Apoptosis plays a major role in various physiological and pathological processes. Cell demise has been investigated in different fields, also including extracorporeal photopheresis (ECP), a therapeutic intervention used in several auto immune disorders, diseases with pathogenic T cell involvement, and chronic graft-versus-host disease (GvHD) (1). Collected cells are reinfused after in vitro exposure to 8-methoxypsoralen and ultraviolet A (UVA) light. One known effect of ECP on PBMC is the induction of apoptosis (2, 3), which can be analyzed by FCM, a generally accepted and reliable tool for detection of apoptosis on single-cell level (4, 5). Several assays for FCM analysis of different stages of programmed cell death are commercially available but all of them include at least one centrifugation step. Hypothesizing that, after apoptosis-inducing treatment, part of the PBMC will be disrupted by centrifugation and thus will not be present for FCM analysis, our intention was to establish a single-platform, no-wash, and multicolor FCM methodology to quantify apoptotic T cells in ECP products.
To discriminate viable from early and late apoptotic/necrotic cells, we tested several markers. We favored the YO-PRO-1®/DAPI combination which allows discrimination of apoptotic from necrotic cells. YO-PRO-1, a DNA-intercalating dye excited at 488 nm, is a reliable marker for FCM analysis of apoptosis (6, 7). DAPI, a nuclear dye excited at 405 nm, which is commonly used in FCM (8), was applied to address late apoptotic/necrotic cells. Camptothecin treatment of Jurkat cells, a popular concept for inducing apoptosis (9, 10), was employed to establish the methodology.
The aim of this study was to identify and quantify apoptotic T cells in ECP products. No cell washing was performed during the whole period, from apheresis over cell culture until cell analysis. In contrast, cell washing was performed with one centrifugation step just before cell staining. Therefore, the comparison between wash and no-wash procedures on each ECP sample was analyzed. Afterwards, the numbers of cells undergoing apoptosis were calculated as cells per μl using Trucount™ beads (11, 12) as internal standard.
We demonstrate that the number of apoptotic cells per μl is clearly influenced by centrifugation both in the Jurkat cell line and in ECP cell samples.
MATERIALS AND METHODS
Induction of Apoptosis
Jurkat ACC 282 cells were cultured in RPMI 1640 (Gibco/Invitrogen, San Diego, CA) containing 10% FBS, 1% L-Glutamine, and 0.5% Antibiotic/Antimycotic (PAA, Pasching, Austria). Apoptosis was induced with 5 μM camptothecin (Sigma-Aldrich, St. Louis, MO) which was added to the cells adjusted to 0.5 × 106 per ml. Jurkat cells in medium without camptothecin served as controls. Cells were incubated at 37°C in a 5% CO2 humidified atmosphere. After 6 and after 8 h, cells were harvested and prepared for FCM analysis.
ECP was performed with the UVAR XTS (Therakos, Exton, PA), an inline instrument in which apheresis and irradiation are performed automatically. Parameters were adapted to the patients' blood volumes, the venous hematocrit, and the clinical performance. The amount of Psoralen (8-MOP; Uvardex™ Therakos, Exton, PA) is calculated based on the volume of the respective apheresis product. An algorithm computes the time of irradiation which is determined to radiate a dose of 2 J/cm2. Samples of the final product were drawn into EDTA tubes.
Culture of Leukocytes After Extracorporeal Photopheresis
The number of WBC per μl was determined with a Sysmex KX21 hematology analyzer (Sysmex, Hyogo, Japan). On the basis of this count, cells were adjusted to 1 × 106 leukocytes per ml with RPMI 1640 supplemented with 10% FBS, 1% L-Glutamine, and 0.5% Antibiotic/Antimycotic. This suspension was plated into 96-well plates (round bottom; Nunc/Thermo Fisher Scientific, Waltham, MA) at a final concentration of 0.1 × 106 leukocytes in 200 μl per well. Aliquots for immediate FCM analysis were taken aside; the other cells were incubated at 37°C in a 5% CO2 humidified atmosphere. Every 24 h cells were harvested (one well per analysis tube) and prepared for FCM. This was done on 3 consecutive days.
