Detecting intracellular thiol redox state in leukaemia and heterogeneous immune cell populations: An optimised protocol for digital flow cytometers

Graphical abstract

Flow cytometric methods for detecting and quantifying reduced intracellular thiol content using fluorescein-5-maleimide (F5M) in viable eukaryotic cells date back to 1983 (Durand and Olive [1]). There has been little development in these methodologies since that time, a period that has witnessed huge technological advances, particularly with the emergence of digital multi-parameter flow cytometric systems. Concurrent advancement in our understanding of redox regulation within eukaryotic cellular systems has also followed, whereby it is now accepted that cysteine thiols partake in redox reactions, which regulate protein activity and function (Groitl and Jakob (2014), Won et al. (2012)). Moreover, we are at the dawn of a new era in redox biology whereby the importance of 'reductive stress' in eukaryotic cellular systems is gathering momentum (Wadley et al. (2018) [4]). It is therefore critical that methods be continually advanced to better understand these concepts in more detail at the cellular level. Flow cytometry is a powerful technique that may be used for this purpose. Henceforth we have rejuvenated these methods to address modern scientific questions. In this paper, essential detail is provided on: The adaption of a protocol initially described by Durand and Olive [1] for use with modern digital flow cytometer configurations. Here we provide optimal conditions for labelling intracellular thiols with F5M for detection using digital flow cytometers. Our modifications avoid the use of methanol fixation thus preserving cell viability in single cell suspension cultures.  TIB-152   TM   ), human acute T-cell lymphoblastic leukaemia. 2 Donor human peripheral blood mononuclear cells (PBMCs).

Specification
Note: full ethical approval and donor consent is required for the isolation of PBMCs from human blood, in strict accordance with the declaration of Helsinki [7].

Experimental procedures
Analysing the effect of glucose oxidase mediated H 2 O 2 on intracellular thiol oxidation using flow cytometry The protocol described is based on a previous method described in Durand and Olive [1], which incorporated various fluorescent stains for the detection of reduced intracellular thiols, including fluorescein-5 maleimide (F5M). The current method is optimised for bench-top digital flow cytometers and allows low level detection of F5M in viable cells, which proved challenging using the previous method. The optimisation of the technique described herein is of particular importance, since some reactive oxygen species (ROS) e.g. hydrogen peroxide (H 2 O 2 ) have the propensity to oxidise proteins with exposed cysteine residues, in turn altering their activity [8]. In the experimental model proposed, exposure of cells to physiological doses of exogenous H 2 O 2 can be modelled in vitro by adding glucose oxidase (GOX) to the culture medium as previously described [5]. The addition of GOX in this way will metabolise glucose in the culture medium, generating H 2 O 2 as defined by the following reaction [9]: β-D-glucose + oxygen → D-glucono-1,5,-lactone + hydrogen peroxide Hydrogen peroxide is membrane permeable and can enter the cell mediating the oxidation of intracellular cysteine residues [10]. To demonstrate this, the example included uses the Jurkat immortalised T-cell line to monitor cellular cysteine thiol redox state over time in response to GOX generated H 2 O 2 . The method details are as follows: 1 Maintain Jurkat cells in 15 ml complete IMDM, at a density between 2 Â 10 5 and 10 6 cells/ml at 37 C and 5% CO 2 . Note: IMDM is particularly suited to this because of the high glucose concentration (4.5 g/l), which is a substrate for GOX. 2 Prepare a 10 mU/ml GOX solution and a 10 mg/ml catalase solution and store at 4 C until required.
