CASPASE ACTIVATION

Caspases comprise a family of cysteine proteases that cleave on the C-terminal side of aspartate residues.^57-59 Several observations have implicated these proteases in the apoptotic process.

Not all caspases participate in apoptosis. Of the 12 human caspases, caspase-1, -4, -5, and -12 are thought to play a role in inflammatory cytokine maturation.^67,68 Whereas caspase-3, -6, -8, -9, and -10 have defined roles in apoptosis,^59,69 the exact functions of caspase-2, -7, and -14 remain to be defined.⁷⁰

Based on their biochemistry and functions, the apoptotic caspases may be divided into two classes.^59,69 The initiator caspases (caspase-8, -9, and -10) transduce other biochemical signals into proteolytic activity. The effector caspases (caspase-3, -6, and possibly -7) are responsible for cleavages that help disassemble the cell. All of these proteases are synthesized as enzymatically intactive zymogens and then activated during apoptosis. The processes of activation, however, differ between initiator and effector caspases.

Initiator caspases are constitutively present as monomers and become activated as a consequence of multimerization.⁷¹ In the case of caspase-8 (and presumably caspase-10), the inactive zymogen monomers⁷¹ bind through their N-terminal protein-protein interaction domains to the adaptor molecule FADD (Fas-associated protein with death domain), which is assembled into multimers on the cytoplasmic surface of ligated death receptors (see Chapter 2).

Caspase-9 activation occurs by a related but distinct process. Treatment of cells with many apoptotic stimuli leads to release of cytochrome c and other intermembrane proteins from mito­chondria.⁷³ After it is released, cytochrome c binds to a cytosolic scaffolding protein called apoptotic protease activating factor-1 (Apaf-1), which undergoes an ATP- or dATP-mediated conformational change and binds procaspase-9⁷⁴ to form a multimeric structure called an apoptosome.^75,76 In a manner that remains incompletely understood, binding to the apoptosome results in realignment of residues at the active site of caspase-9 into a conformation that allows catalytic activity.⁷⁷

In contrast to initiator caspases, effector caspases are synthesized as enzymatically inactive homodimers.⁷¹ Proteolytic cleavage at specific aspartate residues simultaneously produces two fragments (a small subunit and a large subunit that is linked to a short prodomain) and induces a conformational change that allows for realignment of the active site residues into a catalytically competent conformation.⁷⁸ The catalytically active effector caspase then cleaves the larger of the two fragments to remove the short prodomain from the large subunit.⁷⁹

As should be evident from the preceding description, caspase activation can be detected by a variety of techniques, including assays for enzymatic activity (ability to cleave synthetic substrates in vitro or natural substrates in situ), changes in molecular weights of the caspase molecules as they are proteolytically activated, and appearance of new immunoreactive epitopes as the caspases undergo activating conformational changes.⁸⁰’⁸¹ Each of these approaches has strengths and limitations.

Activity Assays

As indicated above, a subset of caspases acquires enzymatic activity during the course of apoptosis. This activity can be detected using suitable fluorogenic or chromogenic synthetic substrates. Because caspase-active sites usually accommodate only four amino acids to the N-terminal side of the scissile bond, most synthetic substrates consist of four amino acids followed by a detectable leaving group. The notable exception is caspase-2, which reportedly recognizes five amino acids.⁸² The sequences of substrates that are widely used for these assays are indicated in Table 3.1.

In preparation for measurement of caspase activity, cells treated with an apoptotic stimulus are washed and lysed in a hypotonic buffer so that a desired subcellular fraction (typically cytosol) can be isolated. Alternatively, cells are lysed in buffer containing a nonionic or zwitterionic detergent. After removal of insoluble material, lysates are incubated at 30 to 37°C with saturating concentrations of substrate. Release of fluorescent or colored product can be monitored over time or at the end of the incubation using a fluorimeter or spectrophotometer, respectively.^80,81 If cells undergoing spontaneous apoptosis are removed from a population before addition of an apoptotic stimulus, a 20- to 100-fold increase in caspase activity can be detected in some model systems.^83-85 A similar approach can be used to study caspase activation under cell-free conditions. When the amount of product released is compared to a standard curve, the result can be expressed in units that can be compared from laboratory to laboratory (e.g., picomoles of product produced/minute/ mg of protein in the extract).

