We show here that AFP cancels XIAP-mediated inhibition of endogenous active caspases in cytosolic lysates of tumor cells, as well as XIAP-induced blockage of active recombinant caspase 3
Trang 1apoptosis protein anticaspase activity and disrupts
XIAP–caspase interaction
Elena Dudich1,2, Lidia Semenkova1,2, Igor Dudich1,2, Alexander Denesyuk3, Edward Tatulov2
and Timo Korpela4
1 Institute of Immunological Engineering, Lyubuchany, Russia
2 JSC BioSistema, Moscow, Russia
3 Department of Biochemistry and Pharmacy, A ˚ bo Akademi University, Turku, Finland
4 Joint Biotechnology Laboratory, Turku University, Finland
Apoptotic dysfunction plays a key role in cancer
pro-gression and leads to chemotherapeutic and
radio-therapeutic resistance [1–3] Many cancer radio-therapeutic
agents operate by inducing apoptosis and are
ineffec-tive in conditions of impaired apoptosis signaling
Novel strategies for cancer therapy are aimed at
dis-covering molecular targets involved in the induction
of apoptosis in normal and tumor cells, and at
selec-tively regenerating the apoptosis propensity in cancer
cells
Apoptosis is induced by two different mechanisms: the extrinsic or receptor-dependent pathway and the intrinsic or mitochondria-dependent pathway [4] Trig-gering of either pathway results in the initiation of caspase cascade activation events Caspases are gener-ally divided into two groups according to their func-tional hierarchy and substrate specificity The initiator caspase family includes caspases 2, 8, 9, 10 and 12, and is characterized by the presence of N-terminal prodomains DED or CARD, which are involved in
Keywords
apoptosis; apoptosome; caspases;
a-fetoprotein; X-linked inhibitor of apoptosis
protein
Correspondence
E Dudich, Institute of Immunological
Engineering, 142380, Lyubuchany, Moscow
Region, Chekhov District, Russia
Fax ⁄ Tel: +7 095 996 1555
E-mail: elena_dudich@mail.ru
(Received 28 February 2006, revised 3 May
2006, accepted 22 June 2006)
doi:10.1111/j.1742-4658.2006.05391.x
Previous results have shown that the human oncoembryonic protein a-feto-protein (AFP) induces dose-dependent targeting apoptosis in tumor cells, accompanied by cytochrome c release and caspase 3 activation AFP posi-tively regulates cytochrome c⁄ dATP-mediated apoptosome complex forma-tion in a cell-free system, stimulates release of the active caspases 9 and 3 and displaces cIAP-2 from the apoptosome and from its complex with recombinant caspases 3 and 9 [Semenkova et al (2003) Eur J Biochem
270, 276–282] We suggested that AFP might affect the X-linked inhibitor
of apoptosis protein (XIAP)–caspase interaction by blocking binding and activating the apoptotic machinery via abrogation of inhibitory signaling
We show here that AFP cancels XIAP-mediated inhibition of endogenous active caspases in cytosolic lysates of tumor cells, as well as XIAP-induced blockage of active recombinant caspase 3 in a reconstituted cell-free sys-tem A direct protein–protein interaction assay showed that AFP physically interacts with XIAP molecule, abolishes XIAP–caspase binding and rescues caspase 3 from inhibition The data suggest that AFP is directly involved
in targeting positive regulation of the apoptotic pathway dysfunction in cancer cells inhibiting the apoptosis protein function inhibitor, leading to triggering of apoptosis machinery
Abbreviations
Ac-DEVD-AMC, Ac-Asp-Glu-Val-Asp-7-amino-4-methyl coumarin; AFP, a-fetoprotein; IAP, inhibitor of apoptosis protein; IBM, IAP-binding motif; RFU, relative fluorescence units; XIAP, X-linked inhibitor of apoptosis protein.
