cells that were growing on the sensors. To test whether the cell index changes can be used as a surrogate marker for selecting optimal time points for performing apoptosis assays, the decision was made to conduct apoptosis assays using two apoptosis inducing agents with different cell-killing kinetics (MG-132 and 5-FU, Figure 3B).

MG-132-mediated cytotoxicity started at four hours post compound addition and reached maximal effect at 10 hours; 5-FU took a longer time before cytotoxicity was observed (24 hours post compound addition), and the maximal effects were not observed until 48 hours.

Observing Outcomes

Correspondingly, apoptotic induction was observed for MG-132 only at 16 hours, but not at 64 hours. This confirms the transient nature of the apoptosis process and stresses the importance of performing the assay at an optimal time-point, to be able to capture the cellular response (Figure 3B). For 5-FU, apoptosis induction was observed only at 64 hours when the maximal cell index changes were observed, but not at 16 hours, when the cell index for the treated samples just started to diverge from the control samples (Figure 3B).

Therefore, the kinetic curves from the cell analysis system can help to determine the optimal time point for conducting apoptosis assays, eg, when cell index value reaches the lowest index (indicative of lowest cell viability).

Another example of real-time monitoring is exemplified by vincristine mediated mitotic arrest. Figure 3C shows the realtime cell index changes that are mediated by vincristine. Mitotic index was also determined at several different time points by examining the percentage of phospho-Histone 3 positive cells, a hallmark of cells that are undergoing mitosis.

A clear time-dependent mitotic arrest phenotype was observed and the mitotic index has an inverse relationship to the cell index at the corresponding time points. The cell index recovery later on is associated with the mitotic escape that occurs after prolonged mitotic arrest, and is correlated with the re-spreading of cells.

Even though different cytotoxic agents have different timedependent cytotoxic kinetics, agents with similar modes of action often have similar Time-Dependent Cell Response Profiles (TCRPs). For example, anti-mitotic tubulin interfering agents vincristine, paclitaxel and motor protein Eg5 inhibitor, S-Trityl- L-Cysteine, all gave similar time-resolved cellular phenotypic changes (Figure 3D).

Other signature profiles include patterns that are mediated by DNA damaging agents, nuclear hormone agonists, and protein synthesis inhibitors. Clustering analysis can be used to classify agents with different signature profiles and compounds that cluster together often possess similar mechanisms of actions. A potential application of this approach would be to understand the potential toxic or off-target effects of certain compounds, and to also to identify the biological activity of compounds with unknown mechanisms of action.

Assessment of Results In addition to compound-mediated cytotoxicity, the cell analyzer is also used in conjunction with the small interfering Ribonucleic Acid (siRNA)-mediated knockdown of specific targets to assess the effect on cell proliferation and cytotoxicity. Real-time monitoring would also yield information as to when to conduct relevant end-point assays to assess the direct cell cycle or cell death mechanisms.

At the same time, TCRP could also implicate mode of actions for the targets of interests. Different targets are likely to have different kinetics of small interfering Ribonucleic Acid (siRNA) mediated target down-regulation and protein half-life. It is therefore more important to monitor phenotypic changes in realtime, thereby allowing the real-time monitoring of the functional down-regulation of the targets.

To demonstrate the utility of the cell analyzer system for functional genomics using Ribonucleic Acid interference (RNAi), the choice was made to target the cellular machinery regulating mitosis. The initial focus was on KIF11, a member of the kinesinlike protein family encoding a motor protein that was required for the bipolar spindle during mitosis. Small molecule inhibitor of KIF11, S-Trityl-L-Cysteine, results in mitotic arrest and apoptosis (Figure 3D).

Transfection of KIF 11 siRNA produced a similar and yet delayed TCRP compared to small molecule inhibitor, likely due to the slower kinetics that are associated with siRNA mediated target down-regulation (Figure 3D & 4A). The impedance profile starts to differentiate from the control at 12 hours post transfection and the maximal effect is observed at 24 hours. Interestingly, the effect appears to be transient, ultimately leading to a recovery of the impedance signal (Figure 4A and B).

