Free thiols play a number of important roles in biology. An example is the predominant cellular antioxidant, reduced glutathione (GSH), which protects against oxidative damage (1). Glutathione exists in two forms, a reduced state (GSH) and in an oxidized state, glutathione disulfide (GSSG). GSH largely determines the thiol-disulfide status of the cell by interchange reactions. Because of the particular metabolism of glutathione a decrease in the GSH/GSSG ratio can be used as an indicator of oxidative stress (2). Agents altering the concentration of GSH have been shown to affect diverse functions such as cell proliferation, transcription of detoxification enzymes, and apoptosis (3–5).
Many current assays for determining the cellular thiol status rely on methods involving lysis of cells before assessing the sample by, for example, enzymatic or HPLC methods (6–8). Besides the high workload involved in these methods, they do not provide information about the thiol redox level on a per cell basis, but only on the overall population. There exist a number of different thiol-reacting reagents such as iodoacetamides, maleimides, and bimanes. Iodoacetamides and maleimides react with thiols by S-alkylation producing thioethers. Iodoacetamides have a high preference for thiols, but they may also react with histidine or tyrosine in the absence of free thiols, whereas maleimides are thiol specific. Monochlorobimane and monobromobimane target GSH in a reaction catalyzed by glutathione S-transferase thereby forming fluorescent adducts. Monochlorobimane can be used for staining rodent (CHO) cells (9); however, glutathione S-transferase present in human cells has a low affinity for monochlorobimane, and GSH is incompletely labeled (10, 11). VitaBright-48 is a cell-permeable maleimide derivative that stains thiols in all cell types independent of glutathione S-transferase activity.
We present here a novel assay for measuring the cellular concentration of cellular free thiols using a thiol-reacting probe. In this assay, the cell population to be investigated is stained using VitaBright-48, which immediately reacts with intracellular thiols, forming a fluorescent compound. Quantification of the fluorescence can then be used to determine the population's cellular thiol level at the single cell level. The assay is extremely fast and easy, because incubation and washing steps are not needed. The method is useful not only for investigating the cellular redox state but, as we show, also for assaying apoptosis when combined with an impermeable stain.
MATERIALS AND METHODS
WEHI-S cells (murine fibrosarcoma cell line) (12) and Jurkat cells (human leukemia cell line, subclone A3, ATCC #CRL-2570) were cultured in DMEM GlutaMAX-1 medium (GIBCO, Life Technologies, Carlsbad, CA) supplemented with 6% heat-inactivated fetal calf serum (GIBCO). The cells were incubated at 37°C, 5% CO2 atmosphere. Apoptosis was induced by treating WEHI-S cells with TNF-α (PeproTech, Rocky Hill, NJ) and Jurkat cells with camptothecin (CPT, Calbiochem, La Jolla, CA) for the time and at the concentrations specified.
The relative fluorescence intensity of solutions of GSH (5mM; Fluka, Sigma-Aldrich, St. Louis, MO), oxidized glutathione (2.5 mM, GSSG; Sigma, St. Louis, MO), VitaBright-48 (5 μM; ChemoMetec, Alleroed, Denmark), and mixtures used hereofwas measured on a Shimadzu (Kyoto, Japan) 5301-PC fluorescence spectrophotometer equipped with a 150-W Xenon lamp using 400-nm excitation.
Cellular fluorescence was quantified using a NucleoCounter NC-3000 image cytometer (ChemoMetec). Details of the NucleoCounter NC-3000 design and capabilities are available at www.chemometec.com. In brief, using large field-of-view optics and a high-speed imaging system, the NC-3000 rapidly images multiple fields-of-view facilitating large cell populations. Field of view is 3.28 mm × 2.45 mm, with a pixel resolution of 2.35 μm/pixel and an approximate magnification of 2×. The NC-3000 used in this study was configured with a dark field light source (used for cell image segmentation) and five different LEDs (used for fluorescence measurements) having peak wavelengths at, respectively, 365, 405, 475, 525, and 625 nm. The NC-3000 software provides an easy-to-use interface that facilitates automated dark field and fluorescent image acquisition, image analysis, and data visualization. Similar to flow cytometry subpopulations can be defined and quantified via gating.