Note: The catalase is used to confirm that H 2 O 2 is mediating changes in thiol redox state. To prepare glucose oxidase solutions for in vitro H 2 O 2 generation, it is recommended to obtain high purity glucose oxidase solid (e.g. Sigma cat no. 49180 or similar quality). Cheaper preparations can contain confounding contaminants such as catalase, which metabolises H 2 O 2 . In a Class 2 biological safety cabinet, dissolve solid GOX in tissue culture-grade water to a concentration of 1 U/ml (U = enzyme units). Sterilise by filtration and aliquot for storage at À20 C and avoid freeze thaw. Prepare 10 mU/ml working solutions by thawing 1 U/ml stock aliquots at 37 C for 1 min and diluting with tissue culture-grade water immediately prior to use. Discard excess. Note that H 2 O 2 production generated by GOX in vitro can be measured and titrated to physiological levels using the H 2 O 2 -specific fluorescent probe Amplex UltraRed (Life Technologies, Hampshire, UK) [5]. 3 Transfer the contents of a 24-48 h Jurkat cell culture flask (60-80% confluence) to a 15 ml tube and centrifuge at 100-300 Â g for 10 min at room temperature in a bench-top centrifuge with swingout rotor. Note: Jurkat cells are non-adherent and do not require Trypsin-EDTA treatment to dislodge from the flask surface. 4 Remove the supernatant and carefully suspend the pellet in 1 ml complete IMDM pre-warmed to 37 C. 5 Perform a viable cell count using trypan blue and dilute the Jurkat cells to a viable density of 10 7 cells/ml. 6 Add 100 ml of the diluted Jurkat cell suspension to five wells in a 96-well tissue culture treated 7 Add 98 ml to two wells in the same plate and label; '0 h +C' and '4 h +C'. Note the "+C" refers to the addition of catalase. Thus there are 7 assay wells in total.
8 Add 2 ml of the 10 mg/ml catalase stock solution (prepared in step 2) to wells labelled '0 h +C' and '4 h +C' only.
9 To start the assay add 100 ml of the GOX solution (prepared in step 2) to wells labelled '4 h' and '4 h +C' and incubate at 37 C and 5% CO 2 . Note: that this assay has been optimised to use IMDM because of the high glucose content (4.5 g/l), rather than RPMI-1640 which has much lower glucose content (2 g/l). This is important since GOX requires a glucose substrate to generate H 2 O 2 .
10 After 1 h, add 100 ml of the GOX solution to the '3 h' well and return to the incubator.
11 Repeat after a further 1 and 2 h for wells labelled '2 h' and '1 h' respectively.
12 Finally, 1 h after the addition of GOX to '1 h', add 100 ml of the GOX solution to wells labelled '0 h' and '0 h + C' and immediately transfer the contents of all 7 assay wells to a fresh 1.5 ml microcentrifuge tube. Centrifuge all at 100-300 Â g for 5 min at room temperature. 13 Carefully remove supernatants and suspend the pellets in 1 ml D-PBS. Centrifuge all tubes at 100-300 Â g for 5 min at room temperature.
14 Remove supernatants and suspend pellets in 1 ml pre-chilled 0.1 mM F5M solution and incubate all tubes for 20 min, on ice and protected from light. 15 Centrifuge tubes at 100-300 Â g for 5 min at 4 C. 16 Remove supernatants and suspend pellets in 1 ml pre-chilled FSB and centrifuge at 100-300 Â g for 5 min at 4 C.  Table 1. Note: normalisation is performed by dividing the F5M signal by the F5M signal of non-GOX treated controls i.e. 0 h. 20 To confirm that GOX treatment is not inducing cell death, viable cell counts should be performed on GOX treated cells before or after flow cytometry acquisition as shown in Fig. 2A. 21 To demonstrate that the GOX generated H 2 O 2 is mediating the loss in reduced thiols, the effect of catalase in '4 h' vs '4 h + C' treated cells can be observed on normalised data as shown in Fig. 2B. The data suggest that H 2 O 2 mediates the loss in reduced intracellular thiols as measured by a loss in F5M (FL1) fluorescent signal.

Method versatility and confocal microscopy confirmation that the F5M signal is intracellular
To illustrate the versatility of our flow cytometric method, we provide a protocol for evaluating intracellular thiol oxidation in heterogeneous cell populations i.e. T-cells within a peripheral blood mononuclear cell (PBMC) fraction. Here we couple the cell surface staining of CD3 (T-cell marker) with F5M. Confocal microscopy was used to verify that F5M signal detected by flow cytometry was intracellular and not due to F5M binding at the plasma membrane. Venous blood from healthy donors following informed consent was collected by a trained phlebotomist into a 6 ml lavender stopper K2EDTA Vacationer 1 tube (BD). All procedures from this point should be performed inside a Class 2 biological safety cabinet: 1 Transfer 6 ml venous blood to a fresh sterile universal tube and add an equal volume of Hank's balanced salt solution (HBS); 12 ml total volume. 2 Add 3 ml Ficoll-Paque PLUS to three sterile 15 ml centrifuge tubes and carefully layer 4 ml of the diluted venous blood on top of the Ficoll-Paque PLUS.