Trang 2interactions with certain adapter molecules to form
death-inducing signaling complex or apoptosome [3–5]
Effector caspases 3, 6 and 7 exist in the cytosol as
inactive zymogens and are activated via a proteolytic
cascade started by the initiator caspases [5]
Molecular pathways leading to apoptosis are
evolu-tionarily conserved and are regulated by specific cellular
proteins Some, such as Bcl-2, control the release of
proapoptotic proteins from the mitochondria [6]
Others, including various cellular inhibitor-of-apoptosis
proteins (IAPs), bind directly to active caspases and act
as natural inhibitors of caspase activity [7] The IAP
family, currently identified in humans consists of
X-linked inhibitor of apoptosis protein (XIAP), ILP-2,
cIAP1, cIAP2, ML-IAP, NAIP, survivin, and livin
[7–9] All IAPs have so-called conservative BIR
domains, which are responsible for their interaction with
caspases XIAP is the most potent of all known IAPs
and contains three BIR domains The third BIR domain
(BIR3) selectively targets caspase 9, whereas BIR2 and
the linker region between BIR1 and BIR2 inhibit
effec-tor caspases 3 and 7 [5,10] This inhibition can be
relieved by IAP antagonists, which bind to IAPs
pre-venting caspase binding [5,8,9,11–13] Recent studies
have revealed that binding of IAP antagonists to IAPs
may stimulate their auto-ubiquitination and
degrada-tion, thereby preventing caspase inhibition [14,15]
Recognition of XIAP as a direct inhibitor of caspases
makes it an attractive therapeutic target This led to an
active search for any suitable molecular inhibitor
capable of easily penetrating a tumor cell to block
XIAP activity in the cytosol [8,9,16] The discovery of
endogenous regulators of IAP activity enhanced these
investigations Several intracellular inhibitory IAPs
have been characterized in humans, namely, Smac⁄
DI-ABLO, Omi⁄ HtrA2, GSPT1 ⁄ eRF3, ARTS, and XAF1
[17–21] However, only the first and best characterized
anti-IAP Smac⁄ DIABLO is currently known to be
directly involved in the regulation of apoptosis Other
anti-IAPs, such as Omi⁄ HtrA2 or GSPT1 ⁄ eRF3, seem
to have a primary physiological role that is not directly
related to XIAP⁄ caspase regulation [8,18,19] Smac ⁄
DIABLO is released from mitochondria into the
cyto-sol during apoptosis, wherein it can bind to XIAP [17]
The main highly conserved functional motif common
to all IAP antagonists, was termed the IAP-binding
motif (IBM), and became a target for finding novel
potential inhibitors of IAP [8,9,16] The motif ATPF⁄
AVPI was first characterized in caspase 9 and
Smac⁄ DIABLO was characterized as being responsible
for binding to BIR3 of XIAP [22]
Smac-derived peptides modeling their XIAP-binding
site, bind to recombinant BIR3 domains in vitro
[18–23] Recent studies have set out to design small molecular drugs carrying the IBMs [5,11–13] or artifi-cial chimerical peptides composed of the IBM sequence fused to a carrier peptide [23] The cell-permeable Smac peptides allowed the apoptosis resistance and chemoresistance of cancer cells with a high level of XIAP to be overcome in vitro and in vivo, as documen-ted [13,23] Despite the strong molecular basis for interaction with XIAP, natural Smac-derived peptides and other artificial IBM-based chimeric constructions have several intrinsic limitations (e.g poor in vivo stability and very low bioavailability) making them unsuitable for the treatment of cancer [8,9,23] The other known natural XIAP-binding proteins cannot act as anticancer drugs because of their exclusive intra-cellular location Therefore, the search for other XIAP-interacting and cell-membrane-penetrating drugs
is a highly desirable goal
Recently, it was discovered that the well-known oncofetal antigen a-fetoprotein (AFP) is able to induce apoptosis selectively in tumor cells without any toxicity towards normal cells and tissues [24–28] AFP is one
of the major serum embryonic proteins involved in the regulation of growth and the development of immature embryonic tissues [29–31] The specific expression and internalization of AFP is restricted to developing cells, such as embryonic cells, activated immune cells and tumor cells, which suggests that it has an important regulatory role in cell growth and differentiation [32– 35] AFP expression is blocked completely after birth and is recovered only after malignant transformation [29–31] Various researchers have documented the existence of specific receptor-dependent mechanisms responsible for the active endocytosis of AFP by malignant cells [34–36] AFP has been well character-ized as a transport protein delivering natural ligands such as fatty acids, hormones, and heavy metals to developing cells [29] The specific expression and inter-nalization of AFP by developing cells, such as embry-onic cells or tumor cells, together with the properties
of the transport protein make AFP very attractive for tumor-targeting therapy [29,30,33] The growth-regula-tory activity of AFP and AFP derivatives has been demonstrated by various authors [24–28,37–43] Spe-cial interest has focused on the tumor-suppressive effects of AFP and its peptide derivatives [24–28,38, 39,42] The growth-suppressive activity of AFP can be realized by inducing apoptosis in many types of tumor
or activated immune cells [24–28,41] AFP can trigger apoptosis in tumor cells via activation of caspase 3, independent of the membrane-receptor signaling [26] AFP stimulates formation of the apoptosome complex, and enhances recruitment and activation of caspases 3
Trang 3and 9 by displacing cIAP-2 from the apoptosome
and from its complex with recombinant caspases 3 and
9 [28]
Based on the molecular mechanisms of
AFP-medi-ated apoptosis, we hypothesized that AFP might
inter-act with XIAP by displacing it from the complex with
caspases, and thus preventing caspase inhibition We
demonstrate here that AFP physically associates with
XIAP in cytochrome c-activated cellular lysates, and
that this complex does not contain the effector
caspase 3 We found that purified human AFP binds
to recombinant XIAP, disrupts the association between
XIAP and activated caspase 3, and antagonizes the
antiapoptotic function of XIAP Our data indicated
that AFP could also bind free XIAP to eliminate it
from the reaction area and prevent caspase binding
