Tumor cell death serves as a useful end-point in the pharmacological profiling of cytotoxic and pro-apoptotic agents (1–4). Most contemporary cell viability assays are, however, performed using an “end-point” approach that reveals the frequency of live versus dead cells at the time of harvesting (5–7). Cell death within cell populations is, however, a stochastic process where cell-to-cell variation in temporal progression through the various stages of cell death arises from subtle fluctuations in the concentrations or the states of regulatory proteins, protein oscillations, the induction of multiple compensatory mechanisms (e.g., autophagy), or “molecular noise” (7–11). Therefore, the ability to continuously track individual cells from the time of encountering a stress signal; through the execution phases of cell death, up to the final point of demise, can provide a kinetic fingerprint of anticancer drug action (3, 6, 7).
It was recently proposed that an ideal approach to monitor cell viability would require the development of noninvasive fluorescent markers that: (i) enable population monitoring and cell tracking over an extended period of time; (ii) do not by themselves modify the viability of the cell system, particularly the structural and bio-physiological properties of the cells; (iii) enable multiparameter analysis in combination with other markers; and (iv) are transferable to high-throughput formats and automation (3, 6–8, 12).
In search for noninvasive fluorescent probes capable of long-term monitoring of cell death in real-time, we evaluated a new anthracycline derivative, DRAQ7. The probe does not penetrate plasma membrane of living cells; however, once the membrane integrity is compromised, it readily binds to nuclear DNA and thus reports cell death. The spectral properties of the molecule provide a detection window in the far-red (>630 nm) (identical to the cell permeant dye DRAQ5). The current study investigated the effects of DRAQ7 on living cells; its intracellular distribution and/or compartmentalization, the effects on cell cycle including DNA replication and possible interaction with genomic DNA that could be detected by the DNA damage response as measured by histone H2AX and ATM kinase phosphorylation. We found that real-time DRAQ7 assay reported the death of cells cultured under variety of perturbation and the overall responses to cytotoxic agents and resulting pharmacological dose-response profiles were not affected by the growth of cancer cells in the presence DRAQ7. Moreover, we for the first time introduced a near real-time microflow cytometric assay incorporating both the DRAQ7 and a mitochondrial membrane potential (ΔΨm) sensitive probe TMRM. In this regard, we provide proof-of-concept evidence that such real-time labeling procedure can provide multiparameter and kinetic fingerprints of anticancer drug action.
MATERIALS AND METHODS
Culture and Treatments
The A549 cells were purchased from American Type Culture Collection (ATCC #CCL-185, Manassas, VA). The cells were cultured in Ham's F12K medium with 2 mM L-glutamine adjusted to contain 1.5 g/L sodium bicarbonate (ATCC) supplemented with 10% fetal bovine serum (ATCC). Dual-chambered slides (Nunc Lab-Tek II) were seeded with 105 cells/ml suspended in 2 ml medium per chamber. The cells were maintained in exponential phase of growth, then treated with 3 μM DRAQ7 (Biostatus Ltd, Shepshed, UK) for a designated time followed by a wash step with phosphate buffered salt solution (PBS) and fixed by transferring slides into Coplin jars containing 1% methanol-free formaldehyde (Polysciences, Warrington, PA) for 15 min. The slides were rinsed with PBS and kept at −20°C in 70% ethanol until required for further staining.
The U937 and THP1α cell lines were cultured in a complete Advanced RPMI 1640 culture medium supplemented with 5% FBS as described previously (5, 13, 14). For quantification of drug-induced cytotoxicity, 2.5 × 105 cells/ml of cells were suspended in 1 ml of medium and treated with cell cycle inhibitors and apoptosis inducers Staurosporine (STS; Life Technologies; 0.01–1 μM), Etoposide (ETO; Merck Millipore; 100–1000 μM), Actinomycin D (Act D; Merck Millipore, Billerica, MA; 0.001–1 μM), Cycloheximide (CHX; Merck Millipore; 100–1000 μM) and small-molecule BH3 mimetics ABT-737 (Selleckchem, Houston, TX; 0–7.5 μM), and TW37 (Selleckchem; 0–25 μM). This was followed by staining with cell impermeable DNA-stains at the end of experiment (end-point assay; 5 min at RT with 1 μg/ml of PI, 100 nM of SYTOX Red or 3 μM of DRAQ7). Alternatively, cells were cultured in the presence of 3 μM of DRAQ7 and regularly examined during the course of the experiment (dynamic assay) by flow cytometry as described below.
