SEARCH

SEARCH BY CITATION

Keywords:

  • SYTO;
  • apoptosis;
  • oncosis;
  • Δψm dissipation;
  • multiparameter flow cytometry

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS AND DISCUSSION
  5. CONCLUDING REMARKS
  6. Acknowledgements
  7. LITERATURE CITED

Background:

SYTO probes are gaining momentum as reliable and easy to use markers of apoptotic cell death, but the phenomenon underlying reduced SYTO fluorescence in apoptotic cells as compared with normal cells is still not fully elucidated. Herein, we attempt to provide further insights into mechanisms of reduced SYTO16 fluorescence during apoptosis.

Methods:

Human follicular lymphoma cell lines were subjected to diverse apoptotic and oncotic stimuli with subsequent multiparametric flow cytometric and fluorescence imaging analysis. SYTO green (SYTO11-16), TMRM, PI, 7AAD, and Hoechst 33342 probes were applied for multivariate analysis of temporal sequence of apoptotic events. Sorting of cells differing in the level of SYTO16 fluorescence and subsequent characterization of obtained subpopulations were also performed.

Results:

Loss of SYTO16 fluorescence (SYTOlow/PI+ events) has been observed in cells exposed to oncotic stimuli, whereas SYTOhigh/PI+ events did not prevail at any treatment scenario. We tracked similarities and discrepancies between SYTO16 and TMRM probes. Often, SYTO16 and TMRM exhibited the same staining profiles, as loss of their fluorescence was detected in a single cell population. However, both mitochondrial uncoupler FCCP and a small-molecule Bcl-2 inhibitor, HA14-1, appeared to induce distinct staining profiles of SYTO16 and TMRM, with the decrease in TMRM fluorescence preceding the loss of SYTO16 fluorescence. Importantly, in both cases (FCCP and HA14-1) the decrease of SYTO16 fluorescence was blocked by pharmacological inhibition of caspases (with z-VAD-fmk).

Conclusions:

The data demonstrate that loss of SYTO16 is caspase-dependent, as is not a mere indicator of Δψm dissipation. Commonly observed similarities between SYTO and TMRM may stem from the fast kinetics of apoptotic events once cell death is initiated. © 2007 International Society for Analytical Cytology.

Apoptosis is a complex, finely controlled, and evolutionary conserved process of great relevance in tissue homeostasis, development, and pathogenesis. The majority of its features can be examined by flow and image cytometry (1, 2) that have proven to be reliable and flexible platforms in a wide spectrum of research and clinical applications (3–5). Analysis of cell demise modes using fluorescently labeled functional probes permits rapid acquisition of quantitative data allowing further biochemical and molecular studies. Multiparameter data obtained by flow and image cytometry permits parallel correlation of different cellular events at a time on a cell-by-cell basis (6–8). Development of novel functional probes and thorough understanding of the exact mechanisms underlying properties of existing ones are of utmost importance for the progress in cell necrobiology (1). This is particularly relevant in view of the growing appreciation of the multitude of cell demise modes, and the proposal to use the term apoptosis only to describe caspase-dependent cell death with stereotypical morphological changes (especially compact chromatic condensation) (9, 10).

Following on from Fray's publication back in 1995, cell permeable SYTO probes have been gaining momentum as reliable discriminators of live, apoptotic, and dead cells, offering cost-effective and easy to perform assays for tracking apoptosis in cultured and primary cells, and proving amenable for the development of multiparameter flow cytometry assays (11–16). Dyes from SYTO families (blue-, green-, orange-, and red-fluorescent) are generally permeable (although at varying levels) to all cell membranes, including bacterial. Still, they have been suggested to differ with respect to cell affinity or excitation and emission spectra in various cell models. They exhibit very low inherent fluorescence, with strong enhancement upon binding to DNA/RNA. SYTO-stained eukaryotic cells display both nuclear and diffuse cytoplasmic staining pattern (17, 18), the latter shown to be abolished after formaldehyde fixation (19). DNA binding properties of SYTO dyes have been exploited to visualize characteristic nuclear morphology during apoptosis or even to assess level of DNA synthesis after exposure to endocrine disrupting chemicals (18, 20, 21). It should be noted, however, that SYTO probes are not exclusive DNA stains, and thus several of them have been successfully applied to visualize translocation of RNA granules in neurons and allow differential discrimination of nuclear/mitochondrial DNA from cytoplasmic RNA using two-photon lifetime imaging (17, 22). Importantly recent reports suggest the reliability of some SYTO dyes as effective substrates in quantitative P-glycoprotein (P-gp) function assessment (14, 19, 23).