Results
AFP promotes caspase activation in cell-free
cytosolic extracts by blocking of XIAP-dependent
inhibition
Recent evidence has shown that AFP promotes the
processing and activation of procaspase 3 in the
pres-ence of low suboptimal doses of cytochrome c in
cell-free cytosolic extracts Simultaneously, AFP induced
the release of cIAP2 from the apoptosome complex
[28] Our recent experimental data allowed us to
hypo-thesize that AFP could operate as a XIAP antagonist
by affecting the interaction of XIAP with active
caspas-es, thus promoting their activity To determine whether
AFP can affect caspase 3 activity in HepG2 cytosolic
extracts in the presence of an inhibitory amount of
exo-genous XIAP, we monitored caspase activation in a
cell-free system Cytosolic cell extracts were activated
by the addition of cytochrome c⁄ dATP together with
AFP or human serum albumin (HSA) in the presence
of an inhibitory amount of rhXIAP Figure 1 shows
that addition of XIAP induced 50% inhibition of
caspase 3 activity in activated cell-free extracts
Addi-tion of AFP in the cytosolic extract induced significant
enhancement of the DEVD-ase activity The same
caspase activity was detected upon the simultaneous
addition of an inhibitory amount of exogenous rhXIAP
together with AFP The data clearly show that AFP
abrogated the inhibitory activity of exogenous rhXIAP
against endogenous caspases and failed to relieve
AFP-mediated caspase stimulation By contrast,
exo-genous HSA did not affect caspase activity in cell-free
extracts (Fig 1) Hence, AFP counteracts with XIAP
by abrogation its caspase inhibition in the
cyto-chrome c-activated cell-free cytosolic extracts
AFP promotes caspase 3 activity exclusively
by abrogation of XIAP-dependent inhibition
To study direct AFP⁄ XIAP ⁄ caspase 3 interaction we used recombinant proteins to form a reaction mixture
in order to avoid the influence of other active com-pounds, which are available in cytosolic extracts An effective amount of rhXIAP was added into the solu-tion of active recombinant caspase 3 to induce 50% decrease of its activity The kinetics of the DEVD-ase cleavage in the reaction mixture was monitored each
5 min intervals rhXIAP in combination with HSA (Fig 2A) or alone (Fig 2B), induced twofold inhibi-tion of caspase 3 activity, but AFP⁄ rhXIAP pretreat-ment significantly reduced inhibition by rhXIAP (Fig 2A) Addition of AFP or HSA alone did not affect caspase 3 activity (Fig 2B) The results show that AFP does not directly affect caspase 3 activity, but targets XIAP by blocking its inhibitory activity against caspase 3 Therefore, AFP antagonizes XIAP function
AFP competes with caspase 3 and caspase 9 for XIAP binding
Functional interference of AFP and XIAP to examine their effect on caspase activity implied a direct physical interaction We further studied whether AFP can com-pete with caspases 3 and 9 for XIAP binding Pure recombinant His-tagged active caspases 3 and 9 were
Fig 1 AFP antagonizes XIAP-mediated caspase inhibition in cyto-chrome c-activated cell-free cytosolic extracts HepG2-derived cytosolic extracts were activated by 1 m M dATP and 5 l M cyto-chrome c and incubated with or without rhXIAP (250 n M ) with the addition of 400 n M AFP or 400 n M HSA for 1 h at 30 C Control lysates incubated without addition of HSA and AFP were taken as controls Caspase activity was measured by DEVD–AMC cleavage The mean data in RFU ± SD from three independent experiments are shown.
Trang 4incubated with rhXIAP and AFP or HSA, and protein
complexes were thereafter immobilized on
Ni–Seph-arose beads After extensive washing the supernatant
and pellets (beads) were blotted and probed with
anti-bodies to XIAP AFP, but not HSA, completely
abro-gated the association of rhXIAP with caspases 3 and 9
(Fig 3, pellet) Western blotting revealed the presence
of XIAP in the supernatants from Ni resin treated
with AFP, but only a negligible amount of free
rhXIAP was detected in the supernatants of
HSA-trea-ted samples (Fig 3) Western blotting of pellets using
antibodies against caspase 3 and caspase 9
demonstra-ted that neither AFP nor HSA was able to modulate
binding of His-tagged caspases on the nickel resin
(Fig 3) The data clearly demonstrated that AFP
cointeracted with XIAP by preventing XIAP⁄ caspase
complex formation
AFP coprecipitates with endogenous XIAP
in cellular extracts
We further studied the ability of AFP to interact with
endogenous XIAP in whole-cell extracts preactivated
with cytochrome c⁄ dATP ⁄ AFP Therefore, we
exam-ined whether AFP might be directly associated with
XIAP or⁄ and caspase 3 in cell-free extracts Protein
complexes were precipitated by the addition of
corres-ponding antibodies and protein A–Sepharose beads
Complex formation was detected by immunoblotting
of the proteins bound to the protein A–Sepharose with
anti-XIAP, anti-AFP, or anti-(caspase 3) IgG Figure 4
shows that AFP coprecipitated with endogenous XIAP
(Fig 4A, lane 3) but not with caspase 3 (Fig 4C, lane
2; C, lane 3), whereas XIAP coprecipitated with both AFP and caspase 3 (Fig 4A, lanes 2, 3) These results show that AFP associates physically with endogenous XIAP in activated cell cytosolic extracts (Fig 4A,
Fig 2 AFP abrogates XIAP-mediated inhibition of caspase 3 activity in vitro Active recombinant caspase 3 (3 n M ) was treated with: (A) a mixture of rhXIAP (200 n M ) with AFP (400 n M ) or HSA (400 n M ); (B) AFP (400 n M ), HSA (400 n M ), XIAP (200 n M ) or without additions Caspase 3 activity was measured by monitoring of DEVD–AMC cleavage at 5-min intervals Data were collected at 30 C for 30 min and expressed in RFU The mean ± SD of three independent determinations is shown.
Fig 3 AFP prevents XIAP ⁄ caspases complex formation Human recombinant XIAP was incubated for 2 h at 4 C with mixed His-tagged active recombinant caspase 3 and caspase 9 in the presence of AFP or HSA as described in Experimental procedures Protein complexes were precipitated by Ni–Sepharose beads Ni–Sepharose-bound proteins (pellet) and supernatants were ana-lyzed by SDS ⁄ PAGE ⁄ immunoblotting with anti-XIAP, anti-(caspase 3) and anti-(caspase 9) sera Input 1: rhXIAP (100 ng); Input 2: recom-binant tagged caspase 3 (50 ng); Input 3: recomrecom-binant His-tagged caspase 9 (50 ng).