3 Transfer the tubes to a bench-top centrifuge with swing-out rotor and centrifuge at 400 x g for 40 min at room temperature (18À20 C) with the brake off. 4 Carefully remove the tubes from the centrifuge and return to the biological safety cabinet. The PBMC layer (white and delicate) will be visible at the Ficoll-Paque PLUS interface. Carefully remove the upper layer (plasma) to waste with a fresh sterile Pasteur pipette without disturbing the PBMC layer. Then using a fresh sterile Pasteur pipette, transfer the PBMC layer to a fresh 15 ml centrifuge tube. For each tube, remove the entire PBMC interface with as little of the Ficoll-Paque PLUS layer as possible. 5 Add 8 ml HBS to each tube and centrifuge at 100-300 Â g for 10 min at room temperature.  Table 1). (C) Chart illustrates normalised FL1 (laser: l488 nm, filter: l530/30 nm) MFI.
An increase in F5M signal is noted at 1 h (suggesting a reductive spike), followed by progressive decrease in signal at 2 h-4 h (suggesting a loss of reduced thiols). Data shown are mean +/À SD of the mean (n = 3). Data analysed by Kruskal-Wallis one-way analysis of variance with Dunn's post-hoc test, *p < 0.05, **p < 0.01 and ***p < 0.001. 6 Using a fresh sterile pipette, remove the supernatant from each tube taking great care not to disturb the PBMC pellet and suspend in 1 ml HBS. The PBMC's can be pooled at this stage. Add sufficient HBS to the pooled cells to make 10 ml. 7 Centrifuge pooled PBMC's at 100-300 Â g for 10 min at room temperature. 8 Remove the supernatant taking great care not to disturb the PBMC pellet and suspend the pellet in 2 ml pre-warmed (37 C) complete IMDM. Transfer the cells to an incubator set at 37 C and 5% CO 2 for subsequent analysis. 9 Prepare the F5M working solution for flow cytometric use in advance. Protect from light and store at 4 C (up to 1 month) or at À20 C for up to 1 year. 10 Remove the PBMC's from the CO 2 incubator and centrifuge at 100-300 Â g in a benchtop refrigerated centrifuge with swing-out rotor for 5 min at room temperature. 11 Remove supernatant and suspend the pellet in 10 ml D-PBS (equilibrated to room temperature) and centrifuge at 100-300 Â g for 10 min. 12 Remove supernatant and suspend the pellet in 1 ml FSB. Prepare a 1 in 20 dilution of the PBMC suspension for cell viability counting (5 ml PBMC in 95 ml FSB is sufficient) and transfer the remaining suspension to ice.  13 Mix 10 ml of the 1 in 20 PBMC dilution with 10 ml trypan blue solution and count viable cells using a haemocytometer. 14 Adjust the PBMC suspension to a density of 10 7 viable cells/ml with pre-chilled (4 C) FSB and add 100 ml (10 6 cells) to three fresh 1.5 ml microcentrifuge tubes and centrifuge for 5 min at 100-300 Â g and 4 C. Note: At this point it is preferable to pellet cells using a microcentrifuge. 15 Remove supernatants and suspend each pellet in FSB to a final volume of 90 ml. 16 Add 10 ml of mouse anti-human CD3 (clone SK7) APC conjugate to two tubes and 10 ml of mouse IgG1 APC conjugate isotype control to the other. This step will surface stain the T-cells allowing intracellular vs extracellular differentiation during confocal microscopy. 17 Vortex all tubes briefly (1 to 2 s) and incubate on ice for 30 min protected from the light. 18 Add 1 ml D-PBS (chilled to 4 C) to each tube and centrifuge for 5 min at 100-300 Â g and 4 C.