The HeLa cells expressing the eGFP-tagged linker histone H1 (15) were cultured on round coverslips (thickness 0.17 mm, 22 mm in diameter, Menzel-Gläser, Braunschweig, Germany) submerged in Petri dishes (40 mm in diameter). The treated cells were cultured and imaged in Dulbecco's MEM (Sigma) supplemented with 10% fetal bovine serum (Sigma), Phenol Red, and antibiotics (penicillin and streptomycin). DMEM was buffered for 5% CO2 in the atmosphere. The concentration of DRAQ7 was 3 μM unless otherwise stated. In cells cultured under suboptimal conditions, DMEM was substituted by DMEM/F-12 with 2% FBS. Concentration of DRAQ7 in these experiments was 3, 5, or 10 μM. Coverslips with cells were mounted in a custom-made steel holder and the sample was placed on a microscope stage and imaged using a Leica TCS SP5 confocal laser-scanning microscope, equipped with stage microincubator.
Detection of H2AX and ATM Phosphorylation
Slides (from above) were washed twice in PBS and the cells on the slides treated with 0.1% Triton X-100 (Sigma) in PBS for 15 min, and with a 1% (w/v) solution of bovine serum albumin (BSA; Sigma) in PBS for 30 min to suppress nonspecific antibody (Ab) binding. The cells were then incubated in a 100 μl volume of 1% BSA containing a 1:300 dilution of phospho-specific (Ser139) γH2AX mAb (BioLegend, San Diego, CA) or 1:200 dilution of phosphor-specific (Ser1981) ATM mAb (Merck Millipore) overnight at 4°C. The secondary AlexaFluor 488 fluorochrome (Life Technologies, at a dilution of 1:100) and incubated for 45 min at room temperature. Before measurement by laser scanning cytometry (LSC), the cells were counterstained with 2.8 μg/ml 4,6-diamidino-2-phenylindole (DAPI; Sigma) in PBS for 15 min. Each experiment was performed with an IgG control in which cells were labeled with the secondary AlexaFluor 488 Ab only; without primary antibody incubation to estimate the extent of nonspecific binding of the secondary antibody to the cells. Other details of cell labeling with the primary and secondary Ab have been previously described (16, 17).
Flow cytometry was performed using a Fishman-R microfluidic flow cytometer (On-chip biotechnologies Co., Tokyo, Japan) equipped with 473 and 640 nm solid-state lasers and integrated digital data processing (18–20). Logarithmic amplification scale using following configuration of band-pass (BP) filters was applied: (i) 473 nm excitation line: FL2 channel (585 BP for collection of PI and TMRM signals); (ii) 640 nm excitation line: FL4 channel (690 BP for collection of: DRAQ7 fluorescence signals). A typical run used a sample with 5,000–10,000 cells. Native data files were converted to FCS 3.0 standard using Fishman-R data converter.
Data and Statistical Analysis
Data analysis and presentation was performed using FCS Express 4 RUO Flow Cytometry (De Novo Software) software. Pharmacological dose response curves and IC50 values were plotted using GraphPad Prism 5.0 (GraphPad Software, La Jolla) software. Student's t and ANOVA tests were applied for comparison between groups with significance set at P < 0.01–0.05.
RESULTS AND DISCUSSION
DRAQ7 Does Not Cause Detectable Cytotoxicity (Acute or Long-Term)
Initial screening of U937 and THP1α cell lines with a dose range of 0.35–10 μM DRAQ7 was undertaken to assess the impact of the probe on cell viability. When analyzed using microflow cytometer, all cells continuously grown in the presence of a DRAQ7, did not show any increase in basal fluorescence after 24 h of exposure (Fig. 1A, upper panel). This feature suggested that DRAQ7 as a viability probe could be particularly suitable for kinetic assays. It was also in contrast to our previously reported data on the spectrally dissimilar plasma membrane permeability marker propidium iodide (PI; 1 μg/ml; excitation at 488 nm, emission at 585 nm) where continuous culture of cells with PI led to a profound increase in overall cell fluorescence, despite having no impact on cell viability (6). Most notably, the fluorescence levels of cells cultured with DRAQ7 were comparable to end-point staining with 1 μg/ml of PI (Fig. 1A, bottom panel). Both U937 and THP1α cells remained viable when grown with the DRAQ7 probe as evidenced by the lack of cell subpopulations exhibiting enhanced fluorescence signal intensity in the top right quadrants (Fig. 1A). Importantly, even a 20 μM concentration of DRAQ7 proved to be nontoxic to both cell lines for up to 72 h, also confirmed by the counterstaining with PI (Figs. 1B–1C). Moreover, no dysfunctions in the mitochondrial function were observed as assessed by the multiparameter labeling with ΔΨm sensitive probe tertamethylrhodamine methyl ester (TMRM; 200 nM) at 24, 48, and 72 h of incubation (Fig. 1D). Most cells featured intact and energized mitochondria indicated by TMRMhigh and DRAQ7neg (deemed live) subpopulation in the upper left quadrant shown in Figure 1D. Cell proliferation and cell cycle distribution were also not affected for up to 72 h growth of cells in the presence of 3–5 μM DRAQ7 (data not shown).