Although mounting evidence show higher sensitivity of SYTO probes as compared with Annexin V based assays (12, 14, 24), the exact mechanisms underlying SYTO differential staining of apoptotic and viable cells still remain ambiguous. The most widely embraced idea is the self-quenching of SYTO dye molecules following changes in interprobe proximity or the decrease in SYTO binding sites as the chromatin condensation and RNA degradation advance in the process of apoptosis (11, 13, 17). Poot et al. reported also that cells treated with camptothecin and subsequently double-stained with SYTO11 and mitochondrial dye CMXRos separate into two sub-populations, with concomitant loss of fluorescence of both dyes (13). Based on the observation that most of the SYTO dyes contain one net positive charge at neutral pH (19) and therefore may resemble mitochondrial membrane permeabilisation (MMP)-sensitive probes, it has been also suggested that alterations in binding of SYTO to mitochondrial DNA or decrease in its Δψm-driven mitochondrial uptake may contribute to the overall reduction of SYTO fluorescence in apoptotic cells (17, 24). These noteworthy but so far scarce data clearly require further studies with aim to elucidate similarities and discrepancies between SYTO and MMP-sensitive probes.

In the present survey we show that SYTO dyes from the green family (SYTO11-16) have overlapping staining characteristics, adding to the findings reported by Poot et al. (13). Applying SYTO16 that is commonly used for the assessment of apoptosis, we compared its staining features to those of Hoechst 33342 and TMRM (lipophilic cation sensitive to MMP). Of utmost importance we report herein certain novel differences in SYTO16 and TMRM staining profiles, revealed upon short-time treatment of follicular lymphoma (FL) cells with HA14-1 (a small molecule Bcl-2 inhibitor), or FCCP (mitochondrial uncoupler).

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS AND DISCUSSION
  5. CONCLUDING REMARKS
  6. Acknowledgements
  7. LITERATURE CITED

Cells

The origin and characteristics of human FL cell lines HF1A3 and HF4.9 have been previously described (25, 26). HF1A3 and HF4.9 cells were cultured in 100 ml (75 cm2) cell culture flasks (Sarstedt, Newton, NC) in RPMI 1640 medium (Cambrex Bio Science, Verviers, Belgium) supplemented with 5% heat-inactivated FCS (EuroClone, Pero, Italy), 2 mM L-glutamine (Fluka Chemie, Buch, Switzerland), 200 μg/μl streptomycin (Cambrex), 240 IU/ml penicilin (Cambrex), 10 mM HEPES buffer (Cambrex), 0.1 mM NAA (Cambrex), 1 mM Na-pyruvate (Cambrex), and 20 μM 2-mercaptoethanol (Fluka Chemie). All cell cultures were maintained at 37°C in a 5% CO2 humidified atmosphere. During experiment cells were always in asynchronous and exponential phase of their growth.

Cell Death Induction

To induce apoptosis, HF1A3 and HF4.9 cell lines were seeded on the 48-well polystyrene cell culture plates (Corning, NY), treated with various concentrations of CD95 cross-linking antibody (clone CH11; Upstate, NY; 1–1,000 ng/ml), dexamethasone (Dex; Sigma; 1–1,000 nM), cycloheximide (CHX; Sigma; 0–10 μg/ml), or a small molecule Bcl-2 inhibitor HA14-1 (Alexis Biochemicals, San Diego, CA; 0–15 μM), and harvested at different times as indicated. To inhibit caspase-dependent cell death induced by CD95 cross-linking and HA14-1 in HF1A3 and HF4.9 cells, respectively, cells were pretreated for 2 h with a pan-caspase inhibitor z-VAD-fmk (Calbiochem, Cambridge, MA) as described earlier (25, 27). For cell sorting, HF4.9 cells were seeded on 24-well polystyrene plates (Corning) and treated with 1 μM Dex for 24 h.

Primary necrosis was induced in HF1A3 cells by hyperthermia (56°C; 5 min or 46°C up to 6 h), treatment with 1% sodium azide (NaN3; Sigma) for up to 4 h (3), or 3% H2O2 for up to 45 min. The density of the seeded cells was less than 5 × 105 cells/ml at the time of all treatments.