Several points need to be kept in mind in order to perform valid assays. First, enzyme assays must be performed under conditions for which the amount of product is a linear function of enzyme activity. This condition is met only if extensive substrate depletion is avoided.

This approach has several advantages over alternative methods for detecting caspase activation described below. The present approach is relatively simple and relies on widely available equipment. It is readily adaptable to microtiter plate format, facilitating high-throughput screening. Finally, it is inherently more quantitative than immunoblotting methods.

TABLE 3.1

Cleavage of Various Synthetic Substrates by Purified Caspases

Sequence

Cleaved by^b

^aX = a detectable leaving group such as p-nitroaniline, 7- amino-4-methylcoumarin, or 7-amino-4-trifluoromethyl- coumarin.

^bCaspases reported to cleave the indicated substrate. Bold indicates substrate with the greatest activity.

Source: Adapted from Earnshaw, W.C., Martins, L.M., and

Kaufmann, S.H., Ann. Rev. Biochem., 68, 383, 1999;

Talanian, R.V., Quinlan, C.

Wong, W.W., J. Biol. Chem., 272, 9677, 1997; Thornberry, N.A., Rano, T.A., Peterson, E.P., Rasper, D.M., Timkey,

T., Garcia-Calvo, M., Houtzager, V.M., Nordstrom, P.A., Roy, S., Vaillancourt, J.P., Chapman, K.T., and Nicholson,

D.W., J. Biol. Chem., 272, 17907, 1997; Margolin, N., Raybuck, S.A., Wilson, K.P., Chen, W., Fox, T., Gu, Y., and Livingston, D.J., J. Biol. Chem., 272, 7223, 1997. With permission.

Nonetheless, this approach also has several limitations. Because of overlap in caspase speci­ficities,⁸⁶ the presence of an activity that cleaves one of the substrates listed in Table 3.1 does not necessarily identify the caspase that has been activated. For example, caspase-3, which cleaves DEVD-AFC, also cleaves LEHD-AFC, albeit less efficiently.⁸¹ Thus, the detection of LEHD-AFC cleavage might reflect activation of caspase-9, caspase-3, or both. In order to specifically detect caspase-9 activity, it might be necessary to deplete the more abundant caspase-3. Although this could, in principle, be done using a caspase-3 inhibitor, currently available inhibitors suffer from the same lack of specificity. Accordingly, caspase-3 must be depleted using the caspase-3 binding domain of X-linked inhibitor of apoptosis (XIAP)⁸⁷ or immunological reagents. Alternatively, immunoblotting assays described below are required to identify the caspases that are proteolytically activated.

Two other limitations must also be kept in mind. First, in some cell types, activated caspase-3 and -9 are sequestered in cytoplasmic aggregates that lack enzymatic activity.⁸⁵-⁸⁸ This aggregation can lead to underestimation of the amount of activated caspases unless other techniques are also employed. Finally, when activity increases, it is unclear whether this reflects caspase activation in all of the cells or only a small fraction. Thus, caspase assays are usually supplemented by some of the techniques described in the next three sections.

Cleavage of Procaspases and Substrates

Caspase-mediated cleavage of more than 400 substrates has been described during the course of apoptosis.⁵⁹-⁶¹ Among the substrates are the procaspases, which undergo cleavage at specific aspartate residues during the course of activation. The precise cleavages sites have also been mapped for most other substrates. This cleavage of procaspases or other substrates into their signature fragments can be used to monitor caspase activation.

When this approach is employed, cells treated with an apoptotic stimulus are washed and solubilized directly in SDS sample buffer or another denaturing buffer. This is preferable to lysing the cells in buffer containing nonionic detergent (Nonidet P-40 or equivalent), as spurious caspase activation has been reported when T lymphocytes are solubilized under nondenaturing conditions.⁸⁹ Lysates containing equal amounts of protein are separated by electrophoresis in the presence of SDS, transferred to nitrocellulose, and probed with antibodies that recognize uncleaved or cleaved antigens. In addition to conventional antibodies that recognize both full-length antigens and their cleaved products, so-called “anti-neoepitope” antibodies are available for some antigens. These reagents recognize epitopes created when the polypeptide of interest has been cleaved (e.g., epitopes consisting of the free C-terminal aspartate of the polypeptide fragment generated by caspase cleavage and the preceding few amino acids). Antineoepitope antibodies are available for several of the caspases, for example.⁹⁰