Labeling: Wash and No-Wash Preparation
The Single-Platform (SP) approach was used throughout to quantify cells by flow cytometric analysis. The difference between the two cohorts was that, in the wash group, cells were washed once, i.e., before cells staining. Cells in the no-wash group were not washed during the whole experiment, i.e., after induction of apoptosis which, for the Jurkat cells, was the addition of camptothecin to the culture. Cells of the apheresis products were not washed at all, from patient to cytometer. Our aim was to quantify the differences in the numbers of apoptotic cells caused by one single-cell washing step. All samples were prepared and acquired in duplicates.
Cells were gently harvested and, for no-wash preparation, transferred into Trucount™ tubes (BD Biosciences, San Diego, CA). For the wash preparation, the same volume was transferred into 5-ml round bottom tubes (BD Biosciences) and washed with PBS (Gibco/Invitrogen, San Diego, CA) supplemented with 2% FBS and 1% antibiotic Antimycotic (PAA), at 400g for 10 min at 4°C, before the supernatants were cautiously decanted. The pellets were resuspended in PBS as earlier, adjusted to a feasible staining volume (100–300 μl), and carefully transferred into Trucount tubes.
The mAb cocktails had been pretitrated for best performance to stain Jurkat cells and PBMC. Working concentrations of the nuclear dyes had been adjusted for optimal staining of camptothecin-treated Jurkat cells and of starved PBMC.
Staining for both preparations was performed using YO-PRO-1 (16 nM; Molecular Probes, Invitrogen, San Diego, CA) for detection of apoptosis. Jurkat cells were addressed using anti CD 7 APC mAb (BD Biosciences). To define leukocytes and subpopulations in ECP samples, a multicolor cocktail was used comprising CD56 PE, CD3 PE-Texas RED, CD 45 PerCP, CD 14 APC, and CD 19+20 APC-Cy7 (all from BD Biosciences). YO-PRO-1 and mAbs were incubated for 20 min at room temperature in the dark before DAPI (0.03 μg/ml; Sigma-Aldrich, St. Louis, MO) was added as a marker for cells referred to as nonviable. DAPI is a highly selective DNA dye, intercalating into the DNA of cells with impaired membranes. Immediately after staining, cells were acquired on a BD™ LSRII flow cytometer.
Flow Cytometric Analysis
Data were compensated using positive controls: for Jurkat cells, manual compensation was performed with camptothecin-treated cells as described earlier; for apheresis products, we used leukocytes starved for 2 days. To discriminate between negative and positive events in the analysis, a nonstained control sample from each culture condition always accompanied acquisition of the stained cells to define their cut off. The LSRII Cytometer was equipped with three lasers (solid state lasers 405 nm and 488 nm, He-Ne laser 635 nm). YO-PRO-1 and DAPI were excited at 488 nm and 405 nm, and band pass filters for light emission were 530/30 and 450/40, respectively. The cytometer performance is checked weekly using SPHERO™ Ultra Rainbow fluorescent particles (8-peak-beads; Spherotech, Libertyville, IL).
For Jurkat cells, an acquisition threshold of 5,000 was set on SSC. For leukocytes, the threshold was set on SSC (800) and CD45 PerCP (5,000). At least 50,000 total events and a minimum of 1,000 beads were acquired per analysis.
Data Evaluation and Gating Strategy
The BD FACSDiVa™ software (Version 5.0.3.) was used for data evaluation. In all analyses, we first defined the beads and inverted this gate to continue with everything but the beads. For discrimination of doublets, cells were gated on SSC-area versus FSC-width. As our aim was to identify apoptotic cells in a defined cell population (Jurkat cells, T cells), the specific lineage antigen was one main criterion, gated in the respective dot plot versus SSC. After this, we addressed viable, apoptotic, and apoptotic/late necrotic cells of the relevant target population.
Apoptotic Jurkat cells were identified by their CD7 expression. CD7+ events with high SSC properties were defined as nonspecific and excluded in the SSC versus FSC dot plot. The CD7 positive population was then analyzed in the YO-PRO-1 versus DAPI dot plot. This enabled clear discrimination between apoptotic (YO-PRO-1+/DAPI−), late apoptotic/necrotic (YO-PRO-1+/DAPI+), and viable (YO-PRO-1−/DAPI−) cells.