SYTO Probes Staining

All probes from SYTO green I family (SYTO11-16) were procured from Molecular Probes (Eugene, OR). SYTO reagents were initially diluted in DMSO (Sigma) to achieve stock concentrations of 1 mM. Aliquots of probes were then stored at −20°C in the dark. After the treatment with apoptosis- or oncosis- inducing agents, cells were collected, rinsed with PBS to remove phenol red containing RPMI medium, and re-suspended in the 100 μl PBS containing selected SYTO dyes and plasma membrane permeability marker, propidium iodide (PI, Sigma, 5 μg/ml), as described earlier (25, 26). Final concentrations of SYTO dyes were as follows: 100 nM (SYTO11 and SYTO13), 200 nM (SYTO12), 250 nM (SYTO16), and 500 nM (SYTO14 and SYTO15). After 20 min incubation at RT in the dark, 500 μl of PBS containing 2% FBS (EuroClone) was added and cells were immediately subjected to the flow cytometric analysis. Above mentioned assay conditions were determined after extensive dose and time-course dye-loading optimization studies. Interestingly, we did not notice substantial differences in the efficiency of SYTO staining when dye-loading was performed in PBS, PBS supplemented with 2% FBS, HEPES or complete RPMI 1640 medium or at 0.25–4 × 106 cells/staining mixture cell density. Moreover, 10–15 min loading at 37°C was as effective as longer incubation times (up to 60 min tested) at 37°C, and as 20 min loading at RT.

As recent reports indicate that SYTO13 and 16 are effective substrates of P-gp efflux pomp, we checked the P-gp status in HF1A3 and HF4.9 cells using cyclosporin A (Sigma), verapamil (Alexis Biochemicals, Lausen, Switzerland), and probenecid (Alexis Biochemicals). P-gp-attributable activity was not detected in any of the cell lines.

Fractional DNA Content Analysis

After cell sorting, 1 × 106 sorted HF4.9 cells from respective subpopulations were collected, fixed in 70% ethanol, and stored at −20°C for at least 2 h. Next, cells were collected, washed with PBS, stained with PI (16 μg/ml) in the presence of RNase A (300 μg/ml), for 1 h at 37°C, and analyzed immediately to assess the percentage of sub-G1 subpopulation.

Fluorescence and Confocal Microscopy

For fluorescence microscopy analysis, 1 × 106 HF1A3/HF4.9 control cells, or cells treated with 1 μM Dex, were rinsed with culture media and stained in RPMI medium with 250 nM SYTO16 and 1.5 μg/ml Hoechst 33342 (Alexis Biochemicals, San Diego) probes for 20 min at RT in dark. Following staining, cells were wet-mounted and imaged by an Olympus AX70 Provis microscope equipped with FVII digital camera (Olympus, Tokyo, Japan) and 100 W Hg burner as an epifluorescent light source. Air objective lenses (40× and 60×) and appropriate fluorescence mirror units (Ex/Em for: SYTO16 490/520 nm, Hoechst 33342 355/465 nm) were applied for obtaining respective images. Cell images were captured by MicroSuite™ FIVE imaging software (Olympus) running under Windows XP Professional (Microsoft, Redmond, WA) and at least 150 cells were analyzed from each sample. Subsequent image analysis was carried out using ImageJ open source platform (developed at the National Institutes of Health, Bethesda and available freely at http://rsb.info.nih.gov/ij/ web page) running under Windows XP Professional (Microsoft).

For SYTO/TMRM co-localization studies, 1 × 106 HF1A3 cells were stained with 250 nM SYTO16 and 150 nM TMRM (Molecular Probes) in PBS for 15 min at 37°C. Following staining, cells were imaged using Nikon Eclipse inverted microscope equipped with UltraVIEW confocal scanning system (Perkin Elmer). Cell images were captured and analyzed by Perkin Elmer Imaging Suite version 5.5 software running under Windows 2000 (Microsoft).

Multiparameter SYTO16/TMRM/7-AAD Labeling

After induction of apoptosis, cells were collected, washed twice with PBS to remove phenol-red containing RPMI media, and stained in 100 μl PBS containing SYTO16 (250 nM) and TMRM (150 nM) for 15 min at 37°C in darkness. Next, samples were briefly cooled on ice to the RT, and plasma membrane permeability marker, 7-AAD (Molecular Probes), was added to a final concentration of 5 μg/ml. Samples were subsequently incubated for 5 min at RT in darkness. Finally 500 μl of PBS containing 2% FBS (EuroClone) was added and cells were immediately analyzed on the flow cytometer.

