Simultaneous Assessment of Kinship, Division Number, and Phenotype via Flow Cytometry for Hematopoietic Stem and Progenitor Cells

For the following protocol, de-identified umbilical cord blood was used as the HSPC source and collected in accordance with the guidelines defined by the Saint-Louis Hospital cord blood biobank (authorization AC-2016-2759) and with the Declaration of Helsinki.

NOTE: Before starting, ensure that all reagents and equipment necessary for this protocol are available, as listed in the Table of Materials and mentioned in the protocol. Prepare the relevant reagents fresh and do not store them, unless explicitly mentioned.

1. Cell dye staining

NOTE: This section describes staining with four combinations of CFSE and violet dye (CTV) cell division dyes. Process all the tubes simultaneously, even if no cell dye solution is added. All steps are performed in sterile conditions to allow the following cell culture step. Time required: approximately 100 min.

1. Process the cord blood unit according to a magnetic sorting protocol²⁹. Ensure that two fractions are available: a large CD34^– fraction and a smaller CD34⁺ fraction. Spin both tubes for 5 min at 300 x g. Aspirate the supernatant without disturbing the pellet.
2. For the CD34⁺ fraction, resuspend it in 1 mL of Dulbecco's modified Eagle medium (DMEM) without fetal bovine serum (FBS). Count the cells using a hemocytometer; the cell density should not be higher than 3 x 10⁶ cells/mL. If this is the case, adapt the volume accordingly. For the CD34^– fraction, resuspend in DMEM w/o FBS, and adjust the volume to a maximum of 6 x 10⁶ cells/mL.
3. Aliquot 250 µL of the CD34⁺ fraction into four 15 mL polypropylene tubes. Label the tubes as follows: CD34⁺/CF (CFSE_only), CD34⁺/CV (CFSE_high CTV_low), CD34⁺/VC (CFSE_low CTV_high), and CD34⁺/VI (CTV_high). Aliquot 250 µL of the CD34^– fraction into another four 15 mL polypropylene tubes. Label the tubes as follows: CD34^–/CF (CFSE_only), CD34^–/CV (CFSE_high CTV_low), CD34^–/VC (CFSE_low CTV_high), and CD34^–/VI (CTV_high). The remaining cells from the CD34^– fraction can be discarded.
4. Prepare two CFSE solutions, named CFSE_high and CFSE_low. For CFSE_high (10 µM), mix 1.1 mL of DMEM w/o FBS with 2.2 µL of CFSE stock (5 mM) solution. For CFSE_low (5 µM), mix 550 µL of DMEM w/o FBS and 0.55 µL of CFSE stock (5 mM) solution.
5. Add 250 µL of the CFSE_high solution to the CF and CV tubes, 250 µL of the CFSE_low solution to the VC tubes, and 250 µL of DMEM w/o FBS to the VI tube. To ensure an efficient mix of cell suspension and cell dye, incline the tube by almost 90 degrees, and deposit the CFSE solutions on the tube wall. Then, hold the tube vertically to mix the two solutions, and pipette three or four times to ensure a fast mixing of the CFSE solutions with the resuspended cells. Incubate at 37 °C for precisely 8 min.
6. After the incubation, add 5 mL of DMEM + 10% FBS. Keep the tubes at 37 °C for 5 min.
7. Spin the tubes for 5 min at 300 x g. Remove the supernatant via aspiration without disturbing the pellet, and wash the pellet with 5 mL of phosphate buffered saline 1x/ethylenediaminetetraacetic acid (PBS 1x/EDTA). Spin again for 5 min at 300 x g. Discard the supernatant without disturbing the pellet, and resuspend the cell pellet in 250 µL of 1x PBS /EDTA.
8. Prepare two CTV solutions, named CTV_high and CTV_low. For CTV_high (10 µM), mix 1.1 mL of PBS 1x/EDTA and 2.2 µL of CTV stock (5 mM). For CTV_low (5 µM), mix 550 µL of PBS 1x/EDTA with 0.55 µL of CTV stock (5 mM).
9. Add 250 µL of the CTV_high solution to the VC and VI tubes, 250 µL of the CTV_low solution to the CV tube, and 250 µL of 1x PBS /EDTA to the CF tube. Use the same technique as described in step 1.5. Incubate at 37 °C for precisely 8 min.
10. After the incubation, add 5 mL of DMEM + 10% FBS. Keep at 37 °C for 5 min.
11. Spin the tubes for 5 min at 300 x g, discard the supernatant without disturbing the pellet, and then wash the pellet with 5 mL of 1x PBS /EDTA. Spin again for 5 min at 300 x g.
12. Discard the supernatant without disturbing the pellet, and resuspend the CD34^– fractions in 1x PBS /EDTA for a final concentration of 1.5 x 10⁶ cells/mL. Resuspend the CD34⁺ fractions in 40 µL of staining buffer, and transfer the cells into 1.5 mL tubes.

