Caspase activation is often considered to be one of the most characteristic features of apoptosis (1–7). Several methods have been developed to detect this event. Because the activation involves cleavage of the zymogen procaspases (reviews, 4–8) the low molecular weight cleavage products can be revealed electrophoretically and identified on Western blots. Another approach utilizes the fluorogenic or chromogenic caspase substrates: nonfluorescent or colorless peptide substrates were developedthat upon cleavage generate fluorescing or colored products (9–12). Another approach that can be applied to study activation of caspases in situ is based on immunocytochemical detection of the epitope that is characteristic of the caspases' active form. Antibodies that react only with the activated (cleaved) caspase form have recently become available and have been used in the cytometric assay (13, 14). Activation of caspases can also be detected indirectly, by immunocytochemical identification of the specific cleavage products such as the p85 fragment of poly (ADP-ribose) polymerase. This method has also been adapted to cytometry (15, 16).
Recently, biotin- (17), or fluorochrome- labeled inhibitors of caspases (FLICA) (18–20), have become available as probes of caspase activation. These reagents were designed as affinity ligands to react covalently with the reactive enzymatic center of activated caspases. During the past two years several articles have been published (19, 21–23), including the papers authored by us (18, 24, 25), in which these reagents have been used to probe for caspase activation. Based on a similar principle, other probes were developed to detect activation of intracellular serine proteases (26).
We have recently carried out additional control experiments to further assess the specificity of caspases detection with fluorochrome-tagged inhibitors. These experiments revealed that binding of these probes, while highly specific to apoptotic cells, may involve mechanisms other than the ones initially anticipated. Results of these studies, which pertain to the interpretation of the binding specificity of these reagents vis-a-vis the active enzyme centers of caspases, are presented herein.
- Top of page
- MATERIALS AND METHODS
- LITERATURE CITED
FLICA were designed as affinity ligands of caspase active centers for the detection of caspase during apoptosis. Their mode of binding was presumed to involve covalent attachment to the cysteine of the active center through the fluoromethyl (FMK) reactive group (18–20). A caspase target sequence was incorporated to provide specificity vis-à-vis an individual caspase, while the fluorochrome (FAM, FITC, sulforhodamine SR) was attached as a fluorescent reporter of their presence/location within the cell. Unlike the DEVD FLICA (FAM-DEVD-FMK) that has a four–amino acid peptide moiety, the three–amino acid VAD FLICA (FAM-VAD-FMK) lacks specificity to individual caspases and therefore is expected to interact with all caspases, perhaps with an exception of caspase-2 (39, 40). As mentioned in the introductory section, these reagents have found utility as either markers of apoptotic cells or as probes of caspase activation (18–26). In the early studies, we observed that cell pretreatment with the unlabeled inhibitors (e.g., z-VAD-FMK) at the time of induction of apoptosis prevented subsequent labeling with FLICA to a large degree (41). It was also observed that FLICA by itself was sufficient to arrest the progression of apoptosis (24). This would suggest that these reagents inhibit caspase activity. Binding of FLICA to proteins of approximate molecular weight to that of the large caspase subunits, detected by Western blotting (42), provided additional evidence of their interactions with active caspases.
During the course of working with these reagents, it became apparent that certain intracellular binding events could not be explained solely by their specific binding to active enzyme centers of caspases. This prompted us to better-characterize these reagents and correlate their binding with other events and probes of apoptosis—the subject of the present study. In the discussion that follows, we first characterize the utility of FLICA as markers of apoptosis as well as detectors of caspase activation and then address peculiarities in their binding properties.
FLICA are specific, sensitive, and convenient-to-use probes for the detection of apoptosis. The sensitivity of the FLICA assay stems from the fact that it detects very early apoptotic stages, the stages that precede loss of plasma membrane integrity, DNA fragmentation, chromatin condensation (Figs. 2 and 5), and externalization of phosphatidylserine (detected by annexin V binding; Fig. 4). Therefore the time-window through which the apoptotic process is detected by the FLICA assay is distinctly wider than by most other methods. Because upon apoptosis induction cells asynchronously enter into and progress through the apoptotic process (43), the wider the detection time-window of the assay, the greater is its sensitivity (ability) to estimate the incidence of these apoptotic events.
As shown in our previous studies (25), a combination of FLICA binding and the assay of plasma membrane integrity based on exclusion of PI reveals three distinct stages of apoptosis. This combination is of particular utility in discrimination between apoptosis and cell necrosis (44). Some of the FLICA assay's other virtues are convenience and protocol simplicity. Incubation of the cultured cells with the reagent, followed by washing out of its unbound fraction is all that is required to label the cells. An additional feature associated with the use of the FLICA assay is that within a single measurement we can combine the ΔΨm collapse, binding of FLICA, and exclusion of PI (Fig. 3) (45).