Trang 5lane 3) As expected, endogenous caspase 3
coprecipi-tated with endogenous XIAP (Fig 4A, lane 2) No
interaction between caspase 3 and AFP could be
detec-ted (Fig 4B, lane 2; Fig 4C, lane 3) This indirectly
proved that AFP and caspase 3 interacted with the
same binding site of XIAP If AFP had been attached
to a binding site on the XIAP molecule other than that
responsible for caspase 3 binding, we would be able to
detect coprecipitation of all three proteins in this
experiment Hence, the results showed that AFP
act-ively binds to endogenous XIAP in cytochrome
c-acti-vated cellular extracts, but that it also prevents
complex formation of XIAP with active caspases The
data also suggested that AFP seems to bind to the
entire XIAP molecule only, because fragmented XIAP,
which was available in the cytosolic extracts (Fig 4A, lane 1), was not recovered on the immunoprecipita-tion⁄ western blotting pattern with anti-(caspase 3) and anti-AFP IgG within the limits of detection (Fig 4A, lanes 2, 3)
AFP physically associates with rhXIAP to form high molecular mass complexes
We then determined whether AFP and XIAP were able to form intermolecular complex rhXIAP was coincubated with rhAFP and then the protein mixture was subjected to native electrophoresis The complex formation was analyzed by western blotting with anti-AFP IgG (Fig 5A, lanes 1, 2) and with anti-XIAP IgG (Fig 5B, lanes 1, 2) Probing with anti-AFP revealed three AFP-specific bands (Fig 5A, lane 2), which correspond to the monomer, natural AFP-dimer and high molecular mass upper band corres-ponding to the AFP-specific macromolecular complex
To identify presence of XIAP in the AFP-specific com-plexes, we probed the blot pattern with anti-XIAP IgG This revealed presence of XIAP in the upper band, corresponding to the high molecular mass AFP-specific complex (Fig 5B, lane 2) We conclude that incubation of pure AFP with XIAP led to the forma-tion an intermolecular complex This AFP⁄ XIAP complex evidently contains more than two proteins, showing the ability of AFP to form multimolecular high-affinity complexes with XIAP
A
B
C
Fig 4 AFP associates physically with endogenous XIAP in the
cel-lular cytosolic extracts HepG2 cytosolic extract was activated by
the addition of 5 l M cytochrome c and 1 m M dATP for 30 min at
30 C in the presence of AFP (6 lg) and thereafter the specific
interaction of AFP, XIAP and caspase 3 was tested using
coimmu-noprecipitation with anti-AFP, anti-(caspase 3) or anti-(rabbit IgG) as
negative control Western blot analysis was carried out with
anti-XIAP (A), anti-AFP (B) and anti-(caspase 3) (C) sera Input: cytosolic
extract activated with cytochrome c ⁄ dATP (20 lg) Molecular mass
markers are indicated on the left.
Fig 5 Direct AFP ⁄ XIAP complex formation in vitro Recombinant human AFP and XIAP were coincubated for 1 h at 4 C and there-after subjected to the native nondenaturing PAGE followed by western blotting with anti-AFP (A) and anti-XIAP (B) IgG Lane 1 on the each pattern corresponds to rhAFP alone and lane 2 corres-ponds to AFP ⁄ XIAP.
Trang 6Search for the potential IBM in the structure of
the AFP molecule
The AFP protein contains a putative IBM-like
sequence ATIF(29–32), which fits the IAP binding
tetrapeptide consensus [9,44] (Fig 6) Similar to the
other IBM proteins, Smac, Omi and GSPT1 [17–19],
AFP requires N-terminal processing to expose the
IBM motif at the newly generated N-terminus
Processing of the first 28 amino acids could allow
exposing of N-terminal motif ATIF that is highly
reminiscent of IBM of caspase 9 and other IAP
antag-onists (Fig 6) In common with other IBMs, the
IBM-like motif of AFP bears Ala at its N-termini The Ala
residue within the IBM is highly conserved (Fig 6)
and has been shown to be essential for the interaction
between XIAP and mature Smac⁄ Diablo [45] This
sequence displays a high degree of similarity to the
IBM of caspase 9 with a single replacement of Pro3 in
caspase 9 to Ile3 in AFP (Fig 6) [22] This position is
variable in different IAP antagonists and does not
seem to be critical in forming the XIAP-binding site
(Fig 6) The AFP IBM-like sequence has Phe at the
P5 position, as in Drosophila Sickle and Grim and
Xenopsis Casp-9 (Fig 6) [22,46] As shown previously
[46], Phe at P5 position is clearly favored for BIR2
binding The requirement for proteolytical processing
of AFP to expose the N-terminal IBM may explain why only part of the total amount of AFP, that which has undergone proteolytical processing, participates in complex formation with XIAP (Fig 5A) Proteolytical processing of AFP is usually observed in cyto-chrome c-activated lysates [28] It has been shown that proteolytical cleavage of pure AFP results in AFP fragments exposing different destabilizing N-terminal residues [27] However, our results show that pure recombinant AFP and XIAP can interact without any requirements for the presence of active caspases in the reaction mixture We tentatively suggest the existence
of another IAP-binding site in AFP, one which does not require N-terminal processing to be activated for XIAP binding
Structural modeling of the AFP(dimer)⁄ BIR2–3 complex
Our results show that AFP can bind to an entire XIAP molecule but not to its fragments It has also been demonstrated that AFP displaces caspase 3 from its complex with XIAP, suggesting that the BIR2 domain
is involved in this interaction The data also suggest that AFP uses at least two different XIAP binding sites to form the AFP–XIAP complex Previous studies suggest that AFP [27], as well as XIAP [47], is able to
Fig 6 Sequence alignment of IBM-bearing proteins Collinear alignment of the N-terminal sequence 1–55 from HSA and 1–60 from human AFP (upper) Sequence alignment of IBM-bearing proteins: human AFP, caspase 9–p12, caspase 7–p20, Smac ⁄ DIABLO, GSPT1; Omi ⁄ Htr2; mouse caspase 9–p12; Xenopus caspase 9–p12; Drosophila ICE, Reaper, Grim, Hid, Sickle, Jafrac2, GSPT1; C elegans GSPT1 Identical res-idues are highlighted in black Resres-idues conserved in several IBM proteins are indicated in grey IBM-like sequence is boxed Protein sequence data have been taken from the Protein Data Bank [63].