Flow Cytometry and Cell Sorting

Multiparametric flow cytometry was performed on a FACScan (Becton Dickinson, San Jose, CA) analyzer, equipped with 15 mW, 488 nm Argon-ion laser as a main excitation source and a standard setting of band-pass (BP) filters. For multiparametric (two-color and tri-color stainings) logarithmic fluorescence signals were collected using following filters: SYTO dyes (FL1 530/30 BP), TMRM (FL2 585/42 BP), PI (FL2 585/42 BP), 7-AAD (FL3 650 LP). As described by others (14) due to the strong spectral overlap of SYTO16 (in FL2 and FL3) special attention has been made to compensate both channels. For evaluation of fractional DNA content propidium iodide fluorescence was collected using linear amplification scale and 585/42 BP filter.

Acquisition of 10,000 events per each sample in 1,024 channels resolution scale was done using CellQuest version3.3 software (Becton Dickinson) running under MacOS 8.1 operating system (Apple, Cupertino, CA). Data analysis and presentation was performed using offline Summit version 3.1 software (Dako Cytomation, Fort Collins, CO) and WinMDI version 2.8 developed by Dr J. Trotter (freely available at http://facs.scripps.edu/software.html), both running under Windows XP Professional operating system (Microsoft).

Sorting of respective subpopulations was performed on EPICS Elite ESP (Coulter, Miami, FL) cell sorter, equipped with 15 mW air-cooled Argon-ion laser operating at 488 nm excitation line. Following BP filter configuration was applied: 525 BP (SYTO16), 610 BP (PI). Sort parameters and data acquisition were controlled by EPICS (R) Elite version 4.02 software (Coulter) running under DOS 3.07 operating system (Microsoft). Sort gates were drawn on bivariate FSC/SYTO16 dot-plots around three apparent subpopulations: SYTO16high (deemed viable cells), SYTO16dim (deemed early apoptotic cells), and SYTO16low (deemed late apoptotic/necrotic cells). These subpopulations correlated entirely to SYTOhigh/PI, SYTOdim/PI, and SYTOlow/PI+, respectively, in accordance with others (23, 24). Sorting was carried out using 3× Sort-Sense Quartz Flow Cell (Coulter) with 100 μm diameter jetting orifice and crystal frequency set at 22 kHz. To avoid destruction of sorted apoptotic subpopulations system pressure operated at 10 PSI and sort rates did not exceeded 3,000 cells/s. All sorts were done using Purity1Recovery2 sort mode (Coulter), which allowed achieving ≥95% cell purity for each subpopulation. For subsequent fractional DNA content analysis at least 1 × 106 cells and for FSC/SSC analysis and sort purity check 5 × 105 cells were sorted into cooled RPMI 1640 medium containing 20% FBS. Each experiment consisted of at least three independent sorts.

Statistical Analysis

Results shown on dot plots and photographs are representatives of at least four independent experiments and SD values between experiments did not exceed ± 5%. Pearson correlation analysis was performed using Excel 2000 software (Microsoft).

RESULTS AND DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS AND DISCUSSION
  5. CONCLUDING REMARKS
  6. Acknowledgements
  7. LITERATURE CITED

Comparative SYTO Family Sstaining

FL HF1A3 cells were cultured in the presence of Dex (1–1,000 nM) or anti-CD95 agonistic antibody (1–100 ng/ml) for up to 48 h, followed by an immediate SYTO/PI staining. As inferred from the bivariate SYTO/PI plots, SYTO11, 13, 14, and 16 allowed a lucid discrimination of viable (SYTOhigh/PI), apoptotic (SYTOdim/PI) and late apoptotic/necrotic (SYTOlow/PI+) subpopulations, whereas SYTO12 and 15 could evidently distinguish the same only on FSC/SYTO bivariate plots (Figs. 1A and 1B). The discrimination of viable, apoptotic and late apoptotic/necrotic cells by SYTO probes was confirmed after back-gating each subpopulation onto FSC/SSC plots and FSC/PI plots (not shown), and the number of SYTOdim cells after staining with any of SYTO11, 13, 14, and 16 dyes correlated well (r2 ≥ 0.98; Pearson and Lee linear correlation test). Similarly, the number of apoptotic (FSClow/SYTOhigh) cells detected by SYTO12 and SYTO15 correlated to a large extent with the results obtained with the use of other SYTO stains (Figs. 1A and 1B, and not shown data for CHX). Still, the apoptotic subpopulation detected by FSClow/SYTOhigh pattern was slightly underestimated (6–8%) compared with that assessed by SYTO11, 13, 14, and 16 (Figs. 1A and 1B), which may be a result of the overlap between FSC-defined subpopulations. In our model the SSC-defined subpopulations were even less clearly separated than the FSC-defined.