Externalization of Phosphatidyl Serine
Changes in the phosphatidylserine symmetry were determined using annexin V conjugated to CF-647 (Biotium, Hayward, CA), SYTOX green (Invitrogen, Life Technologies, Carlsbad, CA), and VitaBright-48 (Solution 1, ChemoMetec). Briefly, 5 × 105 control or treated cells in annexin V binding buffer were incubated with 3 μL of CF-647 conjugated annexin V for 15 min at room temperature according to manufacturer's instructions. After washing, cells were stained with 8 μg/mL VitaBright-48 (Solution 1; ChemoMetec) and 0.5 μM SYTOX green (Invitrogen) and immediately analyzed using image cytometry. Image cytometry was carried out using a NucleoCounter NC-3000 and accompanying software (ChemoMetec). Segmentation was based on events detected using the dark field light source, and aggregates consisting of more than five cells were excluded. Debris was gated out based on area (pixel coverage) and intensity. CF-647 fluorescence was detected using peak excitation at 625 nm and measuring emission at 740/60 nm, SYTOX green using peak excitation at 475 nm and emission at 560/35 nm and VitaBright-48 using peak excitation at 405 nm and emission at 475/25 nm. Images of stained cells were also captured using an IX50 Olympus (Tokyo, Japan) microscope equipped with a Lumenera (Ottawa, Ontario, Canada) camera and the following filter cubes: U-MWG2, U-MWIBA3, and U-MNUA2. For imaging purposes, WEHI-S cells were grown and stained as described above in Lab-Tek chamber slides (Nunc, Thermo Fisher Scientific, Waltham, MA). Jurkat cells were grown and stained using standard conditions and loaded into NC-Slides A2 (ChemoMetec) before microscopic investigations. The experiment was performed twice in its entity (measuring phosphatidyl serine externalization four times during 20 h); besides this, it was also carried out four times with only two measurement points (6 + 20 h).
Fluorochrome-Labeled Inhibitor of Caspases Assay
Caspase 3/7 activity was measured using the fluorochrome-labeled inhibitor of caspases assay (FLICA). Briefly, 5 × 105 control or treated cells in cell culture media were incubated with 5 μL of reconstituted FLICA reagent (SR-DEVD-FMK, Immunochemistry, Bloomington, MN) for 60 min at 37°C according to manufacturer's instructions. After washing, cells were stained with 8 μg/mL VitaBright-48 (Solution 1; ChemoMetec). Stained cells were immediately analyzed using the NucleoCounter NC-3000 system. Based on cell data from the dark field channel, image segmentation was performed as described in the above section. The SR-FLICA probe was detected using peak excitation at 525 nm and emission at 675/80 nm and VitaBright-48 using peak excitation at 405 nm and emission at 475/25 nm. Images of stained cells were also captured using an Olympus microscope equipped with a Lumenera camera as described above. The experiment was performed twice in its entity and four times with two measurements at 6 + 20 h.
VitaBright-48 Forms Highly Fluorescent Adducts with GSH but not with GSSG
Figure 1 shows the relative fluorescence intensities of GSH, GSSG, and VitaBright-48 alone and GSH mixed with VitaBright-48 and GSSG mixed with VitaBright-48. Although the single solutions of GSH, GSSG, and VitaBright-48 only exhibit weak fluorescence intensities, the mixture of VitaBright-48 with GSH, but not with GSSG, exhibits a nine times stronger fluorescence intensity than the sum of the fluorescence intensities of the solutions alone. This synergistic effect suggests the formation of a highly fluorescent product between VitaBright-48 and GSH. It seems that VitaBright-48 is not specific for GSH as it readily reacts with other thiols, for example, DTT also forming a fluorescent product (results not shown). This is in contrast to the stain monochlorobimane that is specific for GSH (at least in rodent cells) and requires the action of the enzyme GSH 5-transferase to form a fluorescent product (9). However, it has been reported that monochlorobimane does not specifically stain GSH in humane peripheral blood mononuclear cells (13). Figure S1 in Supporting Information shows the staining kinetics for VitaBright-48 and monochlorobimane, demonstrating the instant staining of VitaBright-48. Supporting Information Figure S1 also clearly shows the much stronger fluorescence obtained using VitaBright-48 compared with monochlorobimane.