Trang 7dimerize, which could create many possibilities for
interaction stoichiometry We suggest that AFP dimer
forms a complex with XIAP by interacting with both
BIR2 and BIR3 domains The involvement of BIR3 in
the interaction with AFP is supported by recent studies
showing that AFP induced the release of both
caspase 3 and caspase 9, as well as cIAP-2, from the
apoptosome complex [28] It is possible that caspase 9
cointeracts with AFP⁄ XIAP complex similarly to
caspase 3 It is expected that AFP dimer interacts with
the BIR2 and BIR3 domains of XIAP and forms a
2 : 1 stoichiometric complex Native electrophoresis
indicates that the AFP⁄ XIAP complex is formed by
more than two molecules and includes at least three
members (Fig 5) Taking into account that both AFP
and XIAP tend to dimerize, a few interaction models
can be proposed We suggest a simple complex
com-posed of AFP dimer and XIAP monomer
The 3D molecular structure of the AFP molecule
remains unsolved Human AFP and HSA exhibit 39%
amino acid sequence homology [44] The authentic
structural homology of AFP and HSA allowed us to
predict the tertiary structure of AFP based on the
atomic coordinates obtained by X-ray crystallography
for HSA [48] Using the NMR structures of the BIR2
[49] and BIR3 [50] domains, the AFP(dimer)⁄ BIR2–3
complex was constructed (Fig 7) In this complex, the
BIR2 and BIR3 domains show the same local twofold symmetry as two AFP monomers in the AFP dimeric structure Moreover, the IBM-interacting grooves of the BIR2 and BIR3 domains lie close to the ATIF-end
of the first and second AFP monomers, respectively, allowing for the possibility that they belong to the same XIAP molecule
Discussion
Recent evidence has broken the main paradigm of apoptosis, stating that the release of cytochrome c is the point of no return in the apoptotic program [17,18,51,52] It has been shown that certain tumor cells are able to recover after cytochrome c release and survive despite the constitutive presence of cyto-chrome c in the cytosol in the absence of any signs of apoptosis [53] Moreover, caspase activation does not always result in cell death [16–18,52] The ability of the cell to die at the postmitochondrial level depends mainly on the activity of endogenous inhibitors of apoptosis, such as IAPs, sHSPs, or Bcl-2 [6,7,54,55] There is further evidence of a high level of apoptotic activation and the upregulation of IAPs in tumor tis-sue [8,16] Inactivation of XIAP or the cancellation of XIAP inhibition appears both necessary and sufficient for cytochrome c to activate caspases and trigger cell death [9,16] The activity of IAPs is regulated by a group of IAP-regulatory proteins that bind to IAPs and inhibit their antiapoptotic function [17–21] These factors are important research targets to search for new nontoxic drugs with selective pro-apoptotic activ-ity for tumor cells The identification of protein drugs, which can overcome the tumor defense system
by preventing the realization of apoptosis in tumor cells, will have a great potential as tumor therapeutic agents [9]
In this study we showed that AFP could participate
in the regulation of apoptosis in tumor cells by coun-teraction with the most potent endogenous inhibitor of mammalian caspases XIAP Our results show that AFP binds to XIAP and disrupts its interaction with caspase 3 These results are in harmony with the fact that AFP can bind to cIAP-2 and disrupt its interac-tion with caspase 3 and caspase 9 [28] Because pure AFP can bind XIAP in vitro, this interaction appears
to be direct The binding seems to be highly specific, because it did not occur with the nonapoptotic protein HSA, structure and function of which is closely related
to AFP In addition to the direct association with XIAP, AFP could also relive the XIAP-inhibitory effect on the activity of the mature recombinant caspase 3 Moreover, AFP shows the ability to enhance
Fig 7 Hypothetical molecular model of the AFP(dimer)⁄ BIR2–3
complex Each monomer of the AFP dimer is shown in blue and
red, respectively The BIR2 and BIR3 domains are shown in green.
The dashed yellow lines connect the ATIF peptides of each AFP
monomer and the IBM-interacting grooves of the BIR2 and BIR3
domains The figure was produced using MOLSCRIPT v 2.1 [62].