thumbnail image

Figure 1. Discrimination of live, apoptotic and late apoptotic/necrotic cells by SYTO probes. Untreated (control), stimulated with anti-CD95 mAb (100 ng/ml) or dexamethasone (Dex; 1 μM) HF1A3 cells were collected after 24 h, washed and stained either with SYTO16/PI (A) or SYTO12/PI (B), as described in Materials and Methods. Note that SYTO16 could distinguish life (SYTOhigh/PI), apoptotic (SYTOdim/ PI) and late apoptotic/necrotic (SYTOlow/PI+) cells on SYTO vs PI bivariate plots, whereas SYTO12 (and SYTO15, not shown) could only distinguish life, apoptotic and late apoptotic/necrotic cells only on FSC vs SYTO bivariate plots. Cell debris showing extremely low FSC/SSC values were excluded electronically. Plots are representative of four independent experiments. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

Download figure to PowerPoint

Lack of the decrease in SYTO15 fluorescence to dim values in early apoptotic cells is in general accordance with study by Poot et al. (13). To our knowledge, however, this is the first report showing the clear resolution of viable, apoptotic and late apoptotic/necrotic populations based on the bivariate FSC vs SYTO12/15 plots. It is possible, though, that in other cell models the contribution of FSC and SSC examination may differ from that observed herein, depending on the degree of changes in FSC/SSC characteristics during apoptosis.

Apoptotic vs Oncotic SYTO Staining Profile

It has been previously tested by others that loss of SYTO16 fluorescence can be considered as a truly apoptotic feature (12–14, 23, 24), but to the best of our knowledge there is very limited data available to show SYTO behavior in cells exposed to oncotic stimuli. We observed that in the presence of relatively harsh oncotic stimuli (4 h-incubation with 1% NaN3, or 5 min at 56°C) the majority of cells was SYTOlow/PI+, representative of cells with a compromised plasma membrane, whereas the number of SYTOdim/PI events did not significantly increase above background levels (Fig. 2A). The number of cells exhibiting SYTOhigh/PI+ staining pattern, suggested by some reports as an indicator of primary necrosis (14, 19), increased slightly over the background levels in NaN3-treated cells, but not upon heat shock (Fig. 2A). Acknowledging the possibility that the decrease in SYTO fluorescence may follow an initial loss of membrane impermeability to PI, we went on to examine whether an increase in the number SYTOhigh/PI+ events can be revealed upon exposure to milder oncotic stimuli (such as short treatment with 1% NaN3 or 3% H2O2). Clearly, it was always the SYTOlow/PI+ population that prevailed, although a clear increase in the number of SYTOhigh/PI+ events was noticeable, especially in NaN3-treated cells (Figs. 2B and 2C). Similar SYTO behavior was observed when cells were exposed to 46°C (not depicted). To the best of our knowledge this is the first report showing loss of SYTO fluorescence in response to oncotic stimuli.

thumbnail image

Figure 2. Characteristic pattern of SYTO16 staining upon induction of apoptosis and oncosis. (A) HF1A3 cells were exposed to 100 ng/ml anti-CD95 mAb (apoptotic stimuli) or two oncotic triggers: sodium azide (1% NaN3 for 4 h) or heat-shock (+56°C for 5 min). (B) HF4.9 and HF1A3 cells were exposed to 1% NaN3 and 3% H2O2, respectively, for the time indicated. After the treatment, cells were washed and stained with SYTO16/PI. Note the lack of the clearly separated SYTOdim/PI subpopulation in cells exposed to oncotic stress, as compared with cells undergoing apoptosis. Cell debris showing extremely low FSC/SSC values were excluded electronically. Plots are representative of 2–4 independent experiments. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