Strengths of this approach include its simplicity and the widespread availability of equipment. When reagents that recognize caspase substrates are employed, this technique will certainly indicate whether caspases have been activated in situ. In addition, for caspases that are cleaved during activation (e.g., caspase-3, -6, and -7), this method provides a potential way to assess activation status. In contrast, because caspase-9 can be activated without cleavage,^91,92 assessment of caspase-9 cleavage will underestimate the extent of caspase-9 activation. The same might also be true of caspase-8 and caspase-10.⁷²

This approach has additional limitations. First, because immunoblotting is typically performed without controls to confirm its linearity, this approach is best viewed as qualitative rather than quantitative. Second, there is a frequent disparity between the amount of procaspase that has been cleaved and the levels of processed caspase species present in apoptotic cells. The ability of inhibitor of apoptosis proteins to ubiquitylate certain active caspase species,^93,94 thereby facilitating their proteasome-mediated degradation, presumably contributes to this disparity. As a consequence, the decrease of a particular procaspase species overestimates the amount of polypeptide present in the “active” form. Third, this technique is dependent on high-quality antisera. Most of the caspase neoepitope sera cross-react with multiple bands on SDS-polyacrylamide gels. Because of this cross-reactivity, these sera are not suitable for immunolocalization studies. In contrast, commer­cially available “cleaved PARP” (poly [ADP-ribose] polymerase) antiserum is highly specific for the N-terminus of the 89 kDa PARP fragment (T. Kottke and S. Kaufmann, unpublished observa­tions). When this reagent is employed in indirect immunofluorescence, it is possible to determine whether caspases have been activated in all cells or only a fraction of the cells in a population.

Affinity Labeling

Affinity labeling was originally billed as an alternative technique that permitted determination of the percentage of cells with active caspases. As with other enzymes, treatment of caspases with substrate-like molecules fused to reactive groups can result in covalent modification (and inhibition) of the enzymes’ active sites. A wide variety of sulfhydryl proteases can be covalently labeled with substrate-like peptides bound to chloromethyl ketone, fluoromethyl ketone, or acyloxymethyl ketone groups. The inclusion of a reporter moiety (e.g., biotin or a fluorescent group) allows for the detection of the covalently modified enzyme. In the case of caspases, the P2 amino acid side chain points away from the active site, permitting substitution of lysine coupled to biotin at this position in the affinity labeling reagent.⁵⁹,⁹⁵ Alternatively, reporter groups can be coupled to the N-terminus of the substrate-like peptide.

The reaction between an affinity label and its target caspase has been used to probe the active site of caspase-1.⁹⁵ Likewise, when used in conjunction with two-dimensional electrophoresis⁸⁴ or affinity purification followed by mass spectrometry,⁹⁶ affinity labeling has helped identify the active caspase species that are present in lysates of apoptotic cells. Thus, this technique can be very useful under certain circumstances.

It was also suggested that this approach can be used to follow caspase activation on a cell-by- cell basis.^97,98 Treatment of cells with labeled aspartylglutamylvalinylaspartic acid fluoromethyl ketone (DEVD-fluoromethyl ketone), which contains the DEVD peptide recognized by caspase-3 and -7 fused to a reactive moiety, was proposed as a means to identify cells in which caspase-3 or -7 has been activated. In theory, if a population of cells is treated with the reagent and washed to remove unbound label, only cells that contain active caspase-3 or -7 will be labeled. Because the reagent also acts as a protease inhibitor, further apoptotic changes (including activation of caspases downstream of caspase-3)⁹⁹ will be inhibited.