Remove supernatants and then repeat this step. 19 Remove supernatants and add 1 ml D-PBS to the isotype control tube and 1 ml D-PBS to one of the anti-human CD3 tubes and label it 'F5M'.
20 Add 1 ml of 1 mM NEM solution to the remaining anti-human CD3 tube. These cells will be used to set the background for the F5M T-cell stain. Label this tube 'NEM'.
Note: This concentration of NEM is based on 99+% purity. A lower grade can be used, which may require titrating to achieve the optimal concentration. 21 Incubate all tubes for 20 min on ice, protected from the light. 22 Centrifuge for 5 min at 100-300 Â g and 4 C. 23 Remove supernatants and suspend each pellet in 1 ml D-PBS and centrifuge for 5 min at 100-300 Â g and 4 C. 24 Remove supernatants and add 1 ml chilled D-PBS to the isotype control tube. These cells will be used to set the negative T-cell gate. Add 1 ml 0.1 mM F5M to both 'NEM' and 'F5M' tubes.
25 Incubate all three tubes for 20 min on ice, protected from the light. 26 Centrifuge for 5 min at 100-300 Â g and 4 C. 27 Remove supernatants and suspend each pellet in 1 ml pre-chilled FSB and centrifuge for 5 min at 100-300 Â g and 4 C. Repeat this step.
28 Finally remove supernatants and suspend each pellet in 100 ml pre-chilled FSB, keep on ice, protect from light and go to flow cytometer. 29 At the flow cytometer, acquire the 'Ig-control' cells first. These cells can be used to set the positive T-cell gate, which will appear in the FL4 parameter (laser: l640 nm, filter: l675/25 nm). Fig. 3A and B illustrates the gating strategy. 30 Next acquire the 'NEM' cells and create an FL1 (laser: l488 nm, filter: l530/30 nm) histogram plot gated on the T-cell population as seen in Fig. 3B. The peak shown on the FL1 histogram is representative of the F5M background staining and can be seen in Fig. 3C. 31 Finally acquire the 'F5M' cells. As shown in Fig. 3C, these cells fluoresce at an order of magnitude greater than 'NEM' cells. 32 To confirm viability, perform cell counts either before or after flow cytometry acquisition as described in steps 3 and 4. Fig. 3D shows little loss in cell viability post F5M treatment. 33 As shown in Fig. 3E-G, confocal microscopy is used to confirm that the F5M is entering the T-cells and therefore that the flow cytometric FL1 signal generated is representative of intracellular thiol redox status. This supports the notion that the extracellular environment is more 'oxidising', thus cysteine residues here will be to a great extent oxidised i.e. to disulphide or higher oxidation states.
To visualise F5M labelled CD3 + T-cells by confocal microscopy, carefully affix a glass microscope slide to the microscope and pipette 50-100 ml of the final stained cell suspension as a droplet onto the upper surface. To limit evaporation gently place a thin coverslip (19 mm) on top of the droplet. Focus cells using a 63Â oil immersion objective. Use both the argon (F5M) and helium-neon (CD3-APC) lasers at a previously determined appropriate power setting, and adjust the offset, gain, zoom and Z-position as necessary for optimal image quality. If available, opt for sequential scanning modes to allow simultaneous visualisation of both signals with minimal crosstalk between channels. DNA stains such as DAPI can be added to the staining protocol to identify nuclear regions where the appropriate excitation lines and channels are available.  Fig. 4C illustrates pixel peak profiles for NEM + F5M treated (i) and F5M alone treated (ii) extracted proteins. 14 Since F5M irreversible modifies cysteine thiols, identification of redox active proteins is permitted via mass spectrometry based on incremental mass shifts relating to F5M binding. Table 2 shows various intracellular redox active proteins identified from F5M treated Jurkat cells using LC/MS analysis. For the F5M mass shift, the formula C 24 H 13 NO 7 (+427.069 Da) was used, since this relates to the mass difference expected for a Cys-thiol or Cys-thiolate modification by F5M. Note: LC/MS analysis was outsourced to the Bristol Proteomics Facility (University of Bristol, UK).