Download figure to PowerPoint

Cross-Validation with other Apoptosis-Detecting Assays

To assess the relationship between SYTO16 staining, fractional DNA content and light scatter characteristics, HF4.9 cells were treated with Dex, stained with SYTO16 and PI, and SYTOhigh (deemed as viable), SYTOdim (deemed as early apoptotic), and SYTOlow (deemed as late apoptotic/necrotic) cells were sorted as described in Materials and Methods. The analysis of fractional DNA content in the three sub-populations defined by SYTO16 revealed that the majority of SYTOdim cells exhibit oligonucleosomal DNA fragmentation immediately after sort (Fig. 3A), in line with a high correlation between SYTOdim and sub-G1 values stated by Poot et al. (13). The majority of SYTOdim cells had condensed chromatin, nuclear fragmentation (as revealed by Hoechst 33342), and characteristic apoptotic morphology. In contrary, SYTO16high cells did not present apoptotic enhancement of nuclear Hoechst 33342 staining (Figs. 3A and 3B), indicating excellent agreement between the two dyes for identification of apoptotic cells. In accord with Sparrow and Tippett (2005), there was a progressive decrease in FSC values (indicative of cell shrinkage) and broadening of SSC values (indicative of increased cell granularity) as the fluorescence of SYTO16 was declining (Fig. 3B). Our observations are somehow distinct, however, from the reports showing that morphological apoptotic features in SYTOdim cells are apparent only after longer post-sort cultivation times, or that SYTO16 detects early apoptotic changes ahead of morphological features of apoptosis (14, 24). Still, we can not exclude the possibility that the nuclear fragmentation and the decrease in DNA content to sub-G1 levels continue to build-up during the period of cell preparation and sorting, biasing prompt cross-calibration with SYTO staining.

thumbnail image

Figure 3. Cross-validation of SYTO16 and other assays for detection of apoptosis. (A) HF4.9 cells were treated with 1 μM Dex for 24 h, stained with SYTO16/PI, and respective viable (SYTOhigh), apoptotic (SYTOdim) and late apoptotic/necrotic (SYTOlow) subpopulations were sorted as described in Materials and Methods. After sorting, cells were immediately analyzed for nuclear features of apoptosis (Hoechst 33342 staining), fractional DNA content (propidium iodide staining), and FSC/SSC spread characteristics. Note the increase in nuclear fragmentation, sub-G1 levels, and characteristic for apoptosis light scattering changes in sorted SYTOdim cell subpopulations. Insets in histograms showing DNA content represent SYTO16/PI bivariate plots immediately after cell sorting. All results are representative of at least three independent sorts. (B) HF1A3 cells untreated (control) or stimulated with 10 ng/ml anti-CD95 mAb were collected after 24 h and stained with SYTO16 and Hoechst 33342 probes as described in Materials and Methods. Cells with reduced SYTO fluorescence (marked with arrowheads) exhibited extensive nuclear and morphological features of apoptosis in comparison to SYTO bright cells. Similar results were obtained after stimulation of HF1A3 and HF4.9 cells with 1 μM Dex and 5 μg/ml CHX (not shown). Results are representative of at least three independent experiments. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

Download figure to PowerPoint

Similitude of SYTO16 and TMRM Staining in Different Exposure Scenarios

The possibility to concomitantly stain cells with SYTO and TMRM probes at 37°C (see Materials and Methods) has been exploited to compare staining characteristics of SYTO16 and TMRM by multiparametric flow cytometry analysis. To this aim, HF1A3 cells were treated with Dex- or agonistic anti-CD95 mAb, and subsequently stained with SYTO16, TMRM and plasma membrane permeability marker 7-AAD. The sub-populations defined by TMRM and SYTO16 in both 7-AAD and 7-AAD+ cells are presented in Figure 4. The bivariate plots clearly show two cell sub-populations, one with decrease in both TMRM and SYTO16 fluorescence (apoptotic cells), and the other without changes in fluorescence of either of the probes (live cells) (Fig. 4A). This is generally in accordance with previous reports (13, 24), although for the first time we have applied triplicate staining to allow exclusion of cells with markedly distorted cell membrane permeability. As relatively short time between the onset of apoptosis and cell dismantling impedes the successful tracking of the kinetics of apoptotic events, we hypothesized that differences between staining properties of SYTO16 and TMRM can disclose upon shorter treatment scenarios. However, even the shortest exposure of HF1A3 or HF4.9 cells to different inducers of apoptosis (see Fig. 4B depicting CHX-treated HF4.9 cells; results for other stimuli not shown) triggered a concomitant decrease in fluorescence of both probes, and no intermediate events were observed between apoptotic and normal cells. In line, treatment of HF1A3 cells with a broad range of doses of three apoptotic stimuli (Dex, anti-CD95 mAb, CHX) for 6–72 h, followed by triplicate SYTO16/TMRM/7-AAD staining, revealed perfect SYTO16 and TMRM cross-correlation in detecting apoptosis (Fig. 4C).