In ECP samples, T cells were defined as CD45+CD3+ CD56−. Both CD45+ and CD45low cells were accepted as CD3+ cells, because we observed that the intensity of CD45 expression is decreasing in the process of T cell demise. CD45+/low CD3+CD56− cells that showed positivity for CD19+20 or CD14 were defined as false positive and excluded. The CD3+ population was then gated in a plot SSC versus YO-PRO-1. CD3+YO-PRO-1− cells were defined as viable. This was confirmed by back gating to an SSC versus FSC dot plot. CD3+YO-PRO-1+ cells were analyzed in a plot DAPI versus FSC to distinguish between apoptotic (YO-PRO-1+DAPI−) and late apoptotic/necrotic (YO-PRO-1+/DAPI+) cells.
For definition of apoptotic cells in cultured apheresis products, we preferred contour plots to display the cell distribution. Viable, apoptotic, and late apoptotic/necrotic cells were back gated to an FSC versus SSC dot plot in which they could be clearly identified as apoptotic cells due to their lower FSC and higher SSC properties.
For calculation of cell counts per μl, the following formula based on bead number and sample volume was used: Cells per μl = number of cell events × number of beads per tube/number of bead events × sample volume (μl).
Since the cell numbers were not normally distributed for descriptive statistics, the median and minimum/maximum values were used, and the differences between the methods were calculated using the Wilcoxon rank sum test. The differences between the methods were described using the Bland and Altman Plot in which we included the regression curve analysis. WINSTAT (2009 version) was employed for the calculations.
YO-PRO-1/DAPI is a Reliable Combination for Measurement of Apoptosis
Jurkat cells treated with camptothecin and extracorporeal apheresis (ECP) products after 3 days of culture were simultaneously stained with different dyes known to detect apoptotic and late apoptotic/necrotic cells: We used Annexin V (FITC channel) versus PI or versus 7-AAD or versus DAPI and, alternatively, stained YO-PRO-1 (FITC channel) versus the three dyes, respectively. The disadvantage of 7-AAD and PI is that both dyes occupy Channel 3, or Channels 2 and 3, which are thus not available for antibodies, whereas DAPI is excited with the violet laser and emits light at about 450 nm. We obtained very similar results when comparing the combinations Annexin V versus YO-PRO-1 and DAPI versus YO-PRO-1 (Fig. 1). The latter combination in which DAPI served to define dead cells was therefore used for all subsequent multicolor analyses.
Wash and No-Wash Cell Preparation to Analyze Apoptosis in the Camptothecin/Jurkat Model
The camptothecin/Jurkat model was used to establish the single-platform analysis technique to quantify apoptotic cells. In virtually all reports, apoptotic cells are stained using a method that includes cell washing, and results are expressed as percentage of all cells analyzed (13). Our comparison shows dramatic differences between the wash and the no-wash procedures (Figs. 2A and 2C): In four independent experiments, the cell analyses performed 6 and 8 h after camptothecin treatment showed that the differences between the percentage values were significantly smaller (P = 0.2; Wilcoxon Test) than those between cell numbers per μl (P = 0.02), which reflects the selective loss of apoptotic cells due to cell washing (Fig. 2).
Wash and No-Wash Cell Preparation to Analyze Apoptosis in ECP Products
ECP products were cultured and prepared for FCM analysis following multicolor staining as described in “Material and Methods.” With the gating strategy employed (Fig. 3), CD3+ T cells could clearly be addressed as viable (Yo-PRO−/DAPI−), apoptotic (Yo-PRO+/DAPI−), and necrotic (Yo-PRO+/DAPI+).
As described earlier for the Jurkat cell line, a serious loss of apoptotic cells due to cell washing was also observed in the cultured ECP products (n = 8 independent experiments). When examined separately, results from both the wash and the no-wash procedures (Fig. 4) were highly reproducible and correlated significantly (Y = 0.9985* × −1.2511, R2 = 0.9542, and Y = 0.9887* × −0.9188, R2 = 0.9923, respectively). However, the Wilcoxon rank sum test revealed a significant P value of 6.6 × 10−5. Moreover, comparison of both methods using the Bland and Altman calculation (Fig. 5) showed a strongly increasing difference between the results (higher cell loss with increasing cell numbers in the washing method). These differences were highly significant as indicated by the regression analysis resulting in an R2 value of 0.7624. These results confirm our hypothesis that one washing step can dramatically change the analysis results, and they support our approach to establish a no-wash/single platform methodology for the analysis of apoptosis.