Trang 8cytochrome c-dependent activation of caspase 9 and
caspase 3 in the presence of an inhibitory amount of
exogenous XIAP In this respect, AFP behaves in a
similar manner to the IAP antagonists This group of
proteins is characterized by the presence of the
N-ter-minal conserved BIR-binding motif (IBM), which is
required for IAP binding [20] The presence of IBM in
the intracellular protein allowed us to recognize its
possibility of serving as an IAP antagonist [9,13,23]
Mammalian proteins Smac⁄ DIABLO, Omi⁄ HtrA2
and GSPT1⁄ eRF3 are released from mitochondria
upon triggering apoptosis and require processing to
reveal the IBM at the newly generated N-terminus
[17–19] However, other proteins, such as ARTS and
XAF1, which do not contain an IBM-like motif, were
also seen to antagonize IAP function via an unknown
mechanism [20,21] The search for the potential
IBM-like sequence in the structure of AFP revealed the
presence of a similar amino acid sequence ATIF in
human AFP at position 29–32 [44] It can be proposed
that processing of the first 28 amino acid residues
gen-erates the AFP fragment with an N-terminal motif
ATIF that is highly reminiscent of IBM of caspase 9
and other IAP antagonists (Fig 6)
Our results indicate that only entire XIAP could
bind AFP (Fig 5) This means that AFP binds
simul-taneously to at least two BIR domains BIR3 binding
is preferential for an IBM-like motif similar to that
available in caspase 9 Taking into account that AFP
competes with caspase 3 for complex formation with
XIAP, we consider that both BIR2 and BIR3 may be
involved in complex formation with AFP A similar
model has been described previously [56] for a complex
of dimeric Smac protein with recombinant XIAP
frag-ments containing both BIR2 and BIR3 domains, or
for a complex of XIAP with active caspases 3 and -7
[10] Another model, which involves the entire AFP
molecule, could be also proposed, and will introduce
other parts of the molecule in their interaction with
XIAP To identify the molecular mechanism of the
AFP⁄ XIAP interaction additional structural studies
are needed
Although our results show that AFP interacts
phys-ically with XIAP and protects activated caspases from
IAP-induced inhibition, they do not reveal how it
operates There are several possibilities However, a
functional preference of AFP for tumor cells seems
evi-dent [30–41] It has previously been shown that AFP
selectively penetrates tumor cells via specific membrane
AFP receptors expressed on the surface of tumor cells
but not on normal adult cells [32–36] Unlike other
anti-IAPs, such as Smac or Omi, which have an
exclu-sively mitochondrial localization and become available
to interact with IAPs only after apoptosis has been triggered by cytochrome c release, AFP was available whenever it entered into the cell via cellular membrane receptors or was synthesized inside the cell Thus, AFP
is able to regulate the IAP level in the cytosol inde-pendently of whether cell is undergoing apoptosis or not Under conditions of constitutively high levels of XIAP expression in tumor cells [57], AFP could reduce its protein level, presumably by proteosome-mediated degradation
Comparison with normal cell lines and tissues has shown that many tumor cell lines and tissues have constitutively higher levels of active caspase and free cytochrome c in the absence of apoptotic stimuli and yet are not undergoing apoptosis [58,59] Simulta-neously, tumor cells have high levels of expression of survivin and XIAP [57,60] Survival of cancer cells is possible under conditions of pacific equilibrium between pro- and antiapoptotic signals Taking into account that normal cells and tissues do not overex-press apoptotic stimuli and IAPs, whereas cancer cells and tissues do, IAP-targeting drugs will have highly selective proapoptotic activity for cancer cells and lit-tle toxicity towards normal cells [8,9,16] The general obstacle preventing the design of apoptosis-regulating drugs on the basis of known natural anti-XIAPs is their intracellular localization and the inability to use them as internal regulating factors Considering that AFP can penetrate selectively into tumor cells via specific membrane receptors, the molecular mechan-ism of AFP-mediated targeting regulation of apopto-sis could be suggested to be as follows: (a) AFP selectively penetrates tumor cells via specific AFP re-ceptors; and (b) formation of the AFP–XIAP com-plex prevents its binding to activated caspases, increases XIAP instability against ubiquition⁄ protea-somal destruction and reduces the XIAP level to pro-mote apoptosis induction This function of AFP may serve to sensitize tumor cells to weak proapoptotic stimuli by inducing a tumor-specific response to che-motherapeutic or radiotherapeutic treatments The selectivity of the AFP-mediated proapoptotic activity for tumor cells may be explained by its counteraction with IAPs, which are shown to be dominantly over-expressed in tumor cells under conditions of the sim-ultaneous existence of high levels of various active proapoptotic factors