In search of early signs of the dye entry, we also cultured HeLa cells expressing the eGFP-tagged linker histone H1 in the continuous presence of DRAQ7 (3 μM) for up to 72 h. Cells were imaged using confocal microscopy to assess the dye penetration and signs of histone H1 dissociation. Even after 72 h of continuous incubation DRAQ7 was not detected in cell nuclei (Fig. 2A, Supporting Information Fig. 1). The absence of free intracellular DRAQ7 capable of reaching chromatin was also confirmed by the absence of detectable dissociation of eGFP-tagged histone H1 from DNA and subsequent chromatin aggregation. Next to demonstrate the suitability of DRAQ7 for real-time microscopic tracking and detection of cell death, we imaged HeLa cells grown under starvation conditions in culture medium supplemented with reduced serum concentration and 3 μM of DRAQ7 (Fig. 2B). Serum deprivation led to cell death detected by DRAQ7 incorporation within ∼72 h. The DRAQ7 provided thus a straightforward and real-time marker to track the demise of individual cells under stress conditions using time-lapse confocal microscopy.
DRAQ7 Does Not Induce DNA Damage Signaling
Next, we set to explore a possible DNA damage signaling in response to DRAQ7. For this purpose, A549 cells were maintained in cultures either untreated or treated with 3 μM of DRAQ7 for 4, 24, or 48 h. The induction of DNA damage signaling was measured as an increase in phosphorylation of H2AX on Ser139 (expression of γH2AX) and activation of ATM through its phosphorylation on Ser1981 (ATM-S1981P) (Figs. 3A and 3B). These phosphorylation events were detected with the use of respective phospho-specific antibodies, and cellular immunofluorescence measured by laser scanning cytometry (LSC). The exposure of cells to DRAQ7 caused no detectable induction of γH2AX regardless of duration of the treatment (Fig. 3). In fact, a minor decrease in expression of γH2AX was noted in the DRAQ7-treated cells. Also, no detectable effect of DRAQ7 was seen on the cell cycle distribution as revealed by the DNA content frequency histograms (Fig. 3A, insets).
The effect of DRAQ7 on activation of ATM was assessed in an independent set of cultures (Fig. 3B). Similar as in the case of γH2AX a decline of expression ATM-S1981P after 4 and 24 h was seen in cells treated with DRAQ7. However, a minor increase of ATM activation was noted in cells incubated with DRAQ7 for 48 h. The extent of this increase is much lower than in the case of other supravital probes, such as PI, DRAQ5 or Hoechst 33342 (21). In this set of cultures also no evidence of a change in cell cycle distribution in cells exposed to DRAQ7 that would be reflected by the DNA content frequency histograms (Fig. 3B, insets) was apparent.
DRAQ7 Can be Applied in Real-Time Viability Assays
On the basis of the above results, we next determined the applicability of DRAQ7 probe for the quantification of pharmacologically induced cell death based on real-time assay principle and cell sampling using innovative microfluidic flow cytometer (18, 22, 23). Potential interactions between fluorescent probes and cytotoxic drugs can introduce possible bias in estimating the toxic effects in dynamic real-time assays (6, 12, 22). To exclude this possibility, we therefore investigated a panel of cytotoxic drugs and apoptotic inducers on human hematopoietic tumor cell lines. Initially, THP1α cells were exposed to a panel of apoptosis inducers: staurosporine (STS), etoposide (ETO), actinomycin D (AD), and small-molecule Bcl-2 inhibitors ABT-737 and TW37 for up to 24 h, at the drug concentrations described in the Methods (Figs. 4A and 4B). The cells were grown in the continuous presence of 3 μM DRAQ7 and the indicated inducers of apoptosis (Figs. 4A and 4B). Following 24 h of growth cells were aspirated and immediately sampled in complete medium by the microfluidic flow cytometry at the end of the experiment without any additional washing steps. The results were compared with conventional end-point assays where cell samples were probed using DRAQ7 (3 μM), PI (1 μg/ml), and SYTOX Red (SXR; 100 nM) staining at the end of the experiment (Figs. 4A and 4B). Our results indicated that responses of tumor cells to various cytotoxic and pro-apoptotic stimuli were not affected by the growth in the continuous presence of DRAQ7 probe. The calculated IC50 values for STS were 0.05, 0.04, 0.05, and 0.05 for the real-time DRAQ7, end-point DRAQ7, PI, and SXR assays, respectively. The calculated IC50 values for ETO were 4.3, 4.5, 4.2, and 3.9 for the real-time DRAQ7, end-point DRAQ7, PI, and SXR assays, respectively (Fig. 4A and 4B). In both instances, the differences between IC50 values obtain with real-time and end-point protocols were considered not statistically significant (ANOVA; P < 0.01) and corresponded to Pearson linear correlation R2 values greater than 0.98.
