Each electrophoretic (EP), immunohistochemical (IH), and flow cytometric (FCM) technique detects only one specific feature of the process of apoptosis. Gel- and pulse-field EP methods detect small oligonucleosomal DNA strand breaks (multiples of 180–200 bp; DNA laddering) and large DNA strand breaks (more than 30–50 kbp), respectively, which are supposed to be relatively late and early features of programmed cell death (PCD). Unfortunately, EP techniques are not applicable for quantification of PCD. In contrast, IH techniques like in situ end labeling (ISEL) and terminal deoxynucleotidyl transferase nick end-labeling (TUNEL) of DNA and FCM methods like Annexin-V (AnV)-fluorescein isothiocyanate (FITC), TUNEL, and propidium iodide (PI) can detect apoptosis of single cells.
AnV detects phosphatidylserine (PS) transposition on the outer plasma membrane, which occurs at a rather early stage of PCD during the so-called execution phase (1–6). This loss of membrane asymmetry develops downstream of the Bcl-2 checkpoint, after the disruption of the mitochondrial transmembrane potential, the release of apoptosis-inducing factor, and the activation of caspases (1). PS translocation precedes nuclear condensation, loss of membrane integrity (causing PI uptake), and cell shrinkage (6, 7). Whether PS transposition occurs before or after “the point-of-no-return” of PCD is still a matter of debate. ISEL is an IH technique described originally by Wijsman et al. (8) and modified successfully for plastic- embedded BM tissue by Mundle et al. (9). A mix of four nucleotides with DNA-Polymerase-I is used to detect specific 3′-OH ends of single-strand DNA breaks, which are found after endonuclease activation. Therefore, ISEL is detected in the phase of PCD beyond the point-of-no-return. We modified this ISEL technique and made it applicable for FCM (10, 11). We used it in combination with AnV/PI to study the kinetics of different features of PCD. Furthermore, we questioned if PS translocation under every circumstance means inevitable apoptosis by culturing sorted AnV+/PI− CD34+ cells in a single-cell single-well assay.
- Top of page
- MATERIALS AND METHODS
- LITERATURE CITED
In this study, we used Jurkat cells to observe the kinetics of three features of PCD by FCM. We also used CD34+ cells to measure apoptosis characteristics after cell handling and to investigate if AnV positivity after cell handling implied inevitable PCD. PS transposition of cells has been recognized as an evolutionary well-saved and ubiquitous marker of early PCD that is needed for cell engulfment by macrophages before their plasma integrity becomes compromised (5, 12). However, PS translocation is not unique for apoptosis because it was also observed during (secondary) necrosis (13). Therefore, we used the AnV-FITC assay with PI by time-lapse examination to differentiate among viable (AnV−/PI−), early apoptotic (AnV+/PI−), late apoptotic, and secondary necrotic (AnV+/PI++ or AnV+/PI+) cells (3, 4, 6, 7, 13–15). Furthermore, as we needed a method to distinguish between early and late PCD, we developed a modified ISEL technique for FCM (10, 11). ISEL is a variant of the TUNEL technique, but TUNEL is used much more frequently in research (16, 17). DNA-Polymerase-I used in ISEL was tested and it was able to detect DNA fragments generated by endonuclease activity in Jurkat and human BM CD34+ cells during PCD (18–21). To study apoptosis in time, we preferred the ISEL technique because it has been shown to be more specific for PCD than for necrosis compared with TUNEL (20, 22, 23). Furthermore, the TUNEL assay is prone to false positive and negative staining (15, 24). ISEL and PI detect specific DNA fragments and loss of cell membrane integrity, respectively, which are features of inevitable PCD as they occur after each other during the execution phase (1, 2, 16, 25–27). From our background of research with stem cells, we questioned if detection of AnV (within CD34+ cells) means that the point-of-no-return of PCD has been passed. This is still a matter of debate, although some evidence in favor of this hypothesis has been gathered (6, 28, 29).
In line with the above-mentioned considerations, we proved in all Jurkat experiments that membrane PS asymmetry precedes DNA fragmentation and/or membrane perturbation (Figs. 2, 3, and 5), as DNA fragmentation precedes cell membrane leakage (Figs. 4 and 5). Others (1, 4, 6, 13, 16, 17, 30–32) have obtained similar results with various techniques (AnV/PI, AnV/PI with TUNEL, AnV/TUNEL, different DNA binding dyes, morphology, comet assay, and laser scanning cytometry), but not in one setup of serial experiments in time combining three PCD-analyzing techniques with three apoptosis-inducing methods. In the experiment exposing Jurkat cells to CPT, PS expression increased after 2–3 h (AnV+/PI−) and was followed by ISEL increment by only a 1-h difference (Fig. 5). Others also found this relation in time with different techniques (28, 32–35) and it emphasizes the discrete line in time to pass the point-of-no-return regarding PCD (of this cell line under these circumstances). On the other hand, Bacsó et al. (32) proved within a CD95-induced Jurkat apoptosis model that virtually all the AnV+/PI− cells, which increased after 2 h, had apoptotic comets or remnants (as it also detects early 50-kb DNA fragments).