2. Antibody staining

NOTE: Antibody staining can be customized according to the experimental needs. Only the CD34⁺ fractions undergo antibody staining; the CD34^– fractions are used as a single staining control for the cell division dye combinations (CV, VC, CF, and VI fractions). The following panel is tailored for the detection of four types of HSPCs: hematopoietic stem cells (HSCs), multipotent progenitors (MPPs), lymphoid-primed multipotent progenitors (LMPPs), and hematopoietic progenitor cells (HPCs)¹². However, identification of HSCs and MPPs is presented. Time required: 75 min.

1. Prepare the single staining for surface staining, using compensation beads. Mix negative beads and Immunoglobulin G (IgG) beads in a 1:1 ratio, for a total volume equivalent to 20 µL x the number of surface markers (e.g., 120 µL if the staining panel contains six antibodies).
2. Dispatch 20 µL of beads into individual 1.5 mL tubes for each marker. Add the volume corresponding to the dilution factor for each antibody in the corresponding tube (e.g., if the dilution factor is 1:20, add 1 µL).
3. To stain the CD34⁺ cells, prepare a master mix of antibodies¹², based on Table 1. Mix the antibodies into a single 0.5 mL tube. Add 7 µL from the antibody master mix to each of the four CD34+ conditions.
4. Incubate the compensation beads and the CD34⁺ samples at 4° C for at least 30 min.
   NOTE: The incubation time must be adapted to the technical details of the antibodies used for staining.
5. During incubation, prepare the 96-well round-bottom plate to be used for sorting, adding 100 µL of cell culture medium to each well using a multichannel pipette.
   NOTE: Leave wells H8-H12 empty.
6. Label 5 mL polypropylene tubes for the surface staining controls (5, using beads), the cell division dyes controls (4, using the CD34^– fractions), and the CD34⁺ samples (4).
7. At the end of the incubation, wash the cells and beads with 1 mL of staining buffer. Transfer the total volume to the 5 mL polypropylene tubes. Centrifuge the tubes for 5 min at 300 x g, then aspire the supernatant without disturbing the pellet.
8. Resuspend the cells in staining buffer, using approximately 500 µL each for the beads and CD34⁺ cells, and 1 mL for the CD34^– tubes.

Table 1: Template to prepare the antibody mastermix for a cell sorting experiment. Please click here to download this Table.

3. Cell sorting

NOTE: The sorted cell numbers could vary according to the total amount of cells available. In the protocol, a minimum cell number for each control is provided. Time required (for a single plate): 100 min.

