(Nakagawa and Yuan, 2000). Of note, however, the free p20 or p10 subunit of
caspase-12 has not yet been observed in vivo; therefore, the second cleavage step
remains theoretical. Consistent with this model, both calpain inhibitors (they block
the first cleavage event) and the pan-caspase inhibitor zVAD-fmk (predicted to block
the second cleavage step and caspase-12 activity) inhibited OGD-induced apoptosis.
Furthermore, chelation of intracellular Ca2+ with BAPTA-AM has been reported to
block cell death in response to thapsigargin, perhaps by inhibiting the calpain-
dependent activation of caspase-12 (Srivastava et al., 1999). Nakagawa and Yuan
noted that Bcl-xL is also cleaved by m-calpain during OGD, an event that is predicted
to convert Bcl-xL from antiapoptotic to proapoptotic. This raises the possibility that
calpain-cleaved caspase-12 and Bcl-xL cooperate to induce apoptosis.
Rao et al. (2001) reported that in 293T cells exposed to thapsigargin or brefeldin
A, caspase-7 translocates to the surface of the ER where it can bind and cleave
caspase-12 at D94, releasing the prodomain. In vitro, the resulting C-terminal
fragment appears to undergo self-cleavage at D341, separating the p20 and p10
subunits. Recombinant caspase-7 could activate an S-100 extract to cleave caspase-12
at D94 in vitro, but it was not reported whether caspase-7 could directly cleave
caspase-12, and the effect of zVAD-fmk on caspase-12 cleavage was not tested in vivo
or in vitro. One possible explanation that could reconcile the discrepancies between
this study and that of Yuan’s group is that cleavage of caspase-12 by calpain might
be an early initiating event, whereas cleavage of caspase-12 by caspase-7 might be a
late event involving feedback amplification, such has been described in other
apoptotic systems. However, the observed contrasts might be related to the fact that
the two studies were carried out in different species with different inducers of ER
stress. Of note, Nakagawa and Yuan (2000) reported that in mouse cortical neurons,
caspase-12 is expressed as a main isoform of ~50 kDa and a smaller isoform of ~42
kDa. Following OGD, a third larger isoform (-60 kDa) is upregulated. Murine
caspase-12 expressed ectopically in 293T cells has also been reported to be
differentially phosphorylated (Yoneda et al., 2001). The human caspase-12 sequence
has not been formally reported, but antibodies raised against murine caspase-12 detect
a single 60-kDa protein in human 293T cells that is upregulated in response to ER
stress prior to its cleavage (Rao et al., 2001). Therefore, caspase-12 may be expressed
as several isoforms, or as modified forms of a single species, that are differentially
regulated.
Following stimulation, the cytosolic tail of Irel can recruit TRAF2 (Urano et al.,
2000b). When overexpressed in 293T cells, TRAF-2 can also interact with caspase-12
through its TRAFN domain and weakly induce caspase-12 oligomerization and
cleavage (Yoneda et al., 2001). Although the observed interactions need to be
confirmed between endogenous components, it is possible that an activated Irel/
TRAF-2 complex recruits oligomers of casapase-12 that are cleaved by calpain or
undergo autocatalytic activation. This scenario would allow the ER to couple the
UPR survival response with an apoptotic cascade, and the fate of the cell could be
determined by the strength of the two competing signals or by other secondary signals
that modulate one of the responses. This theme is reminiscent of signals elicited by
102 GENETICS OF APOPTOSIS