After carefully analyzing the process of PCD in time, an interesting pattern was observed. AnV+/PI−/ISEL− viable cells developed into AnV+/PI−/ISEL+ and subsequently AnV+/PI++/ISEL++ cells, representing a more progressive phase of apoptosis with the highest ISEL positivity. These cells turned into AnV+/PI+/ISEL+ cells, which are end-stage apoptotic and/or secondary necrotic cells, presumably characterized by more DNA disintegration, nuclear condensation, and leakage of cell and nuclear membranes causing less PI- and ISEL-positive staining (15). These results are comparable with the Nicoletti assay (3, 36) and with the report of MacNamara et al. (37) who studied HL-60 cells under similar conditions. They determined the PCD of these cells as a sub-G0/G1 peak on DNA histograms with forward and sideward scatter features reflecting cell shrinkage and the presence of apoptotic bodies.
The different PCD patterns of Jurkat cells after γ-irradiation, CPT, or Ara-C were explained by the detrimental action upon different cellular targets and by a different dose response. One should realize that the death of these cells is necrotic at high levels of insult, whereas PCD is induced at lower levels. For example, γ-irradiation promptly caused single and double DNA strand breaks, some of which were sublethal and could be repaired, some of which were lethal (causing PCD), and some of which were devastating to the cell (causing necrosis). This explains the initial combination of primary necrosis and apoptosis (of especially cells in S-phase; 38), in which the higher amount of necrosis with more PI positivity (higher than ISEL) gradually declined in order to make place for more apoptotic involvement (ISEL > PI). CPT, a DNA topoisomerase I blocker, induces DNA strand breaks of all cells (although DNA-replicating S-phase cells are more sensitive), whereas Ara-C, a pyrimidine antagonist, predominantly effects S-phase cells. Perhaps this could explain a slower increment in PCD by Ara-C compared with CPT (29, 33–35).
High numbers of AnV+/PI− of CD34+ cells within the life-gate occurred after CD34+ selection with immunomagnetic beads (D1: ±62%) and they also showed a high variability after rapidly freezing and thawing (mean AnV+/PI− of D1-D3: 33 ± 8%). It is a well-known and accepted phenomenon that a highly variable amount of stem cells is lost after CD34 selection and viable freezing and thawing. Cryopreservation in liquid nitrogen of mononucleated BM cells in 10% DMSO leads to absent trypan blue exclusion in approximately 10–15% cells and to a 25 ± 10% loss of stem cells and colony forming unit-granulocyte macrophages (CFU-GM; 39). What are the characteristics of the CD34+ cells within the life-gate regarding cell function versus apoptosis after these cell handling procedures? Cryopreservation of hematopoietic stem cells leads to high and variable AnV positivity (range 5–70%), but whether these cells are destined to die remains to be proven and should not be presumed (40). Membrane alteration or damage could be triggered by the handling during the antibody-coupled immunomagnetic bead selection or could be induced by controlled freezing (DMSO should prevent crystal formation) and rapidly thawing (DMSO can cause osmotic shock) of these cells (41). The big difference in percentage between total AnV+ and ISEL+ CD34+ cells (between 20–55%) in combination with an unchangeable low percentage in ISEL+ cells (±5%) after CD34 selection and after thawing argues strongly for a temporary membrane alteration and not for PCD. In our experiment, at least some of these AnV+ cells repaired their membrane activation-alteration and/or damage during incubation in medium (approximately 6–8% in D1), whereas other cells followed their path of PCD as ISEL was increasing. Furthermore, at least 8% (8 of 96 wells) to 15% (14 of 96) of these thawed AnV+/PI− CD34+ cells (AnV++ fraction) had proliferative capacity. On the other hand, these cells represented at least 30% (8 of 27 wells) to 78% (14 of 18 wells) of the normal plating efficiency found within these two controls in this SCSW assay. Normal plating efficiency (using AnV- CD34+ cells) is defined as the number of wells showing proliferative capacity. In contrast, the CFU-GM capacity (to form colonies of more than 40 cells as being granulocytes and macrophages) was strongly and inversely correlated with the AnV intensity of these PI− CD34+ cells. Therefore, AnV should not be used as a marker or as the only marker of apoptosis in experiments with stem cells in which physical membrane activation or alteration may be expected. PCD should be proved by distinct morphological features or by FCM techniques that detect DNA fragmentation products (15, 32). From these experiments, we conclude that in some experimental settings membrane activation and/or a low apoptotic insult was involved in causing PS exposition in a commitment phase to apoptosis. This phase of “preapoptosis” has been shown to be caspase independent and reversible if the strength of the stimulus is low and of short duration (26). A substantial fraction of AnV+/PI− BM CD34+ cells after different cell handling procedures has certainly not passed the point-of-no-return in the process of PCD, as the cells retained proliferative capacity, although to a lower extent. In analogy, cryopreservation and thawing of human spermatozoa (42) were associated with the induction of membrane PS translocation and high post-thaw levels of AnV were found even in the fractions with high sperm motility.