1. Open the template experiment or set a new experiment. Create a single specimen and multiple tubes, one per each condition.
2. Set the gating strategy detailed in Figure 1, creating six dot plot diagrams. First, visualize the cells on a FSC-A/SSC-A dot plot, and double click on the polygon gating tool to select a population with low side scatter (Figure 1A). In the following dot plot (FSC-A/FSC-H), right click on the plot and select the gate "Cells" from the dropdown menu by clicking on it. Use the same gating tool to select a tight population on the diagonal between the two axis (Figure 1B).
3. In the third dot plot (APC vs. FSH-H), display the population "Single Cells" and gate the cells negative for the expression of APC Lineage (Lin) (Figure 1C). In the fourth plot (CFSE vs. CTV), display the population "Lin^–" and create four separated gates, one for each dye combination (Figure 1D).
   NOTE: These gates must be tight, to select only a small fraction of homogeneously stained cells.
4. Use the fifth and sixth plot (APC-Cy7 vs. BV650 and PE-Cy7 vs. PE) to identify the progenitors of interest. Generously gate the CD34⁺CD38^– population and the CD34⁺CD38⁺ in the fifth plot (Figure 1E). Then, select the CD34⁺CD38^– population in the sixth plot, and draw three gates, according to Figure 1F.
5. Run the single staining tubes containing the compensation beads, clicking on the Acquire button. Adjust the photomultiplier tube (PMT) voltages from the Parameters dropdown menu, particularly for the cell division dyes (between 10⁴ and 10⁵ on a biexponential scale).
6. Refine the compensation matrix accordingly to the panel used for sorting, using the Compensation tab. Record at least 5,000 events in the beads gate, clicking on the Record button.
7. Run the CD34^– fractions and check the compensation matrix again. Record at least 10,000 events in the single-cell gate.
8. Run the CD34⁺ fractions, recording at least 5,000 events in the single-cell gate. Adjust the gate for each dye combination, setting a tight gate to select a homogeneous population (Figure 1D). Similarly, adjust the gating for selecting HSCs and MPPs.
9. Once the analysis is completed and all tubes are recorded, insert the plate in the appropriate holder, after performing the standard Aria calibration for sorting on 96-well plates. Cooling the plate is recommended.
10. Prepare the plate sorting template according to the scheme presented in Table 2, using the experiment sort layout. The wells named "CD34^–" contain 5,000-10,000 cells, sorted on the CF/CV/VC/VI gate. The "Bulk" wells contain at least 500 cells, sorted on the gate CD34⁺CD38^–. Finally, the single-cell wells contain only one event per cell division dye combination per well, so four events per well in total.
    NOTE: The "Bulk" populations can be adapted to specific progenitors' subset; do not sort less than 500 cells.
11. For the sorting, proceed in order, completing every cell division dye combination before moving to the next one. For example, start with sorting the CD34^– CF, in yield purity mode. Click on the acquire button, then on the sort button.
12. At the end of the CD34^– sorting, insert the CF CD34⁺ tube. Acquire, then click on the sort button, making sure to have ticked 0/16/0 as purity grade. Finally, sort the cells of interest, one cell per well, in single-cell purity, making sure to tick the Index Sorting option.
13. Move to the following cell division dye combination, repeating the same order. As a reference, Table 2 provides an example of a sorted plate.
    NOTE: The index sorting function generates individual files for each sorted condition.
14. At the end of the sorting, export the files as .fcs 3.0 files. Put the cells in a 37 °C, 5% CO₂ incubator. The cells are cultured for multiple days, according to the experimental design, for at least 24 h²⁶.

Table 2: Template for a cell sorting 96-well plate, based on the specific requirements for the successive flow cytometry analysis. Please click here to download this Table.

4. Cell sorting data analysis

NOTE: To validate the quality of cell sorting, the FACS data analysis is necessary before moving further. The main output of this step is the generation of a spreadsheet containing the marker intensities of each sorted individual cell.

