Cross-talk between death and survival pathways

Apoptosis plays an important role in maintaining correct cell numbers in multicellular organisms. Under pathologic conditions, such as inflammation and cancer, apoptosis is deregulated and cells exhibit either reduced or increased susceptibility towards apoptotic stimuli. In inflammatory responses, apoptosis is often delayed in the infiltrating cells of the immune system.¹ This process contributes to the accumulation of effector cells, which may then eliminate the pathogen(s) responsible for the immune reaction. However, feedback mechanisms are likely required to avoid uncontrolled expansion of inflammatory cells and subsequent tissue damage. In this article, we would like to draw the attention of the reader to some of the recently reported death receptor–mediated mechanisms, which might play a role in shutting off the function of antiapoptotic cytokines and intracellular molecules.

Apoptosis pathways initiated by death receptors

Death receptors of the tumor necrosis factor (TNF)/nerve growth factor (NGF) receptor family such as Fas, TNF, and TRAIL receptors recruit, upon binding of its cognate ligand, molecules of the death-inducing complex (DISC) that includes procaspase-8.² Caspases are intracellular cysteine proteases that cleave critical cellular substrates during the process of apoptosis. Recruitment of procaspase-8 to the DISC results in its autoactivation. As a consequence, direct activation of procaspase-3 by caspase-8 occurs in many susceptible cells.³ In addition, active caspase-8 cleaves Bid to produce a truncated form (tBid), which either alone or in combination with other molecules induces mitochondria to release proapoptotic factors such as cytochrome c. This event is blocked by antiapoptotic Bcl-2 family members (e.g. Bcl-2 and Bcl-xL).⁴ Upon release from the mitochondria into the cytosol, cytochrome c initiates the formation of a complex known as ‘the apoptosome’ by binding to an adapter molecule called Apaf-1. In the presence of ATP, this leads to the activation of procaspase-9. Activated caspase-9 cleaves procaspases-3, -6, or -7, which rapidly mediate the apoptotic proteolysis in cells.

Activation of caspases might be blocked by members of the inhibitor of apoptosis protein (IAP) family, which bind to and inhibit these death proteases.⁵ Thus, the release of cytochrome c does not necessarily result in apoptosis, and under certain circumstances cells survive.⁶ Members of the IAP family are c-IAP1, c-IAP2, XIAP, and survivin. While it is reasonable to have such blockers of caspases within cells as a safeguard mechanism to avoid unwanted induction of cell death, in particular, in long-living cells such as neurons, it is clear that these inhibitors must somehow be negatively regulated to enable apoptosis. Important molecules able to inhibit IAP functions are Smac/DIABLO^7,8 and HtrA2/Omi,⁹ which are both released from mitochondria in conjunction with cytochrome c upon increase of the mitochondrial membrane permeability.

Activation of inhibitory phosphatases

Both antiapoptotic members of the Bcl-2 family as well as some members of the IAP family are transcriptionally regulated by NF-κB, a transcription factor often activated by survival cytokines. Therefore, a cell may express significant levels of antiapoptotic proteins because of the exposure of survival factors, making the cell resistant to death induced by various stimuli. This raises the question of whether such a cell is able to die under the condition of survival factor stimulation. If the answer is yes, survival signaling must be somehow attenuated.

One potential mechanism has recently been proposed in neutrophils. TNF receptor 1 and the Fas receptor were found to be able to bind inhibitory phosphatases such as SHP-1 due to a conserved YxxL motif present within death domains of all death receptors.¹⁰ The tyrosine residue within the YxxL motif was shown to be essential for SHP-1 binding to the Fas receptor. Interestingly, concurrent engagement of either Fas, TNF, or TRAIL receptors abrogated GM-CSF-induced signaling by dephosphorylation of the tyrosine kinase Lyn, which was previously shown to be crucial for triggering the antiapoptotic pathway induced by the survival cytokine. Moreover, inactivation of the survival pathway occurred even under conditions in which the death pathway was not directly activated. Taken together, death signals may block cytokine-induced signaling cascades, a mechanism, that is likely to be important at least in growth factor-dependent cells. As a consequence of such a signaling block, the expression of antiapoptotic proteins may decline (Figure 1). At present, we do not know whether SHP-1 plays any role in the regulation of apoptosis of cancer cells, which are often resistant to TNF- or Fas ligand-induced apoptosis. Future studies should be performed to answer the question of whether prolonged or repetitive death receptor activation results in cell death even in cells that are originally resistant to apoptosis induction.

Figure 1

Hypothetical scheme showing potential noncaspase and caspase-mediated mechanisms contributing to the functional neutralization of antiapoptotic molecules. Note that not all the suggested mechanisms have necessarily the same importance in each cell type. Death receptors are able to activate SHP-1, which dephosphorylates and thereby inactivates important signaling molecules activated upon exposure of cells to survival factors. This may reduce the expression of antiapoptotic genes after a certain time period, allowing the activation of the mitochondrial pathway. Caspases have multiple intracellular targets (red) including antiapoptotic as well as proapoptotic regulators. It is suggested that cleavage of antiapoptotic molecules is associated with reduced functional activity. In contrast, cleavage of Bid and calcium pumps will promote and/or accelerate the apoptotic process. Calcium overload will activate calpain (blue), which has been shown to cleave XIAP, Bax, and Bid (not shown in this figure)

Disturbing the steady-state caspase activity

Another possibility is that the apoptosis process is favored by direct inactivation of antiapoptotic proteins already present in cells. Apoptotic mitochondrial changes are predominantly prevented by antiapoptotic members of the Bcl-2 family. During apoptosis, it has been shown that both Bcl-xL¹¹ and Bcl-2¹² can be targets of caspases resulting in loss of their antiapoptotic functions and the truncated proteins may even enhance cell death.