Fluorescence Intensities of VitaBright-48 Stained Apoptotic and Nonapoptotic Cells; Correlation with Apoptotic Markers
Extrusion of GSH has been reported to be a key event in the apoptotic process (14–17). The cellular thiol-disulfide status is mainly determined by the level and oxidation status of GSH. Thus, the thiol-reacting probe VitaBright-48 may be used for assaying apoptosis. In order to explore the possibility for using VitaBright-48 for assaying apoptosis, we have investigated the correlation of the fluorescence intensity of VitaBright-48 staining with well-known apoptotic markers such as phosphatidylserine externalization and caspase 3/7 activity. During apoptosis, phosphatidylserine is translocated to the outer layer of the membrane, where it may be recognized and bound by the phosphatidylserine selective protein annexin V. Annexin V will also bind to phosphatidylserine on late apoptotic and necrotic cells; however, necrotic cells lack membrane integrity, and thus, necrotic cells can be distinguished from early apoptotic cells by the use of an impermeant dye. Here, we have used staining with CF-647 conjugated annexin V together with the impermeant dye SYTOX green and VitaBright-48. In Figure 2, Jurkat and WEHI-S cells triple stained with VitaBright-48, CF-647-conjugated annexin V, and SYTOX green are shown. Cells that are apoptotic (stained with CF-647 conjugated annexin V) or nonviable (stained with SYTOX green) or late apoptotic (stained with both CF-647 conjugated annexin V and SYTOX green) exhibit a much lower VitaBright-48 fluorescence intensity, indicating a lower level of reduced thiols (Fig. 2 expanded to show all channels as single channels can be found in Supporting Information as Fig. S2). To investigate the timing of the decrease in reduced thiols versus apoptosis induction, cells were induced to undergo apoptosis and were analyzed 5, 10, 15, and 20 h postinduction using image cytometry. The scatter plots in Figure 3 show that control cells exhibit a much stronger VitaBright-48 fluorescence intensity than cells induced to undergo apoptosis. SYTOX green positive cells, that is, nonviable cells, have been excluded from the scatter plot; however, these cells exhibit very low or none VitaBright-48 fluorescence intensity. Already at the first time point (5 h), there is minor decrease in the VitaBright-48 fluorescence intensity and an increase in annexin V-positive cells of CPT-treated Jurkat cells and TNF-α-treated WEHI-S cells compared with the control. Only very few cells exhibit low VitaBright-48 and low annexin V staining, suggesting that the decrease in reduced thiols either happens simultaneously with phosphatidyl serine externalization or after. After 10 h, the percentage of annexin V-positive cells has increased and so has the subpopulation of cells with low VitaBright-48 staining and high annexin V conjugate staining 15–20 h after apoptosis induction using either TNF-α (WEHI-S cells) or CPT (Jurkat cells) the cells are divided into two main subpopulations; one which has a high VitaBright-48 fluorescence intensity and exhibit low staining with the CF-647 annexin V conjugate and one which shows low VitaBright-48 fluorescence intensity but are intensely stained with the CF-647 annexin V conjugate (Fig. 3). There is also an intermediate subpopulation, which shows both high VitaBright-48 fluorescence intensity and high staining with the CF-647 annexin V conjugate. Early after induction of apoptosis (5–10 h), this population is predominant over the low VitaBright-48/high annexin V population; however, later (15–20 h) the population moves to the left; showing a weaker staining with VitaBright-48. This suggests that the apoptosis-mediated decrease in cellular reduced thiols occurs after externalization of phosphatidylserine.
The correlation between VitaBright-48 and caspase 3/7 activity was also investigated. To measure the activity of caspases, the fluorochrome-labeled inhibitors of caspases (FLICA) assay was used (18). Only cells with active caspases will be stained, and thus, it is not necessary to include a nonviable stain; cells lysed by nonapoptotic means will not be stained with the FLICA probe. In Figure 2, Jurkat and WEHI-S cells stained with VitaBright-48 and the red-fluorescent sulphorhodamin-labeled caspase probe; SR-DEVD-FMK are shown. Cells, which are apoptotic (stained with the SR-FLICA probe), exhibit a much lower VitaBright-48 fluorescence intensity than the nonapoptotic cells, indicating that these cells have a lower level of reduced thiols. Figure 4 shows scatter plots of Jurkat and WEHI-S cells stained with VitaBright-48 and SR-FLICA probe. Cells that stain with the SR-FLICA probe and hence are apoptotic stain weakly with VitaBright-48 indicating that the apoptotic cells have a lower level of reduced thiols. As seen from Figure 4, the timing pattern is analogous to what is seen with annexin V staining; first (5–10 h postapoptosis induction), a subpopulation is stained with the SR-FLICA probe whereof some, but not all, also have a low VitaBrigh-48 fluorescence intensity and then, 15–20 h after induction the SR-FLICA-stained subpopulation exhibit a more pronounced decrease in VitaBright-48 staining. This suggests that the most dramatic decrease in thelevel of reduced thiols happens after activation of caspases 3 and 7.