1. Upload the .fcs 3.0 files in the analysis software.
2. Verify the compensation setting used during cell sorting, using the single staining files recorded before the actual sorting.
3. Set the revised gating strategy using the files corresponding to the different bulk. Copy and paste those gates on the index sorting files.
4. Check that the index sorted cells fell in the set gate. If there are some sorted cells that were incorrectly gated, they can be identified by exporting the plate coordinates recorded during index sorting, and removed later in the analysis.
5. Export the event from the index sorting files as compensated parameters. Export them as .csv files, ticking the options "scale values" and "compensated parameters". These files should be exported in a folder named "Exported files".
6. Combine all the files in a single .csv file, using the script in Supplemental File 1. Set the right path with the function "setwd". The output of this script is a spreadsheet containing all the differently gated events and the relative intensities for all the parameters.

5. After culture antibody staining

NOTE: Perform this part of the protocol in sterile conditions; several reagents are shared with the former steps, and need to stay sterile. For the flow cytometry analysis, use a flow cytometer with a plate reader. This allows performing the staining directly in the tissue culture plate, reducing the cell loss to a minimum by limiting the amount of pipetting and spinning. Prepare the surface marker single color staining using the compensation beads, except for wells A1-A4, which represent the single staining for the CF/CV/VC/VI colors and are already present in the 96-well plate. The bulk populations sorted according to the cell dye help setting the gating strategy for the number of divisions and the general gating. Time required: 120 min.

1. Before starting the protocol, mark the wells that contain at least one cell by checking the plate under an inverted microscope. This step allows to optimize the amount of antibody used for the staining and speeds up the procedure.
2. Prepare the antibody mastermix, according to Table 3. As there is a significant amount of pipetting, the table considers the technical error due to pipetting, including an extra 5% volume. The antibodies described in the table allow to characterize a range of HSPCs from human cord blood samples¹².
3. Centrifuge the plate for 5 min at 300 x g. Rapidly invert the plate under the hood and over a paper towel, to remove the supernatant.
4. Add 8 µL of staining buffer to wells A1-A4. Add 8 µL of the mix to the other wells.
5. Mix the negative beads and IgG compensation beads in a 1:1 ratio, for a total volume equivalent to 120 µL. Dispatch 20 µL in one 1.5 mL tube per each marker. Add the antibody volume corresponding to the dilution factor (e.g., if the dilution factor is 1:20, add 1 µL).
   NOTE: Adapt the total volume to the number of markers used for the staining (e.g., 100 µL if the staining panel contains five antibodies).
6. Incubate the plate and the single staining compensation controls at +4 °C, for at least 30 min.
   ​NOTE: The incubation time must be adapted to the technical details of the antibodies used for the staining.
7. Wash the beads with 1 mL of staining buffer. Transfer the total volume to the 5 mL polypropylene tubes previously labeled. Centrifuge the tubes for 5 min at 300 x g, then remove the supernatant via aspiration.
8. Wash the cells in the plate by adding 100 µL of staining buffer per well using a multi-channel pipette. Centrifuge the plate at 300 x g for 5 min, then rapidly invert the plate under the hood and over a paper towel, to remove the supernatant.
9. Re-suspend the cells in 85 µL of staining buffer, using a multi-channel pipette.
10. Start the analysis on the flow cytometer (Acquisition mode), using the dedicated template and clicking on custom. This customized template takes into consideration the technical features of the 96-well round-bottom plate, particularly the dimensions of each well (diameter, depth and thickness). The probe must reach the bottom of the well, so set it at the exact centre of wells A1 and H12.
11. After selecting the fluorophores of interest from the list proposed by the software, set the plate setup following the plate template of Table 2, corrected for the number of wells containing at least one cell.
12. Select 100 µL as the acquisition volume limit. Tick the agitation option. Set the acquisition rate to 1 µL/s max, as lower speed improves the total volume analyzed per well.
13. Add the appropriate cleaning and washing solutions to wells H8-H12. The template in Table 2 specifically leaves wells H8-H12 empty, as the flow cytometer needs to run a range of wash conditions at the end of the analysis.
    NOTE: This step is adapted to the specifics of the flow cytometer used.
14. In the plots and gates section, first set the single-cell gate, using the FSC-A/SSC-A scatterplot then the FSC-H/FSC-A scatterplot. Create a histogram for each marker of interest.
15. Once the settings are confirmed, proceed to the Analysis section. Analyze the single staining first, recording no less than 5,000 events (optimal range: 5,000-15,000 events), both for the compensation beads and the CD34- stained fractions. Adjust the voltages if required.
16. Once the single stainings are all recorded, it is possible to start the actual acquisition, clicking on the Acquisition function.