thumbnail image

Figure 4. Similarities between SYTO16 and TMRM staining patterns upon induction of apoptosis. (A) HF1A3 cells untreated (control) or stimulated with Dex (1 μM) were collected after 24 h and stained with SYTO16/TMRM/7-AAD probes as described in Materials and Methods. Bivariate SYTO16 vs TMRM plots indicate concomitant and progressive loss of SYTO and TMRM fluorescence. (B) HF4.9 cells untreated (control) or stimulated with CHX 5 μg/ml were collected after 0–20 h and immediately stained with SYTO16/TMRM/7-AAD dyes. Bivariate plots represent cells gated from 7-AAD and 7-AAD+ populations. Note the parallel loss of SYTO and TMRM fluorescence (SYTOdim/TMRMlow) without any intermediate SYTOdim/TMRMhigh or SYTOhigh/TMRMlow populations. (C) HF1A3 cells were treated with Dex (0–1,000 nM), anti-CD95 (0–1,000 ng/ml), and CHX (0–10 μg/ml) for 6–72 h. After stimulation cells were collected and stained with SYTO16/TMRM/7-AAD dyes. Note the excellent correlation (r2 = 0.998) between SYTO16low/7-AAD and TMRMlow/7-AAD in detection of apoptotic cells. Cell debris showing extremely low FSC/SSC values were excluded electronically from each plot. Results are representative of four independent experiments.

Download figure to PowerPoint

To substantiate our findings from multiparameter flow cytometry, HF1A3 cells were examined by fluorescence microscopy after tri-color staining with SYTO16/TMRM/Hoechst 33342 probes, as described in Materials and Methods. Unstimulated stained cells exhibited bright green SYTO16 and punctuate red TMRM fluorescence (compatible with the generally accepted TMRM staining pattern of energized mitochondria) (Fig. 5A). The same cells showed uniform, dim, nuclear staining with Hoechst 33342, and lack of any morphological features of apoptosis (not shown). Upon treatment with Dex, dying cells displayed loss of TMRM staining (considered as a marker of Δψm loss) and reduction of SYTO fluorescence (representative of dim subpopulation previously detected by flow cytometry) (Fig. 5A). All SYTOdim/TMRM cells showed characteristic for apoptosis enhancement of Hoechst 33342 fluorescence with apparent nuclear fragmentation and cell shrinkage when seen in bright-field (data not presented).

thumbnail image

Figure 5. Staining patterns of SYTO16, TMRM, and Hoechst 33342 revealed by fluorescence microscopy examination. (A) HF1A3 cells, untreated (control) or stimulated with 1 μM Dex, were collected after 24 h and stained with SYTO16, TMRM and Hoechst 33342 probes as described in Materials and Methods. Note that cells with complete loss of TMRM fluorescence exhibited reduced SYTO16 fluorescence and extensive nuclear features of apoptosis (cells marked with arrowheads) in comparison to TMRM bright cells. Fluorescence intensity plots (lower panel) were generated using ImageJ software from photographs of Dex-treated cells. Results are representative of at least three independent experiments. (B) HF1A3 cells were stained with SYTO16 and TMRM, followed by confocal microscopy analysis, as described under Materials and Methods. Yellow spots on the composite image indicate points of co-localization, whereas green spots indicate separate staining of SYTO16. Selected green spots are marked with arrows. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

Download figure to PowerPoint

Differences Between SYTO16 and TMRM Staining Profiles

Guided by the results from the preliminary wide-field fluorescent microscopy examination, we performed further confocal microscopy analysis to determine the extent of SYTO16/TMRM co-localization, and observed that in fact SYTO16 co-localizes with TMRM only partially (Fig. 5B). Quantitative analysis revealed that 21–45% of SYTO16 stains independently of TMRM. We therefore presupposed that at least under some conditions it should be possible to observe different staining profiles of SYTO vs TMRM.

Indeed, when HF4.9 cells were exposed to mitochondrial uncoupler FCCP for 15 min, and subsequently stained with SYTO16/TMRM/7-AAD, we observed a decrease in TMRM fluorescence, whereas fluorescence of SYTO16 was preserved almost entirely (Fig. 6A). Longer treatment with FCCP led to a progressive loss of SYTO16 fluorescence, followed by cell membrane permeabilization (Fig. 6A). Interestingly, upon treatment with a pan-caspase inhibitor z-VAD-fmk 1 h prior to FCCP administration a complete blockage of the decrease in SYTO16 fluorescence was observed, whereas loss of TMRM fluorescence was entirely preserved. Of notice, z-VAD-fmk largely inhibited the mortality, as assessed by cell membrane permeability to 7-AAD (Fig. 6A).