Although the concept of using affinity labeling to quantitate the number of apoptotic cells is appealing, there are several potential problems with this approach. First, as indicated in Table 3.1, the cleavage specificities of the caspases are not absolute. Reagents designed for one caspase will label others. This lack of specificity becomes particularly problematic during prolonged incubation.¹⁰⁰ Second, because the reactive groups used in these affinity labels are designed to react with activated serine and sulfhydryl groups, cross-reaction with proteases other than caspases has been detected.¹⁰¹ Because of these potential problems, the use of these reagents to detect caspase activation by immunofluorescence or flow cytometry might be less straightforward than originally envisioned.¹⁰²

Immunodetection of Activated Caspases

Activation of caspase-3,¹⁰³ -7,⁷⁸ -9,⁷⁷ and presumably other caspases is accompanied by conforma­tional changes. These observations have led to the development of conformation-sensitive antisera that can be used to probe for active caspase species.¹⁰³

To apply this approach, cells undergoing apoptosis are fixed with a cross-linking fixative (e.g., formaldehyde or paraformaldehyde), permeabilized, and stained using the conformation-sensitive reagents as primary antibodies followed by peroxidase- or fluorochrome-labeled secondary anti­bodies as detection reagents. Samples can be examined by microscopy or, in the case of cell suspensions stained with fluorochrome-labeled secondary antibodies, flow cytometry.

In order for this approach to work, agents that denature polypeptides and destroy the confor­mational epitopes of interest (e.g., methanol and acetone) must be avoided during fixation and permeabilization. When applied successfully, this approach allows for the detection of caspase activation on a cell-by-cell basis. Because the immunological reagents employed can be highly specific, demonstration that a particular caspase has been activated can be relatively unequivocal. Moreover, this approach is applicable to tissue sections as well as single-cell suspensions.

There are, however, also limitations to this approach. First, conformation-sensitive antibodies to most of the caspases are not currently available. It is important to emphasize that these confor­mation-sensitive reagents are different from the “neoepitope antibodies” (as discussed in the section “Cleavage of Procaspases and Substrates”) that recognize linear epitopes generated during caspase cleavage. Some of the neoepitope reagents currently available are relatively specific, but many exhibit cross-reactivity with multiple polypeptides on immunoblots and should not be used for staining. Second, the approach is inherently qualitative. Finally, because of subtle sequence and conformation differences, reagents that label cells from one species will not necessarily work in cells from another species.

annexin V staining

In healthy cells, certain phospholipids are asymmetrically arranged in the plasma membrane. Sphingomyelin, for example, is present predominantly in the outer layer, and phosphatidylserine is exclusively in the inner layer.¹⁰⁴ During the course of apoptosis, caspase-3-mediated cleavage of protein kinase Cδ separates the N-terminal inhibitory domain from the C-terminal catalytic domain.¹⁰⁵ The resulting constitutively active protein kinase Cδ catalytic domain then activates a lipid scramblase in the plasma membrane,¹⁰⁶ leading to exposure of phosphatidylserine on the outer face of the plasma membrane. Under physiological conditions, this cell surface phosphatidylserine serves as a recognition molecule for clearance of apoptotic cells by phagocytes.¹⁰⁷’¹⁰⁸ The detection of this externalized phosphatidylserine by annexin V, a polypeptide that binds avidly and specifically to phosphatidylserine, forms the basis for widely used histochemical and flow cytometry-based methods of detecting apoptotic cells.^109,110

To perform this assay, cells treated with an apoptotic stimulus are incubated with fluorochrome- coupled annexin V, often in the presence of a vital dye such as propidium iodide, and are examined by fluorescence microscopy or subjected to flow microfluorimetry (Figure 3.1(F)). Alternatively, cells can be fixed with a nonpermeabilizing fixative such as glutaraldehyde before annexin V staining. In either case, apoptotic cells will be labeled with annexin V, whereas intact cells will not.

The advantages of this technique include the ability to examine large numbers of cells and the widespread availability of the required equipment. In addition, if examined in conjunction with propidium iodide (PI) staining, annexin V binding can distinguish apoptotic cells (annexin V⁺, PI^-) from necrotic (PI⁺) cells.¹¹⁰ It is important to realize, however, that after plasma membrane integrity is lost, annexin V will have access to lipids of the inner face of the plasma membrane and will stain all cells. Accordingly, this technique is not suitable for examining tissue sections or cells that have been permeabilized. Moreover, although it was initially thought that annexin V staining detected an early apoptotic event,¹¹⁰ it is now clear that phosphatidylserine becomes accessible to annexin V only after caspase activation.^106,111-113