Normal cells do not undergo AFP-induced apoptosis because they do not express high levels of IAPs, do not contain constitutively activated caspases and do not express membrane AFP receptors AFP seems to
be directly involved in targeting positive regulation of the apoptotic pathway dysfunction in cancer cells by
Trang 9inhibition of IAP function leading to triggering of the
apoptosis machinery
Further studies are required to better understand the
importance of the role AFP in modulating the level of
IAPs in tumor cells Elucidation of the role of AFP
in tumor cell-specific regulation of XIAP function in
apoptosis may have important implications for cancer
treatment and prevention AFP and AFP-derived
pep-tides can potentially be used to overcome drug
resist-ance caused by the differential mechanism of apoptosis
dysfunction in cancer cells
Experimental procedures
Purification of a-fetoprotein
Embryonic AFP was isolated from human cord serum
using ion-exchange, affinity and gel-filtration
chromatogra-phy as described previously [27] The purity and
homogen-eity of the protein were assessed by SDS⁄ PAGE and
western blotting with AFP-specific polyclonal antibodies
as described elsewhere [28] Recombinant human AFP
(rhAFP) was purified from the culture medium of
recom-binant Saccharomyces cerevisiae as described previously [43]
using affinity and gel chromatography
Cells
HepG2 cells originating from the American Type Culture
Collection were grown in Dulbecco’s modified Eagle’s
med-ium (ICN Biomedicals, Inc., Costa Mesa, CA) with addition
of l-glutamine, 10% heat-inactivated fetal bovine serum,
penicillin (100 unitsÆmL)1), streptomycin (0.1 mgÆmL)1) in a
humidified 5% CO2 atmosphere at 37C For a passage,
cells were incubated in 0.25% trypsin solution, then washed
and plated out
Preparation of cell-free cytosolic extracts
Cell-free cytosolic extracts were generated from human
hepatocarcinoma HepG2 as described previously [60] with
minor modifications [28] Cells (4· 108) were collected and
washed three times with 50 mL NaCl⁄ Pi and once with
5 mL hypotonic cell extraction buffer (CEB; containing
20 mm Hepes, pH 7.2, 10 mm KCl, 2 mm MgCl2, 1 mm
dithiothreitol, 5 mm EGTA, 25 lgÆmL)1leupeptin, 5 lgÆmL)1
pepstatin, 40 mm b-glycerophosphate, 1 mm
phenyl-methylsulfonyl fluoride) The cell pellet was then
resuspend-ed in an equal volume of CEB, allowresuspend-ed to swell for 20 min
on ice and then disrupted by passing through a needle The
homogenate was centrifuged at 5000 g for 10 min at 4C
to remove whole cells and nuclei Thereafter the
superna-tant was centrifuged at 15 000 g for 20 min at 4C The
procedure was repeated twice Cytosolic extracts were
assessed for protein content by Bradford assay and stored
in aliquots at)70 C
Analysis of caspase activity Caspase assays were performed with active recombinant caspase-3 (Alexis Corp., San Diego, CA), recombinant full-length rhXIAP (R&D Systems, Minneapolis, MN), purified AFP, and HSA (Sigma-Aldrich, St Louis, MO) All other reagents were from Sigma, unless stated otherwise RhXIAP (200 nm) was incubated with AFP (400 nm) or HSA (400 nm) in IAP buffer (50 mm Hepes, pH 7.5,
100 mm NaCl, 1 mm EDTA, 5 mm dithiothreitol, 0.1% Chaps, 10% sucrose) for 15 min at room temperature Thereafter the active recombinant caspase 3 (3 nm) was added to the reaction mixture, and incubation continued for a further 15 min under the same conditions For the control, caspase 3 was incubated with each of the following compounds separately: AFP (400 nm), HSA (400 nm) or XIAP (200 nm) The kinetics of caspase activity was moni-tored by cleavage of the fluorogeneic substrate [50 lm Ac-Asp-Glu-Val-Asp-7-amino-4-methyl coumarin (Ac-DEVD-AMC), Sigma] at 5-min intervals for 30 min To assess the effects of AFP, HSA and XIAP on caspase activity in cellu-lar extracts in vitro, rhXIAP (250 nm) was incubated with cytosolic extract (40 lg) activated by addition of 5 lm of bovine heart cytochrome c and 1 mm dATP in the presence
of AFP (400 nm) or HSA (400 nm) in 15 lL of a reaction buffer (10 mm Hepes, pH 7.2, 25 mm NaCl, 2 mm MgCl2,
5 mm dithiothreitol, 5 mm EDTA, 0.1 mm phenylmethyl-sulfonyl fluoride) for 1 h at 30C, and the reaction mixtures were analyzed for Ac-DEVD-AMC cleavage Caspase 3 activity was determined by adding 5 lL of cell extracts to 16 lL of substrate reaction buffer (20 mm Hepes, pH 7.2, 100 mm NaCl, 10 mm dithiothreitol, 1 mm EDTA, 0.1% Chaps 10% sucrose, 50 lm Ac-DEVD-AMC) for 40 min at 30C The reaction was stopped by the addi-tion 200 lL of cold NaCl⁄ Pi, and AMC liberation was measured using Victor-1420 Multilabel counter (Wallac, Finland) at excitation 355 nm and emission 460 nm All samples were analyzed in duplicate and the experiments were repeated three times For each sample, caspase activity was expressed in relative fluorescent units (RFU), showing the amount of cleaved substrate normalized for protein content
Direct protein–protein binding assay