We present a new single-platform procedure to quantify apoptotic cells by multiparameter FCM. Many different markers to detect distinct time points of apoptotic pathways are available for detection of apoptosis (14, 15). However, all of the methodologies offered on the market include at least one step of cell washing which, according to our experience, is not in line with quantitative cell analysis. Our assay does not use any cell washing step. Results are thus obtained within a shorter time, and they are very likely to be more accurate. The latter is concluded from our comparison with a cell washing procedure which yields significantly lower results with regard to apoptotic cells (Figs. 2A and 2C).
Using the conventional analysis methodology that includes cell washing, the results are expressed as percentage values. It is generally accepted that cell washing coincides with cell loss, and we assume that this loss is higher for apoptotic than for vital cells. This is supported by our finding that the differences between percentage values are clearly smaller than those between absolute values for apoptotic cells (Fig. 2).
Several Authors used the SP methodology to quantify apoptotic cells. In contrast to our approach, however, they applied cell washing steps at various time points during cell culture or preparation (6, 16, 17) or rinsed cells with PBS after induction of apoptosis (13). Our intention was to demonstrate that in particular apoptotic cells are lost by cell washing.
When starting to establish a single-platform procedure by adapting one of the existing kits, we had to exclude some of them, like the Caspase assay that requires intracellular staining and thus cell washing, or the Annexin V assay which depends on a Ca2+ environment known to interfere with EDTA anticoagulation.
As an alternate to Annexin V, a marker for the phosphatidylserines exposed during the early stages of the apoptotic process (18), we tested YO-PRO-1, another marker for early apoptosis (19) which emits light over the 530/30 nm band pass filter to the FITC channel. To address late apoptotic/necrotic cells, we employed DAPI excited with a violet laser and with a light emission of 450 nm. The advantage of this choice is that there is significant spectral overlap only to the second violet channel (530/30 long-pass filter). In contrast to other dyes like PI or 7AAD, the channels used for the blue and red lasers remain free and can be used for multicolor mAb applications. In our hands, discrimination between live, apoptotic, and late apoptotic/necrotic cells could be easily performed and yielded reliable and reproducible results for both, cell lines and primary human cells. Since there is an obvious need for more complex analysis procedures for different types of cells and a broader fluorochrome panel (20, 21), our approach supports the demand for an extended flow cytometric application.
To obtain PBMC, it is common practice to perform density gradients which require several centrifugation steps. Since we aimed to detect apoptotic T cells in ECP products on consecutive days, but wanted to omit centrifugation, we set up a whole-buffy-culture immediately after apheresis. This approach for culture of PBMC worked very well, because the nucleated cell number in the original products was sufficiently high (8–13 × 106/ml). Because of the relatively low number of red blood cells, erythrocyte lysis reported to induce apoptosis in lymphocytes (22) was not necessary in any of the experiments.
In the described multicolor staining for ECP products, we addressed T lymphocytes using CD3 and employed other markers to define NK cells, B cells, and monocytes. It is known that due to membrane loss several leucocytes downregulate surface lineage antigens with beginning apoptosis (23). This was not the case for CD3 which remained stably expressed (not depicted) and therefore served as a reliable marker also for apoptotic T cells, which showed decreasing FSC signals during demise (24).
Our present data strongly indicate that cells, particularly those in apoptotic degradation, are disrupted and lost during centrifugation. Certainly this is of great importance when dealing with samples from cell culture (both primary cells and cell lines). False results from apoptotic cell enumeration could not only influence subsequent experiments but might also have a negative impact if used for therapeutic treatment of patients by ECP.
In conclusion, the presented no-wash/single-platform flow cytometric approach to quantify apoptotic T cells yields reliable and reproducible results which differ significantly from those obtained after cell washing. On the basis of this knowledge, experiments that aim at the detection of apoptotic cells should be no longer performed with a method comprising cell washing, if quantitative results are important. As outlined, the single platform method is feasible and reliable and not only applicable for fresh blood but also for cultured cells.