When does such cleavage occur? Cleavage of Bcl-xL and/or Bcl-2 may be part of the apoptotic program, including the possibility that cleavage of antiapoptotic molecules is not mandatory within the death pathway. Another possibility is that cleavage of Bcl-xL and/or Bcl-2 is a regulatory event at least in some forms of apoptotic cell death. In this model, active caspases do not directly induce cell death.

The idea of basal caspase activation in viable cells is supported by the presence of IAPs, which inhibit processed and activated caspases.⁵ If the balance between IAP inhibited and activated caspases is no longer maintained, caspases may not directly kill cells but rather decrease the threshold of apoptosis induction by a subsequent death stimulus. For instance, death receptor activation may increase the basal caspase activity in a cell, resulting in decreased levels of Bcl-xL and/or Bcl-2 (Figure 1). Thus, also in this model, prolonged or repetitive death receptor activation results in cell death even in those cells that are originally resistant to apoptosis induction.

We have recently observed that in neutrophils caspases also cleave calpastatin, an endogenous inhibitor of calpains. Decreased levels of calpastatin were associated with the activation of calpain-1 but not calpain-2 (manuscript submitted for publication). Calpain has been shown to cleave Bid¹³ and Bax¹⁴ in apoptosis induced by chemotherapeutic drugs. The calpain-mediated generation of truncated Bid as well as truncated Bax seems to be a proapoptotic event, since both molecules act on mitochondria. Another potential target of calpain might be XIAP.¹⁵ Thus, at least in some cells such as neutrophils, calpain activation is likely to be a proximal event of apoptosis induction that occurs as a consequence of calpastatin proteolysis mediated by activated caspases (Figure 1). Whether calpastatin expression is induced by survival cytokines remains to be investigated. We observed that neutrophils of patients with cystic fibrosis, which exhibit delayed apoptosis because of increased expression of GM-CSF and G-CSF,¹⁶ have increased levels of calpastatin (manuscript submitted for publication). However, it remains unclear as to whether prolonged or repetitive death receptor activation results in a reduction of calpastatin levels in the cells and whether this alteration decreases their life span.

Other potential antiapoptotic targets for caspases are signaling proteins, such as Raf-1 and Akt-1,¹⁷ as well as survival factor receptors themselves. As shown in Table 1, we identified several caspase consensus sites^18,19 within the intracellular part of cytokine receptors. For certain growth factor-dependent cells such as granulocytes, cleavage of the intracellular part of the common β-chain of IL-3/IL-5/GM-CSF receptors might have dramatic consequences. Indeed, it is unlikely that the remaining putative β-chain is able to recruit all the required signaling molecules upon ligand binding.²⁰ The cleavage of intracellular parts of other receptors (e.g. IFN) might be equally important.²¹ Taken together, it can be speculated that death receptor activation not only inactivates important tyrosine kinases in the survival factor pathway (via SHP-1, see above), but also prevents kinase activation because of the generation of functionally inactive cytokine receptors (Figure 1). The biological relevance of this scenario needs to be addressed in forthcoming studies.

Table 1 Putative caspase-3 cleavage sites in human survival factor receptors. The demonstrated tetrapeptide recognition motifs are identical or closely related to the sequences found in other known substrates

Besides the potential cleavage of survival factor receptors, caspases have been shown to cleave other plasma membrane molecules such as calcium pumps, resulting in increases of intracellular free calcium.²² It is likely that calcium-dependent calpain isoforms²³ are activated in this process and that these isoforms are also able to cleave Bax and XIAP (Figure 1). DW Nicholson (Merck Frosst Canada & Co., Kirkland, Quebec, Canada), during the 10th Euroconference on Apoptosis (Paris, October 2002), reported about the caspase cleavage products of several cellular proteins that are able to neutralize the function of IAPs. Thus, it appears that the list of emerging new caspase targets that further amplify the apoptotic process is never ending.

Other potential interactions

It should be noted that there are also potential cross-talks between survival and death pathways, with the result that the death pathway is blocked. One way to inhibit the proapoptotic function of potential death inducers is their phosphorylation by cytokines. Phosphorylated Bad interacts with 14–3–3, resulting in Bcl-2 release,²⁴ and phosphorylated Bid is caspase-8 resistant.²⁵ In addition, phosphorylated caspase-9 cannot be activated,²⁶ but this regulatory event appears to be restricted to the human system, since neither mouse nor rat caspase-9 contain the described phosphorylation site. Moreover, although these mechanisms are crucial by making cells resistant to death stimuli under in vitro conditions, the biological implication of these findings in vivo remains to be determined. Aberrant phosphorylation of proapoptotic proteins by kinases has long been identified as a hallmark of cancer.²⁵

Another strategy for increased cell survival is the forced degradation of proapoptotic molecules via ubiquitination and proteosomal degradation. At least some of the members of the IAP family are not only able to inhibit caspases but also to regulate their ubiquitination. For instance, IAP-2 was found to direct the ubiquitination of caspase-3 and caspase-7,²⁷ and XIAP has been shown to induce the ubiquitination of caspase-3.²⁸ It is likely, however, that the ubiquitinase activity of IAPs is not restricted to caspases and involves a multitude of other intracellular targets.

Conclusions

There are various possibilities as to how survival and apoptosis pathways in a cell can meet. In this article, we addressed potential mechanisms that make sure that inflammatory cells undergo apoptosis even under inflammatory conditions. Although still hypothetical in many cases, one might assume that the same or similar mechanisms are required to limit and resolve inflammatory responses. In other proliferative disorders such as cancer, growth factor stimulation or constitutively active growth signaling may attenuate cell death signaling, resulting in tumor growth because of uncontrolled proliferation and diminished apoptosis.