To determine possible interactions between AFP, caspase 3, caspase 9, and XIAP, we used a direct coprecipitation assay with purified proteins Human recombinant XIAP (350 ng), His-tagged human recombinant caspase 9 (50 ng), anf act-ive His-tagged rat recombinant caspase 3 (300 ng) were mixed with 0.5 mgÆmL)1 AFP or 0.5 mgÆmL)1 HSA in
Trang 1015 lL buffer A (20 mm Hepes, pH 7.4, 100 mm NaCl,
0.5 mm EDTA, 0.5 mm dithiothreitol, 0,1% Chaps, 10%
sucrose) and incubated for 2 h at 4C Thereafter, 15 lL
of Ni-Sepharose beads (Qiagen, Valencia, CA) in 80 lL of
buffer B (20 mm Na2HPO4, pH 7.2, 0.2 m NaCl) were
added to the reaction mixture and incubation was
contin-ued for the next 1 h The beads were separated from
supernatants by centrifugation and both fractions were
collected Protein–bead complexes were washed four times
and boiled in 20 lL of reducing 2· Laemmli sample
buf-fer Samples, protein–beads and supernatants were analyzed
by SDS⁄ PAGE ⁄ immunoblotting in 12.5% polyacrylamide
gel in with b-mercaptoetanol To study direct AFP–XIAP
protein interaction, human recombinant XIAP (1.5 lg) and
human recombinant AFP (2 lg) were incubated for 1 h at
4C in 30 lL of buffer C (20 mm Hepes, pH 7.2, 140 mm
KCl, 5 mm MgCl2, 2 mm dithiothreitol, 1 mm EDTA,
0.1% OVA) Thereafter, protein mixtures (5 lL) were
sub-jected to 8% nondenaturing continuous polyacrylamide gel
(Tris pH 8.7) and separated by Native electrophoresis [61]
Immunoprecipitation
Cytosolic extracts obtained from HepG2 cells were
nor-malized for protein content (500 lg of total protein in
100 lL of buffer C) and activated by addition of 5 lm
cytochrome c and 1 mm dATP for 30 min at 30C in the
presence of AFP (6 lg) The reaction mixtures were
cooled and incubated with 3 lg of the following
antibod-ies: polyclonal rabbit anti-AFP [28], normal rabbit IgG
(Sigma), or with rabbit anti-(caspase 3) (Santa Cruz, Santa
Cruz, CA) for 2 h with a gentle mixing at 4C
Thereaf-ter, 40 lL of protein A–Sepharose bead slurry (Amersham
Pharmacia Biotech) were added to the each sample
Sam-ples were incubated overnight in a rotating shaker at
4C The beads were pelleted by centrifugation and after
intensive washing, were syringe dried The bound proteins
were eluted by boiling in 25 lL of 2· sample buffer
Sam-ples in aliquots of 10 lL were loaded onto the 12.5%
SDS polyacrylamide gel and subjected to SDS⁄ PAGE ⁄
immunoblotting
Immunoblotting analysis
Samples after SDS⁄ PAGE or native PAGE were
electro-blotted onto a poly(vinylidene difluoride) membranes
(Amersham Pharmacia Biotech) using a semidry transfer
apparatus (Bio-Rad Laboratories, Hercules, CA)
Follow-ing blockFollow-ing, the membranes were incubated with primary
rabbit anti-AFP IgG [28] or goat anti-XIAP IgG (R&D
Systems, Minneapolis, MN), or anti-(caspase 3) IgG (Santa
Cruz) and then incubated with donkey anti-(rabbit IgG)
(Amersham Pharmacia Biotech) or anti-(goat IgG) (Imtek,
Russia) each conjugated to horseradish peroxidase The
blots were visualized using ECL or ECL Plus method
(Amersham Pharmacia Biotech) according to manufacturer instructions
Theoretical calculation of protein structure
In order to predict the tertiary structure of the AFP mole-cule [44] the molecular modeling software package sybyl was used (Tripos Associates, Inc., St Louis, MO) A model
of the dimeric structure of the AFP was reconstructed by using of the atomic coordinates obtained by X-ray crystal-lography for HSA [48] and plotted using molscript v 2.1 [62] The atomic coordinates of HSA (code 1AO6) were obtained from the Protein Data Bank [63] The primary structure alignment of AFP and HSA was constructed using multalin [64] The model of AFP was minimized to
an energy gradient < 0.050 kcalÆmol)1ÆA˚)1using the Tripos force field and a combination of Simplex [65] and Powell algorithms [66] Coordinates of the NMR structures of the BIR2 [49] and BIR3 [50] domains were achieved from the Protein Data Bank files: 1C9Q and 1F9X, respectively
Acknowledgements
This work was supported in part by the International Science & Technology Center, ISTC (grant #1878); by the Academy of Finland (Grant # 107762), the Neobi-ology program of the TechnNeobi-ology Development Center for Finland (TEKES), and the Sigrid Juse´lius Founda-tion
References
1 Reed JC (1999) Mechanisms of apoptosis avoidance in cancer Curr Opin Oncol 11, 68–75
2 Hickman JA (2002) Apoptosis and tumorigenesis Curr Opin Genet Dev 12, 67–72
3 Shi Y (2004) Caspase activation inhibition, and reactiva-tion: a mechanistic view Protein Sci 13, 1979–1987
4 Oliver L & Vallette FM (2005) The role of caspases in cell death and differentiation Drug Resist Updates 8, 163–170
5 Fuentes-Prior P & Salvesen GS (2004) The protein structures that shape caspase activity, specificity, activa-tion and inhibiactiva-tion Biochem J 384, 201–232
6 Cory S, Huang DC & Adams JM (2003) The Bcl-2 family: roles in cell survival and oncogenesis Oncogene
22, 8590–8607
7 Salversen GS & Duckett CS (2002) IAP proteins: block-ing the road to death’s door Nat Rev Mol Cell Biol 3, 401–410
8 Wright CW & Duckett CS (2005) Reawakening the cel-lular death program in neoplasia through the therapeu-tic blockade of IAP function J Clin Invest 115, 2673– 2678