Table 3: Antibody mastermix for a flow cytometry experiment, specifically for the identification of HSPCs from human cord blood. Please click here to download this Table.

6. Post-culture flow cytometry data analysis

NOTE: The data analysis described is specific for the software mentioned in the Table of Materials. The main output is the generation of a spreadsheet containing information for surface marker intensity, number of divisions, and kinship per each analyzed cell. Included in this part of the protocol is a script written in R, necessary to this workflow to generate the final analysis spreadsheet.

1. Export files from the flow cytometer as .fcs files. Upload them to the analysis software, grouping them together as "single staining", "bulk", and "single cell".
2. Prepare a compensation matrix using the single staining files, and apply it to the other two groups by drag and drop.
   NOTE: If an automatic compensation tool is used, check the quality by hand before proceeding further.
3. To have a representative gating, concatenate the different bulk wells in a single file. This step rapidly highlights if two colors are overlapping (typically CV and VC) or other anomalies, and therefore need to be excluded. After clicking on the concatenate populations option, select all uncompensated parameters from the "parameters" menu, then click on concatenate.
4. Upload the concatenated file to the workspace, then apply the compensation matrix via drag and drop.
5. Prepare the gating strategy defined in Figure 2 using the concatenated file. In the single-cell gate, display the events on a scatter plot with CFSE and CTV. Create a first gate called Labeled, including all four colors and excluding possible auto-fluorescence (Figure 2C). Then, gate each color individually.
6. Cells labeled with CV and VC need a transformed value, considering that the color is the result of the CFSE and CTV signals. The two coordinated signals are therefore rotated on a logarithmic scale by 45°, to allow the division dilution to proceed in parallel to the x-axis. This transformed value is derived manually, clicking on Tools and then Derive Parameter. Paste the following formula into the formula box:
   
   NOTE: The equation²⁶ assumes that CFSE and CTV are parameters 03 and 17.
7. To correctly visualize this novel parameter named Derived Parameter, set a linear axis ranging ~3-7, clicking on the Axis parameter option and selecting Customize Axis.
8. Apply the gating to each color individually as a histogram plot: for CF and VI, set CFSE-A and CTV-A on the x-axis, respectively. For CV and VC, set the newly Derived Parameter on the x-axis. Set gates corresponding to each peak, as displayed in Figure 3.
9. Apply the gating to each individual single-cell well. Make sure to add the derived parameter to every analyzed well. Manually verify each color gate for each well, to detect events that are incorrectly assigned to a given peak. Examples of gating are displayed in Figure 4.
10. After the analysis is completed, and all wells have been verified, select all the CF/CV/VC/VI gates that contain at least one cell. Export them as .csv files, ticking the options "scale values" and "compensated parameters". These files are exported in a folder named "Exported files".
11. Combine all the files in a single .csv, using the R script in Supplemental File 1. Remember to set the right path with the function "setwd". The output of this script is a spreadsheet containing all the different gated events and the relative intensities for all the parameters.
12. Open the spreadsheet and rename the columns for each parameter, for example, using the following names: CFSE, CTV, CD90, CD123, CD45RA, CD34, CD38. These names will be used to identify the gating threshold to correctly assign to each cell their identity.
13. Add six columns named "Well", "Condition", "Color", "Generation", "Original_cell", and "Culture_time". These variables are the ones defined experimentally and are inferred from each row:
    export_A10 CD34 ⁺ PBS_CV_Peak 1.csv.1 = A10 (well), CD34⁺ (Original_cell), PBS (Condition), CV (Color), Peak_1 (Generation).
14. Export the bulk wells to identify the threshold values for gating: export the compensated population of interest (e.g., CD34⁺CD38^–) as .csv files, ticking the options "scale values" and "compensated parameters." Export these files in a folder named "Exported files".
15. To find the threshold for CD38, identify the greatest numerical value for this parameter. Conversely, to find the threshold for CD34, identify the smallest numerical value for this parameter. Repeat this process for all the parameters of interest.
    NOTE: For the analysis presented in the protocol, the marker CD45RA is used both to identify LMPPs in the CD34⁺CD38^– gate and CMP/GMP in the CD34⁺CD38⁺ gate. This means that two different threshold values need to be extracted for this marker.
16. Copy and paste the threshold values in an excel file called "gating_matrix". This file is organized according to Table 4, and allows the analysis of multiple independent experiments. It is very important to name each column exactly with this scheme: XXYYMMDD_xxh, where XX stands for the two initials of the operator, YY the last two numbers for the year, MM for the month, DD for the day, and xx for the analysis time point.