Analysis of the mitotic index by p-H3 staining at several time points post-transfection indicates again the inverse correlation between impedance changes and mitotic index changes, as lowest cell index corresponds to the maximal phenotypic changes, eg, mitotic index. Since then, cell index recovery occurs as the cells escape mitotic arrest and re-spread (Figure 4B). Analysis of KIF11 mRNA expression by real-time RT-PCR indicates that the phenotypic divergence correlates with the mRNA down-regulation, as 88 precent down-regulation was observed as early as nine hours post transfection and continued throughout the experiment (Figure 4B). Correspondingly, more than 95 percent knock-down of KIF11 proteins were observed at both 24 and 48 hours post transfection.

In addition to the KIF11-7904 siRNA that was described earlier, also used were another KIF11 siRNA (KIF11-7903) and two Polo- Like Kinase 1 (PLK1) siRNAs (PLK1-448 & PLK1-450) which are also involved in mitosis and produced similar TCRP as that of KIF 11 knockdown – although the timing of phenotypic divergence and the extent of phenotypic changes were different, likely due to the different kinetics and extent of siRNA-mediated target downregulation (Figure 4C).

In summary, the advantage of real-time phenotypic monitoring by the RT-CA provides the following advantages:

• Real-time, continuous measurement identifies both subtle and robust phenotypic changes;

• Time-dependent phenotypic divergence from controls identifies the best time points for assaying changes in either mRNA and protein expression;

• Time-dependent cell response profile identifies the best time points for end-point cellular or biochemical assays to understand the direct and specific effect that is mediated by target down-regulation, rather than down-stream, nonspecific effects;

• TCRPs can be predictive of the mechanisms of action during target modulation, providing an opportunity to identify specific pathway-target interactions. It is of particular interest that once a defined TCRP is observed with a compound or pathway modulators, siRNAs can also be utilized in combination to identify targets that modulate the pathway of interest.

Measuring Activity

A Cell Invasion and Migration (CIM) device was used to assess growth factor-mediated migration of endothelial cells in realtime and label-free conditions.

The CIM device is a 16-well modified Boyden Chamber device that is composed of an Upper Chamber (UC) and a Bottom Chamber (BC) (Figure 1). The UC and BC snap together to form a perfect seal. The UC is sealed in the bottom by a micropourous Polyethylene Terepthalate (PET) membrane. The pores allow for the physical translocation of the cells from the upper part of the UC to the lower part or bottom side of the membrane.

The bottom side of the membrane (the side facing the BC) contains interdigitated gold microelectrode sensors, which once bound to migrated cells will generate an impedance signal. The BC contains 16 well reservoirs for the chemoattractant solution which seals to the bottom side of the wells from the UC via pressure-sensitive O ring seals.

To characterize migration of endothelial cells using the CIM system, Human Umbilical Vein Endothelial Cells (HUVEC) from Lifeline Cell Technologies were cultured in Vasculife Vascular Endothelial Growth Factor (VEGF) cell culture medium.

The cells were serum starved in Vasculife Basal medium, detached using Trypsin- Ethylenediaminetetraacetic Acid (EDTA) solution, and the cell density was adjusted to 300,000 cells/mL.

To assess the general migration of HUVEC cells in response to the plethora of growth factors that may be encountered during angiogenesis, VEGF media (containing VEGF, EGF, IGF, bFGF and two percent fetal bovine serum) was serially diluted with a basal medium and transferred to the bottom chamber of the CIM device.

For optimal HUVEC migration experiments, it was determined from previous experiments that Extracellular Matrix (ECM) Protein coating, such as Fibronectin (FN) is necessary and therefore the PET membrane was coated on both sides with 20 μg/mL FN. The Upper Chamber (UC) and Bottom Chamber (BC) of the CIM device was assembled and 100 uL of cell suspension (30,000 cells) were added in each well of the top chamber.

The CIM device was placed in the DP system within the CO2 incubator and cell migration was continuously monitored. Figure 5 shows the time and dose-dependent directional migration of HUVEC cells from the UC to the BC. The combination of all the growth factors and serum provides strong chemoattractant signals which work in concert to induce the directional migration of the cell.