These promising results prompted us to expand our bioassay development approach to incorporate a two-color real-time assay utilizing both the DRAQ7 and a ΔΨm sensitive probe TMRM (Fig. 4C) (1, 24). The latter probe is inert to many human cell lines at concentrations not exceeding 500 nM when grown in continuous presence of the probe for up to 72 h. This offers a unique capability to perform real-time monitoring of mitochondrial membrane inner membrane potential without any additional staining and washing steps. Accordingly, we have combined both probes to provide a multiparameter readout of caspase-dependent apoptosis induced by small-molecule drugs. On treatment with the apoptosis inducer AD, the real-time DRAQ7/TMRM assay provided similar estimates of the three cell populations: TMRMhigh/DRAQ7neg (deemed live), TMRMlow/DRAQ7neg (deemed early apoptotic), and TMRMlow/DRAQ7high (deemed late apoptotic/necrotic) as the data obtained using conventional end-point protocols (Fig. 4C; Pearson linear correlation R2 = 0.98–0.99). The calculated IC50 values based on cell viability estimates for the small molecule Bcl-2 inhibitor ABT-737 were 2.26 and 2.10 using the DRAQ7/TMRM real-time vs. end-point assays, respectively (Fig. 4D). The calculated IC50 values based on cell viability estimates for the small molecule Bcl-2 inhibitor TW37 were 1.79 and 1.46 using the DRAQ7/TMRM real-time vs. end-point assays, respectively (Fig. 4E). In both instances the differences between IC50 values obtained with real-time vs. end-point protocols were considered not statistically significant (ANOVA; P < 0.01 for ABT737 and P < 0.05 for TW37) and corresponded to Pearson linear correlation R2 values greater than 0.98 (Figs. 4D and 4E).
The application of inert DRAQ7 and TMRM probes provided us also with an opportunity to investigate the kinetic properties of small-molecule Bcl-2 inhibitors such as TW37. Figure 4F depicts an on-demand, near-real-time microflow cytometric assay where cells were continuously grown in the presence of both probes on a chip-based device and then sampled (10–20 μl volume) every 5 minutes to provide the kinetic profile of drug action. Most notably, we were able to demonstrate that even though the dissipation of ΔΨm occurs within 10–15 min following challenge with BH3 mimetic TW37, the gradual loss of plasma integrity occurs only after 8 h of stimulation, while nearly 74% of cells can still be considered TMRMlow/DRAQ7neg after 24 h (Fig. 4F). This indicates that THP1 cells were able to compensate for the ΔΨm loss for a considerable amount of time before committing to complete cell demise. These results reinforce our earlier studies on small-molecule BH3 mimetic HA14-1 in follicular lymphoma cells indicating a relatively transient character of Δψm loss and potential involvement of compensatory mechanisms (1, 2, 25). The latter can be based on the multiple switches between slow but also variable in time decision-making processes, involving gradual accumulation of pro-apoptotic molecules (e.g., tBid, Bax complexes), followed by snap-action rapid activation of caspases, and variable kinetics and specificity of proteolytic degradation of endogenous targets as reported recently (7–9).
In search for the noninvasive fluorescent probes capable of long-term monitoring of cell death in real-time, we developed and evaluated a new anthracycline derivative, DRAQ7. DRAQ7 being a far-red fluorescent DNA dye exhibits convenient spectral properties that allow for multiplexing with makers such as GFP, FITC, and Cy3. Once the membrane integrity is compromised, DRAQ7 binds readily to nuclear DNA with high affinity and reports cell death by strong far-red fluorescence. Our data indicated that the growth, cell cycle distribution, and proliferation of several tumor cell lines were unaffected by the continuous presence of DRAQ7 for up to 3 days of culture. Also, unlike other DNA supravital probes (21), continuous incubation with DRAQ7 led to no evidence of either DNA damage, replication stress, or induction of cell senescence (26).
Collectively, data lend a strong support to real-time pharmacological profiling using DRAQ7 (6, 12). In such assays, the fluorescence markers, applied supravitally, should have a minimal effect on structure, function, and survival of the cells being studied. Lack of any noticeable influence on cell responses to or interaction with cytotoxic agents during growth in the presence of DRAQ7 allows for a direct adaptation for the automated and real-time pharmacological profiling of anticancer drugs. Importantly, a substantial reduction of sample processing steps and avoidance of washing protocols achieved with such a kinetic protocol is important for the preservation and retrieval of fragile cell subpopulations. We conclude that the current data provide a foundation for the development of a new generation of cell viability assays with vast applications in both flow and imaging cytometry.