Table 4: Gating matrix for the cell fate assignment, before the statistical analysis. The CD45h refers to the intensity of CD45RA for the HPC subset gating, while CD45l refers to the intensity of CD45RA for the CD34⁺CD38^– subsets. Please click here to download this Table.

7. Statistical analysis

NOTE: The statistical testing of the data generated involves a custom-made analysis pipeline, coded using the programming language Python (Supplemental File 2, Supplemental File 3, and Supplemental File 4). The script is organized in three blocks: the first one for processing the spreadsheet, the second block to generate the heatmap for data visualization, and the last block to generate multiple histograms to analyze and test differentiation and division properties.

1. Starting from the block "0_process_data" (Supplemental File 2), ensure that the gating_matrix and the data spreadsheet paths are correctly defined in the script.
2. Define the "cell_cols" dictionary, to assign the relevant cell fates to each cell. In the specific case, the fates are HSCs, MPPs, LMPPs, common myeloid progenitors (CMPs), granulo-monocytic progenitors (GMPs), megakaryocytic-erythroid progenitors (MEPs), and CD34^–.
3. Using the threshold values defined from the bulk wells (step 6.16), define the "cell_class_exp_time" function. It is essential to be consistent in the column naming, to correctly define these thresholds, using the same name used to define each column in step 6.12.
4. The cell phenotypes are defined in the script using a series of "if-else" statements, based on the thresholds detected during the flow cytometry analysis.
   NOTE: Different phenotypes can be displayed by modifying these statements to accommodate other marker combinations.
5. Specify the experimental specific conditions, using the function "cond_rule" (e.g., different experimental treatments). For the dataset provided, the conditions are named "GT" and "Diff". Describe the two different cell culture media used to culture the cells. This information will be used by the block "1_dot_plot" (Supplemental File 3) for plotting the heatmap.
6. In the block "2_bar_plot" (Supplemental File 4), define the dictionary "class_dct", including the discrete cell fates of interest. For the dataset provided, the cell fates of interest are the same described for the "cell_cols" dictionary.
7. Define "conds" (conditions), "or_cells" (original cell), "sym_labs" (symmetry labels), and "times" (the experimental time point). These are reiterating filters necessary for plotting. "conds" take the conditions defined in "cond_rule" again, "or_cells" are HSCs and MPPs, and "sym_labs" describe the type of divisions.
8. In the block "2_bar_plot", it is possible to plot cells that have progressed up to division 6.
   NOTE: The dataset provided only includes cells up to division 4, so an error message pops up, but this does not prevent the script from working.
9. The figures generated by the script can be retrieved in the folder named "figures" as pdf files. The files named "Test" represent the different statistical tests performed for the corresponding histogram.