We present here a new and fast method for evaluating the cellular level of reduced thiols using the thiol-reactive probe VitaBright-48. Using different cell lines and apoptogenic agents we show that a decrease in the level of reduced thiols correlates with well-known apoptotic markers such as phosphatidylserine translocation and caspase activity. The concentration of GSH has earlier been found to decrease in response to apoptosis induction, also when using nonoxidative apoptogenic agents, due to extrusion of GSH (14–17). Interchange reactions links the level of reduced glutathione directly to the level of other reduced thiols; thus, measuring the total level of free thiols, as we do here using VitaBright-48, may indirectly reflect the level of GSH. However, as we first detect a decrease in VitaBright-48 fluorescence after phosphatidylserine externalization and caspase 3 activation, and hence later than GSH extrusion, we speculate that there may be a delay between the loss of GSH and the subsequent oxidation of the remaining thiol pool detected by VitaBright-48.
We induced apoptosis via the extrinsic pathway in WEHI-S cells using TNF-α and via the intrinsic pathway using the topoisomerase I inhibitor CPT in Jurkat cells. In both cases, the treatments led to apoptosis-related changes, that is, externalization of phosphatidylserine and activation of caspase 3. These changes were accompanied by a decrease in the level of reduced thiols detected by the use of the thiol-reactive probe VitaBright-48. As extrusion of GSH has been proposed to play an important role in the commitment to apoptosis (19); it would be reasonable to expect a decrease in GSH level to happen before the onset of apoptosis. We do find that a decrease in the level of free thiols correlates with the presence of apoptotic markers; however, it seems that apoptosis precedes the decrease in thiol level rather than the opposite. Chromatin fragmentation, which is a late apoptotic phenomenon, has been shown to be accompanied by GSH extrusion and thus a change in the redox balance (14, 17). Instead of the hypothesis that GSH depletion functions in order to trigger the apoptosis commitment, it is possible that loss of GSH is located down stream of caspase 3 activity in the apoptotic cascade. This is also supported by a study by van den Dobbelsteen et al. (17) demonstrating that preincubation of Jurkat cells with the caspase 3 (apopain) inhibitor Z-Val-Ala-Asp-chloromethylketone before exposure to FasL not only prevents caspase 3 activity and chromatin fragmentation but also almost abolishes loss of GSH. Staining with thiol-selective probes provides a very rapid, easy and reliable way of assaying late apoptosis using either flow/image cytometry or fluorescence microscopy. Here we have used VitaBright-48 to detect changes in level of thiols, but monochlorobimane may also be used, however, the latter is a less stable and more complex dye to work with compared to VitaBright-48 and may only be suitable for rodent cells. Cells immediately (within 2 min after addition of VitaBright-48) reach maximal fluorescence intensity, which is stable for at least an hour (not shown), while the staining kinetics for monochlorobimane is slower, and dependent on cell line and temperature (20, 21). Combined staining with an impermeable stain such as propidium iodide enables exclusion of necrotic/late apoptotic cells. In addition, interpretation of the results is very easy; presenting the data as a histogram showing the fluorescence intensity of all viable cells enables division of the cells into different subpopulation depending on their level of free thiols. Cells with a low level of free thiols (apoptotic cells) also have a low intensity score while healthy cells have a high intensity score. Markers can then be used to define the apoptotic subpopulation as shown in Figure 5.
A central feature of this assay is that it enables quantification of the level of thiols (and hence apoptosis) at the single cell level; this is especially important as the concentration of thiols is not necessarily evenly distributed in the cell population, thus conventional biochemical bulk assays involving cell lysis may miss out important information. However, as altered redox levels may cause changes in the level of thiols, care should be taken interpreting results when investigating an apoptotic response under circumstances involving oxidative stress.
We suggest advancing the repertoire of current apoptosis assays (such as reviewed in (22, 23)) with the measurement of changes in the level of thiols. Using VitaBright-48 in combination with PI or other impermeable stains provides an extremely easy and fast way of assaying apoptosis as no washing or incubation steps are required. The assay may be very useful, for example, in high throughput screenings of chemical libraries or following cell health in a batch production.
The authors wish to thank Anna Fossum, BRIC, University of Copenhagen for help with validation of initial results using flow cytometry.
ChemoMetec A/S manufactures the stain VitaBright-48 and has filed a patent for the use of this probe for staining of cells. Mette E. Skindersoe and Soren Kjaerulff are among the inventors of this patent application.