A Cinderella Caspase Takes Center Stage

See allHide authors and affiliations

Science  23 Aug 2002:
Vol. 297, Issue 5585, pp. 1290-1291
DOI: 10.1126/science.1076118

Activation of proteolytic enzymes called caspases is a key step in the apoptotic program. Caspases exist in latent forms in almost all animal cells and become activated in response to apoptotic signals such as those induced by cell stress (for example, DNA damage and withdrawal of trophic support). The first caspases to become activated are so-called “initiator caspases.” These caspases have long amino-terminal prodomains containing specific protein-protein interaction motifs. Through these domains the caspases interact with adaptor proteins that recruit them to specific “death complexes” (large multiprotein complexes that mediate caspase activation). In mammals, these death complexes include the Apaf-1/caspase-9 apoptosome and the FADD/caspase-8 death-inducing signaling complex (DISC). Once the initiator caspases are activated, they process and activate downstream effector caspases, such as caspases 3, 6, and 7 (1). The apoptosome and DISC are thought to account for most caspase-dependent apoptosis. The upstream signaling pathways leading to the assembly of these death complexes are often called the mitochondrial (intrinsic) and death receptor (extrinsic) pathways of apoptosis.

During stress-induced apoptosis, mitochondria release their cytochrome c, which binds to Apaf-1 and promotes apoptosome formation and caspase-9 activation (1). Thus, the most widely held view is that caspase-9 is the initiator caspase in this pathway and that mitochondrial release of cytochrome c is essential for activation of this caspase (1, 2). However, this view has been challenged [see (2) for arguments from both sides] and is now under challenge again from several reports (36), including one by Lassus et al. (3) on page 1352 of this issue. These authors show that in response to cell stress, activation of a neglected caspase, caspase-2, is required before mitochondrial permeabilization and apoptosis can take place (3).

A new model of apoptosis.

Previous studies suggest that caspase-8 and caspase-9 act as initiator caspases in extrinsic and intrinsic pathways, respectively, and activation of these caspases is necessary for the initiation of caspase cascade. In the new model, intrinsic pathways, such as those initiated by cell stress, induce activation of caspase-2, which is required for permeabilization of mitochondria, release of cytochrome c, and apoptosis. Mitochondria may act as amplifiers rather than initiators of caspase activity. Current efforts to develop caspase inhibitors for treating stroke and myocardial infarction have centered on trying to block the activity of postmitochondrial caspase-9 and caspase-3. However, this approach may only inhibit the amplification loop, to delay, but not prevent, cell death.


Caspase-2 (Nedd2/Ich-1) was the first mammalian apoptotic caspase to be identified (7, 8). It closely resembles caspase-9 and CED-3 in the worm, and like them bears a caspase recruitment domain (CARD). Numerous early studies implicated caspase-2 in cell death pathways that could be rescued by Bcl-2. Caspase-2 is ubiquitously expressed, its activation occurs early in apoptosis, and antisense studies suggest that reducing the amount of caspase-2 lessens cell death in response to factor deprivation of cultured cells and sympathetic neurons (711). However, the absence of an overt phenotype in caspase-2-deficient mice (12, 13) led to this caspase becoming the Cinderella of the caspase field.

In their new work, Lassus and co-workers revisit caspase-2 using small interfering RNA (siRNA) to ablate its expression. They demonstrate clearly that caspase-2 is required for DNA-damage-induced apoptosis of E1A-transformed human fibroblasts and some human cancer cell lines (3). Expression of a caspase-2 cDNA construct that is refractory to siRNA restored the ability of cells to undergo apoptosis. Perhaps their most striking finding is that caspase-2 activity is required for translocation of the death protein Bax to the mitochondria as well as for release of the mitochondrial proteins cytochrome c and Smac/Diablo, early steps in the apoptotic program (3). Similar results are reported by two other groups who demonstrate that caspase-2 can induce cytochrome c, Smac, and apotosis-inducing factor (AIF) release from mitochondria directly (4, 5) or through cleavage of the proapoptotic protein Bid, which moves to mitochondria and facilitates cytochrome c release (4). Another group also reports that caspase-2 in the nucleus signals cytochrome c release from mitochondria (6). Using different systems and cell types, all of these studies suggest that caspase-2, not caspase-9, is an apical caspase in the proteolytic cascade initiated by stress signals. If correct, this model might help to explain why apoptosis seems to occur normally in cells other than neurons in mice lacking Apaf-1 or caspase-9 (14). Although caspases cannot be activated by released cytochrome c in these animals, caspase-2 or a similar caspase could still be killing cells independently of the mitochondrial pathway.

This model also gives new breath to the idea that Bcl-2 might act like its worm homolog CED-9, namely by preventing activation of a caspase and so maintaining the integrity of mitochondria indirectly (2). It is generally believed that Bcl-2 blocks apoptosis by preventing loss of mitochondrial membrane potential and release of mitochondrial proteins such as cytochrome c, Smac/Diablo, and AIF. These events are mediated by proapoptotic Bcl-2 family members such as Bax, Bak, and Bid. The alternative view is that Bcl-2 acts like CED-9 by preventing activation of a caspase (2). According to the new model (see the figure), once this caspase is activated it causes damage to cellular components, including the mitochondria. The damaged mitochondria then release cytochrome c, which promotes apoptosome-mediated activation of caspases such as caspase-9 and caspase-3, serving to amplify the caspase activation cascade.

Before this model of apoptosis is accepted, several questions must be answered. Why is the phenotype of the caspase-2 knockout mouse so subtle? Is caspase-2 the only premitochondrial stress-activated caspase, or are there others that can compensate for its absence? How common is the pathway requiring caspase-2, given that some of the cell lines tested by Lassus et al. showed no effects on apoptosis after caspase-2 ablation? How does caspase-2 activation occur upstream of mitochondria, and what factors regulate it? Does it require a CED-4-like adaptor protein that binds to Bcl-2 and is regulated in a manner similar to CED-9-mediated inhibition of CED-3 activation?

To date, Apaf-1 remains the only convincing CED-4 homolog, and it acts downstream of mitochondria and does not bind to Bcl-2 or Bcl-x. Another puzzling observation is that in dying cells, caspase-3 mediates most of the cleavage of procaspase-2, and cells derived from Apaf-1 and caspase-9 knockout mice fail to show processing of caspase-2, which suggests that most of the caspase-2 activation may occur downstream of mitochondria (13). One possibility is that caspase-2 processing by caspase-3 is an amplification mechanism and that the initial caspase-2 activation upstream of mitochondria occurs without any processing.

The answers to these questions are far from academic. Any potential pharmacological use of caspase inhibitors for treating diseases such as stroke and acute myocardial infarction will require targeting caspase activation events upstream of mitochondria, rather than inhibiting caspase-9 and caspase-3. But a drug that binds to caspase-2—dare we say with the fit of a glass slipper—and allows long-term survival of cells would make a very nice conclusion to this tale.


View Abstract

Stay Connected to Science

Navigate This Article