thumbnail image

Figure 6. Differences between SYTO16 and TMRM staining profiles. (A) HF4.9 cells were left unstimulated or treated with FCCP (50 μM) for the time indicated. Note that after 15 min only TMRM fluorescence was reduced, whereas SYTO16 fluorescence was entirely preserved. Longer treatment times led to a progressive loss of SYTO fluorescence and increased cell membrane permeability to 7AAD, both of which were inhibited upon pre-treatment with z-VAD-fmk (20 μM). (B) HF4.9 cells were treated with HA14-1 (7.5 μM) for the time indicated. The loss of TMRM was initially (2 h) accompanied by almost entirely preserved SYTO16 fluorescence. The latter decreased gradually during longer exposure times. (C) HF4.9 cells were treated with HA14-1 alone or with pan-caspase inhibitor zVAD-fmk (100 μM) for 4 h. Note that inhibition of caspases restored SYTO16 fluorescence to values detected in viable cells, without interfering with loss of TMRM fluorescence. Cell debris showing extremely low FSC/SSC values were excluded electronically from each plot. Results are representative of two to four independent experiments.

Download figure to PowerPoint

To determine the staining characteristics of both TMRM and SYTO16 upon stimulation of cells with another mitochondria-targeting compound, we challenged HF4.9 cells with a small molecule Bcl-2 inhibitor HA14-1, previously shown by us to induce apoptosis in FL cells in a single-agent treatment scenario, with caspase activation occurring exclusively as a consequence of the mitochondrial breach (25, 27). Remarkably, the loss of TMRM fluorescence upon HA14-1 treatment appeared to precede the loss of SYTO16 fluorescence, the latter trailing to dim over slightly longer exposure time (Fig. 6B). Lately, we have also observed that z-VAD-fmk, a pan-caspase inhibitor, restored viability, but did not entirely preserve Δψm, following HA14-1 treatment (27). Herein, we re-assessed this finding using triplicate SYTO16/TMRM/7-AAD staining. Indeed, we observed that z-VAD-fmk did not protect against TMRM loss, while fully restoring high fluorescence of SYTO16 (Fig. 6C). Of note, z-VAD-fmk was also able to restore considerable portion of SYTOdim subpopulation (spontaneous apoptosis) in control samples (Fig. 6C).

Overall, our results clearly indicate the dichotomy between the kinetics of TMRM and SYTO16 staining characteristics, and show that the blockade of caspase activation prevents loss of SYTO florescence in cells with mitochondrial rupture. This is an important finding, as a plethora of stimuli that trigger cell death, and lesions affecting different organelles within the cell (e.g. nuclei, endoplasmic reticulum or lysosomes), converge on mitochondria.

CONCLUDING REMARKS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS AND DISCUSSION
  5. CONCLUDING REMARKS
  6. Acknowledgements
  7. LITERATURE CITED

As a growing number of researchers applies SYTO dyes for detection of apoptosis by flow cytometry (12–16,23–26), there is definitively a need for deciphering the phenomena underlying staining characteristics of these probes. We present herein for the first time SYTO staining characteristics upon oncotic stimuli. To our knowledge this is also the first report demonstrating such a distinct behavior of SYTO16 and TMRM, and several conclusions can be made based on the present results: (i) a mere loss of Δψm does not lead to a reduced SYTO16 fluorescence, which appears instead to result from caspase-dependent changes that occur downstream of mitochondrial breach during apoptosis; (ii) many apoptotic stimuli induce mitochondrial rupture and engage downstream apoptotic targets within relatively short time, being the probable reason for the commonly detected resemblance of SYTO16 and TMRM staining profiles. When amplification of protease cascades advances, even more pronounced loss of SYTO fluorescence can be observed; (iii) it is still possible that caspase-dependent changes in mitochondria also contribute to the loss of SYTO fluorescence during apoptosis. With respect to the latter, it will be highly interesting to examine the behavior of SYTO dyes upon induction of alternative cell death pathways, e.g. those where lysosomes and cathepsin-calpain cascade (as proposed by Yamashima (28)), ER or Golgi function as initiators and/or executioners, and in particularly in conditions of blocked caspase-dependent apoptosis (e.g. transfection with the baculovirus p35, a protein that confers highly specific, irreversible caspase inhibition) (10, 28, 29). In closing, the search for mechanisms causing a reduction of SYTO fluorescence in apoptotic cells as compared with viable cells is still being continued.

Acknowledgements

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS AND DISCUSSION
  5. CONCLUDING REMARKS
  6. Acknowledgements
  7. LITERATURE CITED

DW and JS received a fellowship from Centre for International Mobility (CIMO), Helsinki, Finland.

LITERATURE CITED

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS AND DISCUSSION
  5. CONCLUDING REMARKS
  6. Acknowledgements
  7. LITERATURE CITED