Caspase activation appears to be a prerequisite for FLICA binding. In our previous observations, we demonstrated that in all instances when cell death occurred by necrosis, without caspase activation, the cells did not bind FLICA (25, 44). In the present study, we see a good correlation between caspase-3 activation and FAM-VAD-FMK binding among the early apoptotic cells that show condensed chromatin, but are not yet deficit in DNA content (Fig. 2). In the case of late apoptotic (sub-G1) cells, however, activated caspase-3 becomes immunocytochemically undetectable while the cells still remain FLICA-positive (Fig. 2).
While there is evidence that FLICA binds to proteins of molecular weight 18–23 kDa (42), which is in concordance with the molecular weight of large subunits of caspases (4–6), there are certain puzzling observations that cannot be easily explained by the initially proposed mechanism of the binding, i.e., through interactions with the caspase active center only. Thus, if caspase active centers were the only binding sites for these reagents, one would expect a significant level of protection of these sites by the unlabeled caspase inhibitors (z-VAD-FMK, z-DEVD-FMK) via competitive binding. Yet when the latter were added 1 h prior to FLICA addition, even at greater concentrations than the subsequently-added FLICA probes, no significant protection was apparent (Fig. 6; Table 1). It should be noted as well that no protection by z-VAD-FMK was observed when biotinylated caspase inhibitors rather than FLICA were subsequently used as markers of caspase activation (17). Likewise, the unlabeled inhibitors did not prevent the cleavage of fluorogenic caspase substrates in the experiments reported by Komoriya et al. (11). However, when z-VAD-FMK was added along with the inducer of apoptosis, it prevented subsequent binding of FLICA (41), biotinylated inhibitor (14), or substrate cleavage (17). Of course, when z-VAD-FMK was added at the point of induction, all other events of apoptosis were prevented as well (14, 41). Thus, it appears that only during the induction of apoptosis can the unlabeled inhibitor interact with the caspase binding sites, preventing subsequent binding of labeled inhibitors or cleavage of the substrate. This phenomenon is difficult to explain under the assumption that labeled and unlabeled inhibitors (or fluorogenic substrates) compete for the same binding sites.
Another event that is incongruous with the assumption that FLICA binds specifically to the active center of caspase enzymes is the inability of FAM-VAD-FMK to stop the cellular progression of apoptosis. Although in earlier studies we observed that FAM-VAD-FMK was able to arrest cells in apoptosis (24), we were unable to confirm these observations with newer batches of the reagent, even when FAM-VAD-FMK was added 1 h prior to the inducer of apoptosis (not shown). It is possible that early batches of FAM-VAD-FMK contained some unlabeled z-VAD-FMK. The presence in our FLICA reagent of this unlabeled inhibitor could have been responsible for the observed arrest. We also noticed in these earlier studies that the degree of arrest of apoptosis was related to particular batches of serum used in culture medium.
A note of caution should also be added regarding specificity of caspase inhibitors in general, whether they are labeled or unlabeled. Based on their tetra-peptide amino acid sequence these inhibitors are often considered to be specific for individual caspases. This may not be the case, however, particularly when using these inhibitors to treat live cells. For example, the inhibitor with DEVD sequence is designed to be caspase-3 specific. However the inhibitory constant (Ki) of Ac-DEVD-CHO is 0.2–2.2 nM for caspase-3, 0.9 nM for caspase-8, and 1.6 nM for caspase-7 (39, 40). However, to treat live cells these inhibitors are being used at four orders of magnitude higher concentration than their Ki. We have noticed that MCF-7 cells known to be caspase-3–null were labeled with FAM-DEVD-FMK (20, 42). This would indicate that perhaps caspase-7 and caspase-8, if not also other caspases, reacted with FAM-DEVD-FMK in these cells. Other inhibitors also have strong affinity to more than a single caspase (39, 40). Moreover, since little is known about local concentration of the used FLICA within live cells and also about their in situ binding constant to the respective caspase, it is uncertain how specific the binding is.
Unexpectedly, we observed that the fluorochrome moieties of FLICA, either free fluorescein or carboxyfluorescein (FAM), become retained in apoptotic cells. Their mode of binding, however, appears to be different than that of FLICA because the binding is limited to PI-negative cells only. Furthermore, in contrast to FLICA, fluorescein and carboxyfluorescein are not retained within the cell following fixation with formaldehyde, indicating that these associations are probably noncovalent in nature. This fluorochrome binding to apoptotic cells may be simply the result of hydrophobic interactions between the hydrophobic fluorochrome and newly-presented hydrophobic regions on intracellular proteins generated during the apoptotic process. When we compared fluorescein binding to that obtained when a 6-carbon hexanoic acid spacer was added to the fluorescein molecule, the latter more-hydrophilic entity showed reduced apoptotic retention (data not shown). In light of these observations, one cannot exclude the possibility that some hydrophobic interactions from the fluorochrome moieties of FLICA may also contribute, to a minor degree, to the overall retention of FLICA molecules in apoptotic cells.
In conclusion, the present data demonstrate the usefulness of FLICA as a convenient and proven marker of apoptosis, and most likely as a probe of caspase activation. However, in light of some peculiarities in their binding features, as discussed above, one should not interpret their binding by apoptotic cells as evidence of their exclusive interaction with the caspase active centers.